POWER PLANT TESTING WORKS BY PROFESSOR J. A. MOYER Published by McGRAW-HILL BOOK COMPANY Power Plant Testing. A Manual of Experimental Engineering. Svo, xiv + 422 pages, 271 figures. Cloth, $4.00. Published by JOHN WILEY & SONS The Steam Turbine. A Practical and Theoretical Treatise for Engineers and Designers. 8vo, ix+370 pages, 225 figures. Cloth, $4.00. Descriptive Geometry for Students of En= gineering. (Principally "Third Angle" Methods.) 8vo, viii+204 pages, 128 figures. Cloth, $2.00. POWER PLANT TESTING A MANUAL OF TESTING ENGINES, TURBINES, BOILERS, PUMPS, REFRIGERATING MACHINERY, FANS, FUELS, MATERIALS OF CONSTRUCTION, ETC. JAMES AMBROSE MOYER, S.B., A.M. Member American Society of Mechanical Engineers, Mitglied c/es Vereines deutscher Ingenieure, Member of the Franklin Institute, American Society of Refriger- ating Engineers, American Institute of Electrical Engineers, etc. , Assistant Professor of Mechanical Engineering in the University of Michigan, formerly Engineer, Steam Tin- bine Depart- ment, General Electric Company, and Engineer, Westinghouse, Church, Kerr &• Company McGRAW-HILL BOOK COMPANY 239 WEST 39TH STREET, NEW YORK 6 Bouverie Street, London, E.C. 1911 '>* Copyright, 1911, BY McGRAW-HILL BOOK COMPANY • THE SCIENTIFIC PHESS ROBERT DRUMMOND AND COMPANY BROOKLYN, N. V. €"CI.A2 97(>63 PREFACE In the preparation of this book the object in view has been primarily to give in a small volume, somewhat in detail, the generally approved methods of testing engines, turbines, boilers and the auxiliary machinery usually found in power plants, as well as to present more or less complete descriptions of the various kinds of apparatus used and the calibrations required for accurate testing. In addition to this subject-matter, chapters have been prepared on the testing of fuels, refrigerating and hydraulic machinery, as well as on the proper methods and the machinery to be used in making tests of the strength of the materials commonly used in the construction of buildings. As a book for students in laboratory courses it is intended particularly for use in large classes in which at the beginning of the laboratory periods it is necessary to begin at the same time a number of different experiments and tests. On this account care has been taken to state as clearly as possible the descriptions of the apparatus to be used and the precautions to be observed to secure accuracy in the results. Students should be expected, however, to rely to some extent on their own initiative. In most respects the book is probably complete enough in descriptive matter and in general instructions so that very little ' lecture-room work is needed for at least elementary courses. It is the author's opinion that students in experimental engineer- ing laboratories should not receive a great deal of assistance in planning and conducting tests. Sometime they must learn to be resourceful and independent of the "school" type of instruc- tion and obviously the sooner this is appreciated by both instructors and students the greater will be the benefits. At least for very small classes the better plan is the one advocated years ago by VI PREFACE a famous educator, that students working in laboratories when assigned the work of testing a machine, a new type of aeroplane engine, for example, should have very simple instructions such as: "Make tests of this new type of engine, find out what you can about it and report your results." It is to be hoped that the particular method of teaching in laboratories known familiarly as "feeding with a spoon" has disappeared in present-day instruction in technical schools and colleges. Quite a large part of the training required for one to become accurate and reliable in the work of observing and interpreting the results of tests of machinery consists in becoming familiar with the details of the adjustment and calibration of the various instruments, so that they may be used intelligently. Although in the arrangement of the chapters the use of the book by students was given the most careful consideration, yet as a whole the needs of the "practical" man were not lost sight of, and it is hoped that the author's experience when working with this group of readers in testing both large and small power plants has helped to make the book interesting and helpful to them. The book is intended to be also a manual giving useful information in a more or less limited way to those professional engineers having the advantages of a technical training, but who are not thoroughly familiar with the most up-to-date methods of testing. In many cases not nearly all the results that should be calcu- lated to make up a complete report are mentioned. It is the opinion of the author that in a text-book it is desirable that no more than very general instructions should be given regarding the conduct of the test, the quantities to be calculated, and the form and tabulations expected in a report. Such details should be left in the hands of the instructor. Because the size of the book was limited it was necessary to omit explanations of the methods of calculating many interesting and more or less appli- cable results from tests. In the more extended courses it is believed that the instructors can readily fill in these omissions. The author is particularly indebted in the preparation of this book to Professors M. E. Cooley, J. R. Allen, C. J. Tilden, and E. D. Campbell, of the University of Michigan; Professors I. N. Hollis and L. S. Marks, of Harvard University; Professor H. W. Spangler, of the University of Pennsylvania; Professor W. F. M. PREFACE vii Goss, of the University of Illinois; Professor C. H. Peabody of the Massachusetts Institute of Technology; Professor L. V. Ludy, of Purdue University; Professor A. M. Greene, of Rens- selaer Polytechnic Institute; Professor C. C. Lorentzen, of New York University; Professor E. J. Fermier, of the Mechanical and Agricultural College of Texas; Professor E. A. Fessenden, of the University of Missouri; Professor F. H. Sibley, of the University of Alabama; Dr. C. P. Steinmetz and Mr. Richard H. Rice, of the General Electric Company; Mr. H. R. Kent Vice-president, Westinghouse, Church, Kerr & Company; Mr. J. R. Bibbins, of The Arnold Company; Mr. R. A. Smart, of the Westinghouse Machine Company; Mr. St. John Chilton, of the Allis-Chalmers Company; and Messrs. G. E. Wallis, J. E. Emswiler, F. P. Maloney, J. R. Bazley, and E. D. Connell, of Ann Arbor. The short steam tables given in the Appendix have been taken, with permission, from Allen and Bursley's "Heat Engines." In many cases, particularly regarding engineering practice abroad, Pullen's ''Engines and Boilers " has been found very useful. J. A. Moyer. Ann Arbor, Michigan, August, 191 1. INTRODUCTION Tests of the machinery in a power plant are usually made to determine the capacity and efficiencies of its various units when operating under certain definite conditions. In recent engineering practice manufacturers and contractors are generally required to make certain estimates and guarantees of the capa- city and efficiency of the various kinds of machinery supplied. This is exactly equivalent, in other words, to agreeing to pro- vide for doing a given unit of work under specified conditions at a definite cost. The purchaser, on the other hand, for his protection, finds it necessary to determine from the results of reliable tests whether the " guarantees " can be obtained. Obviously, then, the importance of knowing how to make careful and reliable tests, of which the results will not be questioned, cannot be overestimated. Tests of power plants as a whole are also necessary and should be made from time to time in order to determine what results can be obtained from an economic viewpoint when operating under the existing conditions; and also for finding out what saving can be obtained by changes in the operating conditions or by the installation of more efficient auxiliary machinery. From the viewpoint of determining whether or not it is economical to replace old equipment with new, tests of old installations are relatively more important than those of newer ones, because usually it is out of the question to rele- gate practically new machinery to the scrap-heap. The great- est importance of such tests is due, however, to the fact that they show how nearly the existing conditions of operation conform to those of standard engineering practice, and to those obtained in other plants operating with the greatest success. X INTRODUCTION Practice tests in the laboratory are intended to show to students by actual experience the best methods for investi- gating the problems arising in the operation of plants, how to work out in a practical way the doubtful points in designing and constructing machinery, and, above all, to think accu- rately and systematically in such matters. Procedure for making accurate tests may be stated as follows : » i. Procuring a suitable standard testing equipment. Any instruments and apparatus not well known to engineers gen- erally and which are of doubtful accuracy or sensitiveness should always be avoided. Remember that a single element of uncer- tainty may vitiate the acceptance of the results of a test of otherwise undoubted accuracy. 2. Careful calibrating of instruments before a test, so that the greatest possible errors of the tests are definitely known and that proper allowance can be made in the results. 3. Systematic recording of observations. 4. Recalibrating of instruments after a test to determine whether there have been any changes in their accuracy. 5. Preparing of a report embodying data, results and con- clusions. 6. Tabulating and plotting on cross-section paper the important results. This plotting is not only for the purpose of showing the results graphically, but also for the purpose of providing a check or a method of eliminating errors in obser- vations or in calculations. The skill of an engineer in testing is shown more than in any other way, by his ability to check results. If after applying various checks, usually by means of plotted curves, the results for varying conditions are found to agree, the engineer is able to tell definitely whether or not his tests are reliable. CONTENTS PAGE Introduction i I. Measurement of Pressure i (i) Manometer or U-tubes i (2) Bourdon Gage 6 (3) Diaphragm Gage 8 (4) Vacuum Gage 9 (5) Recording Gage 10 (6) Calibration of Pressure Gages 12 Dead-weight Gage Testers 13 Mercury Columns 18 (7) Calibration of Vacuum Gages 21 (8) Calibration of Low-pressure Gages 21 (9) Draft Gages 23 II. Measurement of Temperature 25 (1) Mercurial Thermometer 25 (2) Alcohol Thermometer 26 (3) Calibration of Thermometers 26 Compared with Standard Thermometer . 26 Compared with Temperature of Steam 29 (4) Corrections for Stem Exposure 31 (5) Recording Thermometers 34 (6) Pyrometers 36 a. Thermo-electric 36 b. Mechanical 38 c. Optical 39 d. Calorimetric 41 (7) Seger Cones 43 III. Determination 'of Moisture in Steam 46 (1) Throttling Calorimeter 47 (2) Separating Calorimeter 53 (3) Combined Separating and Throttling Calorimeter for Low-pressure Steam 57 (4) Electric Calorimeter 61 (5) Barrel Calorimeter 62 (6) Calibration of Calorimeters 65 xi CONTENTS PAGE IV. Measurement of Areas 66 (i) Planimeters 66 a. Polar 67 b. Coffin 75 c. Rolling 78 (2) Testing of Planimeters 79 (3) Durand- Bristol Integrating Instrument 80 V. Engine Indicators and Reducing Motions 85 (1) Indicators : 85 a. Watt 85 b. Thompson 86 c. Crosby 88 1. Inside Spring : 89 2. Outside Spring 92 d. Star Brass 93 e. Tabor 93 /. Bachelder 95 g. Cooley-Hill Continuous Indicator 99 h. Optical 101 (2) Calibration of Indicator Springs 105 (3) Tests of Drum Springs 112 (4) Reducing Motions 113 a. Pantograph or Lazy-Tongs , 115 b. Parallel Motions for Indicators 116 c. Reducing Motions Attached to Indicators 118 (5) Calculation of Indicated Horse Power 118 VI. Measurement of Power 122 (1) Absorption Dynamometers 122 a. Prony Brake 123 b. Rope Brake 125 c. Alden Dynamometer 127 d. Water Brake 128 (2) Electric Generators and Motors as Dynamometers. . . 131 (3) Transmission Dynamometers 133 Goss 134 Webber 136 Emerson Power Scales 138 Flather 140 VII. Measurement of the Flow of Fluids '. 142 (1) Air and Other Gases 142 a. Meter ; 142 b. Pitot Tube 143 c. Orifice 146 d. Calorimetric Method 146 e. Anemometers 147 (2) Steam — Wet, Dry, Saturated, and Superheated 148 a. Orifice 148 b. Meter 149 CONTENTS Xlii PAGE (3) Water . 150 a. Meter 151 b. Automatic Weigher 152 c. Venturimeter 155 d. Orifice 158 e. Weir 1 59 VIII. Calorific Value of Fuels 162 (1) Fuel Calorimeters ■ 162 a. Calculation of Water Equivalent 163 b. Mahler 164 c. Atwater 167 d. Emerson 168 e. Parr 169 /. Carpenter 172 g. Junkers 174 (2) Calorific Value from Analysis 177 (3) Proximate Analysis 178 IX. Flue Gas Analysis 181 (1) Sampling Bottles and Tubes 182 (2) Fisher's "Orsat" Apparatus 188 (3) Allen-Moyer "Orsat" Apparatus 190 (4) Calculations 192 a. Weight of Air 194 b. Weight of Gases 195 (5) Recording Apparatus for C0_, 196 X. Boiler Testing 200 A. S. M. E. Rules and Data Sheets 202 XL Steam Engine Testing 229 (1) Mechanical Efficiency and Friction 229 (2) Valve Setting 230 a. Slide Valve 230 b. Corliss Valve 234 (3) Clearance Tests 240 (4) A. S. M. E. Rules and Data Sheets 241 (5) Heat Balance 257 (6) Thermal Efficiency 259 (7) Entropy-temperature Diagrams. . , 262 (8) Willans' Law 274 (9) Steam Engine Lubricators 276 XII. Testing of Steam Turbines and Turbine Generators. . 279 XIII. Methods for Correcting Steam Engine and Steam Turbine Tests to Standard Conditions 296 XIV. Gas Engine and Producer Tests 306 (1) Indicated and Brake Horse Power. 306 (2) Gas Engine Indicators 307 (3) Measurement of Fuel 308 (4) A. S. M. E. Rules and Data Sheets 310 xiv CONTENTS PAGE (5) Abnormal Indicator Diagrams 316 (6) Gas Producers 319 (7) Capacity and Efficiency of Gas Producers 324 XV. Tests of Ventilating Fans or Blowers and Air Com- pressors 325 XVI. Tests of Refrigerating Machines 337 (1) Compression System 338 (2.) Absorption System 345 XVII. Tests of Hot-air Engines 350 XVIII. Tests of Hoists, Belts, Rope Drives, and Friction Wheels 353 XIX. Tests of Hydraulic Machinery 356 (1) Belt-driven Feed Pumps 357 (2) Steam Feed Pumps 358 (3) Impulse Wheels. . . ._ 363 (4) Water Turbines 366 (5) Hydraulic Rams 367 (6) Pulsometers 370 (7) Injectors 372 XX. Tests of the Strength of Materials 376 POWER PLANT TESTING CHAPTER I MEASUREMENT OF PRESSURE The simplest instrument used for measuring pressure is a glass tube bent into the shape of the letter U, as illustrated in Fig. i. When such a tube, called technically a manometer or U-tube, is partly filled with a liquid, usually water or mercury, and is connected at A by means of tubing to the vessel in which the pressure is desired, there will be ob- served a difference in the level of the liquid corresponding to the pressure. If the end of the tube at B is open to the atmosphere, then the difference in the level of the liquid in the two legs measured in inches, multiplied by the weight of a cubic inch of the liquid in pounds, gives the difference in pressure in pounds per square inch between that in the vessel and atmospheric pressure. When the level in the leg B is higher than in A then the pressure measured is greater than atmos- pheric and is called gage pressure to distinguish it from ' the other condition when the level in the leg A is higher than in B; that is, when the pressure is less than atmospheric. In the latter case we speak of vacuum or negative pressure. As such instruments are usually constructed, a scale suitably Fig. i.— A U-tube. The Simplest In- strument for Meas- uring Pressures. POWER PLANT TESTING M graduated for measuring the difference between the levels of the liquid in the tube is placed between the two legs, as shown in Fig. 2. Still another type is illustrated in Fig. 3. In a manometer of this kind one leg can be made very short if it is correspondingly large in diameter. If the scale is adjusted so that the level in the short leg is at the zero of the scale, then the level in the long leg will indicate directly inches of pressure or of vacuum as the case may be. A typical vacuum gage of the same kind is illustrated in Fig. 4. The end of the tube corresponding to the short leg in Fig. 3 is shown at A. When manometers are to be used for pressure or vacuum measure- ments of steam, a condenser (C, Fig. 5) is often employed to prevent the passage of steam into the glass tube in which it would form a water column on the top of the mercury for which a correction l would have to be made. To be effective the condenser, C, must always be partly filled with water. Manometers or U-tubes of very small diameter when filled with mercury may be affected by capillarity to such an extent that in order to obtain the true height corresponding to the pressure, a correction must be added. It is not at all unusual to find manometers used for vacuum gages to be comparatively small in diameter, and unless the gradua- tions of the scale have been corrected for the error due to capillarity the proper allow- ances must be made for all observations. Fig. 5 shows by a curve the values of this as determined by Pullen for mercury columns of W Fig. 2. — A Simple Manometer with a Graduated Scale. correction various diameters. Mercury columns should be read at the top of the meniscus and water columns should be read at the bottom. In this way, 1 Correction for water on the top of a mercury column is most conveniently made by dividing the length of the water column by the specific gravity of mercury (13.6) and adding this equivalent length to the mercury column on which the water rests. MEASUREMENT OF PRESSURE except in very small tubes, the errors due to capillarity may be regarded as negligible. Conversion of Pressures. It is frequently necessary to re- duce pressures in inches of mercury or of water to the equiv- alent in pounds per square inch. Since the weight of a cubic inch of mercury at 70 degrees Fahrenheit is .4906 pound and Fig. 3. Fig. 4. Typical Mercury Vacuum Gages. of water at the same temperature is .0360 pound, pressures in inches of mercury at the usual " room " temperatures can be reduced to pounds per square inch by multiplying by .491 or by dividing by 2.035, and similarly inches of water can be converted to pounds per square inch by multiplying by .0360 or by dividing by 27.78. 4 POWER PLANT TESTING Kilograms per square centimeter are reduced to pounds per square inch by multiplying the kilograms per square centi- meter by 14.223, or by dividing by .0703. A cubic foot of water at 70 degrees Fahrenheit weighs 62.3 pounds and at 30 degrees Fahrenheit, 62.4 pounds. At ordi- nary room temperature the pressure due to 2.31 feet of water is equivalent to one pound per square inch. 1 :: t f r f i 1 t 1 i 1 r V 3 x t 5 K 5 \_ -^ \, ^ ""--, 0.05 0.1 0.15 0.2 0.3 0.4 0.5 0.6 Diameter of Tube in Inches. Fig. 6. — Curve of Capillarity Corrections for Mercury Columns. Tubes used as mercury manometers must be cleaned from time to time by washing the inside surface with nitric acid and afterward thoroughly cleansing them with water. Mercury used in manometers should be free from impurities. Usual 1 The unit pressure of one pound per square inch is equivalent also to that due to a column of air of uniform density, of which the vertical height in feet is approximately, 144.0 divided by the weight of a cubic foot of air at the temperature, pressure and humidity as observed. Tables of the weight of air are given on page MEASUREMENT OF PRESSURE impurities can generally be removed by filtering through a clean cloth of close texture or a thin chamois leather. Air can be removed by boiling, but by far the best method for cleaning mercury is by means of a mercury still. Unfortunately an apparatus of this kind is not available in most engineering laboratories. Pressure Gages. The large size necessary, however, for manometers or U-tubes, even if filled with the heaviest liquids, makes their use unsuitable except for comparatively low pres- sures. Instruments more desirable for high pressures are made by the applica- tion of some kind of elastic material de- signed to produce a uniform deformation for variations of pres- sure. By connecting a suitable auxiliary mechanism to the elastic element it can be made to move a needle to indicate on a graduated dial the degree of pressure. The most common form of such devices is a hollow brass or steel tube bent into the shape of an arc of a circle. It is a well-known principle that when a straight piece of tubing is bent into this shape the sides come nearer together, making the section of the tube a very much flattened oval. A tube of this kind is illustrated in Fig. 7, showing also in the right-hand corner a transverse section. If one end of such a tube is closed and fluid pres- sure is applied to the inside, the parallel sides, as at A and B, tend to separate and consequently there is a tendency for the radius of curvature of the tube to become larger, thus moving the end at E toward F. By connecting a suitable mechanism Fig. 7. — A Typical Bourdon Tube. POWER PLANT TESTING to E, the degree of pressure can be indicated. Instruments of this kind are called Bourdon gages. Fig. 8 shows one of the simplest forms of such gages used in power plants to indicate the pressures. It consists essentially of the curved tube T of oval cross-section closed at one end. By means of suitable levers and gears a pointer or needle P is made to move over a dial graduated or marked to indicate pressures in standard units as, for example, pounds per square in. (English system) or kilograms per square centimeter (Metric system). (Fig. 9.) Fig. 10 shows a form of Bourdon gage in which the amount of vibration of the needle due to the jarring that occurs in Fig. 8. — Typical Bourdon Pressure Gage. Fig. 9. — The Dial of a Pressure Gage. A Modified Bourdon Gage. locomotive and other portable services has been reduced to a minimum by supporting the pressure tube in the middle MEASUREMENT OF PRESSURE instead of at its end as in Fig. 8. This form of tube has also advantages for use in gages exposed to temperatures below freezing, since the arms can be drained of water, while the other form will usually hold the water that has entered. Gages to be used to determine the pressure of ammonia have the oval tube made of steel instead of brass because the latter material deteriorates rapidly in the presence of ammonia. Bourdon gages may be used for indi- cating the pressures of either liquids, steam or gases without observing special pre- cautions if the temperature is never much over 150 degrees Fahrenheit. If, however, the elastic tube in the gage is heated above this limit it is likely to lose some of its temper. When used for steam pressure, therefore, some form of siphon or water seal must always be used to prevent steam from entering the gage. The type of siphon used most commonly is illustrated in Fig. 11. There is always a possibility, however, that air carried in the steam may be entrapped at a, where it forms a cushion, preventing the gage from indicating the true variation in pressure . For this reason the form of siphon shown in Fig. 12 is preferred for accurate measurements. In Bourdon gages any lost motion v y of the parts is taken up by the hair- ] J []>>*^-r-^^ spring attached to the spindle carrying 1^ TT O'V the pointer. Adjustments. The ratio of motion of the pointer with respect to that of the tube can be adjusted in most Bourdon gages by sliding a set-screw in a slot in the short arm of the rack- lever. In the gage illustrated in Fig. 8 when the short arm of the rack-lever is made longer by adjusting the set-screw, the movement of the rack and also of the pointer is reduced for a given deflection of the tube. Fig. 11. — A Circular Siphon for Steam Gages. Fig. 12. — A U-shaped Siphon for Steam Gages. 8 POWER PLANT TESTING Sometimes when used carelessly, especially when subjected to pressures beyond the scale on the dial, the tube of the gage takes a permanent " set "; or, in other words, it does not spring back to its original position, and the pointer does not come back to the zero mark. In such exigencies and also for adjustment after calibration the needle can be forced off from its spindle — preferably by the use of a clamp or " needle-jack " made by gage manufacturers specially for this service — and then set again in position where it should be. Another kind of gage in which there is a metallic disk or diaphragm instead of a bent tube for actu- ating the indicating device is sometimes used. One of this type is well illustrated in Fig. 13. It con- sists of a corrugated diaphragm clamped around its edge by the flanges of an encir- cling chamber. Pres- sure applied on the lower side of the diaphragm deflects it upward, the amount of this upward move- ment being proportional to the pressure. By means of a con- necting strut S the movement of the diaphragm is communi- cated to a rack R connected to a small pinion attached to the spindle of the needle indicating the pressure on the graduated dial. Since the deflection of the center of the diaphragm is pro- portional to the pressure and is inversely proportional to the cube of its thickness, a very slight alteration in the thickness of the diaphragm will cause a considerable change in the reading of the gage. Fig. 13. — A Typical Diaphragm Ga; MEASUREMENT OF PRESSURE 9 Vacuum Gages. For the measurement of vacuum instead of pressure Bourdon gages are also very commonly used. The design used for a pressure gage is altered only in the arrange- ment of the levers moving the needle, which for vacuum meas- urements turn it in the same direction as for pressure (clock- Fig. -A Recording Pressure Gj wise) , when, as in this case, the tube is bent inward or toward the center of the gage instead of outward as for pressure meas- urements. Vacuum gages are usually graduated to read inches of mercury below atmospheric pressure. Absolute pressure in inches of mercury is the difference between the barometer and the reading of such vacuum gages. 1(1 POWER PLANT TESTING Still another type of pressure gages known as a compound gage is used to indicate either pressure or vacuum on the same dial. Recording Gages. In many modern power plants recording gages are used to give a graphic record on a chart of the pres- FlG. :5. — Operating Parts of a Recording Gage with a Helical Tube. (Bristol.) sure or vacuum for 24 hours. The most common type of record- ing gage is shown in Fig. 14. These gages are made with either a circular tube of oval section in the form of a helix as illustrated in Fig. 15, with a metallic Bourdon tube as shown in Fig. 16, or with a dia- MEASUREMENT OF PRESSURE 11 phragm device as in Fig. 17. The first and second of these three types are generally used for cases where the maximum pressure is greater than 3 pounds and the third when it is less. The recording arm is preferably attached directly to the moving element so that no gears, levers, or other multiplying devices are needed. A more compact and less expensive form of such gages is illustrated in Fig. 18. Fig. 16. — A Recording Pressure Gage Operated by a Bourdon Tube. The average pressure corresponding to an irregular curve traced on the circular card of one of these recording gages is obtained with a fair degree of accuracy by integrating the curve by means of a Durand-Bristol integrating instrument described on page 80. Corrections to be applied to the readings of these gages are of course obtained by calibrating in the same way as for an indicating gage. 12 POWER PLANT TESTING Still another type of recording pressure gages is shown in Fig. 19. Calibration of Gages. Until recent years when the so-called "dead-weight " apparatus for testing gages came into general Fig. 17. -A Low-pressure Recording Gage Operated by a" Diaphragm " Device. use, gages used in other places than engineering laboratories were commonly calibrated by comparison with a so-called test gage. Such test gages have usually somewhat finer gradu- ations than the ordinary gages used in practice and are probably also adjusted a little more accurately. They should never be exposed to the severe conditions of service, being intended MEASUREMENT OF PRESSURE 13 only for purposes of comparison. This comparison can be made anywhere by connecting the standard and the gage to be tested to any system of piping in which the pressure can be varied either by pumping a liquid, or by means of valves " throttling " steam, water or air under pressure. The only important precaution to observe is that the two gages shall be at approximately the same level when a liquid is used, and that the velocity of the fluid in the main pipe to which the gages are attached is negligible or is the same at the points where the connections for the gages are inserted in the "main "pipe. Test gages must, of course, be calibrated from time to time with some standard apparatus to insure their accuracy. A bench pump suitable for calibrating by comparison is illus- trated in Fig. 20. Gage Testers. In very many power plants the use of the test gage has been superseded by some form of gage tester and by this means the gages used in the plant can be calibrated directly with an absolute standard. Cali- brations of gages for high pressures by means of mercury columns are for practical reasons suitable only for laboratory work. Dead-weight Gage Testers. The best-known form of this apparatus is made by the Crosby Steam Gage and Valve Co., and is illustrated in Figs. 21 and 22. The latter figure shows a partial section. It consists of a vertical cylinder C, into which is fitted very accurately a piston P, of which the area, when new, is exactly one-fifth of a square inch. A circular platform upon which weights can be placed is attached to the Fig. 18. -A Veyr Compact Tpye of Recording Gage. (Bristol) 14 POWER PLANT TESTING upper end of this piston. The cylinder, C, communicates at its lower end with the reservoir R, fitted with an adjustable plunger working in a screw and is operated by a hand wheel. A pipe, T, attached to the lower part of the reservoir is pro- vided with unions and special fittings for attaching gages of various sizes. In the horizontal portion of this pipe a three- way cock or valve V is provided for either draining the reser- voir or for closing the pipe so that the liquid in the apparatus Fig. -A Combined Recording and Indicating Pressure Gage. will not escape when the gage is removed. In operation, after the gage has been attached securely, the adjustable plunger S (Fig. 22) is screwed down to the bottom of the reservoir R, then with the piston removed, glycerine or heavy oil is poured into the cylinder C at the same time that the plunger P is screwed out. In this way the reservoir can be completely filled with oil witout entrapping any considerable amount of air which would act as a cushion preventing the most satisfactory operation of the apparatus. MEASUREMENT OF PRESSURE 15 Fig. -A Bench Test Pump. If the area of the piston is one-fifth of a square inch then each pound weight added on the plat- form produces a pressure on the liquid of 5 pounds per square inch . The weight of the platform and pis- ton (usually 1 pound) must al- ways be included in the weight pro- ducing the pres- sure. As the load on the platform is increased the plungermust,from time to time, be screwed in to keep the piston and platform " floating." When observations are being taken it is very essential that the loaded platform be given, preferably by hand, a slight rotary motion to reduce to a mini- mum the friction of the piston in its cylin- der. Suggested Proce- dure with Dead-weight Testers. The accura- cy of the gage to be calibrated is deter- mined by subjecting it to known pressures and noting its error. Before the piston has -Crosby Dead-weight Gage Tester. been put into place the reading of the gage, called "zero-reading," should be observed and recorded in a 16 POWER PLANT TESTING form similar to the one on page 20. Then the pressure should be increased 5 pounds per square inch at a time (corresponding usually to a weight of 1 pound) up to the limit of the gradu- ations on the dial, spinning the piston gently when each reading is taken. Commencing then with the highest pressure the Fig 22. — Sectional View of a Dead-weight Gage Tester. same operation should be repeated by decreasing the pressure by the same increments. 1 1 When the pressure is being decreased the movement of the pointer must be always in a counter-clockwise direction just before a reading is taken. In other words, if a weight of 2 pounds has been taken from the load on the piston when only 1 pound should have been removed the pointer will, of course, get below the next point to be calibrated. To secure the reading missed it will not be correct to add 1 pound and MEASUREMENT OF PRESSURE 17 A modification of the dead-weight gage tester is shown in Fig. 23. This instrument is particularly suited for calibrations at high pressures. Its range is from o to 1500 pounds per square inch. Any pressure within these limits can be obtained without shifting heavy weights. Read- ings are taken when the scale beam is balanced. The hand wheel A is used to regu- late the fluid pres- sure by means of a plunger as in the apparatus shown in Figs. 21 and 22. The other hand- wheel B shown in the figure must be kept rotating when observa- tions are taken. The slight jar- ring of the parts due to its rota- tion serves to make the" friction as small as possible. For still higher pressures up to 12,000 pounds per square inch, a heavy stationary type shown in Fig. 24 can be used. Fig. 23. — Crosby Portable Fluid Pressure Scales. take the reading, because the friction and lost motion will now be in the same direction as with increasing pressures; and to overcome this diffi- culty the pressure must be increased again to a value higher than that for which the reading is to be taken. For the purpose of increasing the weight it is not necessary to put on more weights, as additional load in such cases can be put on by the pressure of the hand. The same precautions apply with even greater force to calibrations made with test gages or with a mercury column. With either of these instruments discrepancies may occur with increasing or decreasing pres- sures. In fact the only certain way to get satisfactory results with these instruments is to keep the pointer of the gage or the mercury column, as the case may be, moving continually in the same direction. 18 POWER PLANT TESTING Calibration of Gages with Mercury Columns. The ultimate standard for the determination of reasonably high pressures is the mercury column, but the apparatus required is so com- plicated and occupies so much space that this method is suit- able only for use in laboratories where it will have the attention of skilled observers. For the purpose of calibrating steam gages mercury columns have been fitted up in a variety of ways. Simplest of these is the method of connect- ing, the gage to be tested by means of a short tube to a " closed " mercury well into the top of which a long glass tube has been inserted. The pressure can then be increased either by displacing some of the mercury in the well by means of the plunger in the mercury pump shown atth e right-hand side of Fig. 25, and forcing it up into the glass tube, or else by pouring mercury into the tube from the top as must be done in the ap- paratus shown in Fig. 26. Zero pressure for compari- son is to be taken on the column at the same level as the center of the gage. Beginning then with 5 pounds per square inch pressure on the gage observe the corresponding height of the mercury column and its tem- perature, and then continue the observations, first increasing the pressure and then decreasing by increments of 5 pounds, as indicated by the gage. Equivalent units for calibration can be computed from the height of the mercury column, since 1 inch of mercury at 70 degrees Fahrenheit is equivalent to a pressure of .4906 pound per square inch. When using a mercury-testing apparatus it is necessary to observe the temperature near the mercury column in the Fig. 24. -Fluid Pressure Scales for High Pressures. MEASUREMENT OF PRESSURE 1!) room in which the work is being done, so that the observed height of the mercury column can be corrected to a tempera- ture at which the relation between pressure in pounds per square inch and height is known. The coefficient of cubical expansion of mercury is not constant, as will be observed from the following table : Temp. Deg. Coefficient of Fahr. Cubical Expansion 3 2 .0000998 5° .0001000 70 .0001002 90 .0001004 1 10 .0001007 Fig. 25. — Standard Mercury Column and Fig. 26. — Simple Open Hand Pump. Mercury Column. 20 POWER PLANT TESTING For very accurate work allowance must be made for the expansion of the graduated scale. Coefficients of expansion of metals are given in the Appendix. Instead of connecting the gage directly to the mercury well, it is sometimes connected to one end of a steam drum and the mercury column is connected to the other. The increments of pressure are then obtained by increasing or decreasing the steam pressure in the drum. Observations taken in the calibration of a steam gage should be recorded and the computed errors tabulated in a form similar to the following: CALIBRATION OF PRESSURE GAGE. COMPARISON WITH GAGE TESTER Date Observers Maker of gage Maker's No Laboratory No Limits of Graduation. No. Weight Gage Readings. Lbs. per sq. in. Actual Mean Error of on Tester Pressure. Lbs. per of gage. Lbs. per Remarks Reading. Lbs. Up. Down. Mean. sq.in. sq.in. The error of the gage is determined by the comparison of the mean of the up and down readings with the actual pressure- Curves. From the data tabulated two curves are usually plotted : i . Mean gage readings (abscissas) and actual pressures (ordi- nates). Use a large sheet of coordinate paper for this curve. 2 . Error Curve : Mean Gage Readings (abscissas) and mean corrections, positive and negative (ordinates). See curve in MEASUREMENT OF PRESSURE 21 Fig. 27. Error curves should have the points, if very irregular, connected by a broken line rather than by a " fair " or average curve through them. Never, however, try to draw an irregular curve through each of a number of scattered points when the points are supposed to follow a definite relation between the coordinates selected. A " fair " curve should then be drawn between the irregular points. Calibration of Vacuum and Low-pressure Gages. A vacuum gage is usually calibrated by connecting it to one end of a U-shaped glass tube of which both legs are about 30 inches +4 a w +2 & sr ° -4 LlHJ' ■A 1 Tf-T] T 111 111 111 111 llfttii iififii'iif iPiif ffl f , jif-,; 1 -\ - : - -:■}•!- j i : : EH: f-rH: J :: : TT 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Mean Gage Readings. Lbs. per Sq. In. Fig. 27. — Typical Error Curve for a Pressure Gage. long and are filled to about half their length with mercury. The U-tube and gage are then connected to the receiver of an air pump or else to an aspirator or ejector operated by water or steam pressure, such as chemists use for vacuum filtering. The aspirator is really the more convenient instrument to use. If the readings of the vacuum gage are correct, they will corre- spond exactly with the difference in the level of the mercury in the two legs of the U-tube. In case a condensing engine is operating when the calibra- tion of the vacuum gage is to be made, both the gage and the glass U-tube may be connected to the condenser. A compari- son of the readings taken will show, under the best possible 22 POWER PLANT TESTING conditions, the absolute errors of the gage. A suitable scale about 30 inches long and accurately graduated should of course be provided and placed between the two legs of the U-tube. An apparatus consisting of an air-pump designed for a high vacuum and a mercury column is. illustrated in Fig. 28. It is a very convenient means for testing vacuum gages. Fig. 28. — Air Pump and Mercury Columns for Testing Vacuum A low-pressure gage with a scale from say o to 15 pounds per square inch is very easily and accurately calibrated by using the same glass U-tube mentioned for the calibration of the vacuum gage with air pressure, preferably, or with steam pressure. Otherwise the method for calibration is the same as for a vacuum gage, except that inches of pressure instead of inches of vacuum are observed. MEASUREMENT OF PRESSURE 2:} Draft Gages. Many engineers use an ordinary glass U-tube manometer (Fig. 29) filled with water for measuring small pressures like that due to the draft in a chimney or that produced in air-ducts by ventilating fans or blowers. For such obser- vations in many cases, however, greater accuracy is desired than can be secured by the use of the ordinary U-tube and a special form of manometer is used in which the dis- tance moved by the surface of the liquid in the tube is greater than the vertical change of level. Fig. 30 illustrates a simple device of this kind. It consists of a bottle B, filled with water, having a suitable opening at the bottom to which by means of a short rubber tube the inclined glass tube CD is attached. At the upper end of this tube a piece of rubber tubing T is shown and is intended to be connected to the chimney, duct, or flue in which the pressure is to be obtained. A scale placed behind the inclined tube CD should be graduated so that when the spirit level L is adjusted, vertical differences in level in the bottle will be indicated by the scale. Then differences in the readings of, the scale will give directly the difference in pressure in inches of water just as with an ordinary U-tube. Fig. 20. — Simple U-tube Draft Gage. B c rnT ~-~AD^_^X/ E3^rE==E=5 nm Fig. 30. — Inclined Tube Draft Gage. Very accurate draft gages of this type known as Ellison's, are shown in Figs. 31 and 32. The inclination of the tube in these instruments is usually about 1 to 10. Instead of water a very light oil is used to fill the tube. It is claimed that this 21 POWER PLANT TESTING oil has the advantages of having less capillarity than water and also, being lighter, permits the use of a longer scale for a given difference in level. Graduations on these instruments which are sold commercially 1 are, however, always made to Fig. -Ellison's Improved Draft Gage. read equivalent inches of water. These draft gages are also often used for measuring small differences of pressure. For example, if there are two vessels containing gases at different pressures and one is connected to the left-hand side and the other to the Fig. 32. — Ellison's Differential-direct Draft Gage for High Drafts. right-hand side of the gage, it will indicate the difference in pressure. When calibrating gages it is worth while to notice that when instruments are to be used to observe practically con- stant values it is necessary to calibrate them only near the values to be observed. 1 American Steam Gage and Valve Mfg. Co., Boston and Chicago. CHAPTER II MEASUREMENT OF TEMPERATURE Mercurial Thermometers. Temperatures less than about 500 degrees Fahrenheit are usually measured by means of mercurial thermometers, depending for their action on the expansion of mercury in a glass bulb and a graduated capil- lary tube. 1 Whenever mercurial thermometers are used for any work where reasonable accuracy is expected they should be care- fully calibrated before the test is made ; and . after the test the calibration should be at least roughly checked to be sure that the zero of the thermometer has not changed. Too often it happens in practice when tests are being made, as for example of a boiler or of a steam turbine, that in some way a thermometer not previously calibrated has been used, and before the end of the test is broken. It is then too late to get a calibration and sometimes very important results of tests are made doubtful because of such negligence. Calibrations of thermometers of all kinds must be made often because there is always the possibility that a little of the mercury has become detached from the column and remains unobserved either on the sides or at the top of the capillary tube. In all glass thermometers there is always taking place with use and time a gradual and permanent change in the 1 The best thermometers for ordinary engineering work are those having graduations etched on the tube. The only difficulty with this method is that after considerable, use the ink originally in the etched markings disappears and it becomes difficult to read the scale. When this happens a thick paint made of lampblack and shellac or printer's ink can be rubbed over the etched scale, and when this paint is rubbed off after a few minutes, there will be enough of it left in the etchings to make the scale as legible as when new. A crayon or pencil of soft greasy graphite like those used by glaziers for marking glass or by ship- pers for marking cases is a satisfactory substitute for the paint, although it is not so permanent. 25 26 POWER PLANT TESTING volume of the bulb, more in new thermometers than in old ones, altering the zero point, and of course, also the true values for all the graduations. Alcohol Thermometers. For the measurement of tempera- tures much below zero Fahrenheit thermometers filled with mercury are not satisfactory, and alcohol or " spirits of wine " is used. These liquids, on the other hand, are not suited on account of their high vapor tensions for high temperatures. Conversion of Temperatures and Heat Units. Tempera- tures in Centegrade degrees are converted into Fahrenheit by multiplying by §■ and adding 32. Kilogram-calories multiplied by 3.968 give the equivalent British thermal units (B.T.U.), and kilogranncalories per kilogram X1.8 give British thermal units per pound. A " small " or gram calorie is one-thou- sandth as large as a kilogram-calorie. Calibration of Thermometers. Tests to determine the accu- racy of thermometers are made by subjecting them to known temperatures and noting the errors. This is done usually in one of two ways: 1. By comparison with a so-called " standard " thermom- eter known to be accurate. 2. By comparison with temperatures corresponding to steam pressures. Since the second method is not applicable for temperatures below the boiling point of water, it is not often used for temperatures below 212 degrees Fahrenheit. For "low-read- ing " thermometers, therefore, the first method is generally used. Calibration by Comparison with a Standard Thermometer. For low-temperature calibrations, the thermometers to be tested are usually suspended together with a " standard " thermometer of which the errors are known in a water bath arranged so that the temperature can be varied. This bath may consist simply of a vessel provided with a coil of pipe through which steam can be circulated, and has also a suitable stirring device. If the water is kept well stirred in such an apparatus a uniform temperature can be maintained and three or four thermometers can be calibrated at the same time. Fig. 33 illustrates diagrammatically a very simple apparatus of this kind, except that the water bath is heated by discharg- MEASUREMENT OF TEMPERATURE 27 ing steam directly into the water. This arrangement permits changing the temperature more rapidly than with the coil of pipe mentioned above. " Standard " thermometers for comparison should be pref- erably those which have been calibrated at standardizing laboratories such as at the U. S. Bureau of Standards at Wash- ington, D. C, at the Reichsanstat at Berlin, Germany, or at the Royal Physical Testing Laboratories in London, England. Thermometers a t — Paddles % Fig. 33. — Apparatus for Calibration of Thermometers at Temperatures Less than the Boiling-point of Water. The steam laboratories of nearly all technical colleges have sets of standard thermometers suitable for determining the errors of good thermometers to be used as " secondary " standards. When the method of comparison with a " standard " ther- mometer is to be used for temperatures higher than are obtain- able with the apparatus shown in Fig. 33, the " standard " thermometer and the other thermometers to be calibrated are placed in adjacent cups or wells of the same depth inserted in a suitable cylindrical drum with pipe connections permitting 28 POWER PLANT TESTING a flow of steam through it. The thermometer cups or wells should be filled with cylinder oil or, preferably, for high tem- peratures, with mercury. 1 Temperature is varied by throttling with the valves on either or both the steam inlet and discharge pipes. Usually the necessary adjustment is made more easily by manipulating the discharge valve rather than the inlet. At least five minutes should be allowed after the valves have been adjusted for the mercury in the thermometers to come to rest before readings for comparison are taken. Readings of both the standard and the thermometer being calibrated must be taken as nearly as possible at the same time, and the thermometers should be lifted from their cups when necessary just enough to bring the mercury into view. Observations should be taken as quickly as possible to avoid errors due to cooling and should be made with approximately the same increments. 1 A record of the observations should be made in the form given below. CALIBRATION OF THERMOMETER By comparison with " STANDARD " Date Observers Standard: Thermometer Tested: No. or distinguishing mark No. or distinguishing mark. Range of scale Range of scale. . . Standard Thermometer. True Temper- ature. o p Thermometer Tested. No. Reading, o p _ Known Error. op Reading. o p_ Error. (+ or -) o p Remarks. 1 If oil is used in thermometer cups, precautions should be taken that the oil is absolutely free from the presence of water and that there is no water in the cups. If water is allowed to accumulate there MEASUREMENT OF TEMPERATURE 29 Calibration of Thermometers by Comparison with Tempera- tures Corresponding to Steam Pressures. There is a definite temperature of saturated steam corresponding to every pressure. If, then, the pressure is known, the temperature corresponding can be obtained from " Tables of the Properties of Saturated Steam." * Thermometers to be calibrated are placed in adjacent cups or wells of the same depth inserted in a cylindrical steam drum and filled with cylinder oil or mercury. Pipe connections must be provided for the attachment of an accurately calibrated steam gage and valves are needed at the ends of the drum for regulating the flow of steam through it. Except for the ad- dition of the gage the apparatus is the same as that used for the calibration of thermometers at high temperatures explained in the preceding paragraphs, and the manipulation of the ap- paratus as well as the precautions to be observed are also the same except that the pressure registered by the gage is recorded instead of the temperature indicated by the " standard " ther- mometer. If there is any possibility that the steam supplied to the drum is superheated, then it is necessary to provide a water- jacket around the steam pipe large enough to make the steam at least dry saturated or preferably slightly wet. Another device often used to change superheated steam to the saturated condition is illustrated in Fig. 34. In this apparatus the steam passes down through the vertical supply pipe S closed at the lower end and escapes from perforations near the bottom to bubble up through the water contained in the chamber A and is carried away in the pipe D. In this way the steam can be made to lose enough heat to the water to reduce the superheat. A gage glass G at the side of the drum is serviceable for showing the level of the water. may be a slight explosion, sometimes strong enough, however, to throw the thermometer out of the cup. Thermometers used for measuring high temperatures should be protected from contact with the metal of the cups by wrapping the stems at the mouth of the cup with cotton waste, or, preferably, with rings of cork. 1 Marks and Davis' Steam Tables and Diagrams or Peabody's Steam and Entropy Tables (revised 1909) are recommended. 30 POWER PLANT TESTING A most excellent method for making calibrations of ther- mometers by means of a steam drum is a combination of the last two methods described. That is, the corrections are calcu- lated from the temperature corresponding to the corrected gage pressure; but at the same H — |j~"X time are checked by comparison 1} > V with a •' ' standard ' ' thermometer. In all experimental work in engineering such methods of checking results cannot be too highly commended. Checking not only assists in the elimina- tion of errors of observation as well as in calculations, but for the engineer it is his key to success. 1 Observations should be tabu- lated in the form given on page 31. The column under the head- ing " Standard Thermometer" is to be left blank, of course, when only the steam gage is used for the calibration. To obtain the absolute pressure the barometric pressure must be observed during the cali- bration. Curves. 1. Plot a curve for each thermometer showing observed temperatures of the thermometer tested (abscissas) and the corresponding "true" or "standard" temperatures (ordinates). This curve is of no value, however, unless a reasonably large scale is used. 2. Plot an error curve, taking observed temperatures of thermometer tested for abscissas and " errors " for ordinates. Compare with Fig. 27. Fig. 34. — Apparatus to Reduce the Superheat in Steam. 1 The plotting of curves is still another means which the engineer uses continually for checking his calibrations, his tests, and his conclu- sions. It has been well said that " the physician buries his mistakes but that the mistakes of the engineer bury him." MEASUREMENT OF TEMPERATURE 31 CALIBRATION OF THERMOMETER BY COMPARISON WITH TEMPERATURES CORRESPONDING TO STEAM PRESSURES Record. : i. Date and names of observers 2. No. and type of gage 3. No. of standard thermometer . .' ; 4. Identification of thermometer tested 5. Limits of graduation of both '. 6. Barometer reading ins. mercury 1 2 3 4 S 6 7 8 9 10 _£ a Observed Temperatures °F. Pressures Lbs. per Sq.In. | •V C 00 a a* a ^ 2 H H ■d 1; a - g :°t •a i ^ B * ir! f H - ':- ••.". .'■' ': >:-< Fig. 49. — Seger Cones after Use. has had some experience with them, temperatures can be estimated to about the nearest ten degrees in Centigrade. Four of these cones are shown in Fig. 49. When a series of cones is placed in a furnace the one having the lowest melting point begins to turn over first. The tem- perature corresponding to the cone number is reached when the tip of the cone has bent over and just touches the surface on which it is standing. Hence the highest temperature reached when the cones shown in the illustration were used was about half way between that corresponding to each of the two middle cones. According to the numbers on the cones the temperature, as given by the table on page 45, was between 830 and 860 de- grees Centigrade. The greatest disadvantage with this system is that there is no way of observing a decrease in the temperature, or, in other words, only the maximum temperature is recorded. 44 POWER PLANT TESTING Two types of mercury thermometers protected by heavy metal cases are illustrated by Figs. 50 and 51. It will be observed that a very satisfactory thermometer cup is a part of the casing. The one shown in Fig. 51 has graduations for reading both temperatures and pressures. A thermometer of this type is particularly useful in pipes carrying hot boiler feed-water. When the temperature is above 212 degrees #51* Fig. 50. — Combined Thermometer Fig. 51. — Combined Thermometer Cap and Protective Casing. and Pressure Gage for Boiler Feed-water Pipes. Fahrenheit the thermometer will indicate that the water is being heated at a pressure higher than atmospheric. For water heated in closed vessels or pipes there is for every temperature a corresponding pressure as given in tables of the properties of saturated steam. 1 1 Short and very much abbreviated tables of the properties of saturated steam are given in the Appendix. See also references in foot- note on page 29. MEASUREMENT OF TEMPERATURE 45 The following table gives the temperatures, in degrees Centi- grade, at which the Seger cones will begin to melt : Seger Temp. Seger Temp. Seger Temp. Cone No. Deg. C. Cone No. Deg. C. Cone No. Deg. C. 022 59° 04 1070 I 5 I430 021 620 03 1090 16 145° 020 650 02 1 no J 7 1470 019 680 01 1130 18 1490 018 710 r 1150 19 1510 017 740 2 1170 20 153° 016 770 3 1190 OI5 800 4 1210 26 1650 014 830 5 1230 27 1670 013 860 6 1250 28 1690 012 890 7 1270 29 1710 on 920 8 1290 3° !73° OIO 95° 9 1310 3 1 i75o 09 970 10 1 33° 3 2 1770 08 990 11 !35° 33 1790 07 IOIO 12 1 37° 34 1810 06 1030 13 1390 35 1830 °5 1050 14 1410 36 1850 CHAPTER III DETERMINATION OF THE MOISTURE IN STEAM Unless the steam used in the power plant is superheated it is said to, be either dry or wet, depending on whether or not it contains water in suspension. The general types of steam calorimeters, used to determine the amount of moisture in the steam, may be classified under three heads: i . Throttling or superheating calorimeters. 2. Separating calorimeters. 3. Condensing calorimeters. Throttling or Superheating Calorimeters. The type of steam calorimeter used most in engineering practice operates by passing a sample of the steam through a very small orifice, in which it is superheated by throttling. A very satisfactory calorimeter of this kind can be made of pipe fittings as illus- trated in Fig. 52. It consists of an orifice O, discharging into a chamber C, into which a thermometer, T, is inserted, and a mercury manometer is usually attached to the cock V3, for observing the pressure in the calorimeter. It is most important that all parts of calorimeters of this type, as well as the connections leading to the main steam pipe, should be very thoroughly lagged by a covering of good insulat- ing material. One of the best materials for this use is hair felt, and it is particularly well suited for covering the more or less temporary pipe fittings, valves, and nipples through which steam is brought to the calorimeter. Very many throttling calorimeters have been declared useless by engineers and put into the scrap heap merely because the small pipes leading to the calorimeters were not properly lagged, so that there was too much radiation, producing, of course, condensation, so that the calorimeter did not get a true sample. It is obvious that if the entering steam contains too much moisture the drying action due to the throttling in the orifice may not be sufficient 46 DETERMINATION OF THE MOISTURE IN STEAM 47 to superheat. It may be stated in general that unless there is about 5 to 10 degrees Fahrenheit of superheat in the calorim- eter, or in other words unless the temperature on the low pressure side of the orifice is at least about 5 to 10 degrees Fahrenheit higher than that corresponding to the pressure in Fig. 52. — A Simple Throttling Steam Calorimeter. the calorimeter, there may be some doubt as to the accuracy of results. 1 The working limits of throttling calorimeters vary 1 The same general statement may be made as regards determinations of superheat in engine and turbine tests. Experience has shown that tests made with from o to 10 degrees Fahrenheit superheat are not reliable, and that the steam consumption in many cases is not con- 48 POWER PLANT TESTING with the initial pressure of the steam. For 35 pounds per square inch absolute pressure the calorimeter ceases to super- heat when the percentage of moisture exceeds about 2 per cent; for 150 pounds absolute pressure, when the moisture exceeds about 5 per cent; and for 250 pounds absolute pres- sure, when it is in excess of about 7 per cent. For any given pressure the exact limit varies slightly, however, with the pressure in the calorimeter. In connection with a report on the standardizing of engine tests, the American Society of Mechanical Engineers 1 published the following instructions regarding the method to be used for obtaining a fair samole of steam from the main pipes. It is recommended in this report that the calorimeter shall be connected with as short intermediate piping as possible with a so-called calorimeter nipple made of ^-inch pipe and long enough to extend into the steam pipe to within \ inch of the opposite wall. The end of this nipple is to be plugged so that the steam must enter through not less than twenty |-inch holes drilled around and along its length. None of these holes shall be less than \ inch from the inner side of the steam pipe. The sample of steam should always be taken from a vertical pipe as near as possible to the engine, turbine, or boiler being tested. Good examples of calorimeter nipples are illustrated in Figs. 54 and 59. Never -close and usually do not attempt to adjust the dis- charge valve V 2 without first closing the gage cock, V 3 . Unless this precaution is taken the pressure may be suddenly increased in the chamber C, so that if a manometer is used the mercury will be blown out of it, and if, on the other hand, a low- pressure steam gage is; used it may be ruined by exposing it to a pressure much beyond its scale. Usually it is a safe rule to begin to take observations of temperature in calorimeters after the thermometer has indi- cated a maximum value and has again receded slightly from it. sistent when compared with results obtained with wet or more highly superheated steam. The errors mentioned, when they occur, are prob- ably due to the fact that in steam, indicating less than 10 degrees Fahrenheit superheat, water in the liquid state may be taken up in "slugs" and carried along without ' being entirely evaporated. 1 Proceedings American Society of Mechanical Engineers, vol. 21. DETERMINATION OF THE MOISTURE IN STEAM 49 The quality or relative dryness of wet steam is easily cal- culated by the following method. Using the symbols, pi = steam pressure in main, lbs. per sq.in. abs. p 2 = steam pressure in calorimeter, lbs. per sq. in. abs. t c = temperature in calorimeter deg. Fahr. ri and qj =heat of vaporization, and heat of liquid corresponding to pressure pi, B.T.U. H 2 and t 2 = total heat (B.T.U.) and temperature (degs. Fahr.) corresponding to pressure p 2 . C p = specific heat of superheated steam. Assume 0.5 for low pressures existing in calorimeters. 1 X!=: initial quality of steam, per cent. 1 — Xi = initial moisture in steam, per cent. Total heat in a pound of wet steam flowing into the orifice is xi^+qi, and after expansion assuming all the moisture is evaporated, the total heat of the same weight of steam is, H 2 +c p (t c -t 2 ). Then assuming no heat losses and putting for c p its value 0.5 we have, Xiri+q 1 = H 2 +o.5(t c -t 2 ), .... (3) Xi = H 2+ o. 5 (t f -t 2 ) -q, 1*1 Chart for Moisture Determinations. Using degrees of superheat in the calorimeter as abscissas and initial absolute steam pressures as ordinates, the diagram in Fig. 53 has been constructed. In the calculations it was assumed that the pressure in the calorimeter was atmospheric. In cases where this condition exists, therefore, after determining the degrees of superheat in the calorimeter (t c — 212) and the initial absolute pressure pi, the percentage moisture can be read from the curves without further calculations. 1 Average values for the specific heat of superheated steam for any i/iiiiperatures are given on page 50 POWER PLANT TESTING 170 160 150 nfflltf "WopHiHiiiflti 1 1 111 1 II 11 1 II II 11 111 1 lilt /] ppo?4 | / I / /] 1 / \\\\\ 1 n IflT 140 130 = Steam. Water. Fig. 57. — Carpenter Separating Steam Calorimeter. DETERMINATION OF THE MOISTURE IN STEAM 57 flow of steam through the nozzle N is roughly proportional to the pressure (see page 148), another scale in addition to the one reading pressures is provided at the outer edge of the dial. A petcock C is used for draining the water from the instrument, and by weighing the water collected corresponding to a given difference in the level in the gage G, the scale opposite it can be readily calibrated. Too much reliance should not be placed on the readings for the flow of steam as indicated by the gage, P, unless it is frequently calibrated. Usually it is very little trouble to connect a tube to the nozzle N, and condense the steam discharged in a large pail nearly filled with water. When a test for quality is to be made by this method the pail nearly filled with cold water is carefully weighed and then at the moment when the level of the water in the water gage G has been observed the tube attached to the nozzle, N is immediately placed under the surface of the water in the pail. The test should be stopped before the water gets so hot that some weight is lost by " steaming." The gage P is generally calibrated to read pounds of steam flowing in ten minutes. For the best accuracy it is desirable to use a pail with a tightly fitting cover into which a hole just the size of the tube has been cut Combined Separating and Throttling Calorimeters. Calor- imeters described are effective in removing practically all of the moisture in steam when the pressure is not lower than 25 pounds gage pressure. For lower pressures, particularly around atmospheric, recent experiments show that the efficiency of such carloimeters is in some cases not more than 80 per cent. 1 For this reason in the best current practice for determinations of moisture in low-pressure steam a throttling calorimeter is attached to the discharge of the separating calorimeter. Then if the separating calorimeter has been carefully calibrated for radiation loss and the steam escaping from the separating calorimeter is tested again in a throttling instrument, it is possible to make correct determinations for the percentage of moisture in the steam of almost any degree of wetness. An apparatus of this kind which is reported to have done excellent 1 Proceedings American Society of Mechanical Engineers, Aug., 19 10, page 1 13 2 . The efficiency of the calorimeter is the ratio of the percentage of moisture taken out by the separating calorimeter to the total percent- age of moisture, 58 POWER PLANT TESTING service in tests of Very large low-pressure steam turbines, oper- ating with the exhaust from reciprocating steam engines in New York city, is shown in Fig. 58. The most unique feature of this apparatus is the sampling tube. It was found that for this low- pressure steam the ordinary sampling tube of perforated pipe (see Figs. 54 and 59) did not give a reliable sample. It was also found necessary that the sample should be taken from the main J4" Mercury Column j/ PipeTher Oonuectiou / Fig. 58. — Stott's Combined Separating and Throttling Steam Calorimeter- without changing its direction or velocity until it is actually inside the sampling pipe. If the direction of flow of wet steam is suddenly changed when entering the sampling nozzle, the entrained moisture, because of its greater specific gravity on the one hand and the very slight skin friction between it and the surrounding dry steam on the other, will cause it to continue in its path in a straight line, so that there is a tendency for only dry steam to enter the nozzle, Also if the velocity of the steam in DETERMINATION OF THE MOISTURE IN STEAM 59 the sampler is greater than that in the main, there is a tendency for the dry steam to " accelerate " into the nozzle, leaving the moisture behind. It has been stated that this action has not been observed in tests of steam at high pressures, because (i) of smaller differences between the specific gravity of high pressure steam and water; (2) greater skin friction; (3) the highly divided state of the moisture. 1 As the throttling calorimeter is ordinarily used it would have very little, capacity when used with steam pressures only a little above atmospheric; but by making it discharge into a receiver in which a vacuum of 28 inches was maintained the throttling portion of the calorimeter will evaporate 2 to 3 per cent of moisture. The apparatus shown in Fig. 58 consists of the f-inch brass nozzle on the sampling tube which is bent to point in the direc- tion opposite to that of the flow of the steam. The lip of this nozzle is filed to a knife-edge to avoid disturbing the current of steam around the mouth of the sampler by eddies and impact against a thick lip. This sampling tube is set up so that it ex- tends into the main steam pipe \ of the diameter of the pipe, where it has been observed to give practically the true average flow. When in operation the valve at the sampling tube is opened wide and the flow is regulated by means of the lever cock between the separating and the throttling calorimeters. The necessary throttling action ordinarily produced by an orifice is produced by this cock. A vacuum is maintained in the throttling portion from which the discharge is carried to a small cooling receiver in which the steam is condensed. From this receiver it flows to a " volumetric " measuring tank of which the top is tightly closed and connected by {-inch pipes to the main condenser. A spy-glass shown at the left in the figure is useful for proving that the calorimeter is working properly. It often happens that when the superheat in the calorimeter is less than 5 t'o 8 degrees Fahrenheit there is some moisture passing through and the spy-glass will invariably show it. As the spy-glass is most conveniently made of f-inch gage glass its area is not large enough to carry all the steam, and a by-pass connection is arranged as shown. The large size 1 H. G. Stott, Proceedings American Society of Mechanical Engineers, Aug., 1910, 60 POWER PLANT TESTING of the parts is necessary on account of the very large specific volume of the low-pressure steam. All parts of the apparatus are carefully covered with mag- nesia-asbestos covering 2 inches thick. For the normal rate of flow for the instrument, the radiation can be made less than 0.1 per cent. Calculation of Percentage Moisture for Combination Sepa- rating and Throttling Calorimeter. Quality of steam Xi is calculated for a combination calorimeter as follows: Let Wi= weight of moisture collected in the separating calorimeter in a given time, in pounds. w 2 = weight of dry steam condensed after passing through the throttling calorimeter, in pounds. x 2 = quality of steam discharged from separating por- tion as determined by the throttling calorimeter, then without sensible error the percentage of moisture in the steam is, 1-.,-- ?- + *, ' Wl+W 2 and in terms of "quality," we have also approximately, W 2 1 / n Xi = ■ hxs. 1 (4) Wi+w 2 Still another type of combined calorimeter is illustrated in Fig. 59. In this instrument, known as Ellison's Combined Throttling and Separating Calorimeter, the sample of steam is collected by the perforated tube in the main steam pipe. The temperature before expansion in the throttling plug is indi- cated by the thermometer marked T 1; and another thermom- eter T 2 gives the temperature after throttling. A scale S opposite the glass water gage G is used to show the weight of water separated from the steam. The proportion of moisture 1 If the radiation test shows it is large enough to be appreciable, then if R is weight of condensation due to radiation in pounds in a given time corresponding to that for the other units, then W 2+ R , . s DETERMINATION OF THE MOISTURE IN STEAM 01 separated in relation to the weight of steam passing through the instrument is the percentage of moisture separated. This percent- age is to be added to the percentage of moisture de- termined by throttling as calculated from the read- ings of the thermometers. Electric Steam Calorim- eters. For use with partic- ularly very wet steam, the Thomas electric calorim- eter, Fig. 60, has been designed. It consists es- sentially of a cylindrical vessel B containing a series of resistance coils for heat- ing steam by means of the iT Fig. Perforated Casing Filled with C Gauze. Fig. 60. — Thomas' Electrical Steam Calorimeter. 59. — Ellison's Improved Steam Calorimeter. electric current passing through them.. These coils, are connected to the elec- tric terminals or binding- posts shown in the figure, and are supported in a soapstone cylinder in which there are a large number of £-inch holes through which the coils pass. Steam enters at the bottom of the vessel at A and passing upward through the heated coils the moisture contained in it is evaporated. The steam then passes- up through a perforated casing filled with copper gauze and escapes through the pipe discharg- 62 POWER PLANT TESTING ing at the side at C. A part of this latter pipe is made of a glass tube for observing the condition of the steam. A thermometer is inserted at T for observing the temperature of the steam after this reheating. Although this apparatus is used for steam of high qual- ity as well as low, it has not been generally used to any great extent, probably because throttling calorimeters are preferred because of the greater simplicity and because very often a source of electric current is not conveniently available where tests are to be made. No data are available comparing its efficiency with that cf the combined separating and throttling calorimeters described in the preceding paragraphs, but for accurate tests the latter are generally preferred by engineers. Barrel Calorimeters. There is still another kind of steam calorimeter, known as the barrel type, deserving some attention. It is one of the oldest forms of apparatus for making deter- minations of the quality of steam. In the classification made at the beginning of this chapter it belongs in the group of con- densing calorimeters. Even with expert manipulations, ordi- narily it is much less accurate than any of the calorimeters already described. A typical apparatus of this kind is shown in Fig. 6 1. It consists usually of a weighing barrel B, made of three concentric vessels of galvanized iron with the two annular spaces between the inner and outer vessels filled with pressed sheet cork or hair felt to reduce radiation to a mini- mum. It is usually arranged so that when the inner vessel has been nearly filled with water from the barrel A, a quantity of the steam to be tested can be passed into it. The steam is admitted into the barrel in the most common forms by dis- connecting the water hose R at C and making a temporary connection from the steam pipe out of which a sample is to be taken to a vertical pipe in the barrel of sufficient length to extend nearly to the bottom of the inner vessel. The pipe may be plugged at the lower end and sufficient area for the escape of steam is then secured by drilling into the pipe a number of £-inch holes for some distance from the lower end. This arrangement will make it easier to secure an equal rise in temperature in the different parts of the barrel. A float is usually provided to show the depth of water in the barrel, and a suitable stirring device or agitator consisting of paddles DETERMINATION OF THE MOISTURE IN STEAM 03 attached to a vertical shaft is also needed. This agitator when revolved stirs up the water and brings it to a constant temperature. Briefly the method to be pursued in the operation of the barrel calorimeter may be outlined as follows: First fill the barrel with cold water till the float shows that the water level is within about 6 inches from the top. Then stir well, observe Fig. 6i. — Barrel Steam Calorimeter. the temperature accurately and weigh carefully on a platform scales. The steam pipe should then be connected up to dis- charge into the water after first allowing the steam to blow off into the air, for the purpose not only of removing the condensa- tion in the piping, but also to heat it to as nearly as possible the temperature of the steam. When the temperature has risen to about 1 20 degrees Fahrenheit the steam should be shut off and another weighing made to determine the amount of steam added. While the weighing is being done the water should be 64 POWER PLANT TESTING stirred vigorously and the highest temperature observed. For all the weighings the piping must be in exactly the same position as regards the connection on the barrel and all the pressure in the pipes must be relieved. When the piping between the calorimeter and the steam supply is connected by pipe fitters' unions these should be disconnected to insure the best accuracy. When, however, the connection is made by means of flexible rubber hose the weight can probably be obtained accurately enough without disconnecting the piping if the precaution is taken to relieve the pressure in the piping by opening a petcock located in the steam pipe near the point where it enters the barrel. If, just before making the test for the quality of the steam, the calorimeter is filled with water, heated with steam or other- wise to about 1 50 degrees Fahrenheit and again carefully drained, the barrel will be near its average temperature during the test, and no correction need probably be made for the heat absorbed by the calorimeter. In most cases it is preferable, however, to determine accurately the heat absorbed by the calorimeter and then make the proper corrections ; but unless the work be done very carefully it is valueless. This correction is usually made by calculating the water equivalent or the capacity of the calorimeter to absorb heat measured by the similar capacity of water. This water equivalent is to be added to the weight of water in the calorimeter. Using then the following symbols: w' = weight of water in calorimeter in lbs. w" = weight of water added in lbs. t' — temperature of water in calorimeter, deg. Fahr. t" — temperature of water added, deg, Fahr. t'" = temperature of mixture, deg. Fahr. k = water equivalent in lbs. Then (w'+k)(t'"-t')=w"(t"-t'"), , w"(t"-t''") k= t /„_ t / — -W, • ...» (5) The temperature of the water added should be taken just as it enters the calorimeter and as near to it as possible. The quality of the steam (x ) to be determined is calculated as follows : DETERMINATION OF THE MOISTURE IN STEAM 65 Wi = weight of water in calorimeter in lbs. W2 = weight of steam added, in lbs. k = water equivalent of calorimeter in lbs. t x = initial temperature of water in calorimeter, degs. Fahr. t2= final temperature of water, degs. Fah. Po== pressure of steam, lbs. per sq.in. ro = heat of vaporization of steam (B.T.U.) corresponding to Po- qo and q 2 =^ sensible heat of steam (B.T.U.) corresponding to po and t 2 . Then equating the heat lost by the steam to the heat gained by the water 1 , (x ro+qo-q2)w 2 =(wi+k)(t2-ti). ... (6) (w 1 +k)(t a -t 1 ) + qg-go ( ) w 2 r r The accuracy of this instrument depends principally on the accuracy with which the various temperatures and the weight of the condensed steam are obtained. Usually it is very dif- ficult to obtain accurately the temperature of the mixtures of water and steam. It is not unusual for determinations of moisture with such a calorimeter to vary for the same quality of steam and with expert handling as much as 5 per cent. In the case, therefore, of steam with 10 per cent moisture, the deter- mination of quality might be in error as much as one-half per cent. Calorimeter Calibrations. For a laboratory calibration exercise three calorimeters of different types are connected by means of exactly the same kind of fittings and valves to the same steam main or receiver. A water-jacket or a device like that shown in Fig. 34 should be provided to vary the quality of the steam. Tests should be made simultaneously and for the same length of time on the three instruments. 1 Equation (7) can be made a little simpler for calculations by writing for (q 2 — q ) the difference between the corresponding temperatures Ct 2 — 1 ) without appreciable error. CHAPTER IV MEASUREMENT OF AREAS Planimeters. The most accurate and generally approved method for obtaining the area of irregular figures is by means of integrating instruments called planimeters. Instruments of this kind may differ in many details, yet all of them are based, in theory, on the original Amsler polar planimeter. Polar Planimeters. One of the simplest forms of the polar type of planimeters is shown in Fig. 62. It consists essentially Fig. 62. — Amsler Polar Planimeter. of two arms PO and TO pivoted together at 0. When in use the point P is not to be moved, and is held in place by means of a pin-point upon which a small weight rests. There is a tracing point at T intended to be moved around the border of the area to be measured. Attached to the arm TO is a small graduated wheel W carried on a short axis which must be placed accurately parallel to TO. Any movement of the arm TO except in the direction of its axis will, of course, move the wheel W on the paper or other surface on which it is placed so that the amount of its movement gives a record indicating the area measured. A vernier, V, placed opposite the graduations on the wheel, assists in reading the instrument accurately. The arm TO is usually made of such a length that the movement of the trac- ing point T around an area of one square inch (for English 66 MEASUREMENT OF AREAS 67 units) will move the wheel one-tenth of its circumference. Graduations of the vernier indicate usually one one-thousandth of a revolution of the wheel, or in English units one one- hundredth of a square inch. When the tracing point T is moved around an area in a clockwise direction the wheel will roll in the direction of its graduation, and the area is found by subtracting the final reading from the initial. Amsler planimeters are often constructed with the arm OT adjustable in length, so that it can be set to indicate areas in various units, as, for example, square inches, square feet, square centimeters, etc. The veriner V has ten graduations, and the total length of these ten divisions is one-tenth less than the length of those on the wheel, so that it repre- sents, counted from zero, so many hundredths of an inch. To explain the method of using the vernier, Fig. 63 has been inserted, showing the wheel W and the vernier V, in a drawing of larger scale than in Fig. 62. Readings of the graduations on the wheel W are always taken opposite the zero mark on the vernier, so that the reading indicated in Fig. 63 without the help of the vernier would be little more than 4.7. The graduation on the vernier which is exactly coincident with a graduation on th e roller wheel is the third from zero and indicates three hundredths. The complete reading is therefore 4.73 as determined by the vernier. Theory of Polar Planimeters. As this instrument is con- structed neither of the points T nor W can pass over the arm PO (Fig. 64). If the arms PO and TO are clamped so that the plane of the graduated wheel W intersects the point P ; that is, when the angle TWP is a right angle, and then the arms thus clamped are revolved around this point, the wheel will be continually slipping without any rolling motion in the direction of its axis, and consequently it will not revolve. When, however, the arms are not clamped and if the con- -J xc- I Fig. 63. -Typical Vernier for Planimeter. POWER PLANT TESTING a, P w Fig. 64. — Position of the Arms of a Polar Planimeter to Draw the "Zero" Circle. struction of the instrument will permit the tracing point T to be moved out so far ~~~~---^ that the axis of W will lie in the line PO, then an arc \ described by the movement of T will produce only a rolling motion of the wheel. Obviously with the arms in any position intermediate be- tween that of the clamped right angle and the one with W in line with PO, the wheel will partly slip and partly roll, the amount of slipping and rolling depending on the size of the angle between the arms. It follows, then, that when circumscribing a closed figure, the radial components cause only slipping of the wheel and need not be considered, while the circumferential components produce a result- ant rolling which must be taken into consideration. The path described by the tracing point T when the arms are clamped as indicated in Fig. 64, is called the zero circle for the planimeter. If the tracing point is moved in any path outside the zero circle in a clockwise direction a positive record will be indi- cated on the graduated wheel, while if it is moved in a path in the same direc- tion as before but inside the zero circle, there will be a negative record. According to the theory of polar planimeters, they are Fig. 65. — Theoretical Diagram for a Polar Planimeter. MEASUREMENT OF AREAS 69 designed so that the rolling of the wheel for a given circumferen- tial motion of the tracing point T is proportional to the area included between the path of T, the radial line from P (Fig. 65) to the initial and final points of the path taken by T, and the arc of the zero circle included between these radial lines. In the discussion of this theory, circumferential motion of the tracing point T around the point P, and the angle WOP (marked a) remaining always at a constant value, is to be taken up first. 1 Now let us suppose the tracing-point is moved from T to T' in the figure through a very small angle, TPT' (marked e), keeping, however, the angle a constant, then the graduated wheel W will move through the arc WW, partly rolling and partly slipping. The component of this motion producing rolling will be perpendicular to the axis of the wheel; or, in other words, this component will be perpendicular to OT in all its positions, and without appreciable error for small values it may be represented in this figure by the line WX, making WXW' a right-angled triangle of infinitesimal proportions. When the tracing point has moved from T to T' the point has moved through the arc 00' and the tracing, point subtends in its movement an angle WPW, which is equal to the angle TPT', marked e, which was passed over by T. Then the follow- ing relation is easily obtained: WW' = PW'Xc. The symbol c is constant, expressing the ratio for a given angle WPW between the length of an arc and the corresponding radius for any value of this radius. In other words in terms of the calculus this constant would be expressed in radians. In general for every angle there is a constant value which when multiplied by the radius gives the length of the arc for that radius. The component of WW' corresponding to the rolling of the wheel is WX, which is approximately equal to the arc WW times cos WWX. That is, WX = PW'XcXcos WWX . . . . (0) But if PY is drawn perpendicular to T'W produced PW'cos WWX = WY, (9) 1 In the mathematical discussion following, the graduated wheel will be considered as if it were a part of the arm TO, with its plane exactly at right angles to the axis of this arm. 70 POWER PLANT TESTING and combining (8) and (9) , W'Y = ^ . . . (10) c Since the angle WPW is very small, WW may be taken as being perpendicular to W'P. Now WX is perpendicular to T'Y and the angle W'WX is equal to the angle PW'Y. The trigonometric relations reducing the above to terms of the length of one of the arms of the planimeter and the constant angle a are as follows : WT = — = PW cos PWY = PO' cos PO'Y- WO' c = PO' cos a -WO', then WX = c(PO'cosa-WO') (11) This is an expression for the amount of rolling of the wheel when the tracing-point moves from T to T'. To express the relations required the area will now be expressed trigonometrically in similar units. From geometry the area of sector TPT' = i/2 arc TT'XPT, but arc TT' = PT X c (see page 69) , or area TPT' = 1/2 c XPT 2 . We can write also, PT = v' PO 2 + OT 3 + 2PO X OT cos a, area TPT' = i/2 c (PQ 2 +OT 2 + 2POXOT cos a). . (12) But the area represented by the amount of rolling of the graduated wheel is that part of the sector outside the zero circle (see page 71), and this is the area TT'Q'Q. Now the radius r of the zero-circle, referring again to Fig. 64*, is easily obtained from equations expressing the relations of the sides of the right triangles in that figure for the particular case when there can be no rolling movement. Thus, P0 2 = W0 2 + PW 2 , (13) PW 2 = PT 2 -WT 2 = PT 2 -W0 2 -2W0X0T-0T 2 . . (14) * It will be remembered that with the arms of the planimeter in the position shown in Fig. 64 the tracing point T describes the circumference of the zero circle. MEASUREMENT OF AREAS 71 Combining equations (13) and (14), PO 2 = WO 2 + PT 2 - WO 2 - 2WO XOT - Of 2 . But PT = r, the radius of zero-circle, therefore, r = \ P0 2 + 2W0X0T+0T^ (15) Also from geometry, as explained on the preceding page, Area QPQ' = i/2 rXcXr = i/2 cXr 2 = 1/2 c(P0 2 + 2WOXOT+OT 2 ). (16) Subtracting equation (16) from equation (12). AreaQTT'Q , = cXOT(POcosa-WO). . . (17) Equation (17), which is the expression for the area outside the zero-circle, will be observed to be equivalent to the roll of the graduated wheel as given in equation (11), times the length of the arm OT from the pivot to the tracing-point. If, therefore, for a given area A, we call the reading of the wheel R and the length of the arm from pivot to tracing-point L, then, A = LR (18) It should be noted further that this equation is independent of any other dimensions of the instrument. That this demonstration applies to areas not adjacent to the zero-circle or partly inside and out can be readily shown by subtracting in a given case the area between the zero-circle and the required area. Area of Zero Circle by Experiment. The area of the zero- circle of a planimeter may be found readily by passing the tracing-point around the circumference of two circles each larger than the zero-circle. Preferably for this operation the fixed point of the instrument is placed at the center of the circles. If the calculated areas of these circles are respectively Ax and A 2 , and r is the radius of the zero-circle, then since readings of the graduated wheel show only the areas outside the zero-circle represented by R^ and R 2 , we obtain, Ai=7rr 2 +Ri, A 2 = 7rr 2 +R 2 , 277r 2 = A 1 +A 2 -(R 1 +R 2 ). . . . . . . (19) 72 POWER PLANT TESTING After r has been found 1 it is not difficult to calculate the proper length of thearm OT for any linear units (compare equation 15). In fact very many polar planimeters are constructed with an adjustable arm OT, so that the instrument can be used for any scale or for various units. The exact lengths required for both the English and metric units (inches and centimeters) are usually stamped on the adjustable arm. Mean Ordinate of an Area. 2 If we call m the mean ordinate and 1 the length of a given area A, then A = ml. From equation (18) we have, A = LR, whence ml = LR, m = -R (20) When, therefore, the tracing-point arm is adjustable it may be set as shown in Fig. 66 3 to make it equal to the length of 1 If instead of measuring and calculating the circles both larger than the zero-circle, one of the two is made smaller than the zero-circle, then the reading of the instrument is again the difference between the area of the circle and that of the zero-circle, but the value of this difference is now negative, so that if A t is the area of the circle larger than the zero-circle and A 2 is the area of the one smaller, then using the other symbols as before, A 1 =7rr 2 + R 1 , A 2 = 7rr 2 -R,, 27tr 2 = A t + A, - (R x -R 2 ) . Although this latter method does not fall in with the general demon- stration so Avell, it is, however, usually preferred, as it will give greater accuracy than can be obtained with two circles both larger than the zero-circle, unless one of these is made unusually large. 2 Engineers must calculate mean ordinates most often when deter- mining the mean effective pressure (M.E.P.) of engine indicator dia- grams. 3 To facilitate the adjustment of the arm to the length of the diagram or area measured, sharp points M and N are attached to the back of some planimeters. The point M is often conveniently placed a short distance away from the tracing-point T, and the point N must then be the same distance and in the same direction away from the pivot 0. Then obviously the distance between M and N will be in all cases equal to the length of the adjustable arm. MEASUREMENT OF AREAS 73 the area measured. Then, obviously, the height of the mean ordinate will be equal to the reading of the graduated wheel expressed in the same units. For example, if the subdivi- sions of the wheel are fortieths of an inch, the result will be the mean ordinate also in fortieths. This scale of the wheel is not determined by the diameter of the portion of the wheel which is graduated, but by the diameter of the edge which comes into contact with the surface over which the wheel rolls. If then d is the so-called diameter of "rolling" of the wheel, its circumference is 7rd. Now by dividing the number of divi- sions on the circumference (usually ioo) by nd, the " scale " Fig. 66. — Polar Planimeter with Adjustable Arms for the Rapid Deter- mination of Mean Ordinates. of the wheel is obtained. It may also be found by measuring a rectangular area of the same length as that of the tracer arm and one inch wide, when the reading from the wheel will give the number of divisions per inch. For those instruments of which the radius of the wheel is one centimeter (.795 inch diam- eter) and having 100 divisions, the scale is almost exactly 40 divisions to the inch. Coffin Planimeter and Averaging Instrument. This planim- eter is made commonly in two forms, illustrated by Figs. 67 and 68. As regards details the former is somewhat the simpler and will be explained first. In principle the two are exactly alike. As will be observed in the figures, this instrument has a single arm to which a suitably graduated wheel is attached 74 POWER PLANT TESTING on an axis parallel to the line joining the ends of the arm. One of the ends of this arm is for tracing the outline of the area measured while the other slides up and down in a suitable slot. One of the advantages of this instrument over the polar planim- eter, although it is not so generally adaptable, is that the wheel is made to move over a specially prepared surface, pre- venting unnecessary slipping. On materials having a rough, fibrous, or worst of all, an uneven surface, the movement of the wheel of any planimeter will not be the same as when rolling over a smooth fiat surface. ; i;r. : -i^ J Fig. 67. — Coffin Planimeter. Fig. 68.— Coffin-Ashcroft Averaging Planimeter. The Coffin planimeter may be discussed as a special form of the general polar type in which the pivoting point 0, instead of swinging about the fixed point P (Fig. 62) moves back and forth in a straight line. The angle between the arms PO and OT, as indicated by the dotted lines in Fig. 69, is really invari- able at 90 degrees. Obviously, then, the equation (17) express- ing the area traced by a polar planimeter outside the zero circle becomes, referring to Fig. 65, area = cX0T(-W0), likewise equation (11), expressing the roll of the wheel for the Coffin planimeter, becomes equivalent to, Roll or record of the wheel =c(-W'0')=c(-W0). MEASUREMENT OF AREAS 75 Using, as before, in equation (18), the symbols L and R for, respectively, the length of the arm OT and the reading of the wheel, we have, just as for the polar planimeter, A = LR (21) As an averaging instrument the Coffin planimeter is very much more convenient than the typical forms of polar planimeters. For finding the mean ordinate of an area the use of the polar type of these instruments was explained on page 72. The sliding vertical straight edge shown at the right in Figs. 67 and 68 is p \ Fig. 69. — Theoretical Diagram for a Coffin Planimeter. for the purpose of making the operation of finding the mean ordinate of an area (or the " mean effective pressure " of an engine indicator diagram) as simple as possible. For this operation the straight edges C and K should be adjusted so that when the tracing pin passes over the extreme end of the area to be measured it will just touch both of them. Now if the tracer is started at either end of the area and moved around to the starting point and then moved upward along the vertical straight edge until the reading of the wheel is the same as when starting to trace the area, this last distance traced from the starting point along the vertical straight edge is the mean ordi- nate. To demonstrate this statement the symbols used on page 72 will be continued. Representing the mean ordinate by m, the length of the area A by 1, the reading or rolling of the gradu- 76 POWER PLANT TESTING ated wheel in going around the area by R, and the length of the arm carrying the tracer by L, then as before A = ml. Now, when the tracing-point T moves over a vertical line, the angle DOT, repre- sented by Z in Fig. 70, re- mains constant. If we call the vertical distance moved v, and remember that only the movement of the wheel at right angles to its axis produces rolling, then the reading corresponding to the rolling, R, is, R = v sin z. (22) But for the position shown in Fig. 69 when the tracer Fig. 70.— Theoretical Diagram for T is at the right-hand end Coi'fin Planimeter. of the outline of the area, we have whence sin z = T , vl Substituting this value of R in the general equation (20) for the mean ordinate m of a polar planimeter, then, L vl m= yir v. (23) This relation can be illustrated more simply, however, by referring to Fig. 71, which is typical indicator diagram from a steam engine. In this figure the tracing point of the Coffin instrument is shown at 0, with the tracing arm represented by VO. A rectangle OXYZ, indicated by dotted lines, is shown of which the area is equal to that of the indicator diagram. MEASUREMENT OF AREAS 77 Starting at and moving the tracing point around the indicator diagram once the difference in readings is the area. Now if the tracing point is moved in the opposite direction around the rectangle and again back to the starting point at it will measure a negative area equal to the first area and the reading of the graduated wheel will be the same as when first started around the indicator diagram. The movement of the graduated wheel as the tracing point moves from X to Y is equal and opposite to that in going from Z to 0, so that these two cancel each other. The motion of the tracing point from Y to Z re- quires the axis of the graduated wheel to be parallel to YV and consequently during this move- ment the wheel will not be moved. The only movement that is therefore producing a net change in reading of gradu- ated wheel during the reverse tracing of the rectangle is in going from to X. Conse- quently after going around any irregular area like an indicator diagram in a clockwise direction from the starting point at at the right-hand end of dia- gram, if the tracing point is moved in a vertical direction from the starting point at until the reading of the graduated wheel is the same as when first started, this vertical distance moved, measured from 0, will be equal to the mean height of the indicator diagram. Although measurements of areas may be made with the Coffin planimeter as with the regular polar types with the area in any position as regards its length and breadth, yet when the mean ordinate is to be obtained, its value in a definite position is required and the area must be placed so that its length with respect to which the mean ordinate is to be obtained will lie along the horizontal straight edge shown in the figures. Then Fig. 7 1 . — Diagram Explaining the Method of Mean Ordinates with a Coffin Planimeter. 78 POWER PLANT TESTING the mean ordinate measured along a vertical straight edge will give the result required. Roller Planimeters. For the measuring of very large areas a planimeter differing slightly in theory from the polar type has been designed by G. Coradi, of Zurich, Switzerland. It has the advantage of being adaptable for measuring surfaces of indefinite length and as wide as the length of the tracer arm. This instrument is illustrated in Fig. 72. This instrument is supported at three points — the two rollers R 1 and R 2 and the tracing pin f, or its support s. These two rollers are attached /@}/p=== 1 (f 1 — 1 1 L 1 9 © _ III p' QCz^ 1= t V f B il R 1 lU^=- ' j ,iP '||4b Fig. 72. — A Typical Roller Planimeter. to the shaft A. On the face of one of these rollers is a minutely divided miter-wheel engaging with a small pinion revolving the horizontal shaft carrying the spherical segment K. At the center of the frame B, and in the same vertical plane with the two shafts already mentioned, a vertical shaft carrying the tracer arm is supported. The spherical segment K causes merely by friction contact the movement of the cylindrical " measuring " roller shown at its right. This roller is supported on the auxiliary frame M, of which the tracer arm is a part. The " measuring " roller moves back and forth with respect to the spherical segment to correspond with the movement of the tracing point ; but at the same time MEASUREMENT OF AREAS 79 ■I 1 1 Fig. -Planimeter Testing Rule. the rotation of the segment itself imparts rolling motion of the entire instrument. 1 Testing Planimeters. Tests are made by comparing the readings of the instrument with that calculated for a given area. For such cali- brations it is neces- sary to use an area which can "be gone over accurately with the tracing point pref- erably held mechanically. This is done usually by using a metallic testing rule shown in Fig. 73. It is usually made in the shape of a narrow strip from three to five inches long. At the end marked zero on the graduations a needle point is set which is kept in place by an overlapping screw. At each line of the graduations there is a very small conical hole into which the tracing point of the pla- nimeter can be placed. The beveled end of the testing rule has the index line to be set accurately at the starting mark, so that this point can be very carefully located. "With the tracing-point T in the testing rule and the fixed point P of the planimeter in approximately the position shown in Fig. 74, observe the reading of the instrument corresponding to the area of the circle described by the tracer moving clockwise, in the positions shown. 1 st. When the fixed point P is on the left-hand side of the tracing point. 2d. When P is on the right-hand side. If the reading obtained is greater in the first position than in the second, the end of the shaft carrying the graduated wheel Fig. 74. — Methods of Testing Planimeters. 1 Since this instrument is not often used by engineers, those interested in its theory are referred to Coradi's book of directions (in English) accompanying each instrument, or to Handbuch der Vermes sungskunde, byW. Caville. 80 POWER PLANT TESTING nearest the tracing point must be shifted toward the right to make the instrument accurate, and vice versa. Otherwise the error, if there be one, can be eliminated by taking the mean of the results obtained for the two positions 1 . Another test to be made, if there is doubt about the accuracy of a planimeter after the axis of the wheel has been adjusted, is to determine whether the settings of the adjustable arm marked on the instrument are correct. For this determination circles with several different diameters can be measured with, the testing rule, and if there is a nearly constant percentage error, say x per cent too large, then the adjustable arm must be lengthened x per cent to make the planimeter readings correct, and vice versa. For accurate results the fixed point P should be placed as indicated by the dotted lines in Fig. 74, so that when the tracing point is near the center of the area to be measured the two arms will be approximately at right angles. Durand-Bristol Integrating Instrument. This instrument, illustrated in Fig. 75, has been recently developed by the Bristol Company for obtaining the average radius of records traced on circular charts of uniform graduations like those used in re- cording gages and thermometers, etc. It is a simple device for obtaining quickly the average value of pressure, temperature, draft, watts, volts, amperes and other records generally taken on circular charts. This instrument consists of a wooden base in which there is a metal socket for supporting a rotatable pin slotted for receiving a horizontal shaft to which the integrating wheel is rigidly attached. On this shaft between the integrating wheel and the pin there is an adjustable tracing point and at the opposite end of the shaft there is a triangular support for the shaft, also adjustable. The general principle of this instrument is due to Professor W. F. Durand 2 of Leland Stanford University. Its application 1 For ordinary requirements a testing disk can be used in place of the rule, although it is not usually so accurate. On this disk circles of 1, 2, and 2 J inches diameter are usually engraved; and if neither a testing plate nor a disk is available, rough tests can be made by using circles drawn with a pencil compass on a flat sheet of well-calendered paper. 2 Proceedings American Society qf Mechanical Engineers, October, 1908, pages 1241-1246. MEASUREMENT OF AREAS 81 hinges on the condition that the chart to be measured has a uniform radial scale, the same as there must be a uniform vertical scale for indicator and other similar diagrams in order that they can be averaged with the ordinary planimeters. Ob- viously the mean value of the radius of a circular diagram can- not be determined with ordinary planimeters, since the area of a diagram in polar co-ordinates is proportional to the square Fig. 75. — Bristol-Durand Integrating Instrument for Circular Charts. of the radius and to the angle. 1 In Fig. 76, AB is an irregular curve, considered for this theoretical discussion as traced by a point moving in and out on a straight radial line. The center of the chart is at 0, and at this point there is a socket, in which a rod O'P slides freely back and forth, permitting a tracing point, P, to draw a curve AB. A graduated wheel, W, attached to O'P, serves the same general purpose as the integrating wheel in the ordinary planimeter. Obviously this wheel will be moved only by circumferential motion, and for any radial movement of the rod in the direction of its length it will remain stationary. 1 With the ordinary planimeter the mean square of the radial ordi- nates can-be determined, and we can, of course, take the square root of these values, but in most cases this is not the same as the mean radius. 82 POWER PLANT TESTING The amount of movement will be proportional to the radius WO, which differs from PO by a constant distance PW. Result- ant movement of the wheel W is proportional, therefore, to the angle moved through by the arm O'P and to the radius OW varying from point to point along the curve. Assuming for the present, but as will be shown later the reading for any pare of the curve, as AB, to be proportional to the product of tht angle subtended between the points A and B, AOB, and the mean radius for the curve between these points, then if this reading is divided by the subtended angle expressed Fig. 76. — Diagrammatic Drawing of Bristol-Durand Integrating Instrument. in circular measure the quotient will be proportional to the mean radius. Now if to the value of this mean radius the constant distance WP is added, the true value of the radial ordinate OP is obtained. When, as is usually the case in prac- tice, the curve AB, represents values of radial ordinates with reference to a base circle of constant radius as the datum or " zero line," then if the radius of this base circle is subtracted from OP the remainder will be the true value of the ordi- nate. By making WP equal to the radius of the base circle, as may readily be done by a suitable adjustment of the instru- ment, the two corrections will be "balanced " and the mean value of the radial ordinate will be given directly as the quo- tient of the reading of the wheel and the subtended angle AOB expressed in circular measure. For a chart corresponding to MEASUREMENT OF AREAS 83 twenty-four hours for a circumference, the angular measure to be used as the divisor will be .2618 per hour. The quantity to be determined in such diagrams is the time mean of the quantity measured by the radial ordinate. But since angular motion is made proportional to time, we may represent the desired mean by the following integral formula : fa (24) Now, in Fig. 77, let ABCD denote a curve drawn by a tracing point which moves on the arc of a curve shown by OAV instead of on a straight radial line. Then let OV, ON, OM, etc., denote a series of consecutive positions of the curve OAV, at differential angular inter- vals dd. Then for the actual curved path ABCD substitute the broken line path made up of a series of arcs each rd# in length, and the series of differ- ential bits of the curve OAV as shown. Then at the limit the record of any integrating or averaging instrument will be the same, whether the tracing point is carried along the curve or along the broken line substi- tute as shown. Then suppose an integrating instrument, as shown in Figs. 75 and 76, is applied to such a diagram, and let the tracing point P be carried along the zig-zag path. The record of the wheel will be made up of two parts : 1. That due to the circular arcs rd# and representing by summation the value of ( rdO. Fig. 77. — Theoretical Curves for Bristol-Durand Instrument. /' 2. That due to the differential portions of the arc OAV. Now it is clear that if the diagram extends all the way around from A through BCD to A again the differential elements of the 84 POWER PLANT TESTING curve OAV may be considered as existing in pairs, and that for every element traversed in the outward direction, there will be an equal element traversed in the inward direction. PQ and ST denote the members of such a pair. The record for such a pair will therefore disappear in the summation, and hence for all the pairs, and hence for the diagram as a whole. In such a case, therefore, part " 2 " above becomes zero and the record of the wheel for the entire diagram consists simply of J rd#. This reasoning is seen to be entirely general and independent of the character of the path OAV, and hence must be true whether it be the arc of a circle, a straight line or any other path. In case the curve occupies only part of the revolution, as ABC, then it is clear that in going from A to C the record will involve the two parts, " 1 " and " 2 " above, and that the latter will remain included in the final result and will represent the summation of the record due to the elements of OAV between A and C. This obviously will be the value of / rdf) for the arc GC and it will be canceled by carrying the tracing point of the instrument back from C to G. This method of reasoning is independent of the extent of the arc and is there- fore equally true for an entire revolution, even when the diagram does not end at the same radial distance, as at the beginning. In such cases it is necessary only to trace along the arc OAV so as to " close " the curve, thus canceling part " 2 " above and finding directly the value of I rdO for a whole revolution. In all cases the correction for part " 2 " of the record is made by tracing from the terminal point of the curve along the path, representing no change of time to a point lying in a circum- ference passing through the initial point. This may be stated in other words by saying that to eliminate part " 2 " of the record the tracing point must start and finish at the same distance from the center, and if the diagram is not of the kind to satisfy this condition then the necessary portion of a path of zero change of time must be used to supplement the diagram. This discussion is independent also of the nature of the curve OAV. It may be stated, however, that when OAV becomes a straight line the value of the correction becomes zero. CHAPTER V ENGINE INDICATORS AND REDUCING MOTIONS The engine indicator is simply an instrument showing by graphic diagrams the variations of the pressure in the engine cylinder of steam, gas, air, or whatever the working substance may be. Before James Watt invented the engine indicator (about 1 814) he had already used a steam pressure gage on the cylinder of his engine, and since the movement of the piston in the early steam engines was very slow, he was able to observe with his eyes how the pressure varied during a stroke of the piston. In modern engines the movement of the piston is so rapid, however, that a recording instrument is absolutely necessary. Watt's indicator is illustrated in Figs. 78 and 79. It con- sists of a cylinder CC (Fig. 79) in which the piston P is moved against the resistance of the spring S by steam pressure from the engine cylinder; this pressure being exerted, of course, on the lower side of the piston. A pencil attached to the : upper end of the piston rod traces on a sheet of paper a diagram DD, of which the height on any ordi- nate is' proportional to the pressure. The paper is moved back and forth on a slide by a string E moved in conformity with the piston. The instrument was of great service to Watt in perfecting his steam engines. In the modern indicators, of which a few of the best known makes are to be described, there are many improvements over the instrument used by Watt. Thompson Indicator. Of the engine indicators now in general use the Thompson is the oldest and best known. Fig. 80 shows one view of this instrument and Fig. 81 shows the corresponding sectional drawing. It consists in essential parts of a piston, 8 (Fig. 81) moving in a cylinder 4. This piston is rigidly connected to the rod 12, which passes up through the cap 2. The motion of the 85 86 POWER PLANT TESTING piston rod 12 is transferred to the pencil 23 by means of suitable links designed to make the pencil move parallel to but usually four times as far as the piston 8. The maximum pres- sure of the pencil on the paper used for the diagram is adjusted by the thread and set screw on the handle attached to the bracket X. Fig. 78. — Watt's Original Steam Engine Indicator (Type of 1814). Fig. 79. — Section of Watt's Indicator. The method of changing the springs in the various common forms of engine indicators should be well understood by everyone likely to be called on to " indicate " engines. When the work of changing springs is done clumsily or carelessly, a great deal of time is often wasted by the whole party engaged in the test. The method to be followed in changing springs of a Thompson indicator may be stated briefly as follows: The milled-edged ENGINE INDICATORS AND REDUCING MOTIONS 87 cap 2 should first be unscrewed from the top of the cylinder containing the spring and piston. This cap, together with the piston rod, piston, and link can then be lifted from the main body of the indicator. By unscrewing the small milled-headed screw 19, connecting the piston rod with the pencil arm the spring can then be unscrewed, first from the cap 2 and finally from the piston 8. By exactly reversing the operation another Fig. 80. — Thompson Indicator. spring can be put in the place of the one removed. Changing springs in this instrument is a simple operation. No wrenches or other tools are required. Care should be taken, of course, to screw up the spring firmly against both the cap and the piston. Probably one-half the troubles with indicators in operation arise from loose springs, although not so often probably, with Thompson indicators as with some other types. In selecting a spring for an indicator test it should be of such a scale that the largest diagram to be taken will not be more thai! if inches high ; that is, if the maximum pressure will 88 POWER PLANT TESTING be about 140 pounds, a spring with a scale of 80 pounds per square inch should be selected. The tension of the spring inside the drum carrying the paper for the diagram is varied by loosen- ing the thumb nut and turning the large milled cap till the proper adjustment is secured. 1 Crosby Indicators. For high-speed engines and for accurate results the Crosby indicator has long been a favorite with engineers. This indi- cator is illustrated in Figs. 82 and 83. It consists of a piston 8, moving in the cylinder 4, and is connected by means of the piston rod 10 and the link 14 to the pencil lever 16. All of the pencil mechanism arranged to move the pencil point 23 in a straight line parallel to the motion of the piston 8, is supported by the links 13 and 15 on the sleeve 3. The indicator spring is fastened at its lower end to the piston by a ball-joint and at its upper end it is screwed into the cap 2. The method of attachment of the springs to the piston by means of the ball-joint is shown more in detail; in Fig. 84. In this indicator the spring is changed by first unscrewing the milled cap 2, then this cap, the sleeve 3, the piston rod 10, and the connected parts can be removed from the cylinder 4. 1 Unless there is a very good reason for a change in the tension of the spring in the drum it should not be altered. Particularly in indicators which have been used a long time the pin holding the spring in place is likely to be much worn, so that if adjusted often the spring may get loose, and then there is often considerable difficulty in getting it again into its proper position. Fig. 81. — Section of Thompson Indicator. ENGINE INDICATORS AND REDUCING MOTIONS S9 By unscrewing the spring by hand from the cap, which, of course, must be prevented from turning, and also from the screw on the swivel head 12, the piston, the spring and the hollow piston rod 10 are detached from the other parts. A socket- wrench of the special form provided in every indicator box of this make is to be slipped over the piston rod to engage with the small nut shown in Fig. 83 at the lower end of the piston rod Fig. 82. — Typical Crosby Indicator. 10. Then the piston rod is readily unscrewed from the piston at the same time the spring in the indicator is released from its attachment to the piston. Now with the piston rod still in the socket of the wrench, slip the spring to be used over the piston rod until the head of the spring rests in the concave end of the rod. To do this, the wrench must be held upright, and then if the piston is inverted, or, in other words, if it is held so that the end screwing into the piston rod points downward the piston rod is ready to be screwed into the piston so that the trans- 90 POWER PLANT TESTING verse wire of the spring passing through the bead will be held firmly in the slotted portion of the socket in the piston. Finally screw the piston rod firmly 1 into place. Before the last opera- tion, the lower piston-screw (Fig. 84) should be loosed slightly, and afterward it should be screwed up lightly against the bead to prevent lost motion. It should not be screwed so tightly, Fig. 83. — Section of Crosby Indicator. however, as to prevent the bead from turning, otherwise the desirable qualities of the ball-joint for securing perfect alignment are lost. Now when the piston-rod spring and piston are again assembled, if the sleeve 3 and the pencil motion attached 1 Special care should be taken when putting a spring into a Crosby indicator that the piston rod is screwed into its socket in the piston P (Fig. 84) as far as it will go; that is, until the extreme upper end of the socket a a is brought firmly against the bottom of the corresponding annular channel b b in the piston rod R. ENGINE INDICATORS AND REDUCING MOTIONS 91 Fig. 84. — Section of Crosby Indi- cator Spring and Piston. to it are held in an upright position, the hollow piston rod can be slipped over the threaded portion of the swivel head 11 until the threads on the upper end of the spring engage with those on the cap 2. Then the spring can be screwed securely into the cap 2. Then per- mit the cap to turn in the sleeve 3, and by still turning the spring, screw the piston rod on the swivel head 12, until the top of the rod is nearly flush with the shoulder on the swivel head. The piston and attached spring are now ready to be put into the cylinder by slipping the sleeve 3 into position and screw- ing down firmly the cap 2. 1 The height of the pencil cannot be adjusted to change the position of the atmospheric line without removing the piston from the cylinder of the indicator. It must be done, however, by unscrewing the cap 2 from the cylinder and removing it together with the sleeve 3 and the pencil mechanism. By turning the cap clockwise, the swivel-head 11, and consequently also the atmospheric line is lowered. By turning in the opposite direction both are raised. Never try to make adjustments by • removing or loosening the pins or screws at the joints, 17, 18, 19, 20 and 21. These joints should always be kept tight enough to prevent any lost motion, and occasionally they should be lubricated with refined porpoise oil of the kind usually supplied with indicators. 2 1 Persons in charge of tests should always inspect indicators before the steam pressure is put on the springs to observe whether the cap has been screwed down firmly, and whether the pencil mechanism has been adjusted so as to give with a suitable spring a diagram of the proper height on the drum. 2 Inexperienced testers often put the spring and piston into place by merely slipping on the sleeve 3 and without screwing down the cap 2. Then, as a result, when the steam pressure is put on the indicator the piston, spring and pencil mechanism are thrown off with a great deal of force, and some of these parts are sometimes completely de- molished. When using an indicator having the spring inside the cylinder 4 — and this is true particularly in the Crosby Indicator — all adjustments should be made before the steam is turned on the indicator, because the piston, spring and cap soon become very hot, and unless the parts are cooled, preferably by dipping into cold water, they are difficult to handle. 92 POWER PLANT TESTING The tension of the spring in the drum is changed very much more conveniently than in most other indicators. For high- speed engines the tension must be considerably greater than that required for those running at low speeds. The tension is adjusted by removing the drum (24) by a straight pull, and Fig. 85. — Crosby Outside- spring Indicator. turning the knurled nut at the top of the spring (31) after lifting it from its square seat. Crosby Outside- spring Indicator. Indicators with springs out- side the cylinder (Fig. 85), so that they are not subjected to high temperatures, are particularly desirable for use with engines using superheated steam. There are two principal advantages : (1) The spring can be changed without removing the piston, avoiding an operation often causing confusion and loss of time; (2) the ENGINE INDICATORS AND REDUCING MOTIONS 93 tension of the spring cannot be affected by exposure to very high temperatures. The spring can be changed when the thumb- screw at the top of the central spindle has been unscrewed. Star Brass Indicator — Navy Pattern. The indicator called the " Navy Pattern," manufactured by the Star Brass Co., is shown in Fig. 86. In general principles of construction it is like the Crosby indicator illustrated in Fig. 85. The most essential differ- ence is in the type of straight- line parallel mo- tion for the pencil lever. It will be observed that this is practically the same as that used- in the Thompson indicator (Fig.8o). Tabor Indica- tor. In the form in which it is now manufactured the Tabor indicator, Fig. 87, differs from indicators like the Crosby particularly in the means employed for producing a straight-line parallel motion for the pencil. This is accom- plished by the movement of a roller attached to the pencil lever in curved slots on the inside of the rectangular box-shaped part shown in the figure, attached to the cylinder cap. As regards the point of flexibility in the mechanism, this is not between the spring and the piston, but, more like the Thompson, is between the piston and the piston rod. Details of this construction are shown in Fig. 88. To Change the Spring. The cylinder cap must be first unscrewed, and then this cap, together with the piston, spring, and connected parts can be lifted from the cylinder of the Fig. 86. — Star Brass Indicator — Nav}^ Pattern 94 POWER PLANT TESTING indicator. By removing the small screw under the piston the latter can be unscrewed from the lower end of the spring. The other end of the spring can then be unscrewed from the cylinder cap. Another spring is put into the indicator by slipping it over the piston rod with the end stamped T upper- Fig. 87. — Tabor Indicator. most, screwing this end into the cylinder cap and screwing the piston to the lower end. The pencil mechanism must be moved downward until the piston rod enters the piston and the square shoulder enters the corresponding square socket in the piston. In this last operation care must be taken that ENGINE INDICATORS AND REDUCING MOTIONS 95 the rod is firmly and accurately in the hole, and then the screw at the bottom of the piston should be firmly applied. To Change the Tension of the Spring in the Drum. The drum itself is first removed. Then after loosening the knurled nut on the central shaft and after the drum carriage has been lifted clear of the stops, the carriage can «be turned in the Fig. 88. — Section of a Tabor Indicator. required direction to secure the necessary tension and it can then be replaced by lowering into the stops. Care must be taken also that a firm grasp on the drum carriage is not lost, otherwise the spring will become uncoiled and probably also detached. Bachelder Indicator. Fig. 89 illustrates an engine indicator which is in many essential parts entirely different from the 96 POWER PLANT TESTING general type to which all those already described belong. It is so simple in construction that scarcely any description is necessary. The most radical difference is, however, in. the form of spring used. This is flat and is arranged with a movable fulcrum which can be adjusted to change the scale of the spring. Although a wide range is obtainable in this way, it has been found unsatisfactory to attempt to use the single spring for all the ranges from the highest pressures to low vacuums. On Fig. 89. — Bachelder Indicator. this account at least two springs, one for high and the other for low pressures, are usually supplied. Springs are changed by first removing the taper screw shown at the extreme right-hand side in the figure, and then after unscrewing a circular cap on the side of the cylinder the pin connecting the spring to the piston rod can be withdrawn with a small pliers or similar instrument. Usually before the spring can be withdrawn the thumbscrew attached to the movable fulcrum must be loosened. In the ordinary operation of the instrument the piston is not removed. 1 1 When the spring is calibrated, the piston should be taken out so ENGINE INDICATORS AND REDUCING MOTIONS 97 The spring on the drum is conical in form and is adjusted in practically the same way as in the Crosby Indicator. Precautions for Care of an Indicator. Unless an engine indicator is well taken care of, very soon it will be in a condition in which no reliance can be placed on results obtained with it. That the necessary precautions should be taken is all the more important, because it is one of the most expensive as well as the most delicate instruments used by an engineer in his ordinary practice. The following precautions are particularly important : i. Before an indicator is used all the working parts, espe- cially the piston, should be carefully cleaned. Then after a spring suitable for the pressure has been attached in its proper position and a little cylinder oil has been smeared in a thin coat on the working surface of the piston the parts should be replaced. Moving parts of the pencil mechanism should be oiled occasionally with watchmaker's or porpoise oil. 2. Adjust the screw on the handle provided for moving the pencil so that when the pencil is sharp the application of the usual pressure on the handle will give a very fine line. 3. Adjust the length of the indicator cord so that the drum will be neither too loose nor too tight; or in other words, so that the drum will not strike either of the stops when the engine is operating. On a small engine this is most easily tested by observing the card when the engine is on each of the dead- centers. If the card is either too long or too short the drum will not be moved in either case the required distance, and the indicator card will be correspondingly too short, therefore inaccurate. The cord used should be selected with care. It must be of such a quality as not to be stretched appreciably by the forces to which it is subjected. 4. The atmospheric line should always be taken, preferably after the diagram has been made. It is drawn when the indicator cock is closed. By this order of procedure in tests, the diagram can be more easily taken exactly " on the signal." For calcu- lations the length of the diagram must always be measured on the atmospheric line or on a line parallel to it. The indicator that a little cylinder oil can be put on it. It is not so necessary when in use on a steam engine, as the oil in the steam will usually provide sufficient lubrication. 98 POWER PLANT TESTING cock should be kept closed and the cord to the reducing motion should be unhooked except when a diagram is to be taken. When the cord is unhooked the drum should not be permitted to snap back against the stop. 5. Immediately after a diagram has been taken it should be removed from the drum and examined. If there are unusual irregularities in the lines, unaccountable differences in the areas or in the lengths of different cards, the facts should be noted and the best efforts should be made to remedy the faults. Irregularities are usually due to stretching of the indicator cord, grit on the piston, lost motion in the working parts, usually inside the indicator cylinder, or excessive friction caused by overheating of the piston, particularly when used on gas engines. To correct these faults concerning the piston it must be removed from the cylinder and should then be carefully cleaned and lubricated again with cylinder oil. Before putting the piston and connected parts back into the indi- cator cylinder it should be observed whether or not all the parts are connected firmly and without lost motion. 1 6. After a test, the indicator should be removed immediately from the engine, protecting the hands with waste or thick gloves to prevent burns. All the parts, especially those in the cylinder, should be thoroughly cleaned and then put together again without the spring, which should be put away with the other springs in a box provided for the indicator. An indicator should never be handled by taking hold of the drum, as usually it is fastened to the indicator by only a loose slipjoint, and this comes off easily. 1 One of the causes of errors in results obtained with indicators not so readily detected is due to the pencil motion not being parallel to that of the piston in the indicator. A simple test for this is to draw an atmospheric line on a card placed on the drum. The card should be at least as wide as the height of the drum. Then after taking out the spring raise the pencil to the full height of the card by pressing lightly on the piston. This operation should be repeated several times at several points along the length of the card. To secure the best accuracy it is desirable to " block " the drum in each position. If the lines drawn are exactly perpendicular to the atmospheric lines there is no error in the pencil mechanism. If the test for perpendicularity is made by a triangle and straightedge, it should be done with the triangle first lying on one side and then on the other, to eliminate any inaccuracy in it. Often the triangles used by engineers are very inaccurate. ENGINE INDICATORS AND REDUCING MOTIONS 99 SPECIAL TYPES OF ENGINE INDICATORS Cooley-Hill Continuous Indicator. For many purposes of investigation it is very important to have continuous records showing the variations of the cycles in the operation of an Fig. 90. — Cooley-Hill Continuous Indicator. engine. Many devices have been used for this purpose, but as the motion was taken from the crank shaft there was no simple relation between points on these diagrams and the correspond- ing points in the stroke of the engine. Furthermore, because of the difficult relation, such cards could not be measured with a planimeter. Similar apparatus for the same purpose operated by an electric motor were open to the same objection. To 100 POWER TLANT TESTING overcome these difficulties a continuous indicator was developed in the Engineering Department of the University of Michigan in which the motion was proportional at every instant to the movement of the piston. "With a diagram obtained with this instrument it is not difficult to determine the dead-center follow- ing release, and the conventional indicator card for an engine is then readily obtained by turning the diagram for the complete cycle back on itself by folding the card or ribbon at this dead- center. If transparent paper is used, the complete diagram can be seen with all the points in their true relative positions as regards the movement of the pis- ton. The indicated horse power can then be readily calculated with the aid of a planimeter. This continuous indicator is illus- trated in Fig. 90. The indicator cylinder C, the piston, and the pencil motion may be of any standard make, as the collar M, for attaching the drum mechan- ism, is adjustable in size so that it can be fitted to indicator cylin- ders of different diameters. By this arrangement only one drum motion need be provided for using this indicator motion on a number of types of indicators such as would be required for use with steam engines, gas engines, ammonia compressors, etc. In this apparatus the drum D moves forward a given amount with every stroke of the engine. The indicator cord S is con- nected to the indicator reducing motion, and is driven by being connected in the usual way to the cross-head of the engine. The mechanism operating the drum motion is illustrated in Fig. 91. It consists essentially of two miter wheels B and C, Fig. 91. — Details of Cooley-Hill Indicator. high-pressure air compressors, ENGINE INDICATORS AND REDUCING MOTIONS 101 meshing with a similar wheel E, to which the pulley W, carrying the indicator cord, is attached. At the top of the wheel B and at the bottom of C are so-called silent ratchet clutches, a, a, each of which operates in only one direction to grip the collars concentric with the wheels B and C. Only one of these collars is shown in the figure. Both are rigidly attached to the central spindle J, carrying the indicator drum D (Fig. 90) . For example this central spindle is gripped by the ratchets a, a in the wheel B, during the "forward" stroke of the engine, and is released during the " back " stroke. The ratchet in the wheel C, on the other hand, grips this spindle during the " back " stroke and releases on the " forward " stroke. In this way the drum D is constantly moved on the spindle J in the same direction. Neither of the wheels B nor C is directly attached to the central spindle, and they can move it only when they move in the direc- tion in which they grip their ratchets a, a, engaging in the grooves g, g. The miter wheels B and C are connected to each other by means of a spiral spring enclosed in the casing D. This serves the function of the ordinary drum spring in the usual type of indicator- for bringing the drum and string back when the cross-head moves toward the indicator. Optical Indicators. The usual types of indicators operating with a piston are not suitable for engines running at much over 400 revolutions per minute. For higher speeds optical indi- cators are used. These operate by the deflection of a beam of light from a mirror, the deflection being proportional at any instant to the pressure. When such a device is used on an engine successive indicator diagrams can be readily observed and compared by marking with a pencil the reflection upon a ground-glass plate, and if a photographic sensitive plate is exposed to the beam of light in the place of the ground glass, a permanent impression can be taken, showing at any instant the operation of the engine. Optical indicators are practically the only kind that can be used successfully for indicating the action of modern high-speed automobile engines. Every well-equipped automobile testing plant should be provided with one of these instruments. One of the simplest and best apparatus of this kind is illustrated in Fig. 92. The indicator is shown in the picture vertically above and connected to the head of the engine. 102 POWER PLANT TESTING Steam pressure is communicated to the instrument through the usual type of indicator cock supporting it. A system of levers shown (a simple reducing motion) serves for reducing the length of the stroke of the engine to a suitable size for such a small instrument. A glass mirror moved about a vertical axis by the motion transmitted from the cross-head and about a horizontal axis by the pressure in the engine cylinder, reflects a " ^^\ Fig. 92. — Perry's Optical Indicator. beam of light from a lamp upon a sheet of paper so that the indicator diagram can be traced. Details of the essential parts of this instrument are shown in Fig. 93. Through the indicator cock the pressure in the engine cylinder is communicated to the cored passages marked A, A. This pressure tilts the mirror B, attached to the thin steel diaphragm D. When, therefore, the mirror is still, a ray of reflected light will be seen as a bright spot on the screen ; but when moved both by the pressure and the motion of the cross- head the conventional indicator diagram is traced. It is very ENGINE INDICATORS AND REDUCING MOTIONS 103 interesting to watch the rapid change of shape of such diagrams as load, speed, pressure, cut-off, etc., are changed. With such an instrument these interesting phenome- na in engine operation can be illustrated on a ceiling to a large class of students. Another type of optical indicator in- tended particularly for high-speed auto- mobile engines is shown in Fig. 94. In this instrument the movement of the beam of light is produced by reflection from a small mirror M arranged to move in two distinct planes at right angles to each other. In one plane Ground Glass Plate Fig. 93. -Essential Parts of Perry's Optical Indicator. (o)l )|j — Acetylene Burner Fig. 94. — Section of a "Monograph" Optical Indicator. the movement of the piston is accurately reproduced, and in the other the movement is proportional to the pressure. Either 104 POWER PLANT TESTING of these movements or deflections of the mirror, taken alone, would cause the reflected beam of light to trace on the ground- glass plate a straight line; that due to the pressure being arranged to produce a straight vertical line and that due to the motion of the piston a straight horizontal line. But obviously the two movements taken together trace a diagram indicating at any instant the pressure in the engine cylinder for the corresponding position of the piston. A flexible shaft, attached at one end to the crank shaft of the engine, moves the disk A and with it the crank C, as well as the small lever L attached to it. The free end of this lever is arranged to turn the mirror M about a vertical axis by means of the small strut a, while the pressure exerted on the diaphragm D, as transmitted from the engine cylinder by the pipe P, moves the mirror about a horizontal axis by means of the strut b. In this apparatus the diaphragm takes the place of the piston and spring in the ordinary type of indicator. These diaphragms, like those used in pressure gages (see page 8) can be made of such thicknesses that a diagram of satisfactory size can be obtained for high or low pressures. When the diaphragms are carefully calibrated, a reasonable degree of accuracy can be expected. The relative motions of the Acetylene mirror {n tfae two planes are set in phase by adjusting the milled screw S, operating a small worm wheel ser- ving for chang- ing the angular position of the crank disk A, movement of the mirror about the vertical axis due to the pressure. Fig. 95 shows the A Fig. 95- Top of Tripod Stand Monograph " Optical Indicator Ready for Attachment to Engine. to make the correspond with that apparatus as it would be set up for indicating an engine sample indicator card taken with this apparatus from a gasoline automobile engine is shown in Fig. 96. Calibration of Indicator Springs. The pistons of engine indicators are invariably made of a very definite area, usually ENGINE INDICATORS AND REDUCING MOTIONS 105 one-half square inch ; and it is possible to calibrate the deflection of the springs with respect to this area, so that a certain definite pressure per square inch * in the cylinder will correspond to a definite deformation of the springs. In English units the pressure on the piston in pounds per square inch corresponding to a move- ment of the indicator pencil on the diagram of one inch is called the scale 2 of the spring. Indicator springs should always be calibrated by the makers. The calibration should be made when they are in the indicator in which they are to be used. Fig. 96. — Indicator Card taken from an Automobile Engine with Optical Indicator. Cooley Apparatus for the Calibration of Indicator Springs. An apparatus similar to the one designed by Professor M. E. Cooley of the University of Michigan is most generally used for the calibration of indicator springs. One of the latest and more elaborate forms of this instrument is shown in Fig. 97. In its essential parts this apparatus consists of a small cylinder C, sup- ported on a bracket B, a connection I, at the top of this cylinder for the attachment of the indicator to be tested, and a stuffing- box or gland at the bottom of the cylinder into which a plunger- piston P is fitted. The lower end of this plunger rests on a sensitive platform scales. Any pressure in the cylinder C can therefore be weighed. Steam is admitted to the cylinder through 1 Indicators are always designed to relieve the pressure above the piston due to leakage around it, so that on this side there is always atmos- pheric pressure. 2 Instead of " scale " the word " number is often used. That is, a spring of which the scale is 40 pounds would be called " No. 40." 106 POWER PLANT TESTING a pipe E, and is exhausted through the pipe A. By adjusting the globe valves on the pipes A and E any pressure desired can be secured in the cylinder C, and this same pressure is, of course, exerted both on the piston of the indicator above and on the plunger P below. This plunger is usually made with an area of one-half square inch. For a plunger of this area, then, if for a given pressure the scales balance at 10 pounds, the pressure in the cylinder C, and on the piston of the indicator, is 20 pounds Fig. 97. — Apparatus for Calibrating Indicator Springs. per square inch. To eliminate friction as much as possible, the plunger P should be kept spinning when observations are being taken. For this purpose a hand wheel K with considerable mass, for its " fly-wheel " effect, is provided on the shaft of the plunger. A more uniform motion of the plunger is obtained, however, by having the hand wheel grooved to take a small belt to be driven by an electric motor M. The plunger is sup- ported usually on a ball-bearing joint set in a low pedestal L. By connecting a pipe E to a suitable manifold or similar fitting, to which are attached three separate pipes supplying respectively steam, air, and water under pressure, an indicator ENGINE INDICATORS AND REDUCING MOTIONS 107 can be tested with varying pressures under the actual conditions in service; that is, when used for steam, air or water. A simpler form of the Cooley apparatus intended for the so-called "dry method" of testing is shown in Fig. 99. A Fig. 99. — Apparatus for "Dry" Method of Indicator Testing. suitable fitting for receiving the ind cator I is supported on the bracket B. The legs of this bracket span over a sensitive plat- form scales S. A small rod R rests at its lower end on a small pedestal standing on the platform of the scales S. On the top of this rod there is a cap supported on a small conical bearing to 108 POWER PLANT TESTING give some flexibility. This cap is made to fit easily into the lower side of the piston in the indicator. The indicator itself is attached to the top of the hand wheel W. Then when the hand wheel is screwed downward the indicator comes down with it and compresses the indicator spring. At the same time a pres- sure is exerted on the rod R which can be balanced on the scale beam. When a force is applied to compress the spring in the indicator, the magnitude of the force can be determined by weighing the pressure on the scales. If the area of the piston in the indicator is one-half square inch, then twice the weight on the scales is the pressure exerted in pounds per square inch. Heat can be applied to the indicator by passing steam through a rubber tube wrapped around the cylinder. Method for Calibration of Springs. After cleaning the internal parts of the indicator, inserting the spring to be cali- brated, and oiling the piston with cylinder oil, the indicator is to be attached to the indicator cock on the calibrating appara- tus. Before putting the card on the indicator drum on which the record is to be made, two approximately parallel and vertical lines should be drawn on it about one-half inch apart, similar to the lines AB and CD in Fig. ioo. Meanwhile the indicator should be thoroughly warmed if a calibration with steam pressure is to be made. Then with the indicator cock and the valve on the steam pipe E closed and the one on the exhaust pipe A open, draw the first calibration line on the card. This should be made by setting the pencil point at D, and then by pulling the cord attached to the drum draw a line crossing the vertical line AB. With springs of which the scale is 30 pounds or less, a similar record should be made for incre- ments of every 5 pounds per square inch change in pressure, while for higher scales the increments may be made 10 pounds. If with increasing pressures the lines are drawn toward the left, then with decreasing pressures they should be drawn toward the right with equal increments, beginning at the opposite vertical line AB. By this method the corresponding lines for equal pressures will be immediately over each other between the two verticals. With an accurate scale, graduated preferably to one one-hundredths inch, measure between AB and CD the distance from the atmospheric line first drawn to the various ''pressure lines, "and record the results. Care should be taken ENGINE INDICATORS AND REDUCING MOTIONS 109 that with the increasing increments the pencil rises to the required pressure and that with decreasing increments it falls to these pressures. In other words, if when the lines for increasing pressures are being drawn, the pressure rises too rapidly to draw the line at the proper time when the scale beam is just balancing, then the pressure should be again reduced below the value required, so that the pencil will be again ascending when the line is drawn. Similarly for decreasing pressures, if the pressure gets too low, it must be increased and again brought down to the required value. 1 If the results No_£ Hour. / '. T #-?IVL Which End B. Press Vac. gauge.. ~ Revs. Spring Up INDICATOR NO. c Area Length- M. Orri\. M.E.P.. i.h.p_. —"Atmospheric" Lines D OBSERVER Fig. ioo. — Sample Card Illustrating a Test of an Indicator Spring. obtained do not seem to be consistent, the difficulty is probably due to passing the required pressure so rapidly that the lines have not all been drawn at the proper time. The difference between the lines for increasing and decreasing pressures shows the amount of friction and lost motion in the indicator. 2 The error of the instrument is obtained by comparing the mean ordinates of the card thus obtained with the actual 1 The same precaution must be observed in beginning the test. To be sure that the pencil and piston have not been falling instead of rising, the piston rod should be pushed down lightly before the atmospheric line is drawn. 2 Half of this difference, to be more accurate, represents the friction and lost motion in that position. 110 POWER PLANT TESTING pressures as determined by weighing. From time to time the accuracy of the platform scales should be determined by test- ing with standard weights. For dependable results two cali- brations, each "up and down,'^ should be made for each spring and the results compared. When indicators are used for pressures which are never less than atmospheric, the springs are in compression and appa- ratus of the form described are satis- factory ; but when indicators are used on the low-pressure cylinders of engines the springs are usually in tension. For this service a slightly different de- vice must be used for calibration. A suitable apparatus is shown in Fig. 101. The indicator I is supported on a bracket similar to the one used in the apparatus shown in Fig. 99. A short steel rod (about No. 18, B. & S. gage) is attached to the lower side of the piston in the indicator by screwing into a hole tapped centrally. Now if weights are suspended from the end of this rod * the spring can be calibrated in tension by drawing lines on a paper 1 The weight of this rod and wires or strings supporting the weights must be added to them to get the correct tension. If, however, only the true scale of the spring is desired, as is usually the case, the weight of these parts need not be considered, provided, of course, the atmos- pheric as well as the other lines are all drawn with these parts attached to the piston. Fig. i 01. — Apparatus for Testing Indicator Springs in Tension. ENGINE INDICATORS AND REDUCING MOTIONS 111 card placed on the drum in the same general way as when the spring was tested in compression. A suggested form for the arrangement of data for these calibrations is given below. Calibration of Indicator Spring (Compression) In Indicator No Rated scale of spring Diameter of piston of testing apparatus ins. Area of piston of testing apparatus . . . : sq.ins. Identification marks on spring Observers Date. No. of Weight on Scales, Lbs. Actual Pressure on Piston, lbs./sq.in. Ordinates or " Heights " Measured on Card, Inches. True Scale of Spring, (3) -H6). Up. Down. Average. Remarks. i 2 3 4 5 6 Curves. Results should be shown graphically for calibra- tions of indicator springs by plotting for abscissas the average height, inches, and for ordinates the corresponding actual pressures in pounds per square inch. Calibration of Indicator Springs with the Mercury Column. The method to be followed in calibrating indicator springs with a mercury column is essentially the same as described on pages 18-22 for the calibration of pressure gages. After the indicator has been cleaned and oiled, it should be attached to the testing drum or cylinder by means of an indicator cock. Then while the indicator is being heated to the temperature of the fluid medium used (steam, air, or water), the paper card can be put on the drum after first drawing two vertical lines one-half inch apart, as explained when describing the Cooley apparatus. Following these same instructions the atmospheric and other pressure lines are drawn first with increasing and then with decreasing increments. 112 POWER PLANT TESTINCx Testing the Drum Motion of Indicators. An apparatus for determining the relative accuracy of the drum motion of indicators as regards uniform tension in the cord for a given speed is illustrated in Fig. 102. This device, known as Brown's, ®=S B Fig. —Brown's Apparatus for Testing Drum Motion. consists of a rod R, which is made to take the same movement as the end of the cord in a reducing motion by being attached through a connecting rod to a crank pin on a disk like the face plate of a lathe. At the other end this reciprocating rod is attached to a bell-crank lever F, of which the outer end P carries a pencil for making a diagram on a card attached to a vertical frame 0. The short end A of the bell-crank is connected to a helical spring S and the other end of this spring is attached to an arm fitted on the rod R. To the end of the spring at A the indicator Fig. 103. — Diagrams taken from Apparatus for Testing Drum Motion. cord is attached, being of the same length as when in use on an engine. Now when the reciprocating rod R is in motion and the tension in the spring S is uniform the pencil at P will describe a horizontal line. If, however, the tension in the indicator cord varies and consequently also the tension in the spring is not uniform, the pencil will describe a closed curve. Examples ENGINE INDICATORS AND REDUCING MOTIONS 113 »Cord to Indicator of such curves are shown in Fig. 103. Curve AB was obtained when the apparatus was moving very slowly, EF when operating at about 700 revolutions per minute, and CD when the speed was about 2 50 revolutions per minute. The latter speed is obviously the one for which the stiff- ness and length of the spring in the indicator drum are most suitable. Reducing Motions for Indicators. In the case of most engines the length of the stroke is very much longer than the greatest possible movement of the drum of the indicator. It is therefore necessary to provide some means called a reducing motion, which produces shorter movement, but which at every Cord to Indicator Fig. 104. -Simple Pendulum Reducing Motion. Fig. 105. — Pendulum and Quadrant Reducing Motion. instant corresponds exactly with that of the cross-head. If this correspondence is not secured the length of the indicator 114 POWER PLANT TESTING diagram cannot be accurately reduced nor calculated, and the timing of the events or so-called " points in the stroke " will not be correctly represented. One of the simplest forms of reducing motions is illustrated in Fig. 104. This device is pivoted at one point A to a pedestal Brumbo's Pulley. supported on the frame of the engine, and has a link BH con- nected to the cross-head. The indicator cord rides in a circular arc CD, proportioned to give the required movement to the drum of the indicator. Although this arrangement does not give an exact reproduction of the movement of the cross-head, yet if the pendulum AB and the cross-head are simultaneously ENGINE INDICATORS AND REDUCING MOTIONS 115 at the middle of their strokes the error is insignificant. An improved type of this device is shown in Fig. 105, in which the cord rides in a groove on the circumference of a quadrant pulley. By attaching the pendulum to the quadrant pulley by means of a suitably designed "slip" joint, the pendulum can be disconnected from the quadrant so that the indicator cord will be moved only when the indicator diagrams are to be taken. Fig. 107. — Pantograph or Lazy-tongs Reducing Motion. Brumbo's Pulley is also a form of reducing motion of the pendulum type. It is illustrated in Fig. 106. In this device a guide pulley is placed between the indicator and the quadrant pulley. A modified and simpler device of the same kind consists of an upper portion moving in a vertical direction in a swinging tube and a lower portion pivoted directly to the cross-head. Of the portable devices used for reducing motions the pan- tograph, Fig. 107, is probably the one most used. This device 116 POWER PLANT TESTING is sometimes known as a lazy-tongs. Because of the numerous parts of which it is composed, requiring a great number of joints, it is likely to be troublesome with high-speed engines. A plan view showing one of the methods of attachment of this device to a horizontal engine is given in Fig. 108. This instrument when firmly put together is a perfect reducing motion. Fig. Plan View, Showing Attachment of Pantograph. Parallel Motions, like the one illustrated in Fig. 109, are also very commonly used. They are made usually of rods of iron or of steel nicely riveted together at the joints. The indicator cord is generally attached at B and the ends A and C of the long rod are Fig. 109. — Simple Parallel Reducing Motion. Fig. 1 10. — Simple Parallel Reducing Motion as Attached to a Steam Engine. fastened respectively to the cross-head and to the frame of the engine. It is a necessary requisite that the points A, B, and C shall lie in a straight line as shown. Also DE must be equal in length and parallel to FG. Then AF is in the same ratio ENGINE INDICATORS AND REDUCING MOTIONS 117 to HF as the stroke of the piston is to the length of the indica- tor diagram. Methods of attachment of similar devices to engines are shown in Figs, no and in. Fig. hi. -Simple Parallel Reducing Motion as Attached to an Air Compressor. Fig. ii2 illustrates another interesting parallel motion. It consists of a rod R, moving in a slide S, parallel to the piston- rod. A link BD is attached to the slide R at B and to CE at D, while AE is fastened at one end to the cross-head C. In this case again if A, B, and C are in the same straight line, then the following relation holds: AE : BD and CE : CD as the stroke of the piston is to the length of indicator diagram. The cord is hooked on a pin at H. It is desirable to have cator used. Reducing Wheels which consist simply of a large and a small pulley attached to the same axis, are coming into more or less general use. A typical arrangement is illustrated in Fig. 113. Pulleys D and D' are usually connected by a sliding sleeve so Fig. 112. — Sliding Type of Parallel Reducing Motion. a separate pin for each indi- 118 POWER PLANT TESTING that they can be disconnected when indicator diagrams are not being taken. Fig. 1 14 is a device by Armand Stevart for engines with long strokes. A and B are fixed ends of cord wrapped around a Fig. Fig. 113. — Reducing Motion of Concentric Pulleys. pulley D. The indicator cord g is attached to a small pulley D' and passes around a guide pulley G. D and D' are attached to the cross-head C. Then diameter D-=- diameter D' = stroke of piston-^- by the difference between stroke of piston and length of card. Reducing Wheels are not infre- quently made for attachment di- rectly to the indicator, as illustrated in Figs. 115 and 116. The former shows the Crosby reducing -wheel attachment and the latter a similar device for the Tabor indicator. Calculations of the Indicated Horse Power of an engine show usually the power developed on one side of the piston, which is commonly stated by the formula, I.H.P.^Pi^. ..... 33,000 Where p= mean effective pressure on the piston, lbs. per sq. in.; 1 = length of stroke in feet; a=net area of piston in square inches; 1 n=number of revolutions per minute. x In all piston engines the area of the piston rod must be subtracted from the area of the piston on the side where the rod reduces the area effective for the action of the steam or other working substance. 1 14. — Armand Stewart's Reducing Motion. (*S) ENGINE INDICATORS AND REDUCING MOTIONS 119 Of the terms of this equation only one, the mean effective pressure, is obtained from the indicator cards. If we consider now only one end of the cylinder, the steam does work on the piston during a " forward " stroke, and, on the other hand, the piston does work on the steam on the " return " stroke. Hence to get the mean effective pressure for a stroke the average pressure during the return stroke must be subtracted from the average pressure on the "forward " stroke; and this is ob viou sly the same as the average length of all the ordinates inter- cepted between the u pper and lower lines of the indicator card multi- plied by the scale of the spring. Usually the mean effective pressure is found by means of plan- imeters, the use of which for this purpose was ex- plained on pages 72- 77. An engineer should, however, know how to calculate the mean effective pressure of an indicator diagram with reasonable accuracy without the use of such instruments. In such cases the method of ordinates is very convenient. With suitable triangles draw ordinates perpendicular to the atmos- pheric line at both ends of the diagram as shown in Fig. 117. Lay off on the edge AB of a piece of smooth flat paper, a scale of ten equal divisions so chosen that the total length of the ten divisions is a little greater than the length of any of the indicator diagrams. This scale should then be placed obliquely across the diagram to be measured, so that the beginning and -Crosby Indicator and Reducing Motion Attachment. 120 POWER PLANT TESTING end of the scale will be located on the ordinates at the ends of the diagram. Now mark the diagram opposite the divisions of the scale with fine points, and at the middle of each of these Fig. i i 6. —Reducing Motion Attachment for Tabor Indicator. divisions draw ordinates across the breadth of the diagram. The sum of the lengths of these ordinates divided by ten gives the value of the mean ordinate, and this when multiplied by Fig. Diagram Illustrating Method of Mean Ordinates. the true scale of the spring gives the mean effective pressure. Some time can be saved in summing the ordinates if they are transferred with a dividers one after the other to the edge of the strip of paper. The total length laid off divided by ten is then the mean ordinate. ENGINE INDICATORS AND REDUCING MOTIONS 121 The Engine Constant for Indicated Horse Power. In the use of equation (25), page 118, where I.H.P.= Pi^, 33,ooo considerable time can usually be saved when calculating engine tests if the terms la / a\ , . ..... (26) 33,ooo' called the engine constant, which always remain constant for each end of the cylinder, are first computed carefully and then used as constants throughout the calculations. In other words, the indicated horse power is found for each end of the cylinder by taking the product of the terms, Engine Constant XpXn. CHAPTER VI MEASUREMENT OF POWER— DYNAMOMETERS A dynamometer, according to its derivation, is an instrument for measuring force or " power." These are of two kinds: i. Those absorbing the power by friction and dissipating it as heat. 2. Those transmitting or passing on the power they measure, thus wasting only a small part in friction. Absorption Dynamometers (Prony Brakes). Of the class of dynamometers in which the power received is all absorbed Fig. i i 8. — Simple Prony Brake. in friction, the type generally used is called a Prony brake, named for Rev. John Prony, who many years ago developed a device of this kind for measuring power. 1 One of the simplest forms is shown in Fig. 118. It consists of a lever A, from which a weight w is suspended from one end, and a block B, supported on a revolving drum or pulley, is attached to the other end. A strap to which wooden cleats are fastened is held in place and tightened by the thumb- nuts N, N. When the friction on the strap and block just 1 Strictly speaking, a brake of this kind does not provide means for directly measuring power, as, for example, horse power, because the element of time is not indicated. In other words, it measures the tangential force of which a couple (torque) in linear-weight units, such as foot-pounds, can be computed. 122 MEASUREMENT OF POWER— DYNAMOMETERS 123 balances the weight w, the lever arm A is horizontal and the apparatus is in adjustment. Stops, marked S, S, are provided so as to limit the movement of the lever arm. When the brake is adjusted or " balanced," the work done in a given time in producing the friction (the power absorbed) is measured by the weight moved multiplied by the distance it would pass through in that time if free to move. Then if, r= length of brake arm in feet. 1 n = revolutions of the shaft per minute. w = weight on the brake arm in pounds. Brake Horse-power (B.H.P.) = . . . (27) 33,000 " 2irr In equation (27) the fraction is a constant quantity 33,ooo for a given brake and is called 1he brake constant. When a brake like the one in Fig. 118 is used the effective weight of the brake itself as weighed at the point P must be added to the weight w. A very common variation 2 of the Prony brake is illustrated in Fig. 119. Rotation being in the opposite direction from that in Fig. 118, the knife-edge at E on the arm A will now press on the pedestal T, and the weight w can be determined by weighing the pressure on a platform-scales S. Since the scales receives not only the pressure due to the force pro- ducing friction, but also that due to the weights of the brake and of the pedestal, these weights must be determined and are to be subtracted from all of the readings of the scales to obtain the net weight w, for substitution in equation (27). Weight of the brake and the pedestal, called the zero reading, must be obtained with the brake strap slack, so that the block B will rest as lightly as possible on the pulley. With small engines this zero reading is obtained most 1 The length of the brake arm is measured by the perpendicular distance from the line of action of the weight w to the center of the wheel. When the arm A is horizontal, as in Fig. 118, the length of the brake arm is usually measured by the horizontal distance from P to a line passing through the center of rotation perpendicular to the arm. 2 For information regarding the designing of Prony brakes for absorb- ing large powers the reader is referred to Engine and Boiler Trials, by R. H. Thurston, pages 360-279. 124 POWER PLANT TESTING accurately by observing the weights on the scales when the brake pulley is turned around by hand first in one direction and then in the other. In this way we obtain for both the brake and pedestal, with rotation in one direction the weight plus the friction due to their own weight, and with rota- tion in the other direction the same weight minus the same friction. Half the sum of the two readings is, therefore, the weight corresponding to the pressure on the scales due to gravity alone. With large engines it 4s sometimes difficult to turn them uniformly by hand, so that the zero reading must be obtained by some other method. This is done usually in prac- Fig. -Prony Brake with Platform Scales. tice by placing a very small rod on the drum or pulley D (Figs. 118 and 119) vertically over the center of the shaft. Then if the strap is loose and due care is observed, the pressure on the scales can be obtained with sufficient accuracy without rotation. Other forms of Prony brakes are illustrated in Figs. 120 and 121. The former is called a strap brake and the latter a rope brake. The strap brake is made up merely of a band of steel or leather or of strands of rope placed over or wrapped around a suitable pulley. 1 In this case weights must be suspended from both sides of the brake wheel or pulley. 1 It is desirable to use for Prony brakes pulleys of which the section of the face is a double " U," like Fig. 122. The outside rims are for keeping the brake in position on the pulley and those on the MEASUREMENT OF POWER—DYNAMOMETERS 125 Rope brakes x like the one shown in Fig. 121, are much used for "commercial testing" of engines, as it is easily portable or can be made quickly at a small expense from materials always at hand. Moreover, it is self-adjusting, so that accurate fitting is not required. It consists of a rope doubled around a pulley or fly-wheel on the shaft transmitting the power to be measured. Several U- shaped distance pieces of wood, preferably maple, are provided to prevent the rope from slipping off the pulley and to keep the parts of the rope separated. These dis- tance pieces should be attached to the rope by soft iron or copper belt lacing, drawn in from the outside of the wooden pieces through the center of the rope, instead of being fastened with screws or nails on the inside, which will heat to a high temperature and then char the rope. Sometimes such fastenings when of hard metal will cut grooves into the surface of the pulley. In the case of the strap brake, Fig. 120, the net pull, cor- responding to the weight w, inequation (27), page 123, is Wi — W2. inside for receiving a small stream of water played upon the inside of the pulley. This stream of water by its evaporation will assist materi- ally to dissipate the heat generated. Such brakes are often operated with pipes arranged to discharge water into the pulley and another pipe to carry it away. This is an excellent system provided the latter pipe is used only to carry away a little overflow, but if so much water is used that there is practically no " steaming," the inside rim of the pulley will fill up with water to be spattered around in every direction as well as over the face of the pulley, where it is particularly objectionable, as it produces variable friction. 1 Sir William Thomson (Lord Kelvin) invented in 1872 the first rope brake of which we have any record. Although he utilized this device as a friction brake, it was not used by him as a dynamometer. Fig. 120. 126 POWER PLANT TESTING Now the same relation would hold if, as is often done, a spring balance is fastened to the floor on say the left-hand side; the pull registered by the spring balance would then be Wi 1 , and the net pull is, as before, W1-W2. Brake horse power is cal- culated then by equation (27), substituting for w the net pull Ji.Ji.P. where E.H.P. is the " electrical " horse-power input, as meas- ured with voltmeter and ammeter. If then a pump, air compressor, ventilating fan or a similar machine is to be tested by the electrical method, it should be direct-connected to the shaft of the motor, and its efficiency, E, will be E U-H.P. E.H.P X efficiency of motor' where U.H.P. is the useful work done by the machine, in horse power. 1 When an alternating-current generator is used the power factor must also be measured with a suitable instrument. In this case, the actual E.H.P. is (volts X amperes X power factor) + 746. This method is discussed more in detail on pages 293-295. MEASUREMENT OF POWER— DYNAMOMETERS 133 This last method serves also as a convenient method for obtaining the efficiency of a generator, since by connecting it directly to the shaft of a motor previously " calibrated " (for efficiency) the electrical output of the generator and the input to the motor are readily determined. When a so-called variable speed motor is used as a dyna- mometer, its efficiency must be determined at the particular speed and power at which it will operate when driving the machine to be tested. Transmission Dynamometers. Instruments of this type are used to measure the amount of power transmitted without absorbing any more power than is absolutely needed to move the dynamometer. Goss Belt Dynamometer. One of the simplest forms of transmission dynamometers, designed by Professor W. F. M. Goss, is illustrated in Fig. 127 by a line drawing. Being so simple in construction, it can readily be made in any factory or workshop with the materials available. It is essentially a differential lever measuring the difference in tension between the two sides of a belt. This lever is piv- oted at the point,' D, and to it are attached the shafts car- rying the pulleys A and B. Power transmitted is meas- ured by the product of the speed of the belt and the difference in its tension between the two sides of the dynamometer. The force tending to raise the left-hand end of the lever is assumed to be twice the tension ti, of the tight side of the belt, while that raising the right-hand side is twice the tension t 2 in the slack side of the belt. These tensions on the two sides of the lever can be measured by weights Wi and w 2 suspended from its ends. The force tending to rotate the lever is therefore ffl ^=^ Fig 127. — Goss Belt Dynamoneter. 134 POWER PLANT TESTING twice the difference in tension on the two sides of the belt 2(ti-t 2 ) and this force acts with a leverage AD=DB, when these two arms are made equal. If now the distance from the pivot D to the point of support of the weight Wi is made twice AD, and if the weight Wi is made equal to the difference in the tensions it will balance the lever. Or for these condi- tions we can write the equation, Wi=t 1 -t 2 . Now in any transmission system the difference in the tension of a belt or a rope on the two sides of a driven pulley multiplied by its speed is a measure of its power. If then, d is the diameter of the driven pulley C, plus the thickness of the belt or rope in feet, n is the number of revolutions of this pulley, and Wi is the weight in pounds on the left-hand side required to balance the lever ; when power is transmitted, then, Horse power transmitted = -. . . . (20) v 33,ooo v w To reduce the vibrations of the apparatus a dash-pot is connected to the right-hand side and also to prevent excessive movement of the lever when unbalanced, dash-pots are placed above and below the lever on the left-hand side. Differential Dynamometers. The apparatus illustrated in Fig. 128 is typical of a number of dynamometers indicating by means of a differential lever operated by gearing, the amount of power transmitted. 1 This is a very common form of trans- mission dynamometer. Power is received from the motor (or engine) by the shaft A, which is connected only indirectly by means of gears to the shaft A' opposite, which transmits the power to the work. To the adjoining ends of these shafts bevel wheels B and D are attached. The lever L turns on an axis concentric with the shafts A and A', in a plane perpendicular to 1 Similar forms of differential dynamometers are known as White's, King's and Bachclder's. The first instrument of this kind, it is stated, was invented by Samuel White in 1780. These dynamometers are sometimes called epicyclic, signifying " wheels traveling around a circle or around another wheel." MEASUREMENT OF POWER— DYNAMOMETERS 135 them. It carries bevel wheels C and Ci, gearing with B and D, through which the power is transmitted. There is a tendency then for the left-hand end of L to go downward and for the right- hand end to rise. If, furthermore, the lever L were permitted to revolve, no work would be transmitted from B to D, and there- Weight Shaft to Work. Fig. 128. — Typical Differential Lever Dynamometer. fore D would remain stationary. As these gears are usually pro- portioned so that B revolves with twice as many revolutions in a given time as L, a weight placed on B at a given radius from the center will balance a weight twice as large at the same radius on the lever L. Then the moment of the force applied to the lever L to balance it must be twice as great as the moment of the force transmitted from C and Ci to D. If, therefore, 1 is the length (feet) of the arm at which the weight w (pounds) 136 POWER PLANT TESTING is applied, and n is the number of revolutions per minute of the shaft A, then the power transmitted per minute = — (foot-pounds), (30) and Horsepower (H.P.) = (31) 33,ooo Now if the lever L is to be prevented from turning about its axis, a couple will be required which is twice the driving couple being transmitted. If, then, a weight of w pounds sliding on L as shown in the figure is placed at a distance 1 feet from the axis, so that the lever remains horizontal, the driving couple can be determined; and when the revolutions of the shaft A are known, the power can be found. A dash-pot D is usually attached to the differential lever L to reduce vibrations. This dash-pot should always be kept filled with glycerine or good clean oil. If the dash-pot is sticky consistent results cannot be expected. A Webber Differential Transmission Dynamometer as made commercially is illustrated in Fig. 129. The scale on the lever arm of this instrument is graduated into 100 divisions and a bell is provided which rings at every 100 revolutions. Since the horse power transmitted in one revolution per minute is irlnw . '. ' , , ■- — equation (31), then the horse power corresponding to one division on the scale per 100 revolutions per minute is also 33,ooo for a perfect calibration. It is interesting to observe that if we let v =the vertical force acting at C and Ci ; p = the vertical pressure between the teeth at each point of contact ; d = distance from the center of the rotation to C and Ci; l = the distance from the same center to the weight w. Then from the foregoing discussion it should be clear that 2V=4p and wl = 2vd=4pd. Then if r is the effective pitch radius of the driving gear wheel B, ri is the radius of the small MEASUREMENT OF POWER— DYNAMOMETERS 137 bevel wheels, and the force producing the turning movement in the shaft A is represented by f, we have, and. fr=2pr!, f _2pr 1 _wlri r 2dr (32) Fig. 129. — Webber Differential Transmission Dynamometer. If we know the number of revolutions, then the space passed through by the force f can be calculated, and the work in foot- pounds is the product of the force times the distance passed through. The units given above are of course respectively in feet and pounds. Calibration of a Differential Dynamometer. 1. Examine the 1 dash-pot and observe whether the piston moves freely in the cylinder, particularly without " sticking." After the apparatus has been well oiled the position of the poise to make 138 POWER PLANT TESTING the lever arm horizontal should be observed for " no load." If this is not at zero then all the readings on the scale must be corrected by the amount of this zero reading. 2. At each of the speeds required make a preliminary run without load and observe the reading of the poise when the lever is balanced. 3. Attach a Prony brake to the shaft from which the power is to be transmitted and observe for a series of loads and speeds the readings of the poise on the dynamometer lever. 4. For each speed plot a curve with theo- retical foot-pounds per minute by equation (30) as abscissas and actual foot-pounds per minute as deter- mined by the Prony brake as ordinates. Emerson Power Scales. Another very satisfactory instru- ment for the meas- uring of power trans- mitted by shafting is known as the Emer- son Power Scales. It is illustrated in Fig. 130. It consists of a pulley C keyed to the shaft. To this pulley C a wheel B is connected loosely by studs EE projecting between and bearing against its spokes. The pressure exerted on these studs is proportional to the power transmitted by means of the pulley C, and this pressure is transmitted by a system of levers LL and bell-cranks MM to a sleeve A connected to a "weighing lever " W. The sleeve slides loosely on the main shaft. The amount of the pressure exerted on the studs is indicated for small values by a pointer P, moving over a graduated scale Fig. 130. — Emerson Power Scales. MEASUREMENT OF POWER— DYNAMOMETERS 139 F. For pressures beyond the limits of the graduated scale weights are placed on the scale pan N. A dash-pot D is pro- vided to prevent excessive vibrations and make the pointer " dead-beat." The scale F is calibrated to read the pressure (force) exerted by the torque of the pulleys C on the studs EE in pounds. The Fig. 131. — Flather's Hydraulic Transmission Dynamometer. work or " power " is calculated, therefore, by taking the product of this force times the distance moved through. If d is the diametral distance between the centers of the studs in feet, n the revolutions per minute and w the reading of the scales, then, rdnw Horse power 1 = 33,ooo Compare with (31) for differential dynamometers, page 136. (33) 140 POWER PLANT TESTING A speed counter is attached to the apparatus for counting the number of revolutions. Flather's Hydraulic Transmission Dynamometers. A form of transmission dynamometer which is operated by hydraulic pCI X fe^ Fig. 132.— Diagram Showing Pulleys, Pistons, and Shaft of Flather's Dynamometer. pressure is shown in Fig. 131. The power shaft is keyed to the boss of a pulley B with two or more arms carrying hydraulic cylinders R. Projecting ends or studs from these cylinders bear upon the arms of a loose pulley A on the same shaft. The torque Fig. 133. — Details of Pistons and Cylinders in Flather's Dynamometer. imparted by the driving belt to the loose pulley A is thus trans- mitted to the shaft S through the liquid, and the resulting pressure is conveyed by radial pipes U to the hollow central shaft, and then to a pressure gage G. The hollow shaft is always filled with oil. In the figure an engine indicator I is MEASUREMENT OF POWER— DYNAMOMETERS 141 shown attached to the hollow shaft for recording the pressure. The loose pulley A drives the tight pulley B through its pistons which press on the oil in the cylinders carried by the tight pulley. By means of a worm drive the drum of the indicator receives its motion from the central shaft S. Figs. 132 and 133 show more in detail the construction of the hydraulic cylinders on the pulley B. Fig. 134 shows typical indicator ~^~4lfi0 /^f\f JXrKfx ^j^\ Fig. [34. — Indicator Diagrams from Flather's Dynamometer Attached to a Mining Drill. diagrams from this apparatus. Both were taken from a dyna- mometer connected to a mining drill. The first was taken when the drill was sharp, the second when it was dull. Among the advantages claimed are: (1) its simplicity, (2) that it is not appreciably affected by the velocity of the shafting, (3) that no countershaft is required, (4) by connecting it to a recording gage a continuous diagram of the load can be obtained. CHAPTER VII FLOW OF FLUIDS The flow of fluids will be discussed under these heads: i . The flow of air. 2. The flow of steam. 3. The flow of water. The Flow of Air. When subjected to only a low pressure, air and many other gases are usually measured by a gas meter, of which there are many types sold commercially. There are, however, two general types: (1) wet and (2) dry. The former is by far the more reliable and should be always used in preference to a dry meter when it can be obtained. Wet meters receive their name from the water seal maintained in them. This seal must be always kept at a constant level, determined by calibration, 1 and before using such a meter in a test one should always observe whether the water level is at the standard mark. If it is not, then water must be added or withdrawn as the case may be. A section of a wet meter is shown in Fig. 135. It consists, as usually made, of a series of chambers arranged like an Archimedean screw, which are alter- nately filled and emptied. When air or any gas flows 2 into one of the chambers of the meter it accumulates over the sur- face of the water and by its pressure raises the chamber until it is filled. In the figure the gas enters at the dry-well, V, passes through the drum and out at the front end, then over the drum between it and the case to the outlet. In this way the drum is made to revolve to the left by the pressure on the 1 Gas meters may be calibrated by any apparatus suitable for the displacement of gas as it is withdrawn by water. 1 1 is very necessary, of course, that when the weighings are made the pressure and temperature of the gas be accurately determined. 2 The nature of specific gravity of the gas is not important, as gas meters are calibrated to record volumes, usually cubic feet. 142 FLOW OF FLUIDS 143 surface of the water below and the slanted partition C above forming an ever-increasing pyramidal space between the surface of the water and the plane of the slanted partition. Wet Fig. 135.— Typical Wet Gas Meter. meters are usually very accurate, while dry meters are not " supposed to be instruments of great accuracy." Pitot Tube for Measurement of Air. Probably the most accurate method of measuring air in large volumes is by means of a Pitot tube. A standard instrument of this kind designed for the measurement of air by the American Blower Company is shown in Fig. 136. It consists simply of two tubes, a small one, being placed inside of a larger one as illustrated in detail in Fig. 137. These tubes are arranged so that*each has a separate connection, Fig. ■36. — "American Blower Co. Standard Pitot Tube. 144 POWER PLANT TESTING as at A' and B'. The lower end of the small tube is open at A, while the outside and larger tube has two openings at the opposite sides of B marked in Fig. 137, 2^ inches from the end at A. When the instrument is used it is placed so that the opening at A points against the direction of flow and receives the full effect of the pressure due to the velocity of flow. The side openings in B are subjected to only the static pressure. For convenience let p = velocity pressure and s = static pressure. For example, the difference in the levels in the manometer, a, Fig. 138, is therefore that due to (p + s)— s, or simply p, the velocity pressure . Pitot tubes are usually connected to manometers or prefer- ably to sensitive draft gages, showing the pressure in small fractions of an inch of water. When the end of the Pitot tube 2 Holes at Front and Back of each .03 Diameter in Outer Tube Fig. 137. — Section of Standard Pitot Tube. For Pressures Above Atmospheric For Pressures Less Than Atmospheric Fig. 138. — Arrangement of Connections for Pitot Tube Measurements. at A' is connected to the left-hand end of a draft gage like those in the figures on page 24 and the end at B' is attached to the right-hand end, the instrument acts as a differential gage and the difference between the reading when thus connected and its zero reading is the pressure in inches of water corre- sponding to the velocity alone; that is, (p + s)— s. If we call this velocity pressure P when reduced to feet of water, and if h is the height or " head " also in feet of an equivalent column FLOW OF FLUIDS 145 of air producing the same pressure, then the velocity of the air v in feet per second is where g is the force of gravity (32.2), and wt. of a cu. ft. of water h = Px wt. of a cu. ft. of air 62.3 » P wt. cu. ft. air v=6 3 .3^ : wt. cu. ft. air (34) In the following table the weight is given of dry air and also the weight of the dry air in, a cubic foot of air completely saturated with moisture (100 per cent humidity). The data given are at atmospheric pressure (14.7 pounds per square inch). By interpolating between these tables, the weight of air for any temperature and degree of saturation is easily obtained. Remembering also that the weight per cubic foot is directly pro- portional to the absolute pressure, the weight for any pressure is readily determined. Tables for determining the percentage of saturation by means of wet- and dry -bulb thermometers are given on page 331. Temper- Weight of 1 Cu. Wei jht of Dry Temper- Weight of 1 Cu. Weight of Dry ature, Ft. of Air in 100 % ature, Ft. of Air in 100 % Deg. Fahr. Dry Air. Lbs. Sat arated Air Deg. Fahr. Dry Air, Lbs. Saturated Air O •08635 08623 75 .07424 .07194 10 .08451 08430 80 07355 .07090 20 .08275 08243 85 07288 .06977 3 2 .08075 08024 90 07222 .06861 40 •07944 07874 95 07157 .06773 45 •07865 07780 100 07093 . 066 1 3 5° .07788 07688 110 06968 06310 55 .07712 07592 120 06848 .00025 60 .07638 07496 150 06511 .04790 65 .07566 07396 200 06018 .01200 70 .07494 07298 1 The weight of (about 70 Fahr.). cubic foot of water at about "room " temperature 146 POWER PLANT TESTING Flow of Air through an Orifice. Air under comparatively high pressures is usually measured in practice by means of pres- sure and temperature observations made on the two sides of an orifice. Fig. 139 illus- trates the method with two pressure gages on opposite . sides of the orifice > and a thermometer I for obtaining the tem- 139. — Measuring Flow of Air through an perature ti, at In 1 r Fig. Orifice. initial or higher pres- sure, pi. The flow of air w, in pounds per second may then be calculated by Fliegner's formulas. 1 :w = .53o. when pi is greater than 2p< (35) w = i.o6oXa / P«(Pi- Pa) when pi is less than 2p a j . (36) where a is the area of the orifice in square inches, Ti is the absolute initial temperature in degrees Fahrenheit at the absolute pressure pi in the " reservoir," and p a is the absolute atmos- pheric pressure, both in pounds per square inch. For small pressures it is often desirable to substitute manom- eters for pressure gages. Flow of Air Measured by Cooling. This method depends on taking from the air an amount of heat 2 which can be meas- ured and then computing from the heat units absorbed and the difference in temperature and specific heat of the air, its weight and volume. The arrangement of the apparatus is shown in Fig. 140. A coil of pipes C of which the cooling surface is as equally as possible distributed over the section of the duct D, D' 1 See Peabody's Thermodynamic:;, page 135, and Spangler's Applied 'Thermodynamics, pages 12-13. 2 The method will be equally applicable if heat is added, as for example by passing steam through the coil. This method is often used to cali- brate Pitot tubes and anemometers. FLOW OF FLUIDS ' 147 carrying the air to be measured, is used to absorb heat by. cir- culating water through it. Thermometers are arranged so that the temperatures of the air and of the water can be observed, and a platform scales is shown for obtaining the weight of water. Using the symbols ti and t 2 for the - initial and final temperatures of the air in degrees Fahrenheit, t' and t" for the temperatures of the water entering and leaving in degrees Fahrenheit, w a = weight of air passing through D, D' in pounds per second, Wo = weight of water collected in pounds per second, and 0.2375 = specific heat of the air at con- r mm i t £ ft Fig. 140. — Measuring Flow of Air by Cooling. stant pressure and at temperatures not much above " atmos- pheric," 1 then the heat absorbed by the water per second is Wo (t" — t') B.T.U. and this equals the heat lost by the air, or 0.2375 w a (ti-t 2 ),and w (t"-t') 4 .2iiw (t"-t') 0.2375 (ti-t 2 ) (ti-t 2 ) (37) Anemometers. A very convenient and simple method for measuring directly the volume of the air, or any gases, is by using an instrument called an anemometer. This instrument, Fig. 141, consists in its essential parts of a wheel having fiat 1 The weight of a cubic foot of air is .0765 pound at 62 (522 abs.) degrees Fahrenheit and 14.7 pounds per square inch pressure. Since the volume is directly proportional to the absolute temperature, the weight at any other temperature is easily computed. 148 POWER PLANT TESTING or hemispherical vanes mounted on slender arms. The wheel must be made very light in weight, must be accurately balanced, and should move easily in its bearings. By its own motion it operates a re- cording mechanism indi- cating velocities in feet per minute (in English units) . The Flow of Steam. The flow of the steam from an orifice or nozzle has a very definite critical value when the final pressure is greater than 0.58 of the initial pressure. When the final pressure is less than this critical value the flow is expressed very accurately by the following empirical formula, based on the experiments of Messrs. Emswiler and Fessenden, in the Mechanical Labora- Michigan. Using the following Fig. 41. — A Typical Anemometer for Measuring Velocity of Air. Of tories of the University symbols: pi= initial absolute pressure of the steam in pounds per square inch; -p 2 = final absolute pressure of steam pounds per square inch; a =area of the smallest section of the nozzle or orifice in square inches. Then the weight of the dry saturated steam discharged in pounds per second is approximately, 1 pr 97 a J-± when p 2 is less than 0.580!- 60.5 *" (38) 1 A somewhat simpler formula, known as Napier's formula, which is accurate enough for most calculations, is the following: w = - ! — when p 2 is less than 0.58?! 70 (39) FLOW OF FLUIDS 149 Now since in the theoretical formulas the weight discharged is inversely proportional to the square root of the specific volume v, or w is proportional to -W±-i the formula above corrected for initial quality x of the steam is, Pi" 6o.5\/x when p 2 is less than 0.58 pi- (40) J fe 07 SI 3 B VJ J2 0.5 s SI 0.3 0.1 1.0 09 0.8 0.7 0.6 0,5 0.4 Ratio of Final Pressure to Initial (El ) Pi Fig. 142. — Rateau's. Curve for Flow of Steam giving Values of the Coefficient K. When the steam is superheated the specific volume is con- siderably increased, and for this condition the author has found that the following equation gives very satisfactory results, 1 £1 ^ L * $ A ( ?/ #/ V -c 3 j / / / CURVES FOR DISCHARGE OF STEAM MAINLY WHEN FINAL PRESSURE IS REATER THAN b&i OF INITIAL PRESSU / IE ~ / P 1 (a \ W = 6o.5(i+.ooo65d)' {4I) when, as before, p2 is less than o-58pi and where d is the number of degrees (Fahrenheit) of superheat. When the final pressure p 2 is greater than 0.58P1, the formu- las must be modified to correspond to the reduced flow observed by inserting a coefficient K as a factor in the right-hand member 1 For a more extended discussion of the flow of steam see The Steam Turbine, by the author, pages 52-57. 150 POWER PLANT TESTING of the equations. Values of this coefficient are most conven- iently obtained from the curve in Fig. 142, which was plotted from the experimental results obtained by Professor Rateau of Paris. The Flow of Water. When the quantity of water to be measured is not too large it is most accurately determined by weighing in tanks placed on scales, or by direct measurement of volume in calibrated tanks or barrels. Sometimes it is impracticable to weigh or measure the volume of the water directly, particularly when it must be measured under pressure. For measurements in pipes up to 2 to 3 in diameter a water meter is generally used. A great many types of water meters are sold commercially and not very many are accurate, so that it is absolutely necessary Dial when using a meter to measure water in a test to calibrate it at least before and after the test, under the same conditions of temperature, pres- sure, and rate of flow of the water. In many plants where meters are used con- stantly, suitable con- nections are made to the discharge from the meter, so that at any time the flow through it can be diverted into a tank Pulsating Diaphragm Water Meter. in which it can be measured by volume or weighed. One of the best types of water meters is illustrated in Fig. 143. This belongs to the class operating with a " pulsating diaphragm." The inclined shaft S on this diaphragm traveling around in contact with the peg P on the plate B moves the counting mechanism through intermediate gears. This dia- phragm, in the Thomson-Lambert meter (Fig. 144) is made of hard rubber reinforced with a steel plate, making it FLOW OF FLUIDS 151 much more durable than those made without reinforcing. As the side chambers are alternately filled and emptied, the dia- phragm is moved up and down with a kind of " pulsating " motion. A central spindle on the diaphragm is connected to a set of gear wheels operating the recording mechanism. Fig. 144. — Thomson-Lambert Water Meter. Worthington Water Meter. The Worthington water meter (Fig. 145) is also used frequently. It belongs to the type oper- ating in a cylinder by a reciprocating piston which is driven- back- ward and forward by the pressure of the water. Friction is an important element in meters of this type; but it is not injured by moderately hot water. The readings of a water meter are usually in cubic feet. 152 POWER PLANT TESTING A water meter is essentially a water motor adapted for operating the gearing connected to the counting mechanism. Frequent calibrations of water meters are necessary because they are likely to become more or less clogged with dirt and refuse. The readings are also affected by the temperature, head, and quantity of water flowing, as well as by the amount of air carried in the water. A meter should always be calibrated at least at two or three rates of flow, as it scarcely ever happens that the conditions of the test are so uniform that the meter will be used only for a certain predetermined rate of flow. 1 Fig. 145. — Sections of a .Worthington Water Meter. Willcox Water Weigher. Automatic weighing or measuring devices are often used for determining the weight of condensed steam in engine tests or the weight of feed water in boiler trials. The Willcox weigher is a most satisfactory apparatus of this kind. It consists of a tank (Fig. 146) divided by a partition P into two compartments A and B, one above the other. The upper compartment A receives the inflow of water and the lower one B serves for measuring. Projecting into the lower compartment is a U-shaped discharge pipe C, which is always water-sealed. The upper end of the discharge pipe is covered by a bell float F, which is permitted a short up-and- down movement. In the upper compartment there is a short standpipe S, which is simply a hollow cylinder open at the top 1 Calibration curves are usually plotted with meter readings as abscissas and actual volumes as ordinates. A curve should be plotted for each of the several rates of flow if they are different. Curves of meter readings (abscissas) and correction factors (ordinates) are also useful. FLOW OF FLUIDS 153 Gage Glass and bottom. The bell float F and the standpipe S are con- nected rigidly by a vertical rod (Fig. 147) so that they move together as one piece, and this is the only moving part in the apparatus. The lower end of this standpipe has a corrugated face, and when it is down in its lowest position its corrugated face rests on a soft seat or ring surrounding a circular opening in the partition P. This seat is made of a rubber composition which is not in- jured by boiling ^ ss * sCSSSSs ** s =^ {Vent water. The weigh- er can be used, therefore, with either hot or cold water without risk. In the opera- tion of the appa- ratus, when the standpipe S is down on its seat, water entering through the side inlet accu- mulates in the up- per compartment A until it overflows the top of the standpipe. The water then flows down through the hollow standpipe into the lower compartment until there is a sufficient amount to seal the lower edge of the bell float F. Then as more water accumulates the bell float rises, lifting the standpipe S from its seat and the water in the upper com- partment flows down into the lower one until the volume is that of a " unit charge " for the apparatus, when the " tripping " device discharges the water through the discharge pipe C. The " tripping " is accomplished by a " trip " pipe T, which is normally water-sealed, but which becomes unsealed when a "unit charge" has accumulated. While the water is accu- mulating in the lower compartment B the water in the left-hand leg of the " trip " pipe T is being slowly pushed down because Fig. 146. — Willcox Water Weigher (by Volume). 154 POWER PLANT TESTING Stand Pipe of the increasing pressure of the air under the bell float F, due to the increase of head of water, and a corresponding amount of water spills over the upper end of the right-hand leg R of the " trip " pipe into the discharge pipe C. Due to this action the water level in T is lowered until it reaches the bend in the lower end of the " trip " pipe. Under these conditions the water column in R exactly balances the head of water in the lower compartment B and the air entrapped in the float valve F has a function similar to that of a scale beam, balancing on one side the head of water in the tank and on the other side the head of the standard water column in R, which of course is always constant. At the instant this balance is secured a very small amount of water added in the lower compartment and the corresponding additional spill from R will destroy this equilibrium. Then the air compressed in the bell float F and the upper part of the discharge pipe breaks the water seal in R by suddenly discharging all the water in it. When the air pressure in the bell float F is thus reduced it drops down, carrying down with it the standpipe S in the upper compart- ment A. In this last operation the air is removed from the interior of the bell float F, and water flowing in to replace it will spill over the top of the discharge pipe C and will flow out at the other end till the lower cham- ber is emptied of the " unit charge." At this time, the standpipe S becomes seated, due to the pressure of the water above the bell float and to the downward suction of the syphon. Thus the standpipe is held tightly upon its seat only at the instant when tightness is required; that is, while the "unit charge " is being discharged. After the standpipe has seated water again accumulates in the upper compartment A and the cycle of operations is repeated. A mechanical counter shown at the side of the apparatus is connected to a ball float inside the lower compartment and registers the number of times the apparatus is tripped. An FLOW OF FLUIDS 155 rutomatic weigher of this kind is easily calibrated by weighing several " unit charges," and it can then be used with as great a degree of accuracy as can be expected with rapid weighings in tanks on platform scales. The Willcox weigher may be expected to weigh hot or cold water with a maximum error of not more than one per cent. Venturi Meter. An arrangement of piping in which there is a gradual narrowing of the section to a minimum and then a more gradual enlargement was invented by Mr, Clemens Herschel for measuring the flow of water. This apparatus is called a venturi meter and is shown in Fig. 148. Piesometer tubes (manometers) are arranged to indicate the pressure at the sections shown. Pressures at these sections will be denoted respectively by p m and p ra . Fig. 148. — Herschel's Venturi Meter. From Bernouilli's theorem 1 it follows that the relation between the pressure in pounds per square foot and the velocities in feet per second at the two section's v m and v n of a stream flowing through such a closed horizontal channel is given by 2g £ + 2g' (42) where d is the density of the water in pounds per cubic foot. The volume of water flowing through any section is in cubic feet per second, if a represents the area of a section in square feet, a w ,v TO =a„v B =a, 2g(Vm-Vn) (43) 1 See Jamieson's Applied Mechanics, Vol. II., page 458. 156 POWER PLANT TESTING With suitable manometers or with gages the pressures p m and p n can be obtained, and since all other quantities can be repre- sented by a constant, k, we have Volume per Unit of Time =k(p OT -p n ). . . (44) The exceptional accuracy of this instrument for measuring the flow of water is well illustrated by the curve in Fig. 149, — : :::: — ::: — ::::— IT 14 - > t 13 : t 19 T / 11 2 _r in t 10 /- -9 _1 % J fe « / a -f 7 6 W 5 4 15 20 25 30 35 Discharge in Cubic Feet per Second Fig. [49. — Typical Curve Showing Extreme Accuracy of Venturi Meter. showing the flow as calculated from careful measurements of the head, while the barely perceptible dots shown on the curve indicate the results of actual observations by Pullen with a venturi meter. Flow of Water through Orifices and Nozzles. Theoretically the velocity of flowing water under any pressure is the same as the velocity attained by a body falling freely through a dis- tance equal to that head (h) as in Fig. 150. Furthermore this statement would be the same even if the water had no free surface, provided, however, the pressure at the orifice was that due to a head h. If then there is no loss of head due to friction FLOW OF FLUIDS 157 and eddies formed by the water passing through the orifice, the velocity of discharge, v, in feet per second is, v=\/2gh, (45) where g 1 is the acceleration due to gravity and h is the head in feet. If a is the area of the cross-section of the orifice in square feet, q is the quantity or volume of water discharged in cubic Fig. i 50. — Discharge of Water from an Orifice. feet per second, and assuming the stream is of the same cross- sectional area as the orifice, then, q=a\/2gh (46) Since the actual flow is less than the theoretical in most cases, and considerably less when the discharge is from a hole with sharp edges in a thin plate, a more general form may be written by inserting a suitable coefficient 2 of discharge, k, then, q=kav / 2gh (47) 1 The value of g is approximately 32.2, so that equation (45) can be simplified into v = 8.02V / h. 2 This coefficient is often called the coefficient of contraction. 158 POWER PLANT TESTING Calibration of Orifices and Nozzles. Water under a con- stant pressure is often measured by observations of the flow through either orifices or short nozzles which have been care- fully calibrated. The apparatus required consists usually of a suitably arranged standpipe to which the orifice or nozzle can be attached so that a given head of water can be main- tained 1 and a tank on scales (or calibrated for volumes) to receive the water discharged. A pressure of one pound per square inch is equivalent to a head of water at 62 degrees Fahrenheit of 27.72 inches, or 2.31 feet. A normal atmospheric pressure (14.7 pounds per square inch) is therefore equivalent to a head of 33.96 feet of water. Then for a given pressure or head the quantity of water dis- charged in a given time is readily obtained and the coefficient of discharge can be computed by substituting the values of quantity of discharge q, the head h, and the area a, in formula (47) . Data and results should be tabulated in the form given below. The relative roughness of the edge of the orifice or of the inside surface of the nozzle should be recorded. FLOW OF WATER Form Flow of water through a . Date Form of orifice or nozzle . (Sketch) Observers Formula Diameter, feet .... Area, square feet . No. of Reading. Head in Feet. Time Seconds. Total Pounds Cu. feet. Pounds per Second. Cu. feet per Second. Coefficient of Discharge (k) Remarks. Average 1 In many places a suitable pressure tank is not available, and in such cases the calibration can be made by attaching the orifices or nozzles to pipes carrying water under pressure. The readings of the pressure gage can be reduced to the equivalent head in feet, to which must be added, if the centre of the gage is higher than the orifice or nozzle, the distance in feet from the center of the nozzle to the center of the gage. FLOW OF FLUIDS 159 Curves. Curves should be plotted for each orifice or nozzle with head for abscissas and (i) the discharge (cubic feet per second) and (2) the coefficient of discharge for ordinates. Flow of Water over Weirs. When large quantities of water are to be measured, then orifices are unsuitable and it is customary to pass the whole body of water over a weir or gage notch. This consists of a board placed across the stream so that all the water must pass over it. The length of the notch is usually made less than the width of the stream to give definite conditions. This is accomplished most easily by sawing the notch out of a long board and beveling the edges. A typical arrangement for measuring the head of water on a weir is illustrated in Fig. 151. The head must be determined with great accuracy, and this is done usually by means of a hook-gage, Fig. 152, and a suitable level. Fig. -A Weir for Measuring Water. Fig. 152. — A Hook Gage. The hook-gage consists of a sharp-pointed hook H, attached to a vernier scale V, intended to measure very accurately the amount the hook is moved. Before taking an observation the hook must first be submerged and then raised slowly till the point just breaks the surface of the water. The correct height of the surface is obtained at the instant when the point of the hook pierces it. - The head h of the water flowing over the weir (Fig. 151) is obtained by setting by means of a straight-edge 160 POWER PLANT TESTING SE and the level L the point of the hook at the same level as the crest of the weir. The height observed in this posi- tion is called the zero head. It is to be subtracted from all other readings to get the head of water flowing. The hook- gage must be placed in such a position on the upstream side of the weir where the surface has no appreciable velocity and where there is very little disturbance due to eddies. In terms of the following symbols, q = quantity or volume of water discharged in cubic feet per second ; h-=the head in feet on weir measured in still water; b= breadth of the weir in feet; n=the number of contractions; k = coefficient of discharge. q =2 /3kh 3 / 2 (b — o.inhK^g. (48) Weir with a Triangular Notch. This is the well-known Francis formula for a rectangular notch. The ordinary rectangular notch has two contractions, one at each end. Triangular notches in weirs are sometimes used. One of these in the form of a right-angled isosceles triangle is shown in Fig. 153. It has the advantage of giving the same form of stream whatever the size of the notch or the height of FLOW OF FLUIDS 161 water passing through. It is, therefore, particularly suitable for measuring a flow of water which is somewhat variable. The quantity of water discharged over a triangular weir or notch is, q=4/i5kbh 3/2 \/2g. (49) (50) When the angle is 90 °, b=2h and q =4.26 kh 5/2 . Also when the angle is 6o°, b=2htan30° and q=2.47kh 5 ^ 2 . . . . (51) Any mistake made in determining h will produce a larger percentage error in the results with the rectangular and tri- angular notches than with an or'fice. Where great accuracy is desired and the quantity of water to be handled is not too large, an orifice calibrated and used in the bottom of the tank as shown in Fig. 154 is to be preferred to measurements with a weir. This remark is par- ticularly applicable in connec- tion with the measurements of cooling (circulating) water in tests of large steam engines and turbines. Use the same form for data as given for Calibration of Orifices or Nozzles on page 158. Curves should be plotted with heads for abscissas and (1) the discharge (cubic feet per second) and (2) the coefficient of discharge for ordinates. : 7-:\V'////4;— Fig. 54. — Best Kind of Orifice for Engine Tests. CHAPTER VIII CALORIFIC VALUE OF FUELS— SOLID, LIQUID, AND GAS Calorific Power is a term applied to the quantity of heat generated by the complete combustion of a definite quantity of fuel. In order to insure rapid and complete combustion the fuel is preferably burned in an atmosphere of oxygen under pres- sure. This calorific power of fuels is expressed in the English system as British thermal units per pound, and in the metric system as calories per kilogram. The quantity of heat generated by combustion is measured by the rise in temperature of a given weight of water in a calorimeter, of which the cooling effect or water equivalent k has been determined, and the temperature of the gas escaping has been reduced to that of the room. Now if, W/ = weight of the fuel in pounds, w,„ = weight of the water in pounds, k = water equivalent x of the calorimeter, in pounds, ti = initial temperature of water, degrees Fahrenheit, t 2 = final temperature of the water, degrees Fahrenheit, Q =heat generated, B.T.U., 1 Water equivalent is used to express the heat-absorbing effect of the calorimeter as equivalent to that of a weight of water. This may be found as for calorimeters used for determining the quality of steam by the hot- water method (see page 64), by taking the sum of the prod- ucts of the weights and specific heats of the various parts of the calori- meter (see Calorific Power of Fuels, by H. Poole, pages 14 and 15), or by comparing the results obtained with those that should have been secured, if there had been no absorption of heat, by the combustion of some fuel of which the heat value is known; as, for example pure carbon in oxygen gas. Corrections for radiation can be practically eliminated by having the temperature of the water in the calorimeter before ingition as much below the " room " temperature as the final temperature is above. 162 CALORIFIC VALUE OF FUELS 163 then the calorific value H per pound of fuel in British Thermal Units is, Q ( Wu ,+k)(t 2 -t,) H W/ w/ (52) Bomb Calorimeters. Formerly the calorimeters used for burning fuels in an atmosphere of oxygen were arranged for combustion at constant pressure, but since it was found that more reliable results could be obtained generally with apparatus maintaining a constant volume, the former type is not now much used. When the combustion takes place at constant volume, the vessel receiving the charge of fuel and oxygen must be designed to withstand a great pressure, and therefore on account of the massive con- struction required the vessel is called a bomb calorimeter. The essential part of such a calorimeter is the strong steel vessel or bomb similar to Fig. 155. It consists essentially of a steel shell S capable of resisting with safety a pressure of about 750 pounds. This shell is usually provided with a coat of enamel or a lining of nickel on the inside and is nickel-plated on the outside . The coating or lining on the inside is intended to resist corrosion and oxidiz- ing action during the combustion. The advantage of the nickel lining over the coat of enamel is that when it is worn out or broken it can readily be replaced and at much less expense than the enamel. The shell is closed at the top by an iron cover or cap which is to be made tight by screwing down on a lead washer with considerable force, using a long wrench. At the top of this cover or cap there is a conical seated valve, which is screwed in through the gland and stuffing-box G, by attaching a wrench at P. The valve and its seat are made of good nickel, as this metal is not easily oxidized. A wire electrode, which is well insulated from the cover, extends into the shell and conducts the electric current for firing the charge of fuel, which is placed on a platinum dish Fig. 155. — Section of a Bomb Calorim- eter. 164 POWER PLANT TESTING or crucible supported by another wire attached to the cover on the inside. Usually one gram of fuel is put into the dish to make a test for calorific value. A small iron wire (which was previously weighed) is then suspended over the dish between the electrode and the wire support for the dish. The cover should then be screwed on with a long wrench, the shell itself being held in a vise. The complete Mahler apparatus is shown in Fig. 156, showing the cylinder of oxygen 0, the pressure gage M, the calorimeter vessel D. The end of the conical-seated valve (Fig. 155) is attached by means of pipe connections, preferably Union for Attachment of Bomb when Filling with Oxygen. Fig. 156. — Complete Mahler Apparatus. flexible, to the union U and to the valve W, which because of the high pressure should be opened slowly and carefully, and allow sufficient oxygen to pass into the bomb to provide a consider- able excess above that actually required. The pipes for con- necting the bomb to the oxygen cylinder should connect also with a pressure gage as shown, so that the pressure in the bomb can be regulated. For the combustion of coal a sufficient volume of oxygen is admitted to Mahler bombs of the usual size to make the pressure in the bomb from 200 to 300 pounds per square inch. Now close the valve on the oxygen cylinder and the conical seated valve on the bomb, removing also the connections between the bomb and the oxygen cylinder. CALORIFIC VALUE OF FUELS 165 The fuel, especially if it is coal, should not be too fine, because if accidentally the oxygen should be allowed to go in a little too rapidly, some of sample of the coal will be blown out of the dish and will probably not be burned. The bomb should then be placed in the calorimeter vessel D, which should then be filled with a quantity of water previously weighed to fill it to about the level indicated in the figure. Place the calo- rimeter thermometer T into the vessel, being careful that the end will not be touched and broken by the stirrer or other parts, and then after agitating the water for a few minutes to establish a uniform temperature, the observations can begin. The temperature should be very carefully observed for five minutes and recorded minute by minute, to determine the rate of variation of temperature before combustion. Then the electric circuit should be made and the combustion will, of course, begin immediately; but some little time will be re- quired for the transmission of the heat generated to this water. Now take the temperature at the end of a minute after making the electric circuit; and continue observing the temperature every minute till it reaches its maximum value and begins to fall off regularly. Continue the observations for five minutes more to determine the rate of the fall of the temperature. The stirrer should be worked continuously but not too rapidly throughout the test, being careful, however, that the thermom- eter is not broken. When the observations have been finished, the conical-seated valve should be opened first to relieve the pressure and then the cover or cap can be unscrewed and removed. 1 The method described for the use of the Mahler bomb calo- rimeter can be applied also for determinations of calorific value of liquid fuels. Heavy oils can be weighed directly in the platinum dish or crucible, but light oils which are easily vapor- ized must be put into specially prepared glass bulbs which are broken to allow access of the oxygen, just before the cover is 1 Some engineers wash out the inside of the bomb with a little dis- tilled water to collect the nitric and sulphuric acids formed. Usually, however, this correction for acids is not made, as the heat liberated in the formation of the acids is usually less than one-third of one per cent, which, of course, would be subtracted from the calorific value obtained. If the reader is interested he will find the method explained with the necessary data in the Calorific Power of Fuels, by H. Poole, page 62. 166 POWER PLANT TESTING put on the bomb. If sufficient oxgyen is provided in every case there will be complete combustion in the calorimeter with no other refuse than the cinders remaining. A specimen calculation is given below : Weights, — coal, .0030 lb.; 1 water in calorimeter, 4.85 lbs.; water equivalent of bomb, etc., 1.10 lbs. Weight of iron wire, .0002 lb. Preliminary Observations. Beginning 60.23 F., 3 minutes, 60. 2 6° F., 1 minute, 60. 24 F., 4 minutes, 60. 27 F., 2 minutes, 60. 25 F., 5 minutes, 60. 28 F. Rate of variation before combustion, a — — : : — =.oi° F. Observations during Combustion. 6 minutes, 65. 45 F. 7 minutes, 67.29° F., 8 minutes, 67.38° F., max. 2 Observations after Maximum was reached. 9 minutes, 67.34° F., 12 minutes, 67.28° F., 10 minutes, 67.32° F., 13 minutes, 67.27° F., 11 minutes, 67.30° F. Rate of variation after maximum, a™=— — — = .022° F. 5 The rate of variation of temperature before combustion was for cooling the water and that after combustion was for a loss of temperature by the water. Evidently, then, the two rates are opposed in effect and the true average rate of variation is a„= : '- = +.006° F per minute. 2 1 The coal had been warmed at a temperature of about 240 to 280 degrees Fahrenheit before weighing, in a crucible over a Bunsen burner or an alcohol lamp to drive off the moisture. 2 Some engineers make a curve of temperatures (ordinates) and time (abscissas) and use for the final temperatures in the calculations the value from the curve, when the part of the curve representing the Gooling becomes a straight line. The difference in numerical values is usually very slight. CALORIFIC VALUE OF FUELS 167 Three minutes (5-6, 6-7, and 7-8) were required for com- plete combustion or for the water to reach the maximum tem- perature. Total cooling correction to be added to the observed rise in tempera- ture is therefore, 3X.006-.018 F. The total rise as corrected is, 7.10 + .018 = 7.118° F. The quantity of heat generated is, therefore, Q = (4.85 + 1.10) X7-n8 = 42.35 (B.T.U.) for .0030 lb. of coal; and from this result must be subtracted the heat of combus- tion of the iron wire .0002X3000 1 oro.6o B.T.U. The net value of the heat generated from the coal is, therefore, 42.35 - .60 = 41.75 B.T.U. A modification of the Mahler bomb calorimeter has been designed by Atwa- ter, 2 Fig. 157, and another by Emer- son, Fig. 158. The former consists of the shell of the bomb A, the cap C, screwed on numerous threads to the shell, and holding down the cover B. Into the vertical neck of this cover a screw E, 1 The calorific value of pure iron is about 3000 B.T.U. per pound, 2 Atwater, Bulletin No. 21, U. S. Dept. of Agriculture, |pWll^^^ Fig. 157- -Atwater's Fuel Calorimeter. 168 POWER PLANT TESTING holding another screw F, is fitted and is to be turned down tightly, a lead washer serving as " packing." A small passage for the admission of oxygen from G is opened and closed as required by turning the screw F operating a needle-valve. A wire H of platinum or other non-oxidizable metal passes through the cover B and is insulated from it by a collar of hard rubber. Another wire rod I is attached to the lower side of the cover and electrical con- 1 nection is made between the two wires H and I by a small iron wire stretched between them. A plati- num crucible provided for receiving the fuel is supported by a " screw " ring. Ball bearings of hard steel are sometimes placed between the cover and the cap to reduce friction when screwing down. Holes located in the sides of the cap are for the attachment of a long spanner wrench when turning down the cap. A hand-stirring device S is used for agitating the water in the vessel Q. The usual arrangement of the oxygen tank, pressure gage and tubing for charging an Atwater calorimeter is illustrated in Fig. 159. A pellet press for com- pressing samples of fuel into a suitable size to burn in the cru- cible of this calorimeter is shown in Fig. 160. Fig. 158 shows another form of bomb calorimeter (Emer- son) of which the Mahler is typical. It consists of a nearly spherical shell S, divided into two parts which are screwed together by the ring R. Powdered fuel is placed in the crucible C and is ignited electrically by the current passing through the water in the vessel Q from the terminal at A, then through an insulated contact point P in the bottom of the calorimeter to a small platinum or iron wire in the crucible C, which becomes heated by the passage of the current to a white heat, igniting Fig. 158. — Emerson's Fuel Calorimeter. CALORIFIC VALUE OF FUELS 169 the fuel. One end of this small wire is fastened to make elec- trical contact with the lining of the calorimeter, which in turn is connected electrically with the plug and terminal at B. The outer vessel is to be filled with water to the top. The stirring device consists of small propellers F, on a vertical shaft operated by a small electric motor M. Parr Calorimeter. It is not always convenient to secure a supply of oxygen under pressure for use in a Mahler bomb, and con- sequently another type ! of fuel calorimeter, known as Parr's, has found considerable use, especially for relative determinations in power plants. The results obtained can never be depended on to be as nearly accurate as determinations with one of the bomb type. Fig. 161 illustrates a simple form of Parr calorimeter. Sectional views of the two kinds of calorimeter vessels used are shown in Figs. 162 and 163. In the former the ignition is accomplished by dropping a hot wire through the neck into the shell A of 3 the calorimeter. The cover is attached to the shell by means of a threaded nut F. A charge for the bomb consists of about .004 pound of pul- verized coal from which the moisture has been driven off by warming for about an hour at a temperature of about 240 to 280 degrees Fahrenheit, and eighteen times as much by weight of sodium peroxide, which supplies the oxygen needed for com- Fig. 159. — Apparatus for Charging Atwater's Calorim- eter with Oxygen. Fig. 160. -A Pellet Press for Compressing Samples of Fuel. 170 POWER PLANT TESTING bustion. The charge should be well mixed by shaking the shell after the cover has been securely fastened. The cover must be attached very securely by turning up the nut with a long wrench, while the shell is held in a vise or in some similar manner, because Fig. 161. — Typical Parr Calo- rimeter. Fig. 162. — Parr Bomb for Hot Tube Ignition. Fig'. 163.— Parr Bomb for Elec- trical Ignition. there is a violent explosion when ignition takes place. When the hot wire is put into the tube in the long neck L, the cap R at the top must be struck quickly with a mallet before the wire cools in order to open the valve M, which opens inward into the shell and permits the wire to fall through. To be certain of obtaining a good result the wire should be heated almost CALORIFIC VALUE OF FUELS 171 to a white heat. The rise of the mercury in the thermometer will indicate when an explosion has occurred. The calorimeter is provided usually with wings or small propeller blades for agitating the water in the vessel O. The small pulley P (Fig. 161) shown at the top of the neck is used for turning the calorimeter bodily in the water when supported on Fig Parr Calorimeter with Motor Stirring Device. the pivot F, shown at the bottom of the figure. The water equivalent of the calorimeter is determined in the same way as for other fuel calorimeters. Fig. 164 shows the Parr calorimeter as designed for electrical ignition and with a stirring device operated by an electric motor. Allowance must be made in the calculations for the heat of combustion of the sodium peroxide, which for the propor- tions given is approximately 27 per cent of the heat generated. 172 POWER PLANT TESTING Carpenter's Calorimeter. There are few calorimeters in which the oxygen is supplied at constant pressure which are altogether successful. One of the best forms of an apparatus of this kind has been designed by Carpenter, especially for coal determinations. With this apparatus no thermometers are needed, as the rise in temperature is measured by the expansion of the mass of water sur- rounding the combustion chamber. This apparatus is shown in Fig. 165. It consists of a combustion chamber 15, provided with a removable bottom 17, through which the tube 23, supplying the oxygen, passes into the combustion chamber. Electric current for igni- tion is conducted through the wires 26 and 27. The removable bottom sup- ports also the asbestos cups or crucibles 22, used for holding the sample of coal to be burned. Just beneath the crucibles a silver mirror 38 is pro- vided to deflect the heat. The plug containing the wires and the oxygen pipe 23 is made of alternate layers of as- Products .of combustion leave the spiral tube, the parts of Fig. 165. — Carpenter's Calorimeter. bestos and vulcanite. combustion chamber through which are marked 28, 29, 30, and 31, into the small vessel 39, attached to the outer casing of the instrument, and are finally discharged into the air from a small hole 41, in the side of the vessel. The pressure in the chamber 39 is indicated by a manometer gage 40. The inner casing of the instrument 1, CALORIFIC VALUE OF FUELS 173 containing the water for absorbing the heat generated, is nickel- plated and highly polished to reduce radiation as much as possible. An open glass water gage 10 passes through the casings and extends below the water level. This water gage, with the scale attached to it, replaces the thermometer used in other calorimeters for measuring the rise in temperature. The scale is graduated to read inches, and it is calibrated usually by burning coke in the calorimeter, and determining thus the rise of the water level for a determinable weight of pure carbon. A calibration curve is usually supplied with the instrument. By moving the diaphragm 12, by means of the screw 14, the water level can be regulated, as well as the " zero " level in the glass water gage 10. A funnel 37 is provided for filling the instrument, and by inverting it, this funnel can be used also for draining. The instrument holds 5 pounds of water, and 2 grams of coal is the amount taken usually for a charge, requiring about twenty minutes for complete combustion; powdered coal is used. The asbestos cup should be heated in the flame of a Bunsen burner before it is weighed. The charge of dried coal should then be put into it and weighed again. The difference will be the weight of the coal used. Now put the charge into the combustion chamber 15, place the platinum ignition wire above the coal, connect wires 26 and 27 to the battery, and as soon as the heat generated causes the level of the water to rise in the glass water gage 10, open the valve in the pipe discharging oxygen into tube 23, and then by pulling down the platinum wires to touch the contents of the crucible, the coal will be kindled. At the same time the reading of the glass scale opposite the gage glass 10 must be observed and recorded. Progress of the combustion can be observed through the glasses 33, 34, and 36, arranged vertically over each other for this purpose. As soon as the combustion is complete observe the time and the reading of the scale opposite the glass water gage 10. The difference between this last reading and the one taken at the beginning of the test is called the " actual " scale reading. The correction for radiation is made by observing the reading of the scale of the water gage after the oxygen has been shut off, a length of time equal to that required for the combustion. The difference between this reading and the " actual " reading 174 POWER PLANT TESTING is to be added to the " actual " reading to obtain the corrected reading. By weighing the asbeslos cup after the test is finished and subtracting from this weight that obtained previously for its weight empty, the weight of ash is determined. In order that Carpenter's calorimeter may give determi- nations of heat values that are at all accurate, all the air must be removed from the water used, as the presence of air x will affect the relative level of the water in the gage glass for a given rise in temperature. The oxygen must also be supplied at a constant pressure, maintaining the pressure indicated by the manometer gage at the value for which the calorimeter was calibrated. Most calibration curves are made for a pressure of about 10 inches of water. The apparatus can be made to give good comparative results when operated carefully and " according to directions." In general, the statement is often made that coal calorimeters intended for combustion at constant pressure will usually give nothing more than " faint approxi- mations " to correct results. The same can be said, however, of nearly all the calo- rimeters if they are not very carefully manipulated. When making cal- orific determinations of coal the distinction must be carefully made between results obtained per unit weight of combustible or per unit weight of coal (including mois- ture and ash). Junkers Calorim- eter for Liquids and Gases. An apparatus for determining the calorific power of gases is shown in Fig. 166 and Fig. 167. 1 About 2 inches of kerosene oil is usually put into the glass water gage to prevent air from coming into contact with the water. Jig. 166. — Junkers Calorimeter with Auxil iary Apparatus. CALORIFIC VALUE OF FUELS 175 The gas flowing in pipes at the left (Fig. 166) passes through the meter A, then through the regulator B, and is burned in a type of Bunsen burner C, in the lower part of the calo- rimeter. This instrument consists of a cylindrical copper vessel through which water is constantly circulating. The gases from the Bunsen flame in the calorimeter pass up through the hollow central portion of the instrument and near the top are deflected downward through a group of small tubes arranged in an annular ring between the outside and inside walls of the calorimeter. Around these tubes water is kept circulating continuously to absorb the heat generated by burning the gas tested. After leaving these tubes the products of combustion dis- charge first into a chamber 31 (Fig. 167) and then into the air through the flue D. In order to keep the flow of water as regular as possible it is brought from the supply pipe G into a small reservoir in which the water is kept at a constant level (constant head) by means of an overflow pipe H. The water supplied to the calorimeter passes down through the pipe 6, through a valve at I, and discharges at K, running into a vessel in which it is weighed. A graduated tube Q „ (Fig. 166), is provided to collect the moisture from the steam that is condensed. The condensed steam collects in the combustion chamber 31 and escapes through the tube 35. A thermometer N, in a cup near the valve I, indicates the temperature of the water entering the calorimeter, and one at M shows the temperature of the water leaving. The tem- perature of the products of combustion (burned gases) is indicated by the thermometer 0, in the gas flue. The calo- 67. — Section of Junkers Calorimeter. 176 POWER PLANT TESTING rimeter is provided with an air jacket and is covered with sheets of copper, nickel plated and highly polished so that the radiation loss is considered negligible. If, then, the flow of water and the rate of burning the gas are regulated so that the temperature of the products of combustion 1 as indi- cated by the thermometer at 0, is the same as the tempera- ture of the air surrounding the calorimeter, practically all the heat generated by the burning gas is absorbed by the water. The rise in temperature of the water is observed by reading the thermometers at N and M. Now if the temperatures of the water at the inlet and the discharge have been observed and the weight of the water flow- ing has been determined while, for example, a cubic foot of gas 2 has been burned, then "the difference in temperature in degrees Fahrenheit times the weight of water in pounds gives the heat value in British thermal units per cubic foot of gas. For some calculations relating to the efficiency of heat engines it is desirable to know the number of heat units repre- senting the calorific value of the gas when the steam formed in the combustion is not condensed but is carried off with the products of combustion. To determine this value, some- times called the "low heat value" of the gas, the latent heat at atmospheric pressure of the amount of condensed steam collected must be subtracted from the value obtained by multiplying together the rise in temperature and the weight of water used. This correction is usually about two per cent. This apparatus, although it operates by a constant pressure method, gives very satisfactory determinations. Fig. 168 shows a balance and lamp attachments for a Junkers calorimeter set up for determining the heat value of liquid fuels like gasoline, kerosene, crude oil, etc. The heat 1 Heat lost in products of combustion is explained in Stillman's " Engineering Chemistry," pages 161-165. See also in this book pages 196 and 224—226. 2 In order that results can be compared, it is customary to reduce the calorific power to terms of heat units per cubic foot of gas at a " stand- ard " temperature and pressure as for example, 32 degrees Fahrenheit and 14.7 pounds per square inch, assuming that the volume of the gas is proportional directly to the temperature and inversely to the pressure, or else by determining the specific volume of the gas to calculate the heat units per pound. CALORIFIC VALUE OF FUELS 177 Fig. i 68. — Balance and Lamp for Burning Oils in Junkers Calorim- eter. generated is measured in the same way as when gas is burned and the weight of oil used is determined by weighing on the balance to which the lamp L is attached. Calorific Values from Chemical Analysis. Dulong stated a long time ago that the heat generated by burning any fuel was equal to the sum of the " possible heats " generated by its component elements, less that portion of the hydrogen which combined with the oxygen in the fuel to form water. When hy- drogen and oxygen exist together in a compound in the proper proportions to form water, the combination of these elements has no effect on the calorific value of the compound . No w the calorific value of a pound of carbon is 14,600 B.T.U. and of a pound of hydrogen is 62,000 B.T.U., so that by Dulong's formula, the calorific value of a pound of fuel would be stated, using these values, as x = 14,600c + 62,000 (H— — ) + 4,000 S, o here C, H, and S are respectively the weight of the carbon, hydrogen, oxygen and sulphur in a pound of fuel. As the result of testing the forty-four different kinds of coal with his bomb calorimeter Mahler. developed the following formula, using the same symbols used in Dulong's, x = 200.5 + 67511-5,400. Using this latter formula Lord and Haas 1 determined for a series of 40 Pennsylvania and Ohio coals which they had analyzed and computed the calorific values that the maximum differences between the calculated results and the determinations with a bomb calorimeter were from 2.0 to — 1.8 per cent. With fuels like coke, charcoal, and anthracite coal, in which the Trans. American Inst, of Mining Engineers, Feb., 1897. 178 POWER PLANT TESTING content of volatile matter is small, the calorific values calcu- lated from an accurate analysis are usually in very close agreement with accurate calorimeter tests, but with coals having more than 20 per cent of volatile matter there is likely to be considerable error. Proximate Analysis of Coal. For all tests in which an analysis of the coal or its calorific value is to be determined, it is very necessary that the sample to be tested be selected with the greatest care. The method generally adopted for obtaining a fair sample is known as ' ' quartering, ' ' as explained in the Rules for Conducting Boiler Trials adopted by the American Society of Mechanical Engineers. (See page 210.) The utmost care must be taken that the amount of moisture in the sample received for analysis is the same as that in the original condition, or more specifically in a boiler test, at the time when the coal used in the test was weighed. For this reason samples of coal should be transported and stored in air-tight preserving jars or similar vessels. It is not unusual, moreover, to find that coal containing 10 per cent of moisture will lose as much as 2 or .3 per cent of its moisture in the process of careless sampling, crushing, resampling, etc., while if it is allowed to remain exposed to atmospheric conditions for a considerable time in a warm room as much more may be lost by evaporation. Moisture Determinations. Determinations of moisture are made by careful engineers as soon as permissible after the sample has been procured and with the coal in as large pieces as possible. A good method for making the moisture determinations for anthracite and semi -bituminous coals is to place a weighed sample on top of the hottest part of a boiler setting or a flue and weigh it again after drying for twelve hours. A good laboratory test for the same kinds of coal is to place a sample weighing about .05 pound in an air or sand bath for one hour at a temperature of from 220 to 230 degrees Fahrenheit, and weigh again when the sample is cool. 1 The difference in weight is the amount of moisture in the sample. 1 While the sample is cooling there is the possibility that it may absorb moisture from the air unless it is placed in a desiccator till cool. It is difficult to get accurately the weight of hot bodies on account of the air currents produced. CALORIFIC VALUE OF FUELS 179 When coals taken from the mines west of the Pittsburg dis- trict or other coals containing inherent moisture are to be tested for moisture, a different method must be adopted. The sample of coal is to be spread out in a thin layer and exposed for about four hours to the atmosphere of a warm room. The difference between the weighings before and after this exposure is the weight of surface moisture. Then crush all of this sample to produce coarse grains measuring not more than one-sixteenth inch on a side, mix it thoroughly, and select from it a quantity weighing from .05 to .125 pound, and dry it for one hour, in an air or sand bath in which the temperature is maintained at from 240 to 280 degrees Fahrenheit. Now weigh it again and then continue the heating between these limits, weighing every hour till two of these weighings are the same or the weight begins to increase due to oxidation of the coal. The difference between the original and the minimum weight of this sample is called the moisture in air-dried coal. The sum of the percentage of surface moisture plus the percentage of moisture in air-dried coal is the total moisture. When making the determination for moisture too much care cannot be exercised to remove all of it. Determination for Volatile Matter. The amount of volatile matter in coal is determined usually with a sample as originally received without drying. A suitable sample should weigh about .0035 pound— about 1.5 grams — which should be pulverized in a mortar and put into a clean platinum or porcelain crucible. Then weigh the crucible with the coal it contains in a balance sensi- tive enough to weigh accurately to one-thousandth pound. After this weighing has been done as carefully as 'possible, a cover like the ones usually provided for crucibles of this kind is to be put on to cover it tightly. Now heat the cru- cible for 3^ minutes over a Bunsen burner, keeping the crucible at a bright red heat, and then immediately, without cooling, for 3^ minutes over an air-blast lamp. After cooling weigh, and the difference between the weighings is the sum of the volatile matter and the total moisture. Determination of Fixed Carbon. The determination of the fixed carbon is made by heating again the sample used for the determination of volatile matter. 1 Now, however, the cover 1 If the sample tested is a bituminous coal it will be observed that coke has been formed by the removal of the volatile matter. 180 POWER PLANT TESTING of the crucible is to be removed and heat is to be applied, preferably with a Bunsen burner, until all of the carbon is burned; that is till the weight becomes constant. If the time available for making the test is limited an air-blast lamp may be used instead of the Bunsen burner. The rate of combus- tion can be increased by stirring the sample from time to time with a platinum wire. What now remains in the crucible is the ash. The difference between the weight found after the volatile matter had been driven off and the weight of the crucible and the ash is the weight of fixed carbon. Weight of ash should be determined by weighing the crucible again when empty. If sulphur and phosphorus determinations are required they should be made by an expert chemist. CHAPTER IX FLUE GAS ANALYSIS Flue Gas Analysis. The analysis of flue gases in con- nection with tests of steam boilers gives a valuable means for determining the relative value of different methods of firing and of different types of furnaces. Errors in the analysis of flue gases are most often due to the inability to secure an average sample of the gases in the different parts of a flue or chimney. The composition is likely to vary considerably even during short intervals, and it is therefore desirable to adopt some method of sampling which will permit collecting the sample slowly for a considerable period. A very simple and convenient sampling apparatus is shown in Fig. 169. The sample of gas is taken from the flue or chimney through the pipe shown at the top of the figure. This pipe extends well into the flue and has usually a long slot cut into one of the sides so that a better sample of gas can be taken than if it were taken at the end of the pipe. The end of the pipe outside the flue is connected by means of a short rubber tube to the sampling bottle. A valve V should be put as near as possible to the end of such pipes, so that they can be closed up when the sampling bottle is removed for analysis. If a valve is not provided or is not closed, the suction in the flue will draw air into the pipe, and when again connecting up the sampling bottle this air must be removed before a true sample can be taken. The sampling bottle is preferably one with a wide neck, closed with a cork through which two glass tubes pass into the bottle, one reaching nearly to the bottom and the other entering only a little beyond the bottom of the cork. The longer tube can be connected to an aspirator or ejector, (Fig. 170) a water-jet exhaust or any similar device producing a steady suction. 181 182 POWER PLANT TESTING Small aspirators or ejectors operating on the principle of an injector with a small stream of water which entrains the gases is very convenient for collecting samples con- tinuously. Water enters through a vertical nozzle, entrains air or gas drawn in through the side open- ings and the mixture of atomized water and air is discharged with consider- able velocity through the forcing tube at the bottom. If an aspirator is not available, the bottle may be filled with mercury and by making a siphon of the rubber tube attached to the longer of the two tubes in the bottle, the mercury can be gradually drawn out and gases drawn in. By adjusting the valve V the jrate of flow of the gases into the bottle can be regulated. Mercury is too heavy to use in a very large sampling bottle and therefore water is often used instead, with the dis- advantage, however, that the water will probably absorb some of the con- stituents of the gas. On this account very little water should be left in the bottle, with the sample of gas. If the water is saturated with gases, as it will be from long use, this precaution need not be observed. P = Pipe V= Valve Fig. 169. — Sampling Bottle for Collecting Flue Gas. FLUE GAS ANALYSIS 183 ^1H The bottle and the tubes must .be completely filled with water before beginning to take the sample, because any air left in them will remain in the bottle to be mixed with the sample of the gas. If the end of the cork going into the bottle is made slightly conical it will be easier to avoid entrapping bubbles of air at the top of the bottle. This type of sampling bottle can be used also very con- veniently by reversing the connections, of its tubes; that is, by attaching the long tube to the pipe entering the flue, and then turning the bottle upside down. The water will then run out through the shorter tube and the gas will be drawn in to fill the Water bottle. A portion of the gas can be re- moved from the sampling bottle into the measuring burette or tube required for making the analysis of the flue gas by connecting the short tube of the sampling bottle to the burette or to some other part of the gas analysis apparatus as may be required. If now the rubber tube connected to the longer tube of the sampling bottle when disconnected is put into a pail well filled with water, then as the gas is withdrawn from the bottle, water will be drawn from the pail to displace it. One of the advantages of this appa- ratus is that it can be easily made from the materials obtainable in almost any town or village. A two-quart preserving jar with a rubber cork to fit and tubes of glass, brass or iron can be used to make up a very good sampling bottle. Since there is nearly always a great variation in the com- position of the gases in the various parts of a flue or chimney it is not very likely that a tube open at the end and having a long slit like the one described in the preceding paragraphs will give a " fair " sample. Obviously most of the gas will enter the slot in that portion of its length nearest the col- lecting apparatus. Another device often used for a sampling Discharge Fig. 170. — Water-jet As- pirator or Ejector. 184 POWER PLANT TESTING tube consists of a horizontal pipe into which a number of branch tubes are fitted. These branch tubes are arranged so that the openings at their ends will take samples from the different parts, of the flue or chimney in which they are placed. Sometimes these branch pipes are also slotted or are perforated with small holes drilled into their walls. Fig. 171 shows an arrangement of sampling tubes for col- lecting flue gas recommended by the American Society of Mechanical Engineers. It consists of a series of standard one-fourth inch pipes, all open and otherwise alike at the ends and of equal lengths. Each pipe is to be placed with one end in a shallow box or receiver made of gal- vanized sheet iron. It is convenient usu- ally to make the depth of this sheet -iron box about the same as that of a course of bricks. These tubes should be arranged so that the open ends will be at points well distributed over the area of the flue which in the figure is marked A. The other ends, which are also open, are to be enclosed in the receiver B. The receiver is connected by four tubes C, C with a mixing box D. The flue gases drawn from it should be well mixed and should represent an average sample from the flue of chimney from which they are taken. Tests have shown that two such sampling devices placed in the same flue one above the other about a foot apart, will furnish samples of the flue gases showing the same composition when analyzed. A very convenient type of sampling bottle is shown in Fig. 171. — A. S. M. E. Arrangement of Sampling Tubes for Flue Gas. FLUE GAS ANALYSIS 185 Fig. 172. It consists of a bottle with an opening at the bottom (tubulated), and is provided with a cork at the mouth through which a glass funnel F and a tube are passed. The bottle con- tains water and light oil, and when it is filled there will be a layer of about 4 inches of the oil over the water. The tube O at the top is to be connected to the sampling tubes in the flue and the sample is taken in by opening the valve in this tube and also the one at the bottom of the bottle. The water drains off at the bottom and is replaced by the sample of gas. The glass funnel is used for pouring water into the bottle and in this way expelling the gas needed for analysis. The gas is thus made to pass out through the same tube through which it is drawn in. Another type of sampling apparatus used by many engineers consists of two galvanized-iron tanks, each about 2 feet high, and about 5 inches in diam- eter. On the side of each of these tanks and close to the bottom a valve is attached by soldering. These two valves are connected by a piece of heavy rubber tubing. One of the tanks is closed at the top, and a small stopcock or valve is attached to the cover. The other tank is open at the top. The apparatus is used for collect- ing gas by filling to ' ' overflowing ' ' the tank with the closed top with water from the other tank by raising the latter so that the level of the water in it is above that of the water in the closed tank. By means of rubber tubing the stopcock or valve on the closed tank is then connected to the sampling tubes in the chimney or flues. Meanwhile the open tank is held at such an elevation that the water will not run back into it and create a vacuum in the closed tank. After this connection has been made the stopcock and valves are to be opened again, so that when the open tank is placed below the level of the closed one, the water will flow into the open tank and fill the other one with gas. This operation should be repeated several times before the sample is carried away Fig. 172. — Another Type of Sampling Bottle. 186 POWER PLANT TESTING to be analyzed, so that there can be no doubt that none of the air in the sampling tubes entered the sample to be analyzed. This water can be used over and over again, and when it has become saturated with gas it is practically as good as mercury for use in collecting the gases. When a sample of flue gas is taken from the flue at a considerable distance from the furnace it is likely to become mixed with air leaking through the brickwork of the boiler setting, and the analysis will not show the true relations between the volumes of the so-called flue gases and the excess of air. To prevent as much as possible this leakage of air the joints in the masonry must be examined and repaired if necessary and the sample must be taken as near as possible to the fire, bearing in mind, however, that they must be drawn very slowly from the hot flue in order that they will be cooled down gradually to avoid dissociation. For hot gases an earthen- ware collecting vessel may be used if a glass bottle is likely to be broken. If dissociation occurs in the sample the analysis may show results entirely different from the true composition of the gas in the flue. It is also difficult to prevent the entrance of air into a flue through the bearings of dampers, and whenever it is possible the sample of flue gas should be obtained between the furnace and the damper. At high tem- peratures sampling tubes of other metals than platinum 1 or nickel are not quite satisfactory, since by their oxidation they abstract the oxygen from the gases passing through them. Apparatus for the Analysis of Flue Gases. Samples of flue gases contain in varying amounts carbonic dioxide (car- bonic acid), oxygen, carbonic oxide, nitrogen, unburned hydro- carbons, and occasionally some free hydrogen. For the data which an 'engineer usually requires it is not necessary to deter- mine by direct analysis more than three of these; carbonic dioxide, C0 2 , oxygen, 2 , and carbonic oxide, CO. The determination of carbonic oxide, CO, with the facilities and the portable apparatus ordinarily available in engineering laboratories is often somewhat doubtful. Some authorities state that there is rarely more than a trace of carbonic oxide to be found in the gases from combustion in the ordinary 1 Porcelain and annealed glass are also satisfactory materials to use for making sampling tubes for very hot flues. FLUE GAS ANALYSIS 187 types of furnaces. When more than one per cent of carbonic oxide is shown by the analysis and the carbonic dioxide deter- mination is not over 14 per cent, it may usually be assumed that a large part of what is taken to be carbonic oxide is oxygen which was not absorbed by the proper reagent. In the following table a set of analyses of flue gases is shown. The determinations were made by Scheurer-Kestner with coal from Ronchamp. Other analyses of flue gases may be checked by a comparison with this table. Thus when the analysis shows about 8.2 per cent CO2, the sum of the per- centages of C0 2 and 2 will probably be between 19 and 20. PERCENTAGE COMPOSITION OF FLUE GAS C0 2 O2 CO N Hydrocarbons 8.2 "•3 . 2 79.8 •5 10.8 9.0 . 2 79-7 •3 12.9 5-5 . 2 80.3 1 . 1 13-4 4-4 . 2 80.2 1 .8 14.6 2.8 ■3 80.6 1.7 In the portable apparatus used by engineers for the analysis of flue gases a separate pipette or treating tube is provided for each reagent, and the chemicals used are of greater strength than the reagents used by some chemists. The following reagents give satisfactory results in a portable apparatus: (1) For absorbing CO2 a solution of one part of potassic hydrate (KOH) or caustic potash dissolved in two parts by weight of water is generally used. (2) For absorbing 2 either an alkaline solution of pyro- gallic acid or sticks of phosphorus are employed. The alkaline solution of pyrogallic acid is prepared by mixing together preferably in the absorption pipette or treating tube, to prevent access of air, 5 grams of pyrogallic acid powder and 100 cubic centimeters of potassic hydrate (KOH) solution prepared as explained above. 1 (3) For absorbing carbonic oxide a hydrochloric acid solu- tion of cuprous chloride is used. This is prepared by dis- solving about 10 grams of cupric oxide in from 100 to 200 1 When the temperature is lower than about 55 degrees Fahrenheit this reagent does not give satisfactory results. 188 POWER PLANT TESTING cubic centimeters of concentrated hydrochloric acid. This solution must be allowed to remain in a bottle tightly closed and well filled with copper wire or gauze, until the cupric chloride is reduced to cuprous chloride. In this latter state the liquid will be colorless. Exposure to the air produces a brown color, indicating the cupric state. After a time these reagents must be re- placed by new solu- tions. The potassium hydrate solution may be used till each vol- ume has absorbed forty volumes of C0 2 . Py- rogallic acid solution deteriorates rapidly and each volume should be expected to absorb only one or two volumes of O2. Cuprous chloride will absorb an equal volume of CO. Portable devices for the analysis of flue gases are generally known as " Orsat " apparatus. Of these there are various types. The one devised by Fisher, shown in Fig. 173, has been used extensively. It consists of a measuring-tube M surrounded by a water- jacket, and set of absorption pipettes, A, B, C, each filled with a reagent. Each of these pipettes (Fig. 174) consists of two glass vessels connected by a U-shaped glass tube at the bottom. One end of these pipettes is joined by means of a short piece of rubber tubing to a glass yoke T, which is designed for attachment at one end e to a tube leading to the sampling-bottle and at the other end to the measuring- Fisher's "Orsat" Apparatus. FLUE GAS ANALYSIS 189 tube M. A water bottle W is connected by a flexible rubber tube P to the bottom of the measuring tube. Before a sample of gas is taken into the apparatus for analysis certain adjustments must be made. In the first place, the reagents in the pipettes must all be brought to a standard level at some arbitrary point, usually indicated by a scratch on the glass tube just below the short rubber tube connecting it to the yoke. This adjustment is accomplished by opening one at a time the valves at the tops of the pipette and removing the air (or the gas as the case may be) from it by lowering the water level in the measuring tube. The position of the water bottle determines, of course, the level of the water in the measuring tube. When all the air and gases remaining from a previous test have been expelled from the apparatus by filling the measuring tube and the tubes comprising the yoke with water, one of the tubes in the sampling bottle should be connected to the apparatus at e and by opening the valve in the yoke at that end and lowering gradually the level of the water in the measuring tube M a sample is obtained for analysis. This sample must be measured by the scale on the measuring tube at atmospheric pressure 1 to the nearest tenth of a cubic centimeter. 2 After the measurement has been made and recorded if the cock in the tube leading to the absorption pipette containing the reagent for absorbing C0 2 is opened and then the water bottle is raised, all of the measured sample of gas can be forced over into Fig. 174. — Pipette of Fisher's "Orsat" Apparatus. 1 The pressure of the gas is " atmospheric " when the water bottle is held so that the water in it and that in the measuring tube are at the same level. 2 The scales of practically all measuring tubes used for gas analysis apparatus are graduated in cubic centimeters. 190 POWER PLANT TESTING the pipette. The reagent acts more rapidly on the gas if the water bottle is raised and lowered a few times. This movement of the water in the measuring tube agitates the gas and also the reagent and exposes more of the gas to the direct action of the absorbent. To increase the surface over which the reagents can act, the pipettes are filled with small glass tubes. When the gas has been in the first pipette for about a minute, it should be drawn back into the measuring tube with the level of the reagent brought back to the mark where it was originally, and the cock closed. The pressure of the gas is again made atmospheric and its volume measured. Now repeat this operation until two or three measurements are obtained which are alike, showing that all the CO2 has been absorbed. Then the cock on the tube leading to the pipette containing the absorbent for oxygen can be opened, the gas forced over, and measured several times till a constant volume is observed. Finally the gas is passed into the third pipette for absorbing CO, repeating the operation of measuring as with the other pipettes. The absorption of oxygen will usually require considerably more time than for the determinations of the carbonic acid (C0 2 ) and the carbonic oxide (CO), so that it is unnecessary to make a measurement of the volume of the gas till after the gas has been exposed about three minutes to the reagent. Soft rubber bags (see Fig. 174) should be attached by means of glass tubes to the corks shown in the pipettes on the farther side in Fig. 173 and are provided to protect the reagents from absorbing oxygen and carbonic oxide from the air. Both of these reagents will absorb oxygen from atmospheric air, so that the access of fresh air must be prevented. The rubber bags are useful also for the purpose of producing alternately, with the pressure of the hand, suction and pressure for agitating the reagents. A form of Orsat apparatus particularly suitable for portable use, and in which renewals of broken parts can be cheaply and easily made, is illustrated in Fig. 175. This apparatus, designed by Professor John R. Allen and the author, 1 is also particularly suitable for the use of engineers because the pipettes containing the reagents can be removed from the apparatus very easily 1 Made commercially by the Bausch & Lomb Co., Rochester, N. Y. FLUE GAS ANALYSIS 191 tor changing solutions. They can be emptied, refilled, and replaced in a very short time. The absorption pipettes (Fig. 176) are made simply of two glass test-tubes, the smaller one inside the larger one. The small test-tube is held as it were inverted, and has a glass " capillary " tube fused into its closed end. The outer tube is closed at the top by a rubber stopper through which the capillary portion of the inner tube passes. Very small glass tubes are placed in the inner test-tube to Fig. 75. — Allen-Moyer Gas Apparatus. Fig. 176. — Absorption Pipette of Allen-Moyer Gas Apparatus. increase the surface for the action of the reagent. The pipette is held in place 'by means of a hard-rubber disk supported on brass screws. The level of the reagent in the pipette is estab- lished when the air in the inner tube is drawn out and the level of the liquid rises to a mark on the glass capillary tubes. In the ustial forms of the Orsat apparatus the pipettes invariably become leaky at the stopper provided for emptying. In the Allen-Moyer apparatus there is no opportunity for such leakage. When the sample of the gas is passed through the capil- lary tube into the inner test-tube, the reagent is displaced 192 POWER PLANT TESTING and raises the level in the outer test-tube. Similarly when the gas is passed back into the measuring tube the level falls in the outer tube, rises in the inner one, and is brought back to the original level at the mark on the capillary tube. Otherwise the method of operation is the same as described for Fisher's apparatus (Fig. 174). In this apparatus the measuring tube M and water bottle W are of the same type as those used in Fisher's design. The yoke is also similar, although usually made of hard rubber to avoid breaking it in transportation. It has also spring pinch- cocks instead of ground-glass cocks. When glass cocks are used by inexperienced persons all sorts of difficulties are likely to result, particularly that it often happens that they are not pressed into their seats tightly enough to prevent the loss of gas or the entrance of air. Sometimes the glass cocks will be put into their seats so tightly that it is impossible to move them without breaking. These difficulties, although met often enough in laboratory work, are still more frequently observed in practice. Coefficient of Dilution. The coefficient of dilution is the ratio of the volume of the air supplied to the volume theoreti- cally necessary to provide the oxygen required for combustion. It will now be shown how this coefficient can be calculated from an analysis of the flue gases. Oxygen when combining with carbon to form carbonic dioxide produces a volume equal to itself, thus, c+o 2 =co 2 , and in forming carbonic oxide produces twice the volume 2C+0 2 = 2C0. Now if we use symbols to designate the percentages by volume of the gases in a sample of flue gas as follows : a is the percentage by volume C0 2 , b is the percentage by volume 2; c is the percentage by volume CO, d is the percentage by volume N (nitrogen). Then the volume occupied by the free oxygen in the air before combining with the carbon was a+b+|c per cent, while that required is obviously a+lc per cent. FLUE GAS ANALYSIS 193 The coefficient of dilution is therefore, a + b + |c -TJ+- •••;••• (53) In a little different form the reactions given in the last paragraph may be stated (i) for carbon burned to CO2, 2C + 20 2 = 2C0 2 , (54) (2 vols.) (2 vols.) (2 vols.) 24 64 88 and (2) for carbon burned to CO, 2C + 2 = 2CO. ..... (55) (2 vols.) (1 vol.) (2 vols.) 24 32 56 The ratio of the volume of the carbon vapor burned to CO2, to the volume burned to CO is the same as the ratio of the volume of CO2 in the products of combustion is to the volume of CO. Further since the ratio of volumes of the carbon vapor is obviously the same as the ratio of the corresponding weights, we may say that in a mixture of gases, the ratio of the weight of carbon required to produce the C0 2 to the weight needed for the CO is equal to the ratio of the volume of C0 2 in the mixture to the volume of CO. The atomic weights of carbon and oxygen show (equation (54)) that a volume of oxygen is 2§ (64/24) times as heavy as an equal volume of carbon vapor. It follows then that for burning one pound of carbon to CO2, 2§ pounds of oxygen are required. The other reaction (55) showing the combination of carbon and oxygen to form CO shows with the same reasoning that 1^ pounds of oxygen are required to burn one pound of carbon to CO. In the general case we are considering and using symbols a and c respectively as before to represent the volumes of C0 2 and of CO in the flue gases, then the weight of the carbon burned to C0 2 is to the weight burned to CO as a is to c and if a +c Q represents the weight of carbon burned to C0 2 and the a -He 194 POWER PLANT TESTING weight burned to CO, then the weight of oxygen required per pound of carbon is 2§( ) + I 3\ ) and the weight of air 1 per pound of carbon is in pounds, f hfc)+^fc)] (56) If z is the percentage by weight of carbon then the weight of air in pounds per pound of coal is ^K^c+'^c)}- •■■■■' W Volumetric analyses of the flue gases can be used also to calculate the weight of the products of combustion (flue gases) per pound of coal burned and also the heat units lost in these gases. For this calculation the relations of the molecular weights are important. Molecular weight of co 2 = = 44 2 = = 32 co = = 28 N 2 = = 28 c = = 12 Now in a sample of x pounds of flue gases in which the percentages by volume are represented by the symbols a, b, c, d, the relative percentages by weights of the constituents will be co 2 = 44a. x ' 2 = = 32b. X co = 28C_ X ' N 2 = 2 8d_ X and we can write further 1 Weight of air may be checked with Peabody's and Jacobus' equa- tions (61, 62 and 63 1 ), pages 225-228. FLUE GAS ANALYSIS 195 Weight of carbon burned to CO2 in x pounds of gas 1 ^ = — X44a= 12a. 44 Weight of carbon burned to CO in x pounds of gas Total weight of carbon burned in x pounds of gas= 1 2 (a +<;). Total weight of carbon burned per pound of gas = 12 -. x Total weight of gas generated per pound of carbon (a+c) Of this total weight of gas as expressed by the last equa- tion the constituents are distributed in percentages by weight as follows : Weight of CO2 in samples per pound carbon burned, 44a. x or we may write and similarly, ixj X - 123 ■■(a+c)' co 2 = = W\ 44a 12(0+1 - lb . o 2 = =w 2 - ft , ib -^ 1 2 (a+c) co = = wz -■ 28c n. 1 2 (a + c) N 2 = = w 4 -- 2&d -lb i2 (a+c) Total weight of gases iv g per pound of coal burned, if there is z per cent ? of carbon in the coal, is in pounds 2(440 + 326 + 28c + 28d) w e = -, (58) g 1 2 (a +c) 100 ° ; 1 It may be assumed for very approximate values that z= i — y where y is the per cent of ash in the coal. 196 POWER PLANT TESTING Now if we represent by t f and t a the temperatures respect- ively in degrees Fahrenheit of the gases in the flue and of the air entering the furnace, then the heat lost in the flue gases Q g per pound of coal is, inserting values of specific heats, l Q == (.2IJWi + .2I']W2+.245W 3 +.244W 4 )(t i -Q. fi IOO The total heat generated Q by the more or less incomplete combustion of one pound of coal when there are a and c per- centages by volume respectively of CO2 and CO in the flue gas is Qo =— (-^7X14,600+-^- X 44 oo)(B.T.U,). . (59) ioo\a+c a +c / Since the heat of combustion of carbon when burned to C0 2 is approximately 14,600 and when burned to CO is about 4400 B.T.U. Finally, if Q p is the heat from perfect combustion or the " calorific " value in B.T.U. of a pound of coal, then the efficiency Q of the furnace 2 = — . The percentage of heat from perfect com- ~p (i bustion lost in the flue gases =— . Qp Recording Apparatus for Determining C0 2 . A typical apparatus for making a continuous record of the percentage by volume of carbonic dioxide in gases is shown in Fig. 177. The gas is taken to the instrument from the side flue or last com- bustion chamber of each boiler or furnace to the inlet pipe D and is drawn through the machine by a special water aspirator 0, fixed to the top of the instrument by means of the standard T. After actuating the aspirator 0, a portion of the water flows to the small tank L, which serves as a pressure regulator, and is provided with an overflow tube R. From this tank the water 1 This method of finding the heat escaping in the flue gases can be used to correct determinations made with the Junkers calorimeter (page i-7 6) when the products of combustion are discharged at a temperature different from that of the rodm. 2 For the calculation of related quantities see "Heat Balance" (A.S. M.E. Rules), page 214; also Allen and Bursley's Heat Engines, pages 64-65, FLUE GAS ANALYSIS 197 enters the tube H in a fine stream, which is adjusted by the cock 5 and gradually fills the vessel K. This vessel consists of an upper and a lower compartment, the two being in communi- cation through a tube erected in the upper chamber and reaching nearly to the top. Water, which enters this vessel K through the tube //, gradually fills the upper chamber and thus com- presses the air contained in it. This pressure is transmitted to the lower compartment through the communication tube mentioned above, and acts upon the mixture of glycerine and water with which this is filled, driving it out into the calibrated tube .C. When the rising liquid in C has reached the inlet and outlet to this vessel, no more gas can enter the calibrated tubes and the aspira- tor will now draw the gas through the seal F. Before the liquid can close the central tube in C, the gas must overcome the slight resistance offered by the elastic bag P, and is thereby forced to assume atmos- pheric pressure. When the liquid has sealed the lower open end of this central tube, exactly ico cubic centimeters of flue gas are trapped off in the outer vessel C and its companion tube, under atmospheric pressure. As the liquid rises, the gas is forced through the thin tube Z into the vessel A, which is filled with a solution of caustic potash for absorbing carbonic dioxide. The gas remaining gradually displaces the potash solution in A, sending it up into the vessel B. This has an outer jacket, filled with glycerine and supporting a float N. Through the center of this float reaches a thin tube, through which the air in B is kept at atmospheric pressure. The float is suspended from the pen gear M by a silk cord and counter-balanced by the. Fig. 177. — Recording C0 2 Apparatus. 198 POWER PLANT TESTING weight X. The liquid in B forces a portion of the air through the central tube in the float, and then raises the latter, causing the pen lever to swing upward, carrying the pen, Y, with it. The mechanism is so calibrated and adjusted that the pen will travel to the top, or zero line, on the chart when only- atmospheric air is passing through the machine, and nothing is absorbed by the potash in A. When there is any carbonic dioxide in the gas it is absorbed by the potash in A, and not so much of this liquid would be forced up into the vessel B. .A4 ^ From Boikr Flee Fig. 178. — C0 3 "Weighing" Apparatus (Econometer). The float would not then cause the pen to travel up so high on the chart, in proportion to the amount of CO2 absorbed. Another apparatus for making continuous determinations of C0 2 in flue gases is shown in Fig. 178. Gas from the boiler flue enters at M , passes through an excelsior filter where dust is removed, and then goes on through tubes leading it through glass vessels containing cotton wool and calcium chloride. After being cleaned it passes in the direction of the arrows through the valve B into the weighing apparatus. On account of the greater specific gravity of C0 2 the larger the percentage of this gas the greater the tendency will be to pull downward FLUE GAS ANALYSIS 199 the vessel G so that the pointer 5 on the balance can be adjusted to make the scale over which it travels indicate the percentage of CO2. For such a method of determination, obviously, the gas must be clean and dry. The cleaning is done by the excelsior and wool filters and the drying is done by the calcium chloride. CHAPTER X BOILER TESTING Tests of steam boilers are made to determine usually the following principal results : (i) Quantity of steam evaporated or furnished per hour. (2) Efficiency as a heat user, or weight of water evap- orated per pound of combustible (fuel less moisture and ash) . (3) Weight of water evaporated per hour per square foot of water-heating surface. (4) Weight of fuel burned per hour per square foot of grate surface. Leakages of any kind are always a lurking enemy for those engaged in any kind of accurate testing, and work with boilers is no exception. Poor results with boilers are due more often to air leakage than to any other fault. Air entering the setting and flues instead of the furnace does not assist com- bustion, but, on the contrary, absorbs from the hot gases a quantity of heat which otherwise might pass through the boiler-heating surfaces into the water in the boiler. When the object of a series of tests "is, for example, to compare one kind of coal with another, or one type of grate or mechanical stoker with another, the losses due to air leakage would not be of much consequence in what 'are only comparative results; but if a boiler is to give the best possible efficiency and capacity, air leaks must be stopped. In practice the importance of closing air leaks in the boiler setting is forcefully presented when patented devices for fuel saving are installed in boiler plants. Important economies in many cases are guaranteed if the new device is adopted, and then the claims of the agent are made good by instructing his workmen to go over the boiler, closing up all cracks in the setting through which cold air could enter, and at the same 200 BOILER TESTING 201 time covering the outside surface of the setting with a coating impervious to air. By such means the owner of the plant pays a high price for results that could have been obtained much more cheaply. The first of the principal objects of a boiler test stated above is to determine its capacity " rating." 1 The unit of capacity most generally used in steam boiler practice is the " boiler " horse power. Now this term horse power has two very distinct meanings in engineering practice. Usually it is taken to mean the rate of doing work or the work done in a definite period of time. In this sense it means, as in the case of engines, turbines, water-wheels, etc., 33,000 foot-pounds per minute. 2 In the case of a steam boiler, however, where the work done must be measured by the conversion of water into steam, a horse power is taken as the evaporation of 30 pounds of water at a temperature of 100 degrees Fahrenheit into steam at 70 pounds pressure above the atmosphere. When this unit was adopted it was considered that 30 pounds per hour was approximately the requirement per indicated horse power of an average engine. The Committee on Boiler Tests of the American Society of Mechanical Engineers (page 218) have adopted what is in effect the same unit, stating it, how- ever, somewhat differently — that a boiler horse power is equiv- alent to evaporating 34.5 pounds of steam per hour from feed- water temperature of 212 degrees Fahrenheit into steam at the same temperature. According to the latest steam tables this is equivalent to approximately 33,480 B.T.U. per hour, or 558 B.T.U. per minute. Unit of Evaporation. For reducing the results of the boiler tests to a common standard the term "unit of evaporation" is used. It is the heat required to evaporate a pound of water 1 When the steam supply is small or if, for any reason, the noise due to escaping steam from the calorimeters used is objectionable, a calorim- eter may be shut off sometimes between readings. The length of time the calorimeter is in operation must then be carefully noted in order to determine the weight of steam lost through it by calculating the flow through its orifice (see page 148). The flow of steam, however, through the calorimeter must always be started before observations of the tem- peratures are to be taken in order to get constant conditions. 2 This unit of horse power was adopted by James Watt, who considered it equivalent to the work done by a good London draft horse. 202 POWER PLANT TESTING from and at 2 1 2 degrees Fahrenheit, which according to standard steam tables is approximately equivalent to 970.4 B.T.U. 1 Graphical Log Sheets of boiler tests similar to the one shown in Fig. 179 are very serviceable for checking the obser- vations when made during the test as the data are taken. In Fig. 179. — Graphical Chart of a Boiler Trial. the report of a test it shows also the relative irregularity or regularity of the conditions affecting the results. Standard Methods for Boiler Trials. The American Society of Mechanical Engineers has adopted rules for conducting boiler trials which are generally accepted in America and are also considered with favor in England. 2 These rules are so complete that they will be given here with practically no abridgement. 3 1 Marks and Davis' Steam Tables and Diagrams, see also Peabody's Steam Tables, 1909 Edition. 2 Engines and Boilers by W. W. F. Pullen, pages 466-475. 3 Transactions American Society of Mechanical Engineers, vol. 31, pages 34-1 1 1 (including discussions). BOILER TESTING 203 RULES FOR CONDUCTING BOILER TRIALS ABRIDGED CODE OF 1899. I. Determine at the outset the specific object of the pro- posed trial, whether it be to ascertain the capacity of the boiler, its efficiency as a steam generator, its efficiency and its defects under usual working conditions, the economy of some particular kind of fuel, or the effect of changes of design, proportion or operation; and prepare for the trial accordingly. II. Examine the boiler, both outside and inside; ascertain the dimensions of grates, heating sufaces, and all important parts; and make a full record, describing the same, and illustrating special features by sketches. The area of heating surface is to be computed from the surfaces of shells, tubes, furnaces, and fire boxes in contact with the fire or hot gases. The outside diameter of water tubes and the inside diameter of fire tubes are to be used in the computation. All surfaces below the mean water level which have water on one side and products of combustion on the other are to be considered as water-heating surface, and all surfaces above the mean water level which have steam on one side and products of combustion on the other are to be considered as superheating surface. III. Notice the general condition of the boiler and its equip- ment, and record such facts in relation thereto as bear upon the objects in view. If the object of the trial is to ascertain the maximum economy or capacity of the boiler as a steam generator, the boiler and all its appurtenances should be put in first-class con- dition. Clean the heating surface inside and outside, remove clinkers from the grates and from the sides of the furnace. Remove all dust, soot, and ashes from the chambers, smoke connections, and flues. Close air leaks in the masonry and poorly fitted cleaning doors. See that the damper can be opened wide and closed tightly. Test for air leaks by firing a few shovelsful of smoky fuel and immediately closing the damper, observing the escape of smoke through the crevices, or by passing the flame of a candle over cracks in the brickwork. 204 POWER PLANT TESTING IV. Determine the character of the coal to be used. For tests of the efficiency or capacity of the boiler for comparison with other boilers the coal should, if possible, be of some kind which is commercially regarded as a standard. For New England and that portion of the country east of the Alle- gheny Mountains, good anthracite egg coal, containing not over 10 per cent of ash, and semi-bituminous Clearfield (Pa.), Cumberland (Md.), and Pocahontas (Va.) coals are thus regarded. West of the Allegheny Mountains, Pocahontas (Va.) and New River (W. Va.) semi-bituminous, and Youghiogheny or Pitts- burg bituminous coals are recognized as standards. 1 There is no special grade of coal mined in the Western States which is widely recognized as of superior quality or considered as a standard coal for boiler testing. Big Muddy lump, an Illinois coal mined in Jackson County, 111., is suggested as being of sufficiently high grade to answer these requirements in districts where it is more conveniently obtainable than the other coals mentioned above. For tests made to determine the performance of a boiler with a particular kind of coal, such as may be specified in a contract for the sale of a boiler, the coal used should not be higher in ash and in moisture than that specified, since increase in ash and moisture above a stated amount is apt to cause a falling off of both capacity and economy in a greater measure than the proportion of such increase. V. Establish the correctness of all apparatus used in the test for weighing and measuring. These are : i . Scales for weighing coal, ashes, and water. 2. Tanks, or water meters for measuring water. Water meters, as a rule, should only be used as a check on other measurements. For accurate work, the water should be weighed or measured in a tank. 3. Thermometers and pyrometers for taking temperatures of air, steam, feed- water, waste gases, etc. 4. Pressure gages, draught gages, etc. The kind and location of various pieces of testing appa- 1 These coals are selected because they are about the only coals which possess the essentials of excellence of quality, adaptability of various kinds of furnaces, grates, boilers, and methods of firing, and wide dis- tribution and general accessibility in the markets. BOILER TESTING 205 ratus must be left to the judgment of the person conducting the test; always keeping in mind the main object, i.e., to obtain authentic data. VI. See that the boiler is thoroughly heated before the trial to its usual working temperature. If the boiler is new and of a form provided with a brick setting, it should be in regular use at least a week before the trial, so as to dry and heat the walls. If it has been laid off and become cold, it should be worked before the trial until the walls are well heated. VII. The boiler and connections should be proved to be free from leaks before beginning a test, and all water connec- tions, including blow and extra feed pipes, should be discon- nected, stopped with blank flanges, or bled through special openings beyond the valves, except the particular pipe through which water is to be fed to the boiler during the trial. During the test the blow-off and feed-pipes should remain exposed to view. If an injector is used, it should preferably receive steam directly through a felted pipe from the boiler being tested. 1 If the water is metered after it passes the injector, its temperature should be taken at the point where it leaves the injector. If the quantity is determined before it goes to the injector, the temperature should be determined on the suction side of the injector, and if no change of temperature occurs other than that due to the injector, the temperature thus determined is properly that of the feed water. When the temperature changes between the injector and the boiler, as by the use of a heater or by radiation, the temperature at which the water enters and leaves the injector and that at which it enters the boiler should all be taken. In that case, the weight to be used is that of the water leaving the injector computed from the heat units if not directly measured, and the temperature, that of the water entering the boiler. 1 In feeding a boiler undergoing test with an injector taking steam from another boiler, or from the main steam pipe from several boilers, the evaporative results may be modified by a difference in the quality of the steam from such source compared with that supplied by the boiler being tested, and in some cases the connection to the injector may act as a drip for the main steam pipe. If it is known that the steam from the main pipe is of the same pressure and quality as that furnished by the boiler undergoing the test, the steam may be taken from such main pipe. 206 POWER PLANT TESTING Let w =weight of water entering the injector, pounds. x = weight of steam entering the injector, pounds. hi =heat units per pound of water entering injector. h2 =heat units per pound of steam entering injector. h 3 =heat units per pound of water leaving injector. Then, w + x=weight of water leaving injector. I13 — hj X=Wr^— j- 1 . . . . . . n 2 — n H (60) See that the steam main is so arranged that water of con- densation cannot run back into the boiler. When coal is used for fuel it is usually weighed in wheel- barrows which can be pushed upon large platform scales. The time when beginning to fire from each loaded wheel- barrow should be noted. VIII . Duration of the Test. — For tests made to ascertain either the maximum economy or the maximum capacity of a boiler, irrespective of the particular class of service for which it is regularly used, the duration should be at least 10 hours of continuous running. If the rate of combustion exceeds 25 pounds of coal per square foot of grate surface per hour, it may be stopped when a total of 250 pounds of coal has been burned per square foot of grate. In cases where the service requires continuous running for the whole 24 hours of the day, with shifts of firemen a number of times during that period, it is well to continue the test for at least 24 hours. When it is desired to ascertain the performance under the working conditions of practical running, whether the boiler be regularly in use 24 hours a day or only a certain number of hours out of each 24, the fires being banked the balance of the time, the duration should not be less than 24 hours. IX. Starting and Stopping a Test. — The conditions of the boiler and furnace in all respects should be, as nearly as pos- sible, the same at the end as at the beginning of the test. The steam pressure should be the same; the water level the same; the fire upon the grates should be the same in quantity and condition, and the walls, flues, etc., should be of the same temperature. Two methods of obtaining the desired equality BOILER TESTING 207 of conditions of the fire may be used, viz., " the standard method " and " the alternate method," the latter being em- ployed where it is inconvenient to make use of the standard method. 1 X. Standard Method of Starting and Stopping a Test. — Steam being raised to the working pressure, remove rapidly all the fire from the grate, close the damper, clean the ash pit, and as quickly as possible start a new fire with weighed wood and coal, noting the time and the water level 2 while the water is in a quiescent state, just before lighting the fire. At the end of the test remove the whole fire, which has been burned low, clean the grates and ash pit, and note the water level when the water is 'in a quiescent state, and record the time of hauling the fire. The water level should be as nearly as possible the same as at the beginning of the test. If it is not the same, a correction should be made by com- putation, and not by operating the pump after the test is completed. XL Alternate Method of Starting and Stopping a Test, — The boiler being thoroughly heated by a preliminary run, the fires are to be burned low and well cleaned. Note the amount of coal left on the grate as nearly as it can be estimated; note the pressure of steam and the water level. Note the time, and record it as the starting time. Fresh coal which has been weighed should now be fired. The ash pits should be thor- oughly cleaned at once after starting. Before the end of the test the fires should be burned low, just as before the start, .and the fires cleaned in such a manner as to leave a bed of coal on the grates of the same depth, and in the same condition, as at the start. When this stage is reached, note the time and record it as the stopjDing time. The water level and steam 1 The Committee concludes that it is best to retain the designations " standard " and " alternate," since they have become widely known and established in the minds of engineers and in the reprints of the Code of 1885. Many engineers prefer the " alternate " to the " standard " method on account of its being less liable to error due to cooling of the boiler at the beginning and end of a test. 2 The gage-glass should not be blown out within an hour before the water level is taken at the beginning and end of a test, otherwise an error in the reading of the water level may be caused by a change in the tem- perature and density of the water in the pipe leading from the bottom of the glass into the boiler. 208 POWER PLANT TESTING pressures should previously be brought as nearly as possible to the same- point as at the start. XII. Uniformity of Conditions. — In all trials made to ascertain maximum economy or capacity, the conditions should be maintained uniformly constant. Arrangements should be made to dispose of the steam so that the rate of evaporation may be kept the same from beginning to end. This may be accomplished in a single boiler by carrying the steam through a waste steam pipe, the discharge from which can be regulated, as desired. In a battery of boilers, in which only one is tested, the draft may be regulated on the remaining boilers, leaving the test boiler to work under a constant rate of production. Uniformity of conditions should prevail as to the pressure of steam, the height of water, the rate of evaporation, the thickness of fire, the times of firing and quantity of coal fired at one time, and as to the intervals between the times of cleaning the fires. The method of firing to be carried on in such tests should be dictated by the expert or person in responsible charge of the test, and the method adopted should be adhered to by the fireman throughout the test. XIII. Keeping the Records.— Take note of every event con- nected with the progress of the trial, however unimportant it may appear. Record the time of every occurrence and the time of taking every weight and every observation. The coal should be weighed and delivered to the fireman in equal proportions, each sufficient for not more than one hour's run, and* a fresh portion should not be delivered until the previous one has all been fired. The time required to consume each portion should.be noted, the time being recorded at the instant of firing the last of each portion. It is desir- able that at the same time the amount of water fed into the boiler should be accurately noted and recorded, including the height of the water in the boiler, and the average pressure of steam and temperature of feed during the time. By thus recording the amount of water evaporated by successive por- tions of coal, the test may be divided into several periods if desired, and the degree of uniformity of combustion, evapo- ration, and economy analyzed for each period. In addition to these records of the coal and the feed- water, not less frequently BOILER TESTING 209 than every half hour, observations should be made of the tem- perature of the feed-water, of the flue gases, of the external air of the boiler room, of the temperature of the furnace when a furnace pyrometer is used, also of the pressure of steam, and of the readings of the instruments for determining the moisture in the steam. A log should be kept on properly prepared blanks con- taining columns for record of the various observations. When the " standard method " of starting and, stopping the test is used, the hourly rate of combustion and of evaporation and the horse power should be computed from the records taken during the time when the fires are in active condition. This time is somewhat less than the actual time which elapses between the beginning and end of 'the run. The loss of 'time due to kindling the fire at the beginning and burning it out at the end makes this course necessary.. XIV. Quality of Steam. — The percentage of moisture in the steam should be determined by the use of either a throttling or a separating steam calorimeter. The sampling nozzle should be placed in the vertical steam pipe rising from the boiler. It should be made of one-half-inch pipe, and should extend across the diameter of the steam pipe to within half an inch of the opposite side, being closed at the end and perforated with not less than twenty one-eighth-inch holes equally distrib- uted along and around its cylindrical surface, but none of these holes should be nearer than one-half inch to the inner side of the steam pipe. The calorimeter and the pipe leading to it should be well covered with felting. Whenever the indica- tions of the throttling or separating calorimeter show that the percentage of moisture is irregular, or occasionally in excess of three per cent, the results should be checked by a steam separator placed in the steam pipe as close to the boiler as convenient, with a calorimeter in the steam pipe just beyond the outlet from the separator. The drip from the separator should be caught and weighed, and the percentage of moisture com- puted therefrom added to that shown by the calorimeter. Superheating should be determined by means of a ther- mometer placed in a mercury well inserted in the steam pipe. The degree of superheating should be taken as the difference between the reading of the thermometer for superheated steam and the readings of the same thermometer for saturated steam 210 POWER PLANT TESTING at the same pressure as determined by a special experiment, and not by reference to steam tables. For calculations relating to quality of steam and corrections for quality of steam, see pages 46-60. XV. Sampling the Coal and Determining its Moisture. — As each barrow load or fresh portion of coal is taken from the coal pile, an average shovelful is selected from it and placed in a barrel or box in a cool place and kept until the end of the trial. The samples are then mixed and broken into pieces not exceeding 1 inch in diameter, and reduced by the process of repeated quartering and crushing until a final sample weighing about 5 pounds is obtained, and the size of the larger pieces is such that they will pass through a sieve with one-quarter-inch meshes. From this sample two one-quart air-tight glass preserving jars, or other air-tight vessels which will prevent the escape of moisture from the sample, are to be promptly filled, and these samples are to be kept for sub- sequent determinations of moisture and of heating value and for chemical analyses. During the process of quartering, when the sample has been reduced to about 100 pounds, a quarter to a half of it may be taken for an approximate deterrnination of moisture. This may be made by placing it in a shallow iron pan, not over 3 inches deep, carefully weighing it, and setting the pan in the hottest place that can be found on the brickwork of the boiler setting of flues, keeping it there for at least 12 hours, and then weighing it. The determination of moisture thus made is believed to be approximately accurate for anthracite and semi-bituminous coals, and also for Pitts- burg or Youghiogheny coal; but it cannot be relied upon for coals mined west of Pittsburg, or for other coals containing inherent moisture. For these latter coals it is important that a .more accurate method be adopted. The method recommended by the Committee for all accurate tests, whatever the character of the coal, is described as follows : Take one of the samples contained in the glass jars, and subject it to a thorough air-drying, by spreading it in a thin layer and exposing it for several hours to the atmosphere of a warm room, weighing it before and after, thereby determining the quantity of surface moisture it contains. Then crush the whole of it by running it through an ordinary coffee mill BOILER TESTING 211 adjusted so as to produce somewhat coarse grains (less than one- sixteenth-inch), thoroughly mix the crushed sample, select from it a portion of from 10 to 50 grams, weigh it in a balance which will easily show a variation as small as 1 part in 1000, and dry it in an air or sand bath at a temperature between 240 and 280 degrees Fahrenheit for one hour. Weigh it and record the loss, then heat and weigh it again repeatedly, at intervals of an hour or less ; until the minimum weight has been reached and the weight begins to increase by oxidation of a portion Of the coal. The difference between the original and the min- imum weight is taken as. the moisture in the air-dried, coal. This moisture test should preferably be made on duplicate samples, and the results should; agree within 0.3 to 0.4 of one per cent, the mean of the two determinations being taken as the correct result. The sum of the percentage of moisture thus found and the percentage of surface moisture previously deter- mined is the total moisture. XVI. Treatment of Ashes and Refuse. — The ashes and refuse are to be weighed in a dry state. If it is found desirable to show the principal characteristics of the ash, a sample should be subjected to a proximate analysis and the actual amount of incombustible material determined. For elaborate trials a complete analysis of the ash and refuse should be made. XVII. Calorific Tests and Analysis of Coal. — The quality of the fuel should be determined either by heat test or by analysis, or by both. The rational method of determining the total heat of com- bustion is to burn the sample of coal in an atmosphere of oxygen gas, the coal to be sampled as directed in Article XV. of this code. The chemical analysis of coal should be made only by an expert chemist. The total heat of combustion computed from the results of the ultimate analysis may be obtained by the use of Dulong's formula (with constants modified by recent determinations), viz., 1 4,6006" + 62,000 lH ] +4000S, in which C, H, O, and 5 refer to the proportions of carbon, hydrogen, oxygen, and sulphur respectively, as determined by the ultimate analysis. 1 1 Favre and Silberman give 14,544 B.T.U. per pound carbon; Berthe- 212 POWER PLANT TESTING It is desirable that a proximate analysis should be made, thereby determining the relative proportions of volatile matter and fixed carbon. These proportions furnish an indication of the leading characteristics of the fuel, and serve to fix the class to which it belongs. As an additional indication of the characteristics of the fuel, the specific gravity should be determined. XVIII. Analyses of Flue Gases. — The analysis of the flue gases is an especially valuable method of determining the relative value of different methods of firing, or of different kinds of furnaces. In making these analyses great care should be taken to procure average samples — since the composition is apt to vary at different points of the flue. The composition is also apt to vary from minute to minute, and for this reason the drawings of gas should last a consider- able period of time. Where complete determinations are desired, the analyses should be intrusted to an expert chemist. For approximate determinations " Orsat " * apparatus may be used by the engineer. For the continuous indication of the amount of carbonic acid present in the flue gases, an instrument may be employed which shows the weight of the sample of gas passing through it. XIX. Smoke Observations. — It is desirable to have a uni- form system of determining and recording the quantity of smoke produced where bituminous coal is used. The system commonly employed is to express the degree of smokiness by means of percentages dependent upon the judgment of the observer. The Committee does not place much value upon a percentage method, because it depends so largely upon the personal element, but if this method is used, it is desirable that, so far as possible, a definition be given in explicit terms, as to the basis and method employed in arriving at the per- centage. XX. Miscellaneous. — In tests for purposes of scientific research, in which the determination of all the variables entering into the test is desired, certain observations should lot 14,647 B.T.U. Favre and Silberman give 62,032 B.T.U. per pound hydrogen; Thomsen 61,816 B.T.U. 1 See R. S. Hale's paper on " Flue Gas Analysis Transactions A.S.M.E., vol. 18., page 901. BOILER TESTING 213 be made which are in general unnecessary for ordinary tests. These are the measurement of the air supply, the determina- tion of its contained moisture, the determination of the amount of heat lost by radiation, of the amount of infiltration of air through the setting, and (by condensation of all the steam made by the boiler) of the total heat imparted to the water. XXI. Calculations of Efficiency. — Two methods of defining and calculating the efficiency of a boiler are recommended. They are: „„ . , „ 1 ... Heat absorbed per lb. combustible i. Efficiency of the boiler = 7^—; — r^- — : ^ — r , : -rr-r ; Calorific value of 1 lb. combustible „,„ . r,i , -1 1 Heat absorbed per lb. coal 2. Efficiency of the boiler and grate = ^n — =-f : —. — ^ — — , . Calorific value of 1 lb. coal The first of these is sometimes called the efficiency based on combustible, and the second efficiency based on coal. The first is recommended as a standard of comparison for all tests, and this is the one which is understood to be referred to when the word " efficiency " alone is used without qualification. The second, however, should be included in a report of a test, together with the first, whenever the object of the test is to determine the efficiency of the boiler and furnace together with the grate (or mechanical stoker), or to compare different furnaces, grates, fuels, or methods of firing. The heat absorbed per pound of combustible (or per pound coal) is to be calculated by multiplying the equivalent evap- oration from and at 212 degrees per pound combustible (or coal) by 965.7. 1 XXII. The Heat Balance. — An approximate " heat balance," or statement of the distribution of the heating value of the coal among the several items of heat utilized and heat lost may be included in the report of a test when analyses of the fuel and of the chimney gases have been made. It should be reported in the following form : 1 This value is the one given in the accepted steam tables in 1899 when this code was published. According to the more recent deter- minations it should be 970. (See Marks and Davis' and Peabody's revised Steam Tables.) 214 POWER PLANT TESTING Heat Balance, or Distribution of the Heating Value of the Combustible. Total Heat Value of i lb. of Combustible B.T.U. Heat absorbed by the boiler = evaporation from and at 2 1 2 degrees per pound of combustible Xq-65-7- 1 Loss due to moisture in coal = per cent, of moisture referred to combustible -^ioo><[(2i2— if) + 966 1 + o. 48 (T — 212) ](/ = temperature of air in the boiler-room, T =that of the flue gases). Loss due to moisture formed by the burning of hy- drogen = per cent, of hydrogen to combustible -4-100 X9X[2i 2 (-0 + 966 1 + 0.48(7-212)]. Loss due to heat carried away in the dry chimney gases = weight of gas per pound of combustible xo. 24 x(r-o. Loss due to incomplete combustion of carbon = CO per cent C in combustible C0 2 + CO 100 Loss due to unconsumed hydrogen and hydrocar- bons, to heating the moisture in the air, to radiation, and unaccounted for. (Some of these losses may be separately itemized if data are obtained from which they may be calcu- lated). Totals 1 This value is the one given in the accepted steam tables in 1899 when this code was published. According to the more recent determinations it should be 970. (See Marks and Davis' and Peabody's revised Steam Tables.) 2 The weight of gas per pound of carbon burned may be calculated from the gas anal- yses as follows : , , iiC02 + 8 0+7(CO + N ) . ,. , ' „„ „ ,, r Dry gas per pound carbon = n ~ ,nn\ • m which CO2, CO, O and N are the 3(,OU2 + OU,) percentages by volume of the several gases. As the sampling and analyses of the gases in the present state of the art are liable to considerable errors, the result of this calculation is usually only an approximate one. The heat balance itself is also only approximate for this reason, as well as for the fact that it is not possible to determine accurately the per- centage of unburned hydrogen or hydrocarbons in the flue gases. The weight of dry gas per pound of combustible is found by multiplying the dry gas per pound of carbon by the percentage of carbon in the combustible, and dividing by 100. 3 CO2 and CO are respectively the percentage by volume of carbonic acid and carbonic oxide in the flue gases. The quantity 10, 150 = Number of heat units generated by burning to carbonic acid one pound of carbon contained in carbonic oxide. XXIII. Report of the Trial. — The data and results should be reported in the manner given in either one of the two fol- lowing tables, omitting lines where the tests have not been made as elaborately as provided for in such tables. Additional BOILER TESTING 215 lines may be added for data relating to the specific object of the test. The extra lines should be classified under the headings provided in the tables, and numbered as per preceding line, with sub letters a, b, etc. The Short Form of Report, Table No. 2, is recommended for commercial tests and as a con- venient form of abridging the longer form for publication when saving of space is desirable. For elaborate trials, it is recom- mended that the full log of the trial be shown graphically, by means of a chart. (Fig. 179.) TABLE NO. 1. Data and Results of Evaporative Test. Arranged in accordance with the Complete Form advised by the Boiler Test Committee of the American Society of Mechanical Engineers. Code of 1899. Made by of boiler at to determine Principal conditions governing the trial Kind of fuel l Kind of furnace State of the weather Method of starting and stopping the test (" standard " or " alternate," Art. X. and XL, Code) 1 . Date of trial 2. Duration of trial hours. . .-.. Dimensions and Proportions. (A complete description of the boiler, and drawings of the same if of unusual type, should be given on an annexed sheet.) 3. Grate surface width length area. . . . sq. ft. 4. Height of furnace ins. 5. Approximate width of air spaces in grate in. 6. Proportion of air space to whole grate surface per cent. 7. Water-heating surface sq. ft. 8. Superheating surface 9. Ratio of water-heating surface to grate surface — to 1. 10. Ratio of minimum draft area to grate surface 1 to — . Average Pressures. n. Steam pressure by gage lbs. per sq. in. 12. Force of draft between damper and boiler ins. of water 13. Force of draft in furnace 14. Force. of draft or blast in ash pit I The items printed in italics correspond to the items in the "Short Form of Code " 216 POWER PLANT TESTING Average Temperatures. 15. Of external air deg. 16. Of fireroom 17. Of steam 18. Of feed water entering heater 19. Of feed water entering economizer 20. Of feed water entering boiler 2 1 . Of escaping gases from boiler 22. Of escaping gases from economizer Fuel. 23. Size and condition 24. Weight of wood used in lighting fire. . . 2 5 . Weight of coal as fired x 26. Percentage of moisture in coal 2 27. Total weight of dry coal consumed 28. Total ash and refoise 29. Quality of ash and refuse 30. Total combustible consumed. . 3 1 . Percentage of ash and refuse in dry coal . lbs. per cent, lbs. , lbs. per cent. Proximate Analysis of Coal. Of Coal. 32". Fixed carbon per cent. 33. Volatile matter 34. Moisture 35- Ash Of Combustible . . . .per cent. 36. Sulphur, separately determined r 00 per cent. 1 00 per cent. Ultimate Analysis of Dry Coal. (Art. XVII., Code.) Of Coal. Of Combustible. 37. Carbon (C) 38. Hydrogen (H) . 39. Oxygen (O) . . . 40. Nitrogen (N) . . 41. Sulphur (S) . . . 42. Ash per cent per c 100 per cent 100 per cent. 43. Moisture in sample of coal as received. . . 1 Including equivalent of wood used in lighting the fire, not including unburnt coal withdrawn from furnace at times of cleaning and at end of test. One pound of wood is taken to be equal to 04 pound of coal, or, in case greater accuracy is desired, as having a heat value equivalent to evaporation of 6 pounds of water from and at 212 degrees per pound. (6X965 7 = 5794 B T.U ) The term "as fired" means in its'actual condition, including moisture. 8 This is the total moisture in the coal as found by drying it artificially, as described in Art. XV. of Code. BOILER TESTING 217 Analysis of Ash and Refuse. 44. Carbon 4.5. Earthy matter Fuel per Hour. 46. Dry coal consumed per hour 47. Combustible consumed per hour 48. Dry coal per square foot of grate surface per hour 49. Combustible per square foot of water-heating surface per hour per cent. lbs. Calorific Value of Fuel. (Art. XVII., Code.) 50. Calorific value by oxygen calorimeter, per lb. of dry coal 51. Calorific value by oxygen calorimeter, per lb. of combus- B.T.U. 52. Calorific value by analysis, per lb. of dry coal ' 53. Calorific value by analysis, per lb. of combustible.. Quality of Steam. 54. Percentage of moisture in steam 55. Number of degrees of superheating 56. Quality of steam (dry steam = unity") . per cent- deg. Water. Total weight of water fed to boiler 2 Equivalent water fed to boiler from and at 2 1 2 degrees Water actually evaporated, corrected for quality of steam Factor of evaporation 3 Equivalent water evaporated into dry steam from and at 212 degrees 4 (Item 59 Xltem 60) lbs. Water per Hour. 62. Water evaporated per hour, corrected for quality of steam 63 . Equivalent evaporation per hour from and at 2 1 2 degrees 2 64. Equivalent evaporation per hour from and at 212 per square foot of water-heating surface 2 1 See formula for calorific value under Article XVII. of Code. 2 Corrected for inequality of water level and of steam pressure at beginning and end of test. 3 Factor of evaporation = — — , in which H and 4 are respectively the total heat in 965.7 steam of the average observed pressure, and in water of the average observed temperature of the feed. According to the newer steam tables the divisor in this fraction should be 970. * The symbol " U. E." meaning " Units of Evaporation," may be conveniently sub- stituted for the expression " Equivalent water evaporated into dry steam from and at 212 degrees." its definition being given in a footnote. 218 POWER PLANT TESTING Horse Power. 65. Horse power developed. (34 J lbs. of water evaporated per hour into dry steam from and at 2 1 2 degrees, equals 011c horse power) l H. P. 66. Builder's rated horse power 67. Percentage of builders' rated horse power developed. . . . per cent. Economic Results. 68. Water apparently evaporated under actual conditions per pound of coal as fired. {Item 57 -j- Item 25) . . 69. Equivalent evaporation from and at 212 degrees per pound of coal as fired. 2 (Item 61 -hltem 25.) 70. Equivalent evaporation from and at 212 degrees per pound of dry coal. 2 (Item 61 -hltem 27.) 71. Equivalent evaporation from and at 212 degrees per pound of combustible. 2 (Item 61 -s- Item 30.) (If the equivalent evaporation, Items 69, 70, and 71, is not corrected for the quality of steam, the fact should be stated.) lbs. Efficiency. (Art. XXL, Code.) 72. Efficiency of the boiler; heat absorbed by the boiler per lb. of combustible divided by the heat value of one lb. of combustible 2 73. Efficiency of boiler, including the grate; heat absorbed by the boiler, per lb. of dry coal, divided by the heat value or one lb. of dry coal per cent. Cost of Evaporation. 74. Cost of coal per ton of lbs. delivered in ooiler room $ 75. Cost of fuel for evaporating 1,000 lbs. of water under observed conditions $ 76. Cost of fuel used for evaporating 1,000 Ibs^ of water from and at 2 1 2 degrees $ Smoke Observations. 77. Percentage of smoke as observed . per cent. 78. Weight of soot per hour obtained from smoke meter. ounce. 79. Volume of soot per hour obtained from smoke'meter cu. in. 1 Held to be the equivalent of 30 lbs. of water per hour evaporated from 100 degrees Fahrenheit into dry steam at 70 lbs. gage pressure. (See Introduction to Code, page 201.) 2 In all cases where the word combustible is used, it means the coal without moisture and ash, but including all other constituents. It is the same as what is called in Europe " coal dry and free from ash." BOILER TESTING 219 Methods of Firing. 80. Kind of firing (spreading, alternate, or coking) 81. Average thickness of fire 82. Average intervals between firings for each furnace during time when fires are in normal condition 83 . Average interval between times of leveling or break- ing up Analyses of the Dry Gases. 84. Carbon dioxide (C0 2 ) 85. Oxygen (O) 86. Carbon monoxide (CO) 8 7 . Hydrogen and hydrocarbons 88. Nitrogen (by difference) (N) per cent. :oo per cent. TABLE NO. 2. Data and Results of Evaporative Test. Arranged in accordance with the Short Form advised by the Boiler Test Committee of the American Society of Mechanical Engineers. Code of 1899. Made by on .... -. boiler, at to determine Kind of fuel Kind of furnace Method of starting and stopping the test (" standard " or " alternate," Art. X. and XL, Code) Grate surface sq. ft. Waterheating surface Superheating surface " . Total Quantities. 1 . Date of trial 2 . Duration of trial 3. Weight of coal as fired x 4. Percentage of moisture in coal ' 5. Total weight of dry coal consumed 6. Total ash and refuse 7 . Percentage of ash and refuse in dry coal 8. Total weight of water fed to the boiler ' 9. Water actually evaporated, corrected for moisture or superheat in steam 10. Equivalent water evaporated into dry steam from and at 2 1 2 degrees 1 i See footnotes of Complete Form, pages 216 ancj 217. hours. lbs. per cent. lbs. per cent, lbs. 220 POWER PLANT TESTING Hourly Quantities. ii. Dry coal consumed per hour . lbs. 12. Dry coal per square foot of grate surface per hour. . . 13. Water evaporated per hour corrected for quality of steam 14. Equivalent evaporation per hour from and at 212 degrees l 15. Equivalent evaporation per hour from and at 212 degrees per square foot of water-heating surface l . Average Pressures, Temperatures, etc. 16. Steam pressure by gage lbs. per sq. in. 1 7 . Temperature of feed water entering boiler deg. Temperature of escaping gases from boiler Force of draft between damper and boiler ins. of water. Percentage of moisture in steam, or number of degrees of superheating , per cent, or deg. Horse Power. 21. Horse power developed (Item 14-^34^) ' H. P. 22. Builder's rated horse power 23. Percentage of builder's rated horse power developed. per cent. Economic Results. 24. Water apparently evaporated under actual conditions per pound of coal as fired. (Item 8 h- Item 3) lbs. 25. Equivalent evaporation from and at 212 degrees per pound of coal as fired. 1 (Item 10 -h Item 3) 26. Equivalent evaporation from and at 212 degrees per pound of dry coal. 1 (Item 10 -j- Item 5) 27. Equivalent evaporation from and at 212 degrees per pound of combustible. 1 [Item 10-^ (Item 5— Item 6)] •• (If Items 25, 26, and 27 are not corrected for quality of steam, the fact should be stated.) Efficiency. 28. Calorific value of the dry coal per pound B.T.U. 29. Calorific value of the combustible per pound 30. Efficiency of boiler (based on combustible) 1 per cent. 31. Efficiency of boiler, including grate (based on dry coal) Cost of Evaporation. 3 2 . Cost of coal per ton of lbs delivered in boiler room $ ^2- Cost of coal required for evaporating 1000 pounds of water from and at 2 1 2 degrees $ 1 See footnotes of Complete Form, pages 217 and 218. BOILER TESTING 221 EFFICIENCY OF THE BOILER The efficiency of the boiler, including the grate, or the efficiency based on coal, is the quotient arising from dividing heat absorbed by the boiler by the heating value of the total amount of coal supplied to the boiler, including the coal which falls- through the grate. It may be conveniently calculated by multiplying the number of pounds of water evaporated from and at 212 degrees Fahrenheit into dry steam per pound of dry coal by 965.7, 1 and dividing the product by the heating value in B. T. U. of one pound of dry coal. The efficiency of -the boiler, not including the grate, or the efficiency based on combustible, is the quotient arising from dividing the heat absorbed by the boiler by the heating value of the combustible burned. It may be calculated by multiplying the number of pounds of water evaporated from and at 212 degrees Fahrenheit into dry steam per pound of combustible by 965. 7, 1 - and dividing the product by the heating value in B. T. U. of one pound of combustible; the term " combustible " being defined as coal dry and free from ash, or the coal supplied to the boiler less its moisture and the ash and unburn ed coal which falls through the grate or is otherwise withdrawn from the furnace. The efficiency of the boiler, not including the grate (or the efficiency based upon combustible) is a more accurate measure of comparison of different boilers than the efficiency including the grate (or the efficiency based upon coal), for the latter is subject to a number of variable conditions, such as size and character of the coal, air-spaces between the grate bars, skill of the fireman in saving coal from falling through the grate, etc. It is, moreover, subject to errors of sampling the coal for drying and for analysis, which affect the result to a greater degree than they do the efficiency based upon combustible, for the reason that the heating value per pound of com- bustible of any sample selected from a given lot (such as a car- load) of coal is practically a constant quantity and is inde- pendent of the percentage of moisture and ash in the sample; while the sample itself, upon the heating value of which the efficiency based on coal is calculated, may differ in its per- *See page 201 and footnote, page 213. 222 POWER PLANT TESTING centage of moisture and ash from the average coal used in the boiler test. When the object of a boiler test is to determine its efficiency as an absorber of heat, or to compare it with other boilers, the efficiency based on combustible is the one which should be used; but when the object of the test is to determine the efficiency of the combination of the boiler, the furnace, and the grate, the efficiency based on coal must necessarily be used. It has been proposed that in reporting the efficiency of a boiler when the fuel used contains hydrogen, the efficiency should be considered to be the sum of the percentage of the heating value of the fuel which is utilized by the boiler in making steam of the percentage of that heating value which is lost in the shape of latent heat in the moisture in the chimney gases, which moisture is formed by the burning of the hydrogen. This latent heat may amount to over three per cent of the total heating value of the fuel. The reason assigned for this pro- posal is that, since it is impossible for this heat to be utilized by the boiler because the gases are discharged at a temperature above 212 degrees Fahrenheit, it should not be charged against the boiler. It is not considered advisable that this method of reporting the efficiency should be adopted (1) because it is opposed to the generally accepted definition of effi- ciency, which is the useful work received from an apparatus divided by the work (or heat value of the fuel) put into it; (2) because in order to calculate it, it is necessary to know both the percentage of hydrogen in the coal and whether or not all of this hydrogen has been burned to H 2 0, the first requiring an analysis of the coal, which is not always obtainable, and the second an analysis of the gases for hydrogen, which cannot be obtained with any approximation to accuracy with our present methods of sampling and analyzing gases; and (3) because it is opposed to the almost universal custom of reporting boiler tests. It is true that the latent heat of the water in the chimney gases cannot be utilized (unless an econ- omizer which discharges its gases below 212 degrees is used), and it is not the fault of the boiler that it cannot be utilized. It may be considered the misfortune of the boiler, when tested with hydrogenous coal, similar to the misfortune under which BOILER TESTING 223 an engine labors when it is tested while supplied with a con- denser which gives a vacuum of less than 30 inches of mercury. The engine might give a higher efficiency with a vacuum of 30 inches than it would with one of 27 or 28 inches; but it is not customary to credit the engine with the efficiency which it loses on account of the imperfect vacuum. Since it is well understood that a boiler cannot show quite as high an efficiency (as commonly defined) when using bitu- minous coal high in hydrogen as when using anthracite nearly free from hydrogen, no harm is done, and much confusion is avoided, by reporting the efficiency as the. percentage of the heating value of the coal which is actually utilized in making steam. The fact that bituminous coal is used is always stated in the report of a test made with that coal. If desired, a state- ment may also be made in the " heat balance " of the approxi- mate or estimated percentage of heat which is lost in the latent heat of the moisture in the chimney gases, together with the loss- due to moisture in the coal. DISTRIBUTION OF THE HEATING VALUE OF THE FUEL In the operation of a steam boiler the following distribu- tion of the total heating value of the fuel takes place : 1 . Loss of coal or coke through the grate. 2. Unburned coal or coke carried in the shape of dust or sparks beyond the bridge wall. 3. Heating to 212 degrees the moisture in the coal, evap- orating it at that temperature, and evaporating the steam made from it to the temperature of the flue gases ^weight of the moisture in pounds X [(212 degrees — /) +966 1 +0.48(7"— 212)], in which T is the temperature (Fahrenheit) of the flue gases and t the temperature of the external air. 4. Loss of heat due to steam which is formed by burning the hydrogen contained in the coal, and which passes into the chimney as superheated steam =9 times the weight of the hydrogen x[ (21 2— /) +966 1 +0.48(7— 212)]. 5. Superheating the moisture in the air supplied to the furnace to the temperature of the flue gases ^weight of the moisture X 0.48(7"—/). 'See footnote, page 213. 224 POWER PLANT TESTING 6. Heating of the gaseous products of combustion (not including steam) to the temperature of the flue =their weight xo.2 4 (r-o. 7. Loss due to imperfect burning of the carbon of the coal and to non-burning of the volatile gases. 8. Radiation from the boiler and furnace. 9. Heat absorbed by the boiler, or useful work. Item 1 depends upon the size of the spaces between the grate bars; upon the kind of grate, as a plain, shaking, or traveling grate; upon the size of the coal; upon the char- acter of the coal, as it requires to be more or less distributed on the grate in order to get a sufficient supply of air through it; upon the rate of driving of the furnace, rapid driving with some coals requiring more frequent shaking or cleaning of the grate than slow driving ; and upon the skill of the fireman. Item 2 depends upon the nature and fineness of the coal and tipon the force of the draft. It is usually so small as to be inappreciable in its effects upon the results of the trial of a stationary boiler driven with natural draft, but in locomo- tives, with rapid rates of combustion, it often becomes quite important. Item 3 depends upon the amount of moisture in the coal. Item 4 depends upon the amount of hydrogen in the coal. Item 5 depends upon the amount of moisture in the air. The moisture in the air may be obtained from its temperature and relative humidity, as determined by a wet-and-dry bulb thermometer by reference to hygrometric tables. The loss of heat due to the moisture in the air will rarely exceed 0.25 per cent, of the heating value of the fuel, and it may usually, therefore, be neglected. (For hygrometric tables see page 403.) Item 6 depends chiefly upon the type and proportions of the boiler, and upon the rate at which it is driven. This item is usually the largest of all the heat losses. Items 3, 4, 5, and 6 depend also on the temperature of the flue gases. Item 7 depends upon the character of the coal and of the furnace, and upon the skill of the fireman. This loss may be very large, 20 per cent or more of the heating value of the coal, when highly bituminous coals are used in a furnace not adapted to them. BOILER TESTING 225 Item 8 depends chiefly upon the type, size, and setting of the boiler, and, when expressed as a percentage of the total heat of the fuel, upon the rate at which it is driven. Item 9 is the heat absorbed by the boiler, or the useful work. It is also the difference between the total heating value of the coal and of the sum of the losses of items i to 8 inclusive. COMPUTATION OF THE WEIGHT OF THE CHIMNEY GASES FROM THE ANALYSIS BY VOLUME OF THE DRY GAS. Two methods of calculating from the analysis by volume of the dry chimney gases the number of pounds of dry chimney gases per pound of carbon, or the weight of air supplied per pound of carbon, have been given by different writers. These may be expressed in the shape of formula? as follows : ,., „ A A An nC0 2 + 80 + 7(CO + N) „, (A) Pounds dry gas per pound C = /y '- (6i) 3(uu 2 + uuj , D , D , . An _"(C0 2 + 0)+CO (B) Pounds air per pound C = 5-8— — ^r ^ . . (62) Formula A may be derived from the method of computation given in Mr. R. S. Hale's paper on " Flue-Gas Analyses," Transactions American Society of Mechanical Engineers, vol. 18, page 902, and formula B from the method given in Peabody's and Miller's " Treatise on Steam Boilers." Both are based on the principle that the density, relatively to hydrogen, of an elementary gas (O and N) is proportional to its atomic weight, and that of a compound gas (CO and C0 2 ) to one-half its molecular weight. Both formulas are very nearly accurate when pure carbon is the fuel burned, but formula B is inaccurate when the fuel contains hydrogen, for the reason that that portion of the oxygen of the air supply which is required to burn the hydrogen is contained in the chimney gas as H 2 0, and does not appear in the analysis of the dry gas. The following calculations of a supposed case of combustion of hydrogenous fuel illustrates the accuracy of formula A and the inaccuracy of formula B. Assume that the coal has the 226 POWER PLANT TESTING following analysis: C, 66.50; H, 4.55; O, 8.40; N, 1.00; water, 10.00; ash and sulphur, 9.55 — total, 100. Assume also that one-tenth of the C is burned to CO, and nine-tenths to C0 2 ; that the air supply is 20 per cent in excess of that required for this combustion; that the air contains 1 per cent by weight of moisture ; and that the S in the coal may be considered as part of the ash. We then have the following synthesis of results of the combustion of 100 pounds of coal: O from Air N = ox 15 Total Air C0 2 CO H02 59.85 lbs. C to C0 2 X2f 6.65 " CtoCO Xi| 3.50 " HtoH,OX8 T59.60 8.87 28 .OO 534-3 1 29.70 93-74 693.91 38.57 121.74 219.45 15-52 31-5° 196.47 657-75 854.22 1.05 " HtoH 2 0\ 8.40 " HtoHjO/ 9-45 1. 00 " N 1 .00 9.55 ' Ash and S Excess of air 20 per cent. 39-29 i3i-55 170.84 1025.06 Moisture in air 1 per cent. 10.25 Total wt. gases, 11 2 5. 67 lbs. = Total dry gases, 1064.56 lbs. Per Cent. Total dry gases, by weight, Total dry gases, by volume, 39-29 3-69 3-5o8 790.30 N 74.24 80.656 219.45 CO2 20 . 61 14.252 15-52 CO 1-546 1-584 61 . 20 Total gases 1 125.76 + ash and S 9.55 = 1135.3 1 lbs. total products. Total air 102 5.06 + moisture in air 10.25 + coal 100 = 1135.31 lbs. Dry gas per lb. coal 10.6456; per lb. carbon = 10.6456 -h .665 =16.008 lbs. Dry air per lb. coal 10.2506; per lb. carbon = 10.2506 -^.665 = 15.414^3. Computation of the weight of dry gas and of air per lb. carbon. Formula A: Dry gas per lb. C Formula B : .252X11+3.508X8 + 82.240X7 3(14.252 + 1.584) = 16.008. pounds. .. . An 2(14- 252+3. 5o8) + i584 Air per pound L = 5.8 — = 13.589 pounds. 14. 252 + 1. 584 The error in the last result is 15.414 — 13-589 = 1.825 pounds. BOILER TESTING 227 228 POWER PLANT TESTING Professor D. S. Jacobus gives another formula for the air per pound of carbon, in which the error of formula 62 is almost entirely avoided. It is Formula C: 7N N Air per pound C= 3(C02 + CO) ^o. 77) or O , 33(c o 2+C0) , (63) in which N, C0 2 , and CO are the percentages by volume of these gases. Making the computation from the data of the above analysis, we have: 80.656 Air per pound C = ^ —— = 15.434 pounds, P F 0.33(14.252+1.584) .^** the true value being 15.414 pounds. Fig. 180 is a diagram showing the heat distribution and losses in a steam boiler and engine plant, due mainly to Parsons. The method of showing graphically the percentage of losses is particularly interesting. CHAPTER XI STEAM ENGINE TESTING Most important of the tests made of nearly all classes of machinery is that for mechanical efficiency; meaning the comparison of the useful work performed with the amount of work theoretically possible to obtain with a perfect machine. In other words, in an engine the mechanical efficiency, E m , is the ratio of the brake horse power to the indicated horse power, or ■p TT p Em= L ' H 'p' ($4) ■ The difference between the indicated horse power and the brake horse power is called the friction horse power. In many cases with very large engines, it is not readily possible to obtain the brake horse power directly, and in such cases it is customary to obtain approximately the horse power lost in friction from a so-called " friction indicator diagram," obtained from the areas of indicator diagrams when the only work done is that required to overcome its own friction, or in common parlance, when the engine is " running light." The brake horse power is then taken to be the difference between the indicated horse power and the friction horse power. Such a determination of friction horse power and of mechanical efficiency by calculation cannot be considered very accurate, because the friction of an engine increases slightly with in- creasing loads. Observed and calculated data of mechanical efficiency may be tabulated in the following form : 229 230 POWER PLANT TESTING STEAM ENGINE TESTING TESTS FOR MECHANICAL EFFICIENCY AND FRICTION Date Description of engine tested Test made by . Tare of Brake lbs. Length of brake arm feet . Engine and Brake Constants (see pages 121 and 123.) No. of Read- Time. Weight on Brake.lbs. Areas of Indicator Cards, sq.ins. Indicated Horse power. Brake Horse power. Fric- tion Horse power. Mech. Effic. ing. ■z, Head End. Crank End. Head. End. Crank End. Total % Valve Setting (Slide Valve Engines). In order that steam may be used economically in an engine, it is necessary that the valve be set carefully and accurately, so that when an indicator card is taken the diagram obtained will be as nearly as possible like the ideal. Adjustment of a slide valve on an engine is accomplished in two different ways, with different effects : (1) By moving the valve on its stem; (2) By adjusting the eccentric. » To Set the Valve for Equal Leads. The valve should be placed first in such a position on its stem (valve rod) that the amount of its travel will be the same on the two sides of its mid-position. In other words, the valve is to be set on its stem so that its movement will be symmetrical with respect to the ports. Typical slide valves are shown in mid-position in Figs. 181 and 182. Setting the valve symmetrically or in mid- position is easily accomplished by turning the engine (or by moving the eccentric on the shaft) until the valve is brought to the farthest point of its travel on one side of its mid-position. Then measure the width 1 of the part of the steam port 'which is uncovered. The valve should then be shifted to the limit of its travel at the other end so that the width of the uncovered 1 It is assumed of course that corresponding dimensions of the ports are the same at the two ends of the valve seat. STEAM ENGINE TESTING 231 portion of the steam port can be measured at that end. If the widths measured at the two ends are not the same, the valve must be moved on its stem a distance equal to half the difference between the widths of the uncovered ports. This movement of the valve will be in a direction away from the Exhaust Lap _J L_ -*| («- Exhaust Lap Steam Lap Ordinary D-slide Valve in Mid-position. port having the smaller opening. Measurements should be repeated and if the widths of the uncovered ports are still unequal this method of adjustment must be continued. By the method described the two leads of the valve will have been made the same; in other words, the distance the Exhaust Lap Steam Lap steam. Lap j**i Exhaust Lap Fig. 182. — Piston Type of Slide Valve in Mid-position. valve uncovers the steam ports when the engine is on the dead- centers will be the same at both ends of the cylinder. Now move the engine accurately to the dead-center * toward the 1 An engine can be put on dead-center quite accurately by the " method of trammels." When the engine is just a little off the center to be deter- mined, make small scratch-marks opposite each other both on the cross- head and on one of the guides. Now set a pair of dividers or trammels with one end resting on the bedplate of the engine, its foundation, or some convenient stationary object near the fly-wheel, and with the other end 232 POWER PLANT TESTING head end, for example, and move the eccentric on the shaft in the direction in which the engine is to run 1 until the steam port at the head end of the cylinder has been opened a distance equal to the lead required and in the position defined in general by stating that a further movement of the valve in the same direction will open the port wider. In this position the eccentric should be firmly fixed to the shaft. The engine should then be turned over to the other dead-center so that the lead can be measured at that end. If the leads are not the same, the difference should be halved, one part to be taken up by moving the valve on its stem, and the other by moving the eccentric. To set the valve for equal cut-offs, the valve is first set on its stem so that the travel will be the same on both sides of the mid-position as explained at the bottom of page 230 to make its movement symmetrical with the ports. Now an adjustment of the eccentric must be made so that steam will be cut off at each end of the cylinder when the piston has moved the same distances or the same per cent of the stroke from the two ends. To perform this adjustment, mark on the cross-head as accurately as possible the limits of the stroke, and set the cross-head at the per cent of the stroke for cut-off appearing to be most suitable for the conditions of load. Move the eccentric on the shaft in the direction in which the engine is to run until it can be seen that the valve would be just closing the steam port at the end of the cylinder from which the piston is moving. Fasten the eccentric securely in this position and turn the engine over to observe whether the valve will be just closing the other steam port when the piston has moved the same distance, measured on the cross-head, mark a point on the fly-wheel. The engine should then be moved over or beyond the dead-center until the marks made on the cross-head and on the guide come together again. With the dividers set with the points the same distance apart as before again put a mark on the fly-wheel. Then if the engine is turned back so that the end of the dividers used to mark on the fly-wheel is at a point half way between the two marks, it will be set quite accurately on the dead-center required. In all these adjustments care must be taken to turn the engine each time in the same direction with respect to the dead-center so that the lost motion or back lash is taken up in the same direction. 1 This applies only to a valve like the one in Fig. 181, which takes steam on the "outside." When the valve takes steam on the inside (Fig. 182) the eccentric must be moved in the opposite direction. STEAM ENGINE TESTING 233 from the other end of the cylinder. If the setting is not correct, the error should be halved, correcting for one half by moving the valve on the stem and for the other by moving the eccentric on the shaft. This operation, which is a " cut and try " process, must be repeated until the required setting is secured. Methods similar to those described are preferred usually for the accurate setting of the slide valves of slow- and medium.- speed engines. High-speed engines as well as slow-speed engines of the Corliss type have their valves set usually on the basis of the information secured from indicator diagrams taken on the engines, showing approximately the " timing " of the events of the stroke. To set a slide valve successfully Cut Off Compression Atmospheric Line Fig. 183. — Indicator Diagram Illustrating the Point of Cut-off. by the " indicator " method, the valve and ports should be measured to determine the " lap " dimensions and port openings indicated in • Fig. 181, page 231, and the valve travel by direct measurements. "With these data a Zeuner 1 valve diagram should be constructed, showing a good steam dis- tribution for assumed lead or cut-off. Then construct the theoretical indicator card from the Zeuner diagram and adjust the setting of the valve on t*he stem and the eccentric on the shaft till a close approximation to the theoretical card is obtained. In this adjustment the first thing to be done is to equalize the valve travel by locating it on its stem so that the travel will be the same on both sides of its mid-position. 1 It is beyond the scope of this book to take up a discussion of valve diagrams. The theory and construction of the Zeuner diagram is given in nearly all books on the steam engine. 234 POWER PLANT TESTING Use a spring light enough to give an indicator diagram about 1 1 inches high so that events of the stroke, — admission, cut-off, release, and compression, will be shown as clearly as possible. Sometimes it is difficult to determine these events on a diagram on account of the curves gradually running into each other without the point separating the different curves being clearly defined. A good method for such cases is to produce along their regular trend both of the curves of which the intersection is required and take for the intersection the point where these curves cross each other. The method is illustrated on an indicator card in Fig. 183 showing the point of cut-off. In a slide valve engine it is not possible to set the valve to secure at the same time equal cut-offs and equal leads. Ideal and imperfect indicator diagrams taken from slide valve and Corliss engines are shown in Fig. 184. A little study of such diagrams may help to solve many difficulties in valve setting. Setting Corliss Valves. A brief desciiption 1 of the essential parts of the valve gear of a Corliss engine will assist in obtain- ing a clearer conception of the subject. In Figs. 185 and 186 similar capital letters of reference indicate the same parts of the mechanism. Fig. 185 shows all the essential parts of the valve gear. The steam valves work in the chambers S, S and the exhaust valves work in the chambers E, E. The double-armed levers AC, AC, work loosely on the hubs of the valve-stem brackets and the lever arms B, B; the former are connected to the wrist plate W by the rods M, M ; the levers B, B are keyed to the valve stems V, V, and are also connected by the rods 0, to the dashpots D, D. The double-armed levers carry at their outer ends, C, C hardened steel catch plates, which engage with arms B, B, making the two arms B and C work in unison until steam is to be cut off; at this point another set of levers H, H, connected by the cam rods G, G to the governor, come into play, causing the catch plates to release the arms B, B, the outer ends of which are then pulled downward by the weight of the dashpot plunger, causing the steam valves 1 This description is from American Machinist, vol. 18, page 391. For clearness the article is considered unusually good. STEAM ENGINE TESTING 235 236 POWER PLANT TESTING to rotate on their axes and thus cut off steam. These are the essential features of the Corliss gear, although the design of the mechanism is greatly modified by different builders. The exhaust valve arms F are connected to the wrist plate by the rods N, N, and it is seen that all the valves receive their motion from the wrist plate; the latter receives its motion from the hook rod I; this rod is generally attached to a rocker arm, not shown; to this arm the eccentric rod is Corliss Valve Gear. also attached. The carrier arm is usually placed about mid- way between the wrist plate and eccentric, and in the center of its travel stands in a vertical position. The setting of the valves is not a difficult matter when, on the wrist plate, its support, valves and cylinder, the cus- tomary marks have been placed for finding the relative positions of wrist plate and valves. Now referring to Fig. 186, when the back bonnets of the valve chambers have been taken off, there will be found a mark or line a on the end of each steam valve s, s, coinciding STEAM ENGINE TESTING 237 with the working or opening edges of each valve ; another line b will be found on each face of the steam valve chamber coin- ciding with the working edge of the valve, and the line h, on the face of each exhaust valve chamber, coincides with the working edge of the exhaust port. On the hub of the wrist plate will be found a line d, coinciding with the center line d, k; lastly, there are three lines f, c, f, on the hub of the wrist- plate support, placed in such a way that when the line d coincides with the line c, the wrist plate will stand exactly in the center of its motion, and when the line d coincides with either of the lines f, f, the wrist plate will be at one of the extreme ends u or v of its travel. It should be noticed that since the lines f , e, f are drawn on the periphery of the hub of Fig. i 86. — Diagram of a Corliss Valve Mechanism the wrist-plate support, and the line d is drawn on the periph- ery of the wrist-plate hub, these lines cannot stand in a vertical line, as shown; we have adopted this way of showing them, simply for the purpose of making the matter plain. In setting the valves, the first step will be to set the wrist plate in its central position, so that lines c and d will coincide, and fasten the wrist plate in this position by placing a piece of paper between it and the washer L on its supporting pin. Now set the steam valves so that they will have a slight amount of lap, that is to say, the lines a, a must have moved a little beyond the lines b, b; the amount of this lap depends much on individual preference and experience; it ranges from ^ to -J inch for small engines, and from | to ^ inch for comparatively large engines. This lap is obtained by length- 238 POWER PLANT TESTING ening or shortening the rods M, M by means of the adjusting nuts. Now place the exhaust valves e, e, by lengthening or short- ening the rods N, N by means of the adjusting nuts in a position so that the working edges will just open the exhaust ports, or, in other words, place the lines g and h nearly in line with each other. Some engineers prefer a slight amount of lap, others prefer a slight opening of the exhaust ports when the valves are in this position; under these conditions the lines g and h cannot be in line, but will stand apart, as indicated in the illustration; the distance between these lines will, of course, be equal to the desired amount of opening; for small engines it is about fg inch, and for larger engines may be increased to ^ inch, but in any case the amount of this opening should be less than the lap of the steam valves, other- wise there will be danger of steam blowing through. The paper between the wrist plate and the washer on the supporting pin should now be taken out, so that the wrist plate connected to the valves can be swung on its pin. The next step will be to pay some attention to the rocker arm. Set this arm in a vertical position by means of a plumb- line, and connect the eccentric rod to it ; then turn the eccentric around on the shaft, and see that the extreme points of travel are at equal distances from the plumb-line. To secure this a little adjustment in the stub end of the eccentric rod may be necessary. Now connect the hook rod I to its pin on the wrist plate, and again turn the eccentric around on the shaft, and thus determine the extreme points of travel of the wrist plate. If all parts have been correctly adjusted, the line d will coincide with the lines f, f at the extreme points of travel; if this is not the case, the hook rod will have to be adjusted at its stub end so as to obtain the desired equalized motion of the wrist plate. The next step will be to set the valves correctly with the position of the crank; to do so the lengths of the rods M, M, N, and N must not be changed, but the following mode of pro- cedure should be followed : Place the crank on one of its dead- centers, and turn the eccentric loosely on the shaft in the direction in which the engine is to run, until the steam valve nearest to the piston shows an opening or lead of ^ to £ inch, STEAM ENGINE TESTING 239 according to size of engine, the smaller lead, of course, being adopted for small engines. After the proper lead has been given to this valve, secure the eccentric, and turn the shaft with eccentric in the same direction in which the engine is to run until the crank is on the opposite dead-center, and notice if the opening or lead at this end of the cylinder is the same as on the other steam valve; if not, shorten or lengthen slightly, as may appear necessary, the connection between wrist plate and eccentric; of course much adjust- ment in the length of these connections is not admissible with- eout rsetting the valves with reference to the wrist plate. The only thing which now remains to be done is to adjust the cam rods G, G. To do so, secure the governor balls in their highest position, and disconnect the hook rod from wrist pin; lengthen or shorten the cam rods G, G, so as to bring the detachment apparatus into action, swing the wrist plate back and forward and make such adjustment in the rods G, G, as to permit the steam valves to be released when the steam port has been opened about J inch. This adjustment is for the purpose of keeping the engine under the control of the governor, in case, for some reason or another, the load on the engine is suddenly thrown off. After this adjustment the governor balls should be placed in their lowest position, in which the releasing gear should not detach the steam valves, but allow the steam to follow nearly full stroke. Sometimes the releasing gear is constructed in such a manner as to close the steam valves automatically, in case the belt leading to the governor should be broken, or the load on the engine suddenly thrown off. In cases of this kind the governor balls need not be placed in their highest position, but should be placed in their lowest position, and the wrist plate moved to either end of its extreme travel; the steam port opposite this end of travel of wrist plate will then-be wide open; now adjust the corresponding cam rod so that the releasing gear is nearly on the point of releasing the valve; then move the wrist plate to other end of its extreme travel, and adjust the other cam rod in the same manner. To prove the correctness of the cut-off adjustment, raise the governor balls to about a position where they would be when at work, or to a medium height, and block them there ; then, with the connection made 240 POWER PLANT TESTING between the eccentric and the wrist plate, turn the engine shaft slowly in the direction in which it is to run, and when the valve is released measure upon the slide the distance through which the cross-head has moved from its extreme position. Continue to turn the shaft in the same direction, and when the other valve is released, measure the distance through which the cross-head has moved from its extreme position, and if the cut-off is equalized, these two distances will be equal to each other. If they are not, adjust the length of the cam rods until the points of cut-off are at equal dis- tances from the beginning of the stroke. Replace the back bonnets and see that all connections have been properly made, which will complete the setting of the valves. Wherever convenient, it is desirable that an indicator be applied to the engine when at work, and the setting of the valves tested. If necessary, they should be readjusted for the best possible condition for economical work. Clearance Determination of an Engine. This test is made usually to determine the clearance volume of a steam or a gas engine. It is sometimes important to know the clearance volume of an engine, as it materially affects the expansion curve of the engine. If it is too large it causes an excessive loss in the engine. The clearance volume is also necessary if a theoretical expansion curve is to be constructed. The engine is first set on dead-center with the piston at the head end of the cylinder. This is done by the " method of trammels " (see footnote, page 231). Then the steam chest cover and valve are to be removed and a rubber gasket under a block of wood is placed over both steam-port openings in the valve seat and bolted on. Usually candle wicking must be packed around the piston to stop excessive leakage. For this purpose the cylinder head must be removed and again replaced. Two vessels filled with clean water should be provided and weighed. The clearance space is to be filled from one vessel, the time required being taken. As soon as the space is filled the first vessel is removed and the space is kept filled with water from the other vessel for five minutes. The vessels are then again weighed and the water used from each of them deter- mined. The average rate of leakage while filling the space is STEAM ENGINE TESTING 241 usually assumed to be one-half the rate of leakage when full of water as during the leakage test. If w x = weight in pounds to fill the clearance space, / minutes = the time required to fill the clearance space, and w 2 = the weight of water in pounds necessary to keep it full for one minute ; then the leakage during filling is w' = —Xt, 2 and the clearance = (wi—w f ) in pounds of water, which can be readily reduced to cubic inches. The clearance for the crank end is found in the same general way as that for the head end. Most engines have small holes at the top of the cylinder at each end (for double-acting engines) which lead into the clearance space. Holes which are covered by the piston on the dead-center would obviously be of no value. All water must of course be drained from each end before filling. Remov- ing the cylinder head for packing the piston is the best method for observing with certainty that there is no water in the head end. The drip pipes in the cylinder can usually be relied on to remove water from the crank end. Determinations should be repeated several times, and the average value is to be used in calculations. RULES FOR CONDUCTING STEAM ENGINE TESTS An elaborate report has been published by the American Society of Mechanical Engineers entitled " Rules for. Conducting Steam Engine Tests." 1 In this report the rules regulating standard commercial tests of engine plants, including auxiliary machinery, are given in great detail. Every engineer should have available a copy of these rules. It is impracticable to print in this book the complete report, but space will be taken to give the most important parts. 1 This report is issued in a pamphlet of 78 pages and can be obtained at at small cost from the Secretary of the American Society of Mechanical Engineers, Engineering Society's Building, New York city. It is printed complete in the Transactions of the Society, Vol. 24, pages 713- 846. 242 POWER PLANT TESTING One of the most important subjects discussed is in regard to the definition of the " Heat consumption " of an engine plant. This is to be determined by measuring the quantity of steam consumed by the plant, calculating the total heat of the entire quantity and crediting this total with that portion of the heat rejected by the plant which is utilized and returned to the boiler. " Engine plant " as used in this report should include the entire equipment of the steam plant producing power ; that is, the main cylinder or cylinders, the steam jackets and reheaters, the air, circulating and boiler-feed pumps if steam driven, and any other machinery driven by steam required for the operation of the engine. That the engine plant should be charged with the steam used by all the auxiliary machinery in determining the plant economy is necessary because the steam consumption of the engine is finally benefited, or at least it should be, by the heat they return to the system. It is, of course, now the general practice in commercially operated plants to pass the exhaust steam from auxiliaries operated non-condensing, through a feed-water heater, thus carrying back to the boiler a great deal of the heat. When a plant is operating non-condensing, discharging the steam into the atmosphere, or with a jet condenser, the steam consumption of the engine cannot be determined by weighing or measuring the steam used as can be done when a surface condenser is used. The method followed in this case is to determine the steam used by the weight of water supplied to the boiler, assuming, of course, that all the steam from the boiler or boilers used goes to the engine tested. It can usually be arranged for a test of one of the engines in a large plant that one or more boilers can be segregated or cut off in the piping connections so that these boilers alone supply the engine. When, however, this method is to be used it is necessary to determine by a separate test the leakage of the boiler and of the piping from the boiler to the throttle valve of the engine. This leakage, of course, is the amount of feed-water pumped into the boiler to keep the level in the water gage constant without taking away any more steam than is lost in this way. When determining this leakage, the pressure in the boiler must be maintained the same as that at which steam is to be supplied to the engine for its test. The feed-water pumped STEAM ENGINE TESTING 243 into the boiler supplying the engine less the boiler and pipe leakage will be the net amount of steam used by the engine. Surface condensers are usually perferred for accurate engine tests because the steam used by the engine can be determined directly by weighing or measuring the condensed steam. A surface condenser is essentially a vessel of considerable size in which there are a great many brass tubes. It is usually designed so that exhaust steam passes on its way through the condenser into contact with the outside surface of these tubes while cold water for condensing the steam circulates inside the tubes. Steam condensed in this way accumulates in the bottom of the condenser and is removed by an air pump used for pro- ducing a vacuum, or by gravity if the pressure in the condenser is atmospheric, as it will be if the engine is operating " non- condensing," that is, without a vacuum. It is very essential, however, that surface condensers be tested for leakage, preferably before and after every important test is made ; and the leakage should be determined with the same vacuum in the condenser as there is when it is used in the engine test. Probably the best method to determine the amount of condenser leakage is to pass cooling water through the tubes at the normal rate of flow, maintaining at the same time with the air pump the required vacuum. Then the amount of water removed by the air pump is the leakage of circulating water through the tubes into the steam space. Under normal operating conditions there would be no leakage of steam into the circulating water, because the water will be at higher pressure than the steam. If only a test is to be made to determine whether or not there is any leakage in the condenser, the test most generally applied is to close all connections to the condenser and observe whether a vacuum once established can be maintained a reason- ably long time. A more rapid test, applicable, however, only where salt water is used for cooling, is to test the condensed steam several times during a test by adding a little silver nitrate to a small amount of the condensed steam. If there is no appreciable precipitate x it may be assumed that the condenser does not leak. Some of the important considerations to be observed in an 1 The white precipitate formed with the salt in sea water is of course silver chloride, thus, AgN0 3 + NaCl = AgCl + NaN0 3 . 244 POWER PLANT TESTING accurate engine test will now be given as stated in the rules adopted by the American Society of Mechanical Engineers. General Conditions of Plant. Examine the engine and the entire system of piping concerned in the test; note its general condition and any points of design, construction, or operation which bear on the objects in view. Make a special examination of the valves and pistons for leakage by applying the working pressures with the engine at rest, and observe the quantity of steam, if any, blowing through per hour. Dimensions, etc. Measure the dimensions of the cylinders of the engine when it is hot. If they are much worn, the average diameter should be determined. Measure also the clearance, which should be done, if possible, by filling the spaces with water previously measured, the piston being placed at the end of the stroke (see page 240). If the clearance cannot be measured directly, it can be determined approximately from the working drawings of the cylinder. Calibration of Instruments. All instruments and apparatus should be calibrated and their reliability and accuracy verified by comparison with recognized standards. Leakages of Steam, Water, etc. In all tests except those of a complete plant made under conditions as they exist, the boiler and its connections, both steam and feed, as also the steam piping leading to the engine and its connections, should, so far as possible, be made tight. All connections should, so far as possible, be visible and be blanked off, and where this cannot be done, satisfactory assurance should be obtained that there is no leakage either in or out. Duration of Tests. The duration of a test should depend largely upon its character and the objects in view. The standard heat test of an engine, and, likewise, a test for the simple deter- mination of the feed-water consumption, should be continued for at least five hours, unless the class of service precludes a continuous run of so long duration. It is desirable to con- tinue the test, the number of hours stated to obtain a number of consecutive hourly records as a guide in analyzing, the reliabil- ity of the whole. The commercial test of a complete plant, embracing boilers as well as engine, should continue at least one full day of twenty- four hours, whether the engine is in motion during the entire STEAM ENGINE TESTING 245 time or not. A continuous coal test of a boiler and engine should be of at least ten hours' duration, or the nearest multiple of the interval between times of cleaning fires. Starting and Stopping a Test. (a). Standard Heat Test and Feed-water Test of Engine. The engine having been brought to the normal condition of running, and operated a sufficient length of time to be thoroughly heated in all its parts, and the measuring apparatus having been adjusted and set to work, the height of water in the gage glasses of the boilers is observed, the depth of water in the reservoir from which the feed water is supplied is noted, the exact time of day is observed, and the test held to commence. Thereafter the measurements deter- mined upon for the test are begun and carried forward Until its close. If practicable, the test may be commenced at some even hour or minute, but it is of first importance to begin at such time as reliable observations of the water heights are obtained, whatever the exact time happens to be when these are satisfactorily determined. When the time for the close of the test arrives, the water should, if possible, be brought to the same height in the glasses and to the same depth in the feed-water reservoir as at the beginning, delaying the conclusion of the test, if necessary, to bring about this similarity of condi- tions. If differences occur, the proper corrections must be made. (b). Complete Engine and Boiler Tests. For a continuous running test of combined engine or engines, boiler or boilers, the same directions apply for beginning and ending the feed- water measurements as those just referred to. The time of beginning and ending such a test should be the regular time of cleaning the fires, and the exact time of beginning and ending should be the time when the fires are fully cleaned, just prepar- atory to putting on fresh coal. For a commercial test of combined engine and boiler, whether the engine runs continuously for the full twenty-four hours of the day, or only a portion of the day, the fires in the boilers being banked during the time when the engine is not in motion, the beginning and ending of the test should occur at the regular time of cleaning the fires, the method followed being that given above. In cases where the engine is not in continuous motion, as, for example, in textile mills, where the working time is ten or eleven hours out of twenty-four, and the fires are cleaned 246 POWER PLANT TESTING and banked at the close of the day's work, the best time for starting and stopping a test is the time just before banking, when the fires are well burned down and the thickness and condition can be most satisfactorily judged. Measurements of Heat Units Consumed by the Engine. The measurements of the heat consumption require the measure- ment of each supply of feed-water to the boiler — that is the water supplied by the main feed pump, that supplied by auxil- iary pumps, such as jacket water, water from separators, drips, etc., and the water supplied by gravity and other means; also the determination of the temperature of the water supplied from each source, together with the pressure and quality of the steam. The temperatures at the various points should be those applying to the working conditions. The heat to be determined is that used by the entire engine equipment, embracing the main cylinders and all auxiliary cylinders and mechanism concerned in the operation- of the engine, including the air pump, circulating pump, feed pumps, also the jacket and reheater, when these are used. The steam pressure and the quality of the steam are to be taken at some point conveniently near the throttle valve. The quantity of the steam used by the calorimeter must be determined and properly allowed for. Measurement of Feed-water or Steam Consumption of Engine, etc. The method of determining the steam consump- tion applicable to all plants is to measure all the feed-water supplied to the boilers, and deduct therefrom the water dis- charged by separators and drips, as also the water and steam which escapes on account of leakage of the boiler and its pipe connections and leakage of the steam main and branches con- necting the boiler and the engine. In plants where the engine exhausts into a surface condenser the steam consumption can be measured by determining the quantity of water discharged by the air pump, corrected for any leakage of the condenser, and adding thereto the steam used by jackets, reheaters, and auxiliaries as determined independently. The corrections or deductions to be made for leakage above referred to should be applied only to the standard heat-unit test and tests for determining simply the steam or feed-water consumption, and not to coal tests of combined engine and STEAM ENGINE TESTING 247 boiler equipment. In the latter, no correction should be made except for leakage of valves connecting to other engines and boilers, or for steam used for purposes other than the operation of the plant under test. Losses of heat due to imperfections of the plant should be charged to the plant, and only such losses as are concerned in the working of the engine alone should be charged to the engine. Measurement of Steam Used by Auxiliaries. It is highly desirable that the quantity of steam used by the auxiliaries, and in many cases that used by each auxiliary, should be deter- mined exactly, so that the net consumption of the main engine cylinders may be ascertained and a complete analysis made of the entire work of the engine plant. Indicated Horse Power. The indicated horse power should be determined from the average mean effective pressure of diagrams taken at intervals of twenty minutes, and at more frequent intervals if the nature of the test makes it necessary, for each end of the cylinder. With variable loads, such as those of engines driving generators for electric railroad work, and of rolling-mill engines, the diagrams cannot be taken too often. The most satisfactory driving rig for indicating seems to be some form of well-made pantagraph, with driving cord of fine annealed wire leading to the indicator. The reducing motion, whatever it may be, and the connections to the indicator, should be so perfect as to produce diagrams of equal lengths, when the same indicator is attached to either end of the cylinder, and produce a proportionate reduction of the motion of the piston at every point of the stroke, as proved by test. The use of the three-way cock and the single indicator con- nected to the two ends of the cylinder is not advised, except in cases where it is impracticable to use an indicator close to each end. If a three-way cock is used, the error produced by the increased clearance should be determined and allowed for. Testing Indicator Springs. To make a perfectly satis- factory comparison of indicator springs with standards, the calibrations should be made, if practicable, under the same conditions as those pertaining to their ordinary use. Brake Horse Power. This term applies to the power deliv- ered from the fly-wheel shaft of the engine. It is the power 248 POWER PLANT TESTING , absorbed by a friction brake applied to the rim of the wheel, or to the shaft. A form of brake is preferred that is self adjust- ing to a certain extent, so that it will, of itself, tend to main- tain a constant resistance at the rim of the wheel. One of the simplest brakes for comparatively small engines, which may be made to embody this principle, consists of a cotton or hemp rope, or a number of ropes, encircling the wheel arranged with weighing scales or other means of showing the strain. An ordinary band brake may also be constructed so as to embody the principle. The wheel should be provided with interior flanges for holding water used for keeping the rim cool. Quality of Steam. When saturated steam is used, its qual- ity should be obtained by the use of a good throttling calo- rimeter attached to the main steam pipe near the throttle valve. When the steam is superheated, the amount of superheating should be found by the use of a thermometer placed in a thermometer-well filled with mercury, inserted in the pipe. The sampling pipe for the calorimeter should, if possible, be attached to a section of the main pipe having a vertical direction, with the steam preferably passing upward, and the sampling nozzle should be made of a half -inch pipe, having at least twenty one-eighth inch holes in its perforated surface. Speed. There are several reliable methods of ascertaining the speed, or the number of revolutions of the engine crank- shaft per minute. The most reliable method, and the one recommended, is the use of a continuous recording engine register or counter, taking the total reading each time that the general test data are recorded, and computing the revolutions per minute corresponding to the difference in the readings of the instru- ment. When the speed is above 250 revolutions per minute, it is almost impossible to make a satisfactory counting of the revolutions without the use of some form of mechanical counter. Recording the Data. Take note of every event connected with the progress of the trial, whether it seems at the time to be important or unimportant. Record the time of every event, and the time of taking every weight, and every observation. Observe the pressures, temperatures, water heights, speeds, etc., every twenty or thirty minutes when the conditions are practically uniform, and at much more frequent intervals if the conditions vary. STEAM ENGINE TESTING 249 Uniformity of Conditions. In a test having for its object the determination of the maximum economy obtainable from an engine, or where it is desired to ascertain with special accuracy the effect of predetermined conditions of operations, it is impor- tant that all the conditions under which -the engine is operated should be maintained uniformly constant. Analysis of Indicator Diagrams, (a) Steam Accounted for by the Indicator. The simplest method of computing the steam accounted for by the indicator is the use of the formula ■.M~^[(C+E)XW C -(H+B)XW*], . . (65) which gives the weight in pounds per indicated horse power per hour. In this formula the symbol M.E.P. refers to the mean effective pressure. In multiple-expansion engines this is the " combined " mean effective pressure referred to the cylinder in question. C is the proportion of the stroke completed at a point on the expansion line of the diagram near the actual cut-off or release, the symbol H'to the proportion of compres- sion, and the symbol E to the proportion of clearance; all of which are determined from the indicator diagram. The symbol W c refers to the weight of one cubic foot of steam at the cut-off or release pressure; and W /( to the weight of one cubic foot of steam at the compression pressure, these weights being taken from steam tables. The points at cut-cff and release on the expansion line and the point at the beginning of compression are located as shown on the sample diagram, Fig. 187. On the expansion and com- pression lines they are the points which mark the complete closure of the valve. The point at cut-cff lies where the curve of expansion begins after the rounding of the diagram due to throttling or " wire-drawing " which occurs when the valve is closing. This point of cut-off may be located on the diagram by finding the point where the curve is tangent to a hyperbolic curve. Should the point selected in the compression curve be at the same height as the point in the expansion curve, then W c =W/ t , and the formula becomes ^|x(C-H) X W c) .... (66) 250 POWER PLANT TESTING in which (C — H) represents the distance between the two points divided by the length of the diagram. When the load and all other conditions are substantially uniform, it is unnecessary to work up the steam accounted for by the indicator from all the diagrams taken. Five or more sample diagrams may be selected and the computations based on the samples instead of on the whole. (b) Sample Indicator Diagrams. In order that the report of a test may afford complete information regarding the con- ditions of the test, sample indicator diagrams should be selected from those taken and copies appended to the tables of results. In cases where the engine is of the multiple-expansion type Atmospheric Line Fig. 187. — Typical Indicator Diagram. these sample diagrams may also be arranged in the form of " combined " diagrams. (c) The Point of Cut-off. The term "cut-off" as applied to steam engines, although somewhat indefinite, is usually con- sidered to be at an earlier point in the stroke than the beginning of the real expansion line. That the cut-off point may be defined in exact terms for commercial purposes, as used in steam- engine specifications and contracts, the Committee recommends that, unless otherwise specified, the commercial cut-off, which seems to be an appropriate expression for this term, be ascer- tained as follows: Through a point showing the maximum pressure during admission, draw a line parallel to the atmos- pheric line. Through the point on the expansion line near the actual cut-off, draw a hyperbolic curve. The point where these two lines intersect is to be considered the point at the com- mercial cut-off. The percentage is then found by dividing the STEAM ENGINE TESTING 251 length of the diagram measured to this point, by the total length of the diagram, and multiplying the result by ioo. The principle involved in finding the commercial cut-off is shown in Figs. 188 and 189. The first represents a diagram G f Fig. 188. — Indicator Diagram from a Slow-speed Corliss Engine. from a slow-speed Corliss engine and the second a diagram from a single -valve high-speed engine. In the latter case, owing to the inertia or " fling " of the indicator pencil, the steam line is irregular and the maximum pressure is found by taking the mean of the vibrations at the highest part of the curve. Fig. H G F — Indicator Diagram from a High-speed Single-valve Engine. The commercial cut-off, as thus determined, is situated at an earlier point of the stroke than the actual cut-off referred to in computing the " steam accounted for " by the indicator on page 249. 252 POWER PLANT TESTING (d) Ratio of Expansion. The commercial ratio of expansion for a simple engine is determined by dividing the volume cor- responding to the piston displacement, including clearance, by the volume of the steam at the commercial cut-off, including clearance. In the multiple expansion engine it is determined by divid- ing the net volume of the steam indicated by the low pressure diagram at the end of the expansion line, assumed to be con- tinued to the end of the stroke by the net volume of the steam at the maximum pressure during admission to the high pressure cylinder. The ideal ratio of expansion is the quotient obtained by dividing the volume of the piston displacement by the volume of the steam at the cut-off (the latter being referred to the throttle-valve pressure) less the volume equivalent to that retained at compression. In a multiple expansion engine the volumes to be used are those pertaining to the low pressure cylinder and the high pressure cylinder, respectively. (e) Diagram Factor. The diagram factor is the proportion borne by the actual mean effective pressure measured from the indicator diagram to that of a diagram in which the various operations of admission, expansion, release and compression are carried on under assumed conditions. The factor rec- ommended refers to an ideal diagram which represents the maximum power obtainable from the steam accounted for by the indicator diagrams at the point of cut-off, assuming first that the engine has no clearance; second, that there are no losses through wire-drawing the steam either during admission or release ; third, that the expansion line is a hyperbolic curve ; and fourth, that the initial pressure is that of the boiler and the back pressure that of the atmosphere for a non-condensing engine, and of the condenser for a condensing engine. In cases where there is a considerable loss of pressure between the boiler and the engine, as where steam is transmitted from a central plant to a number of consumers, the pressure of the steam in the supply main shpuld be used in place of the boiler pressure in constructing the diagrams. Standards of Economy and Efficiency. The hourly con- sumption of heat, determined by employing the actual temper- ature of the feed-water to the boiler, as pointed out in the STEAM ENGINE TESTING 253 section entitled " Measurement of Heat Units Consumed by the Engine " on page 246, divided by the indicated and brake horse power, that is, the number of heat units consumed per indicated and per brake horse power per hour, are the standards of engine efficiency recommended by the Committee. The con- sumption per hour is chosen rather than the consumption per minute so as to conform with the designation of time applied to the more familiar units of coal and water measurement which have heretofore been used. The British standard where the tem- perature of the feed-water is taken as that corresponding to the temperature of the back-pressure steam, allowance being made for any drips from jackets or reheaters, is also included in the tables. It is useful in this connection to express the efficiency in its more scientific form, or what is called the " thermal efficiency ratio." The thermal efficiency ratio is the proportion which the heat equivalent of the power developed bears to the total amount of heat actually consumed, as determined by test. The heat converted into work represented by one horse power is 1, 980,0 co foot-pounds per hour, and this divided by 778 equals 2545 B.T.U. Consequently, the thermal efficiency ratio is expressed by the fraction 2 545 B.T.U. per I.H.P. per hour' Heat Analysis. For certain scientific investigations it is useful to make a heat analysis of the indicator diagram, to show the interchange of heat from steam to cylinder walls, etc., which is going on within the cylinder. This is unnecessary for commercial tests. Entropy-temperature Diagram. The study of the heat analysis is facilitated by the use of the entropy-temperature diagram in which areas represent quantities of heat, the co-ordi- nates being the entropy and absolute temperature. Ratio of Economy of an Engine to that of an Ideal Engine. The ideal engine recommended for obtaining this ratio is that which was adopted by the Committee appointed by the Civil Engineers of London, to consider and report a standard thermal efficiency for steam engines. This engine is one which follows the Rankine cycle where steam at a constant pressure is admitted 254 POWER PLANT TESTING into a cylinder having no clearance, and after the point of cut- off is expanded adiabatically to the back-pressure. In obtain- ing the economy of this engine the feed-water is assumed to be returned to the boiler at the exhaust temperature (page 261). returned to the boiler at the exhaust temperature. The ratio of the economy of an engine to that of the ideal engine is obtained by dividing the heat consumption per indi- cated horse power per minute for the ideal engine by that of the actual engine. Miscellaneous. In the case of the tests of combined engines and boiler plants, where the full data of the boiler performance are to be determined, reference should be made to the directions given by the Boiler Test Committee of the Society, Code 1899. (See pages 203-228.) In testing steam pumping engines and locomotives in accord- ance with the standard methods of conducting such tests, recom- mended by the Committees of the Society, reference should be made to the reports of those Committees in the Transactions American Society of Mechanical Engineers, vol. 12, page 530, and in vol. 14, page 131 2. Report of Tests. The data and results of the test should be reported in the manner and in the order outlined in one of of the following tables, the first of which gives, it is hoped, a complete summary of all the data and results as applied not only to the standard heat -unit test, but also to tests of a com- bined engine and boiler for determining all questions of per- formance, whatever the class of service. It is recommended that any report of a test be supplemented by a chart in which the data of the tests are graphically pre- sented. (As example of such a chart as applied to a boiler test see page 202.) DATA AND RESULTS OF STANDARD HEAT TEST OF STEAM ENGINE Arranged according to the Short Form Advised by the Engine Test Committee of the American Society of Mechanical Engineers. Code of 1902. 1. Made by on Engine located at Test made to determine. Date of Trial STEAM ENGINE TESTING 255 3 . Type and class of engine ; also of condenser 4. Dimensions of main engine: (a) Diameter of cylinder, inches (b) Stroke of piston, feet (c) Diameter of piston rod, inches (d) Average clearance, per cent (e) Horse power constant for one pound mean effective pressure and one revolution per minute 5. Dimensions and type of auxilaries TOTAL QUANTITIES, TIME, ETC. 6. Duration of test, hours 7. Total water fed to boilers from main source of supply, lbs. . . . 8. Total water fed from auxiliary supplies: (a) " . . . (b) " ... (c) " . . . 9. Total water fed to boilers from all sources " 0. Moisture in steam or superheating near throttle. ... % or deg. 1. Factor of correction for quality of steam 2. Total dry steam consumed for all purposes lbs. . . . HOURLY QUANTITIES 13. Water fed from main source of supply per hour, lbs 14. Water fed from auxiliary supplies: " " " (a) " (b) " (c) " 15. Total water fed to boilers per hour " 16. Total dry steam consumed per hour " 17. Loss of steam and water per hour due to drips from main steam pipes and to leakage of plant lbs 18. Net dry steam consumed per hour by engine and auxiliaries, lbs PRESSURES AND TEMPERATURES (CORRECTED) 19. Pressure in steam pipes near throttle by gage lbs. per sq. in 20. Barometric pressure of atmosphere in inches of mercury, ins 21. Pressure in receivers by gage lbs. per sq. in 22. Vacuum in condenser in inches of mercury ins 23. Pressure in jackets and reheaters by gage lbs. per sq. in 24. Temperature of main supply of feed- water deg. F 25. Temperature of auxiliary supplies of feed-water: (a) " (b) " (c) " 26. Ideal feed- water temperature corresponding to the pressure of steam in the exhaust pipe, allowance being made for the heat derived from the jacket or reheater drips deg. F 256 POWER PLANT TESTING DATA RELATING TO HEAT MEASUREMENTS Heat units per pound of feed-water, main supply B.T.U. . . Heat units per pound of feed- water, auxiliary supplies: (a) (b) (c) 43- 45- Heat units consumed per hour, main supply Heat units consumed per hour, auxiliary supplies : (a) (b) (c) .... Total heat units consumed per hour for all purposes Loss of heat per hour due to leakage of plant, drips, etc. Net heat units consumed per hour: (a) By engine alone (b) By auxiliaries Heat units consumed per hour by engine alone, reckoned from temperature given in item 26 B.T.U INDICATOR DIAGRAMS Commercial cut-off in per cent of stroke Initial pressure lbs. per sq. in. above atmosphere Back pressure at mid stroke above or below atmosphere, in pounds per square inch Mean effective pressure in pounds per square inch Equivalent M.E.P. in pounds per square inch: (a) Referred to first cylinder (b) Referred to second cylinder (c) Referred to third cylinder Pressures above zero in pounds per square inch : (a) Near cut-off (b) Near release (c) Near beginning of compression Percentage of stroke at points where pressures are measured: (a) Near cut-off (b) Near release (c) Near beginning of compression r Steam accounted for by indicator in pounds per I.H.P. per hour: (a) Near cut-off (b) Near release Ratio of expansion (page 252): (a) Commercial (b) Ideal SPEED Revolutions per minute POWER Indicated horse power developed by main engine cylinders : First cylinder • Second cylinder Third cylinder Total Brake horse power developed by engine STEAM ENGINE TESTING 257 STANDARD EFFICIENCY AND OTHER RESULTS 46 Heat units consumed by engine and auxiliaries per hour: (a) Per indicated horse power, B.T.U (b) Per brake horse power, B.T.U 47. Equivalent standard coal in pounds per hour: (a) Per indicated horse power, pounds (b) Per brake horse power, pounds 48. Heat units consumed by main engine per hour corresponding to ideal maximum temperature of feed water given in item 26: (a) Per indicated horse power, B.T.U (b) Per brake horse power, B.T.U 49. Dry steam consumed per indicated horse power per hour: (a) Main cylinders, including jackets, pounds (b) Auxiliary cylinders, pounds (c) Engine and auxiliaries, pounds 50. Dry steam consumed per brake horse power per hour: (a) Main cylinders, including jackets, pounds (b) Auxiliary cylinders, pounds (c) Engines and auxiliaries, pounds 51. Percentage of steam used by main engine cylinders accounted for by indicator diagrams, near cut-off of high pressure cylinder, %. ADDITIONAL DATA Add any additional data bearing on the particular objects of the test or relating to the special class of service for which the engine is used. Also give copies of indicator diagrams nearest the mean, and the corresponding scales. Heat Balance. The importance of checking tests of engines cannot be too strongly stated. An important aid for deter- mining the correctness of such tests is obtained by calculating a heat balance, which is making a balance sheet showing the heat received and rejected by the engine. Only such tests should be considered satisfactory which show a reasonable agreement in the heat balance. In ordinary commercial test- ing by experienced engineers a heat balance is not often calcu- lated, but for accurate laboratory work it should always be made out. It shows at a glance how much heat energy is received by the engine and what disposition is made of it. In a heat balance the heat supplied to the engine will be accounted for in the following items: (1) Heat equivalent of useful work as calculated from brake horse power. (2) Heat equivalent of engine friction. (3) Heat discharged in the con- densed steam (into hot well). (4) Heat absorbed by cooling water. (5) Heat radiated and other losses (by difference). 258 POWER PLANT TESTING Heat supplied to the engine per minute Q =pounds of steam supplied to engine per minute times the total heat in a pound of steam (q x +xi ri), for wet steam and Hi +c p (t' — ti) for super- heated steam. The terms q 1; x 1} r 1} Hi and ti represent as in other equations in this book respectively the heat of the liquid, the quality, the heat of vaporization, the total heat, and the temperature, all corresponding to the pressure of the steam supplied. The other term t' is the temperature of the super- heated steam as observed by a thermometer in the steam pipe. Heat equivalent of the useful work done (Q M ) per minute is the product of the brake horse power (B.H.P.) and the constant 33, ceo divided by the mechanical equivalent of a B.T.U. ; that is, B.H.P X33.000 778 Similarly, the heat equivalent of the engine friction (Q/) per minute is expressed by F.H.PX33.000 778 where F.H.P. is the friction horse power or the difference between the indicated 1 and the brake horse power. Heat discharged in the exhaust to the hot-well may be repre- sented by Q e = pounds of steam per minute times the heat of the liquid at the temperature of the exhaust. Or for practical purposes this is the same as taking the product of the pounds of steam per minute and the temperature of the exhaust (condensed steam) in degrees Fahrenheit less 32. The heat absorbed in the cooling water per minute is then Q c =pounds of cooling water per minute times the difference between the heat of the liquid at the outlet and inlet. For the small difference in temperature occurring in the cooling water of engine tests we may take in place of the difference of the heats of the liquid the difference of the corresponding tem- peratures, or approximately, 1 Often it is preferred, to express the indicated work in the heat bal- ance instead of the two items useful work and engine friction. Then the r , • -,- , , LH.P.X 3^000 heat equivalent of the indicated work Q,= ;STEAM ENGINE TESTING 259 Q c = pounds cooling water per minute times the difference between outlet and inlet temperatures. Heat radiated and other Q r losses is found, by taking the difference between the heat supplied Q and the sum of the several items Q u , Qf, Q e and Q c . There is no "standard" method for calculating a heat balance. Some engineers prefer to take for the base of com- parison; that is, the heat supplied, the total heat of the steam less q 2 , the heat of the liquid in the condensed steam (at the temperature of the exhaust) . With this value for the heat supplied the item Q e drops out and the net heat supplied, Q M = q 1 +x 1 r 1 -q 2 = Q w +Q / + Q f +Q r . HEAT BALANCE B.T.U. B.T.U. per min. per min. Per cent. Heat supplied to the engine, Q> Heat equivalent of Useful work Q u Heat equivalent Engine friction Q/ Heat discharged in exhaust G>. . . . Heat absorbed by Cooling Water Q c . Heat Radiated and Other losses Or. . . Total, . Total, ioo Obviously the totals on the left- and right-hand sides of this balance sheet must be the same. Thermal Efficiency. The ratio of the heat converted into work to the heat supplied to the engine is called the thermal efficiency. The first of these quantities is calculated from the indicated horse power and the latter from the weight and total heat per pound of the steam supplied. The only uncertainty arises as to the proper base from which to calculate. The total heat of steam given in steam tables is calculated from 32 degrees Fahrenheit, but it is obvious if this base is adopted the engine may be charged with more than its share of heat. If, for example, the exhaust steam from the engine passes through a feed-water heater and that the engine returns the condensed steam to the boiler as feed-water at say 1 50 degrees Fahrenheit, 260 POWER PLANT TESTING then there will be 1 50 — 3 2 , or 1 1 8 B . T. U . in every pound of steam passing continually from the engine to the boiler and from the boiler back to the engine without doing any work. More accurately, then, the thermal efficiency of an engine (E,) should be stated as E <=— n — j (09) V« where Q u and Qf as before are the heat equivalents respectively of the useful work and of the engine friction, while Q n should be defined as the net heat supplied to the engine. From this discussion it follows that the more efficient the feed-water heater is the higher the efficiency of the engine will be, and if an efficient heater is not used there is no reason for charging a loss to the engine. Obviously the limit to the amount of heat returnable to the boiler by means of a heater is the amount of heat in the exhaust steam, and this limit can be nearly approached under actual practical conditions. It is, therefore, very reasonable that the "'datum " for the calculation of the heat supplied to the engine should be the heat of the liquid corresponding to the temperature of the steam in the exhaust pipe from the engine. All this applies, of course, equally well to engines whether operat- ing condensing or non-condensing. In other words, the net heat supplied to the engine is the total heat of the steam enter- ing the engine, less the heat of the liquid at the temperature of the engine exhaust. The temperature of the exhaust must be taken by a ther- mometer in a suitable thermometer cup placed in the exhaust pipe close to the engine; and, similarly, the temperature and quality of the steam supplied to the engine must be determined by thermometers and a steam calorimeter placed close to but on the boiler side of the engine throttle valve. Engine Performance Expressed in British Thermal Units. A method which is rapidly gaining in favor among practical engineers is to express the performance of a steam engine or turbine by the number B.T.U., supplied, per indicated horse power per minute, the heat units supplied per minute being determined, as explained in the preceding paragraphs ; that is, the total heat in the steam entering at the throttle less the heat of the liquid at the temperature of the exhaust. (See page 287.) STEAM ENGINE TESTING 261 Engine Performance Compared with the Rankine Cycle. In order to know how much the efficiency of an engine can be improved it is most desirable to compare the actual thermal efficiency as defined on page 259, with the highest possible efficiency. For steam engines the standard cycle for comparison is now generally taken as the Rankine or Clausius cycle, 1 in which the operation of the engine is assumed to be perfect ; that is, without clearance in the cylinder, initial condensation, leakage or radiation. The indicator diagram for the Rankine cycle is represented by Fig. 190. The steam is assumed to be supplied to the engine cylinder at constant pressure until the point of cut-off, after which it is expanded adiabatically down to the back pressure at which the engine is operated on the return stroke Fig. >■ Volume .90. — Indicator Diagram for the Ideal Rankine or Clausius Cycle. when the exhaust steam is swept out of the cylinder and returned as feed-water to the boiler at the temperature of the exhaust. The same Rankine cycle represented in Fig. 190 when shown by a so-called entropy-temperature diagram can be made simpler both for analysis and calculations. This other kind of diagram, the details of which are somewhat more difficult to understand, is universally used by steam turbine engineers and has for the problem in hand . particular advantages. In this diagram, which will be now described, any surface represents accurately 1 Years ago it was not unusual to make this comparison with the efficiency of the Carnot cycle as a basis. This efficiency of a heat engine, it will be remembered, is expressed by the ratio of T l — T 2 to 7\ where 7\ is the absolute initial temperature and T 2 is the absolute final tem- perature. 262 POWER PLANT TESTING V ,-.800 4 jj| 3 600 - lll« 2 11111 a> a S ^§§■§1^ £400 ^llllll 2 HHH^ 3 J 100 < - III I) , , c 1.0 Entropy ( ji ) Fig. 191. — A Simple Entropy-tem- perature Diagram. to given scales, a quantity of heat. Absolute temperatures (T) are the ordinates, and entropies 1 ((f)) are the abscissas. Fig. 191 shows a simple heat diagram laid out with absolute temperature and entropy for the co-ordinates. Steam at a certain condition of tempera- ture and entropy is repre- sented here by the point A. Then if some heat is added, increasing both temperature and entropy, the final condi- tion is represented by the point B, and the area ABCD represents the heat added in passing from the condition at A to the condition at B. Such a diagram is usually called an entropy-temperature diagram, although the name heat diagram would probably be more appropriate, since every area represents a definite amount of heat. Another entropy-temperature diagram is shown in Fig. 192, representing by the various shaded areas the heat added to water at 32 degrees Fahrenheit to completely vaporize it at the pressure P^ The unshaded area under the irregular curve AB represents the heat in a pound of water at the freezing point (32 degrees Fahrenheit or 492 degrees in absolute tem- perature). The area OBCD is the heat added to the water to bring it to the temperature of vaporization, or in other words, this last area represents the heat of the liquid (q) given in the steam tables for the pressure P^ Further heating after Entropy, which Perry calls the "ghostly quantity," has no real physical significance, so that complete definition is not possible. If dQ is a small amount of heat added to a body and T is the absolute temperature at which the heat is added; then the change in entropy of that body is dQ/T, or d — dQ/T. Entropy of saturated steam above the entropy of water at the freezing point is easily calculated. For saturated steam at any pressure, then <£ =xr/T + 6, where % is the quality of the steam, r is the heat of vaporization, T is the absolute temperature, and d is the entropy of the liquid (water). The symbols used here are those given in Peabody's Steam and Entropy Tables (1909). STEAM ENGINE TESTING 263 vaporization begins is at the constant temperature Ti corre- sponding to the pressure Pi, and is represented by an increas- ing area tinder line CE. When ' ' steaming ' ' is complete, the latent heat, or the heat of vaporization (r), is the area DCEF. If after all the water is vaporized more heat is added, the steam becomes superheated, and the additional heat required would be represented by an area to the right of EF. The use of the entropy-temperature diagram in exhibiting the behavior of steam during expansion will now be discussed and illustrated with a practical example. 1.0 1.5 2.0 Entropy ($) Fig. 192. — Entropy-temperature Diagram Showing Total Heat in a ; Pound of Dry Saturated Steam. Fig. 193 illustrates the heat process going on when feed-water is received in the boilers of a power plant at 100 degrees Fahren- heit, is heated and converted into steam at a temperature of 400 degrees Fahrenheit, and then loses heat in doing work. When the feed-water first enters the boiler its temperature must be raised from ico to 400 degrees Fahrenheit, before any " steaming " begins. The heat added to the liquid is the area MNCD. This area represents the difference between the heat of the liquid of steam at 400 degrees Fahrenheit (q c ) and at 100 degrees Fahrenheit (q n ) and is about 306 B.T.U. The hor- izontal or entropy scale shows that the difference in entropy between water at 100 and 400 degrees Fahrenheit, is about .436. 1 1 As actually determined from Peabody's Steam Tables, pp. 2 and 264 POWER PLANT TESTING Every reader should understand how such a diagram is con- structed and especially how the curves are obtained. In this case the curve NC is constructed by plotting from the steam tables the values of the entropy of the liquid (usually marked with the symbol d) for a number of different temperatures between ioo and 400 degrees Fahrenheit. If now water at 4C0 degrees Fahrenheit is converted into steam at that temperature, the curve representing the change is necessarily a constant temperature line and therefore a Entropy {) Fig. 194. — Entropy-temperature Diagram Illustrating the " Available Energy '•' in Steam. diagrams like those in the preceding figures, it will now be shown how the available energy of dry saturated steam for any given 1 That the steam might be dry and saturated, the expansion would have had to follow the curve ET and G would have appeared at G'. The heat of the liquid, q, of a pound of steam at 100 degrees Fahrenheit is represented by OBNM, and the heat of vaporization (r) is MNG'F' , so that the total heat (q + r or H) is OBNG'F'. The total heat of wet steam is expressed by q+xr, where x is the quality or relative dryness. In the case of this adiabatic expansion, then, q is as before OBNM and xr is MNGF. It is obvious also that the lines NG and NG' have the same relation to each other as the areas under them, so that line NG area MNGF xr line NG' area MNG'F' NG NG' (70) showing that the quality of the steam at any point, G, on a constant temperature line (which for saturated steam is also a constant pressure line) is determined as in this case by the ratio of NG to NG', STEAM ENGINE TESTING 267 £ c/ T, and P 1 \ lE 1 a g B nA T.,and P., G v P ]p' conditions can be readily calculated from the data given in steam tables. Fig. 195 is an entropy-temperature diagram representing dry saturated steam which is expanded adiabatically from an initial temperature T x corresponding to a pressure Pi to a lower final tem- perature T 2 corre- sponding to a pres- sure P 2 . The other initial and final conditions of total heat (H) and en- tropy (0) are rep- resented by the same subscripts 1 and 2 . The avail- able energy or the work that can be done by a perfect engine under these conditions is the area NCEG. It is now desired to obtain a simple equation expressing this available energy E rt in terms of total heat, absolute temperature and entropy. Explana- tions of the preceding figures should make it clear that H x =area OBNCEGF, H 2 =areaOBNG'F', E a =areas (OBNCEGF -fFGG'F') - OBNG'F', E =H 1 -H 2 +FGG'F / , th ere fore E a =R 1 -R 2 +(^ 2 -^ 1 )T 2 1 . ........ (71) An application of this equation will be made at once to determine the heat energy available from the adiabatic expan- sion of a pound of dry saturated steam at an initial pressure 1 It should be observed that this form is for the case where the steam is initially dry and saturated. For the case of superheated steam a slightly different form is required which is given on page 272. Fig. Entropy -Practical Example of Adiabatic Expansion. 268 POWER PLANT TESTING of 165 pounds per square inch absolute to a final pressure of 15 pounds per square inch absolute. Example. Pi =165 T 1 =. . . P2 = 15 T 2 =673.0 from steam tables. 1 Hj =1193.6 from steam tables. H 2 =1146.9 from steam tables. i =1.5605 from steam tables. <|>2 =i-7499 from steam tables. Substituting these values in equation (71), we have E a =ii 9 3.6-ii46.9 + (i. 7499-1. 5605)673.0 = 174.2 B.T.U. per pound of steam. The important condition assumed as the basis for the deter- P=° 'P X

2 -c|, 1 )T 2 -^(i-x 1 )(T 1 -T 2 ).. . . (72) Example. Calculations of the heat energy from adia- batic expansion for the same conditions given in the preceding example on page 268, except that the steam is initially 5 per cent wet, are given below. Pi =165 lbs. abs. T L =825.9° F. P 2 = 15 lbs. abs. T 2 =673.o°F. Hi =1193.6 B.T.U. H 2 =1146.9 B.T.U. 4>i=i-5 6o 5- 4>2 =i-7499- r L =855.9 B.T.U. Xi=i.co-.c5 =.95. E a u, =II 93-6-ii46.9 + (i-7499-i-5 6o 5)673-°-^^- 825.9 X.o5(825. 9-673.0), Here because E BW = 166.3 B.T.U. 'In general terms, = xr ■hO. &■ + ?i 4*- _x t r r, l + e v % M g a 53 * .50 •4S .46 .44 .42 200 250 300 350 400 450 500 550 600 650 700 750 Temperature" I' Fig. 197. — Mean Values of Specific Heat (C p ) of Superheated Steam Integrated from Knoblauch and Jacob's Data. values of the mean and the true specific heat at constant pres- sure (C p ). True specific heat represents the ratio of the amount of heat to be added to a given weight of steam at some particular condition of temperature and pressure to raise the temperature one degree to that required to raise the temperature of water at maximum density one degree. The mean specific heat is almost invariably used in steam engine and turbine calculations. 1 Zeit. Verein deutscher Ingenieure, Jan. 5, 1907. Values of mean specific heat are taken from Mechanical Engineer, July, 1907, and Professor A. M. Greene's paper in Proc. American Society of Mechanical Engineers, May, 1907. STEAM ENGINE TESTING 271 Entropy Diagram of Superheated Steam. The graphic representation of the heat added during the superheating of steam is easily shown with entropy-temperature diagrams. Fig. 199 shows a diagram similar to the one representing dry saturated steam with the added area EHJF to show the super- q0.70 2 ^ 0.G0 20 1 i\ \ 16 u/ 12/ ' 10/ \ s/ \ Pressu sq. c re in Kgr m. absoli i.per te. u ^^ ^^ 100° 200 250 300 Temperature = 0. Fig. -Values of the " True " Specific Heat of Superheated Steam. heating from the temperature, Ti corresponding to the pressure Pi to the temperature of the superheated steam, T s . The total heat in a pound of steam above the freezing point is now repre- sented by the area OBCEHJFO. For adiabatic expansion of superheated steam at the temperature T s and pressure Pi to a pressure P 2 the available energy is the area CEHKL. 272 POWER PLANT TESTING Too much calculation is involved in the construction of entropy diagrams to make a new diagram for every particular case from the proper- ties usually found in steam tables; but the construction of such diagrams should be understood. From the explanations that have preceded, the construction of all the lines except EH should be obvious. This line is obtained by calculating the en- tropy of superheated steam for various val- ues of temperature from the following well-known relations: , , fdQ f T °C p dT a -l = Cpm \log e ~ =2.3028 Cpm(log 10 (T s -log 10 T } ) V (73) Entropy — Entropy-temperature Diagram for Superheated Steam. and E to ,=H 1 -H 3 +C pm (T.-T 1 )-C rm .(T,'-T 2 ) (74) where E os is the available energy of the Rankine cycle for superheated steam and T g ' is calculated from the equation above where cj> s and (f> 2 ( = i) are now both known quantities. £pm ^ s ^ e mean value taken from the curves in Fig. 197 for the temperature T s , and C' pm for T/. Approximate Steam Consumption Calculated from an Indi- cator Diagram. It is often very convenient to be able to calculate the approximate steam consumption of a steam engine from the data obtainable from an indicator card, the size of the piston, STEAM ENGINE TESTING 273 the stroke, and the speed. Using a double-acting engine, the following symbols * may be used: p =mean effective pressure, pounds per square inch from indicator diagram. 1 = length of the stroke of the engine in feet. a =area of the piston in square inches. b =percentage of clearance to the length of the stroke. c ^percentage of stroke at any point in the expansion line. 2 n = number of revolutions per minute; and 120 n = number of strokes per hour. w =weight of a cubic foot of steam having a pressure as shown by the indicator diagram corresponding to that at the point in the expansion line selected for c, pounds, w' =weight of a cubic foot of steam corresponding to the pres- sure at the end of compression, pounds. la(b +c) Then the number of cubic feet per stroke = — in 144(100) the clearance and piston displacement volumes (at c). Weight of steam per stroke, pounds = -. ^. (75) fa r ' r 144(100) Wvjy Volume of the clearance, cubic feet — [44(100) Weight of steam in clearance, pounds remaining in the r a law'(b ) cylinder = -. r . 144(100) Approximate net weight of steam used per stroke law(b+c) law'(c) la [0. +c)w— cw' . (76) 144(100) 144(100) i4,4oo| Approximate weight of steam from diagram per hour 120 nlal 14,400 (b+c)w-cw' . . . . (77) 1 Compare with Power, September, 1893. 2 In other words this is the percentage of the entire stroke which has been swept through by the piston corresponding to the point in the expansion curve selected for measurements. It is preferable, how- ever, to take this point not very far from the point of cut-off, since the assumption must be made that the product of pressure times volume in the expansion curve is constant, which is, of course, not accurate, and the error becomes greater as the expansion increases. 274 POWER PLANT TESTING i Indicated horse power for a double acting engine _ 2plan ~33>ooo' (78) Steam consumption per indicated horse power is (77) divided by (78) or *=^^f(b+c)w-cw , l (79) The difference between the theoretical steam consumption calculated by the formula and the actual consumption as determined by tests represents steam " not accounted for by the indicator," due to cylinder condensation, leakage through ports, radiation, 700 1 I I I I . I I I I I I — p-r-n — r-|-| — 1 ' i /f I 1 1 1 I e t c - If the steam supplied to the engine is very wet, corrections for this moisture should be made in the value of w. Willans Law. One of the most serviceable checks that can be applied to engine tests is plotting the Willans line of total steam consumption per hour. Curve sheets illustrating this as plotted from ac- tual tests by Bar- raclough & Marks 1 are shown in Fig. 200. It will be observed that the points representing the weight of steam used per hour when plotted for the horse power corresponding are on a straight line. In other words Willans law is usually stated thus: " With a fixed cut-off \F? j£ /^ 7 7 7 _y _j S 7* * y S ~K 7 B 7 Z 7 -*£ S ^ S 7- 7 f + ' /r s y 7 S y ^r ^ y 7 /* J* ^ ^ >c J. A 1 y. , ' ^ ^ * s ct 7 10 20 30 Indicated Horse Power Fig. 200. — " Willans " Lines for an Engine with a Throttling Governor. Proceedings Institution of Civil Engineers, vol. 20, page 323. STEAM ENGINE TESTING 275 and a throttling governor the total steam used by the engine per hour at different loads can be represented by a straight line upon a mean effective pressure base or upon a horse-power base." It will be shown also in the following paragraphs how, theoretically, this relation holds for an engine operating at a fixed cut-off and with a throttling governor and that the steam consumption per hour is proportional to the mean effective pressure and also to the horse power developed. If we assume that the expansion curve is hyperbolic, which is usually near the truth, then the mean " forward " pressure given by an indicator diagram is x I ^), (80 where pi is the initial pressure of the steam and r is the ratio of expansion. With a throttling governor r is of course constant. The terms in the parentheses can then be represented by a constant c and the mean forward pressure p/ then can be written as pic. Volume of steam used per hour for a double-acting engine can be expressed in cubic feet as 120 (nV), (8i) where n is the number of revolutions per minute and V is the volume of steam admitted to the cylinder per stroke. Now if we use the symbol Vi for the specific volume, that is, the volume in cubic feet of a pound of steam at the pressure pi, and assum- ing for a small range of pressures that piVi =a constant k, then we can write, if W is the weight of steam used per hour in pounds, w=i 2o i nV) = i2o(nVp 1 ) Vi k v Now with a throttling governor and constant cut-off all these quantities are constant except pi, and writing a constant z for the term — ; we have W=zpi, but it was shown above k that the mean forward pressure p/=cpi, so that W=z^- f (83) 1 Compare with Perry's " Steam Engine," page 2S6. 276 POWER PLANT TESTING So that the curve -representing this equation is a straight line and passes through the origin of co-ordinates. If, however, we use the mean effective pressure instead of the mean forward pressure, then M.E.P.= P/ -p 6 , (84) where p& is the mean " back " or exhaust pressure. In these last terms then W=|(M.E.P.+p fe ) (85) This last equation may be stated as W = a constant XM.E.P. + another constant which, when plotted to a scale of mean effective pressure for abscissa and weight of steam used per hour for ordinates, is also a straight line, intersecting the axis of ordinates above the origin at a distance corresponding to the steam con- sumption per hour at no lead. Since the indicated horse power (I.H.P.) is proportional to mean effective pressure, a straight line will result when the steam consumption and indicated horse power are plotted; and the same holds true also when steam consumption is plotted with brake horse power (B.H.P.) instead of indicated horse power. Curves Showing Results of Tests Graphically. One of the best checks of an engine test is to plot the principal observations graphically as the test proceeds. This is particularly important as regards the total weight of steam used per hour. For a series of tests each made with a different load the points plotted with horse power (either indicated or brake) as abscissas and total steam per hour as ordinates should lie along a straight line, known as the Willans line (see page 274). This statement applies accurately only for engines operating with a throttling governor at, of course, constant speed; but is generally applicable to steam turbine tests, irrespective of the type of governor. Steam Engine Lubricators. The proper oiling of an engine is most important. The operation of nearly all types of lubri- cating devices is easily understood, particularly when operated by the gravity of the oil or by a pump. Another type of lubri- cator for cylinder lubrication which is operated by the weight STEAM ENGINE TESTING 277 of a column of condensed steam is shown in Fig. 201. The pipe C, in which the condensed steam accumulates, must be made at least 2 feet long to give a sufficient head or pressure to feed the oil. The oil reservoir is in the cylindrical vessel below the condenser pipe. This is filled with oil through an opening in the top as the water which has accumulated in the apparatus is drained through the cock D. Water from the condenser Fig. 201. — Engine Cylinder Lubri- cator Operated by Pressure of Column of Condensed Steam. Fig. 202. — " Detroit " Cylinder Lubricator. pipe C is carried down to the bottom of the oil reservoir by the pipe shown in dotted lines on the left-hand side. Oil, being lighter than the water, remains in the upper part of the reservoir and is forced down through the pipe on the right- hand side and through the needle-valve I by which the flow of oil can be regulated so that as small an amount as one drop in two or three minutes passes into the steam pipe at H to mix with the steam going to the engine cylinder. Through the 278 POWER PLANT TESTING gage glass S the number of drops passing through can be observed. Another gage glass L on the side of the reservoir shows the relative amounts of oil and water. When an engine is not operating all oil cups and lubricators should be carefully closed. The feed of oil from this lubricator is stopped by closing the valves V and F. A valve is usually provided on the bulb B, which should be closed when draining from D, so that the water in the condenser pipe C will not be lost, and thus prevent the operation of the apparatus till a sufficient amount of condensed steam has accumulated to produce the pressure necessary to force the oil into the steam pipe. A slight modification of the lubricator described is illustrated in Fig. 202. The condensed steam is brought to the bottom of the reservoir through the pipe P, which is open at its lower end. On the other hand, the pipe S is open at the top and the oil is forced by the pressure due to the head of water in the pipe above (not shown) through the gage glass on the left- hand side, then through the horizontal pipe T to which is attached the valve and nipple connected to the steam pipe supplying the engine. In this figure at the top of the reservoir a plug with a nicely finished handle is shown, which is to be removed for filling with oil. CHAPTER XII TESTING STEAM TURBINES AND TURBINE GENERATORS Testing Steam Turbines. 1 In every power plant the means should be available for making tests of the steam equipment to determine the steam consumption. Usually tests are made to determine how nearly the performance of a turbine approaches the conditions for which it was designed. The results obtained from tests of a turbine are to show usually the steam consump- tion required to develop a unit of power in a unit of time, as, for example, a horse power hour or a kilowatt hour. In such tests a number of observations must be made regard- ing the condition of the steam in the passage through the tur- bine and of the performance of the turbine as a machine. To get a good idea of what these observations mean, it may be profitable to follow the steam as it passes through the turbine. The steam comes from the boilers through the main steam pipe and the valves of the turbine to the nozzles or stationary blades as the case may be. It then passes through the blades and finally escapes through the exhaust pipe to the condenser. It is preferable to have a surface condenser for tests, so that the exhaust steam can be weighed. The weighing is preferably done in large tanks mounted on platform scales. Methods for Testing. The important observations to be made in steam turbine tests are : i . Pressure of the steam supplied to the turbine. 1 Tests of the turbines alone in a modern station may be only a rough indication of the over-all economy of the plant. Recently steam turbines were installed in a large power plant where they replaced steam engines of an excellent make. Tests of the turbines and of the engines made without considering the losses in the rest of the plant showed very little gain in efficiency by this change, although it was found that the fuel consumption was reduced twenty per cent. Parts of this chapter and the chapter following are taken from the author's work on " The Steam Turbine," 279 280 POWER PLANT TESTING, 2. Speed of rotation of the turbine shaft, usually taken in revolutions per minute. 3. Measurement of power with a Prony or a water brake, if the power at the turbine shaft is desired; or with electrical instruments (ammeters, voltmeters, and wattmeters), if the power is measured by the output of an electric generator. 4. Weight, or measurement by volume, of the condensed steam discharged from the condenser. Unless a surface con- denser is used it is very difficult to obtain the amount of steam used by the turbine. All leakages from pipes, pumps, and valves, which are a part of the steam which has gone through the turbine, must be added to the weight of the condensed steam. The accuracy of a test often depends a great deal on how accurately leaks have been provided against, or measured when they occur. 5. Temperature of the steam as it enters the turbine. If the temperature is higher than that due to the pressure of the saturated steam given in steam tables, the steam is superheated : if, however, the temperature is not higher the steam may be wet, and a calorimeter must be attached as near the turbine steam chest as possible. 1 All gages, electrical instruments, and thermometers should be carefully calibrated before and after each test, so that observa- tions can be corrected for any errors. The zero readings of Prony and water brakes for measuring power should be carefully observed and corrected to eliminate the friction of the apparatus with no load. Unless all these precautions are taken the dif- ficulties in getting reliable tests of turbines are greatly increased. In all cases tests should be continued for several hours with absolutely constant conditions if the tests are to be of value. The most valuable test of a steam turbine or of a reciprocating steam engine is made when varying only the load ; that is, with pressures, superheat, and speed constant. When the steam consumption is then plotted against fractions of full load, a 1 The most satisfactory tests of turbines are made with steam slightly superheated rather than wet. When steam is very wet (more than about 4 per cent moisture for ordinary pressures) the determination of the quality is difficult. There is also a danger that steam showing only a few degrees of superheat by the reading of the thermometer is actually wet. The high temperature is due in such cases to heating from eddies around the thermometer case or in steam pockets near it. TESTING STEAM TURBINES AND TURBINE GENERATORS 281 water -rate curve is obtained. For such a curve a series of tests are needed, each for some fraction of full load; and in each separate test the power as well as all the other conditions must be held constant. Another important test of the performance of a steam turbine is made by varying both the speed and the power and keeping the other conditions constant. The observations of speed and power from such a test give a power parabola as illustrated in Fig. 203. This curve shows at what speed the tur- bine gives the greatest output. Tests may also be made with varying initial steam pressure, but keeping other con- ditions including ex- haust pressure and load constant. Calculations of the steam consumption and efficiency of turbines made by allowing for the different losses as calculated separately and then added to- gether as is often done to determine the losses in electrical apparatus are of very little value except when made by experienced designers. 1 Commercial Testing. The methods used by the New York Edison Company in commercial tests of steam turbine-generator units may well be explained briefly. During a test the load on the turbine unit is maintained as constant as possible by " remote control " of the turbine gov- .50 f.45 l j ".DO k^ l^w ^ / V a! g, 400 &35 M.30 vk; 2\ || M 300 *i ^ !! \ 1! $ 26 \ «j V ^25 1 24 e <*^ ¥ .. > a Yc x & % § |21 J 20 ■f M -30 £18 Fig. 203.- )0 800 1200 1600 2000 2100 Speed E.P.M. -Results of Tests of a Turbine at Various Speeds. 1 This method of calculation of steam consumption is explained in detail in "The Steam Turbine," by the author, pages 86-93. Steam con- sumption of a turbine can be predicted by calculations much more accurately than for a steam engine, 282 POWER PLANT TESTING ernor by the switchboard operator. The maximum variation in load is to be held within 4 per cent above and below the mean. For some time previous to the test the turbine is run a little below the load required for the test, but at least ten minutes before the starting signal is given the test load must be on the machine. Three-phase electrical load is measured by the two-wattmeter method, 1 using Weston indicating wattmeters of the standard laboratory type. These instruments are calibrated at a well- known testing laboratory immediately before and after the test. Power factor is maintained substantially at unity and all electrical readings are taken at one-minute intervals. When the turbine is supplied with a surface condenser, the steam consumption, or water rate, is determined by weighing in a large tank supported on platform scales the condensed steam delivered from the condenser hot well. Above the tank on the scales a reservoir is provided which is large enough to hold the condensation accumulating between the weighings, which were made at intervals of five minutes. By using a loop connection for the gland water supply (of Westinghouse turbines) or the water from the step bearing (of Curtis turbines, using water for this bearing) the necessity for connecting the weighings for these amounts is avoided. Because the circulating water at the stations of this company is usually quite salty, any condenser leakage is detected by test- ing the condensed steam by the silver-nitrate test with a suitable color indicator. This color method is said to be a decided advantage over the usual method of weighing the leakage accumulating during a definite period when the condenser is idle and tested only with full vacuum. By taking samples of circulating water and condensed steam at the same time, it is possible to detect any change in the rate of condenser leakage. The water level in the hot well is maintained at practically a constant point by means of a float valve in the well, auto- matically controlling the speed and, therefore, the amount of the delivery of the hot-well pump. This device avoids the 1 Cf . Kent's "Mechanical Engineer's Pocket-Book," 8th ed., page 1396, or Foster's " Electrical Engineer's Pocket-Book," 4th ed., pages 51 and 325. TESTING STEAM TURBINES AND TURBINE GENERATORS 283 necessity for the difficult correction to be made in a test when the levels in the hot well are not the same at the beginning and end of a test. Temperatures and pressures of the admission steam are determined by mercury thermometers and pressure gages located near the main throttle valve of the turbine; the amount of superheat is determined by subtracting from the actual steam temperature after making thermometer connec- tions the temperature of saturated steam corresponding to the pressure at the point where the temperature is measured. All gages and thermometers are calibrated before and after the test. Vacuum is measured directly at the turbine exhaust by means of a mercury column with a barometer alongside for reducing the vacuum to standard barometer conditions (30 inches). By this latter arrangement the necessity for temperature connec- tions between the two mercury columns not at the same place is avoided. It has sometimes happened that split condenser tubes have caused a leakage of steam which was extremely difficult to measure. Cases are reported where the split opened up only when the condenser was heated with a large volume of steam. On this account it is preferable not to use a leaky condenser for accurate tests; in other words, the condenser should be thor- oughly repaired before tests are made. The effect of split tubes causing an irregular amount of leakage is usually shown in tests by inconsistent results in the weight of condensed steam. In that case the leakage will be greatest with largest flow of steam through the condenser and it will be observed, for an engine operating with a throttling governor or with a steam turbine, that when the " Willans line " page 274, is drawn to check the tests that it will be curved instead of straight. It should be noted, however, that a curved " Willans line " does not necessarily indicate this phenomenon in condenser leakage, as the irregularity may be due to faulty design of the engine 1 or turbine. Tests of condensers for leakage should be run long enough so that the quantity of water coming through can be determined with accuracy. 2 Usually a leakage test run for less than an hour or two is of no use at all. 1 The "Willans" line for a reciprocating engine operating with an automatic cut-off governor is usually a curve slightly concave upward. 2 To determine accurately the weight of condensed steam the air- 284 POWER PLANT TESTING, When the method of determining the weight, of condensed steam by weighing the boiler feed-water is used the chances of error are very great and every possible precaution to insure accuracy must be observed. In the first place valves no matter how good should not be relied on to prevent the passage of steam through them. For this reason careful engineers insist on disconnecting from the line of steam piping between the boilers and the engine or turbine tested all other piping con- nected to it, and then blanking off with flanges all the sections disconnected. If flanged pipe fittings have been used in the pipe lines, blanking off sections in the various pipes is very easily accomplished by disconnecting the flanges and inserting a thin iron or copper plate with holes around the edge to fit the bolt-holes in the flanges. The plate is then easily bolted in place. Another important precaution to observe is that the outlets of all drain or drip pipes and of all blow-off valves must be visible. It is equally important that all the piping between the boiler feed-pumps and the boilers is exposed with all branches blanked off or plugged. Boiler leakage should be determined before and after each test with preferably the pipe supplying the turbine blanked off at the throttle valve, although if the throttle valve is reasonably tight the precaution of blanking off this valve is not considered so important as the others men- tioned. In tests for boiler leakage the required steam pressure must be maintained on both the piping and the boilers. Measur- ing feed-water with water meters should not be thought of unless only approximate results are expected, and in that case such an arrangement is only allowable if the temperature of the boiler feed is not over 80 to 90 degrees Fahrenheit for most meters, and a by-pass at the meter is provided, so that at fre- quent intervals the meter can be calibrated by actual weighing of the flow through it, with the rate of flow and temperature of the water the same as in the tests. Boiler leakage is often as much as from ten to fifteen per cent of the. weight of feed- water, and in some reliable tests a still greater leakage has been observed. pump, piping, and tanks must be free from leaks, and. the condenser and pump should be so arranged with respect to each other that the condensed steam will flow in a continuous stream to the pump and into the tanks. TESTING STEAM TURBINES AND TURBINE GENERATORS 285 Steam Consumption Determined by a " Heat Balance " Method. There is still another method sometimes used for determining the steam consumption of engines and turbines operated with jet condensers or condensers of a similar type where the cooling water and condensed steam are mixed and dis- charged together from the condenser. This method is based on the measurement of the amount of heat absorbed by the cooling water from the condensed steam. The weight w c and temper- ature of the cooling water leaving the condenser t 2 , the quality of the exhaust steam x and the temperature of the mixture of condensed steam and cooling water t" are determined as accu- rately as possible and from these data the weight of condensed steam w s is of course readily calculated by a simple algebraic equation as follows: w c (t 2 -32)+w s (q+xr) = (w c +w s )(t"- 3 2) . . (86) where q and r are respectively the heat of the liquid and the heat of vaporization corresponding to the temperature of the exhaust steam. In this equation " heat contents " are measured for each term from 32 degrees F. The method is, however, unreliable, and at best can be depended on for only very approx- imate results. The reason for this inaccuracy is the difficulty of measuring, especially in large plants, the quantity of cooling water and the true average temperature of a large volume of water flowing in a pipe or channel. It is found usually that the temperature of the water discharged from the condenser will vary from one side of the pipe to the other, and small errors in the determination of this temperature, because the rise in tem- perature is small, will make large discrepancies in the calculated weight of condensed steam. Stage pressures should be observed and recorded when tests are made of steam turbines having a series of pressure stages. These data are often extremely useful, both for checking the weight of condensed steam if the turbine is of the nozzle type * 1 Usually the nozzles discharging the steam into the second stage of the turbine are always open, so that the total area is always constant. If, therefore, the areas of the smallest sections of these nozzles are measured and the pressure is observed in the first stage, the weight can be calculated with a considerable degree of accuracy by using the formula for the flow of steam on page 148. 286 POWER PLANT TESTING ^ and also for showing any abnormal conditions in the several stages. Results calculated on a basis of kilowatts output should be net; that is, the power required for excitation should be subtracted from the generator output. If, however, the gen- erator is self-exciting the net output can be measured directly at the terminals of the machine. "Guarantee" tests of steam engines and turbines should be made under conditions as nearly as possible the same as that for which the turbine was designed. Different machines will have different correction factors for varying conditions of pressure, superheat, vacuum, etc., so that water rates cor- rected for large variations are always likely to be more or less inaccurate. This is particularly true in respect to vacuum corrections. Some turbines will give a very good efficiency with a low vacuum, but at a high vacuum because of an insuf- ficiently large steam space the efficiency will be low. HEAT UNIT BASIS OF EFFICIENCY A thermal efficiency can be calculated readily by deter- mining what percentage the heat equivalent of the work is of the heat " used by the turbine," assumed to be the difference between the total heat in the steam at the initial conditions and the heat of the liquid in the condensed steam at the temperature of the exhaust. By this method the full load test of a Westinghouse-Parsons turbine reported by F. P. Sheldon & Co., will be calculated from the data given in an official report. In order to make the results of such calculations of steam turbine tests comparable with the usual heat unit computa- tions of reciprocating steam engine tests the results are often expressed in terms of indicated or " internal " horse power. It was assumed the mechanical efficiency of a reciprocating engine of about the same capacity at this load was about 93-3 per cent. TESTING STEAM TURBINES AND TURBINE GENERATORS 287 THERMAL EFFICIENCY OF A 400-KILOWATT STEAM TURBINE Brake horse power 660 Corresponding indicated or "internal" horse power of a recip- 660 rocatmg engine = 708 •933 Total steam used per hour, pounds 9169 Steam used per "internal" horse power per hour, pounds 12 .96 Steam used per "internal" horse power per minute, pounds. ... 0.216 Steam pressure, pounds per square inch, absolute 166 . 9 Superheat, degrees Fahrenheit 2.9 Vacuum, referred to 30 inches barometer, inches 28 .04 -Temperature of condensed steam, degrees Fahrenheit (at .96 pound per square inch absolute pressure) 100 Total heat contents of one pound of dry saturated steam at the initial pressure, B.T.U 1 193 Heat equivalent of superheat in one pound of steam, B.T.U. (C p from Fig. 197, page 270) 1 Total heat contents of one pound of superheated steam, B.T.U. 1195 Heat of liquid in condensed steam, B.T.U. 68 Heat used in turbine per pound steam, B.T.U 1127 Heat used in turbine per "internal" horse power per minute, B.T.U. (n27.2X0.216) 243 Heat equivalent of one horse power per minute, B.T.U. =— : . 2 778 Thermal efficiency, per cent (42.42 h- 243.5) J 7 CALCULATION OF EFFICIENCY (SHAFT AND BUCKET) OF A STEAM TURBINE GENERATOR COMPARED WITH THE RANKINE OR CLAUSIUS CYCLE. 1. " Electrical" kilowatts. 2. R.P.M. 3. Steam per hour (corrected for moisture). 4. Water rate per " electrical " kilowatt, pounds per hour. (3) -f- (1) 5. PR, loss in generator, kilowatts. 6. Rotation loss of generator alone, kilowatts. 7. Rotation loss of wheel and generator, kilowatts. 8. "Shaft" kilowatts. (i) + ( 5 ) + (6) 9. "Bucket" kilowatts. (i) + (s) + (7) 10. Water rate per "shaft" kilowatt, pounds per hour. 11. Water rate per "bucket" kilowatt, pounds per hour. 12. Steam-chest pressure, pounds per square inch, absolute. 13. Exhaust pressure, pounds per square inch absolute. 14. Available energy, B.T.U. 15. Theoretical water rate, pounds per kilowatt hour, B.T.U. = 44200 Avail. En. (14) 16. " Shaft ' ' efficiency = (1 5) -=- (10) . 17. " Bucket" efficiency = (15) -r(n). Notes. — For calculating rotation loss of a new design, stage pressures are of course used. Steam per hour is usually calculated from the area of the nozzles in the first stage if the governor is not operating. For a speed-torque test the flow of steam is constant and KW. for determining items (8) and (9) are read from this curve I KW. R.P.M. 288 POWER PLANT TESTING In the case of steam turbines the net over -all efficiency or the heat equivalent of the "shaft" kilowatts compared with the available energy in the Rankine or Clausius cycle is the only one of any practical value to operating engineers. It shows the engineer how his engine is working in comparison with an assumed perfect engine. A comparison of different turbines on the basis of this net over-all or " shaft " efficiency is the most satisfactory way of considering their relative merits. Curves in Fig. 204 are given to compare the steam consump- tion of a standard turbine generator and a 4-cylinder compound reciprocating steam engine of the type used by the Interurban and Metropolitan Companies of New York, assuming both I y> / !/ 21 •1 S20 A / "lin ae with 1 V Ni ri u6y 1 1 |% , s 5 5 1S 9 °: y (if IR n \ f" h ~5 " or F \ L Cyliude 16 v. 15 L'u rt e Fig. 204. — Comparative Water Rate Curves for Engines and Turbines. units operating under the same conditions. These curves illustrate the good overload economy or the turbine, showing that at 50 per cent overload the engine designed for equal work in the cylinders requires for the same output 43 per cent more steam than the turbine. These results are particularly interesting because the peak capacity of a station with a given equipment of boilers and auxiliaries is increased in proportion to the reduction of steam consumption at overload. For a given investment the turbine gives a much larger range of load and, moreover, affords the means by which the peak capacity of existing stations can be greatly increased. The speed output curve (Fig. 203) is very useful to engineers to determine if a turbine is running at its best speed. If the TESTING STEAM TURBINES AND TURBINE GENERATORS 289 correspond' ng curves of steam consumption per kilowatt output (usually called water rate per kilowatt) and efficiency curves are calculated according to the form on page 287, a great deal of information is obtained about the operation and economy of a turbine. The torque line in Fig. 203 is always drawn straight, just as a " Willans line." A curve of total steam consump- tion is usually a straight line for the normal operating limits of a turbine, but usually becomes curved when a by-pass valve opens on overload, or when the turbine is over its capacity so that the pressures are not normal in the stages. The torque line shows why a turbine engine is not adaptable to automobiles. The starting torque of a small commercial turbine is not large, so that starting would be difficult with a small wheel, and reversing and speed reduction would be as difficult as with a gasoline engine. The reciprocating steam engine as well as the gasoline engine has, therefore, advantages over the steam turbine for this service. Method of Making Tests to Determine Wheel and Blade Rotation Losses of a Steam Turbine. The simplest method for making such a test, and the one commonly employed, is to attach an electric motor to the turbine shaft (sometimes in a direct-connected set the generator is used as a motor) and run it at a number of different speeds. In taking a series of speeds, no observations are made until conditions have become " steady," and the speed must be held constant for several minutes so that a number of readings can be taken on the electrical instruments measuring the input of the motor. The results give the rota- tion loss of the wheel and blades in steam as well as bearing friction and the rotation or " windage " and electrical losses of the motor. Then the turbine wheel is removed, leaving the packing at the generator end of the turbine on the shaft, and the motor is run alone. The power now measured is that required to overcome the rotation and electrical losses of the generator and the bearing friction. Curves of power and speed as vari- ables are plotted for each set of observations, and the disk and blade loss is determined by subtracting the ordinates of one curve from those of the other. It may be assumed with sufficient certainty that the weight of the turbine wheel itself would not alter the bearing losses to any considerable extent. 1 1 It may be interesting to observe that since disk and blade friction 290 POWER PLANT TESTING The important fact that all results given here are for disks and blades revolving in a stagnant medium must not be over- looked, and it must be not assumed that the results will be the same under actual operating conditions. It may be a coincidence • that the losses are the same in both cases. Under operating conditions, the spaces between the wheel blades are filled with steam flowing from the nozzle over the blades and then to the condenser. Now it has been shown by a series of experiments by Lasche of the Allgemeine Electricitat Gesellschaft (Berlin) that increasing the number of nozzles around the turbine wheel reduces the disk and blade rotation losses. These losses in the blades are very largely due to the fan action of the blades which start currents of steam just as a centrifugal fan does. In other words, this is what Stodola calls " ventilation." With steam flowing through the blades, this fan action is largely prevented and the losses are consequently reduced. Another reason why the disk and blade rotation losses should be less when the tur- bine is operating than they are in stagnant steam is that they are really friction losses, or a conversion of kinetic energy into heat, with the effect of either superheating or drying the steam. In a turbine with more than one stage a part of the heat energy gained as the result of the friction is converted in the next expansion into kinetic energy or velocity. It is usually assumed that about 1 5 per cent of the disk and blade losses are regained by reheating, and that therefore the actual friction losses in an operating turbine are about this amount smaller than in stagnant steam. In cases of full admission true blade friction disappears; and a proportionate reduction will also take place, according to the degree of admission, when it is partial. Investigation of wheel and blade friction losses by the author, using a modification of the method first suggested by Lasche of Berlin, did not show the reduction in these losses to be expected when determined under operating conditions. These results, however, cannot be considered conclusive, as the type of machine used was not well suited for the purpose, and only 25 per cent of the blades were filled with steam. It has been stated that is proportional to the density of the medium, the friction is therefore greater in air than in dry saturated steam at atmospheric pressure. This is shown by experiments published by Lewicki in Zeit. Verein deutscher Ingenieure, March 28, 1903. TESTING STEAM TURBINES AND TURBINE GENERATORS 291 when a large quantity of steam passes into the casing through a suitable opening without passing through nozzles and escapes through the exhaust (without increasing the pressure), the disk and blade rotation losses are increased as much as 20 per cent. This apparently is an influence to counteract the effect of filling the blades. In all the analysis that has preceded there are so many uncertain variables entering that it is impossible to get agreement, although, apparently, we have a large amount of data from which to draw. It may be stated, however, that all in all, the best data on disk and blade friction seem to show that it is smaller and of less significance than the results of most investiga- tors would show. A little space should be given to Lasche's very interesting method. A turbine-generator set was used in which the num- ber of nozzles discharging into the turbine could be regulated and the output of the generator was observed for each setting of valves, and tests with varying loads were made at a number of different speeds. The turbine wheel was then removed from the shaft, and by running the generator as a motor the friction losses in the stuffing-box at the generator end of the turbine and in the bearings, as well as the windage loss of the generator, were determined. The resistance of the armature and brushes was also measured to calculate the heating (I 2 r) loss. The sum of these losses was calculated for a number of loads (kilowatts) , showing the electrical output at each speed. Another curve representing the power delivered to the shaft by the turbine was obtained by adding to the generator output for each set of nozzles open the corresponding generator losses (windage, heating, and bearing friction). The lower portions of both curves are practically straight lines, and by producing the latter curve to the horizontal axis, its intersection represents on the- scale of abscissas the disk and blade rotation losses of the turbine at the speed of the test and under actual operating conditions. By making a series of such tests at different speeds curves of rotation losses can be made. Although this method requires very careful experimenting, the same must be said of any other method of obtaining these losses. At least it must be admitted that by this method a number of uncertain factors to be con- sidered in the " stagnant steam " method are eliminated. 292 POWER PLANT TESTING The curves obtained by this method are really the same as " Willans lines" (page 274), and might just as well be plotted for total "flow" of steam per hour as for nozzles open. In fact in turbines where there are no nozzles the ' ' flow ' ' of steam must be used. It is obvious that any load curve of brake horse power giving the total steam consumption can be used to deter- mine the rotation loss by producing the " flow " line to the axis on which the output is scaled. A good check on the results of such rotation loss tests is secured by observing whether the lines for the speeds near the rated speed cross each other at about the rated output. In a good design the speed-output curve will give nearly the same output at speeds considerably above or below the rating. 1 The no load steam consumptions of 2000, 5000 and 9000 kilowatt Curtis turbine generators are respectively about 14, 12.5 and 8 per cent of that at full load. In other words these percentages are only from 1 to 2 per cent greater than the sum of the disk and blade rotation and generator windage losses. Generator windage loss is probably about equal to the sum of all the turbine losses. It is generally assumed that the no load steam consumption of a Parsons turbine (without the generator) is about 1 2 per cent of that at the normal maximum output. It is stated 2 that at no load the steam required for very large reciprocating engines and generators is probably in no case less than 1 5 per cent of that used at full load. Leakage Loss. The other important mechanical loss in a steam turbine is that due to the leakage of steam through the passages of the turbine without doing work. In impulse tur- bines of more than one stage this loss is chiefly caused by the leakage of steam between the shaft and the diaphragms. In a great many turbines no satisfactory packing is provided at these places and the loss is sometimes more than 10 per cent of the total anount of steam supplied to the turbine. In reac- tion turbines the loss is due to leakage through the radial clear- ance passages and is large or small in proportion to the size 1 For a more detailed description and illustrative figures the reader is referred to the author's book on The Steam Turbine, pages 120-123. 2 Kruesi, Proc. Am. Street and Interurban Railway Engineering Association, 1907. TESTING STEAM TURBINES AND TURBINE GENERATORS 293 of these clearances. The loss is usually assumed to be about 5 per cent in good Parsons turbines. Future improvements in the economy of all types of steam turbines will depend largely on the success of designers in reduc- ing these leakage losses. Analysis of Losses. The following table shows how the losses in a De Laval 200 -kilowatt turbine generator have been divided up by Stevens and Hobart : Nozzle losses 12 per cent Radiation losses and leakage 1 " Rotation losses due to the turbine wheel revolving in steam 4 ' ' Losses due to the steam traveling over the blades 9 ' ' " Bearing friction losses 1 " Losses in speed-reduction gearing 2 " Generator losses 4 " Losses due to residual kinetic energy in the steam passing to the condenser 8 " " Electrical output 59 " Total 100 Electrical Output of Turbine Generators. Measurement of Direct Current. Careful engineers will not ordinarily use the instruments on the switchboard of a power station for measur- ing the electrical output of a generator, because, unless excep- tional precautions have been taken to avoid "stray " magnetic fields and the instruments have been calibrated in place under operating conditions with a sufficient interval of time between observations of current (amperes) at different loads so that the shunts of the ammeters will reach a constant temperature for the particular value of current flowing there may be con- siderable error in the observations. Switchboard voltmeters are usually satisfactory if they are carefully calibrated; but the shunts of the type of ammeters ordinarily used have approximately only 60 millivolts drop, so that the indicating part of the ammeter must be almost entirely a circuit of copper wire. It is for this reason that such instruments are likely to be affected considerably by varying room temperatures, and with some shunt arrangements they are susceptible to errors, also from variations in the value of the circuit itself. For accurate measurements, it is therefore best to use only the port- able types of indicating ammeters having shunts of 200 milli- volts 1 drop. In these latter instruments the indicating part 'This value for the drop in shunts is an arbitrary value selected 294 POWER PLANT TESTING is made up largely of resistance wires having practically no tem- perature coefficient. Portable voltmeters are also to be pre- ferred to those on the switchboards. Unless standard shunts of 2co millivolts drop as provided for good portable ammeters are used the influence of " stray " magnetic fields must be guarded against. When on the other hand switchboard instruments are used, such influences must be investigated and arrangements must be devised so that " stray " fields will not affect the measurements. The influence of very weak magnetic fields can be eliminated from the final results by turning the instruments between successive readings. Observations of current (amperes) made with the switchboard type of instruments, are also often in error due to thermo- electric effects producing a small electromotive force sufficient, however, to alter the readings of the millivoltmeter. The error due to this cause can be observed by reading the millivolt- meter at the close of the test immediately after the current has been shut off in the main circuit. It should be, of course, the object of the observer to take this reading before the shunts and leads have cooled appreciably. If there is an error due to this cause there will be a small position or negative deflection of the needle from the correct zero, which should be applied as a correction to all the observations of current. Measurement of Alternating Current. The same general precautions outlined above for direct -current instruments, must be observed in the use of those for alternating current. Although steady magnetic fields are not often a cause of much trouble, it happens often, particularly in the case of large generators, that there are large magnetic fields influencing the measuring instruments which have the same frequency as that of the cur- rent measured. To eliminate the effect of such " stray " fields shielded types of instruments shotild be used. Only with the most expert handling can accurate results be expected when unshielded instruments are used. For measuring large values of alternating current, instrument transformers are generally used. These should be of the precision type and should be sent to a standardizing laboratory before and after a series of tests by a number of makers of electrical instruments because it gives the best compensation of all the temperature errors. See General Electric Review, February, 191 1, TESTING STEAM TURBINES AND TURBINE GENERATORS 295 to be calibrated, and a certificate of accuracy should be obtained. The transformers should be calibrated at as nearly as possible the values of the current to be measured in the tests. Whenever it is possible tests of generators should be made with a non-inductive load, water rheostats being usually the most satisfactory apparatus for providing such a load. At least tests should be made under conditions giving a low power- factor, so that there can be no error in the readings of the instru- ments due to phase displacements in the instrument trans- formers. With a purely non-inductive load the readings of the ammeters and the voltmeters can be used to check the watt- meters. Although the readings of the wattmeters should be taken as the correct value of the output the apparent power as indicated by the ammeters and voltmeters should agree with the wattmeter readings within one per cent. If a non-inductive load cannot be secured the switchboard ammeters and voltmeters will be satisfactory for readings to indicate whether or not the load on the circuits is properly balanced. Watt-hour meters are not usually satisfactory for the accuracy expected in most tests, and the use of these instruments should be generally avoided. It is only in the case where tests must be made under extremely variable service conditions, where it is difficult to obtain a true average from the readings of indicating instru- ments, that watt-hour meter, either for direct or for alternating current, may sometimes give more accurate results than the portable indicating types of instruments. Whenever watt- hour meters are used in tests they should be checked in place for a series of constant loads at the frequency, voltage/ etc. which are to be used in the test. Single-phase indicating instruments are to be preferred for measurements of polyphase current to the standard types of so-called polyphase instruments. The reason for this pref- erence is that the indications of a polyphase instrument are produced by two influences from separate phases of the current in such manner that a correction cannot be applied to obtain true values unless the division of the load is determined by the use of single-phase instruments. Obviously, then, if it is neces- sary to have single-phase instruments in the separate circuits, it is desirable to have them of the precision type, and polyphase instruments are not needed. CHAPTER XIII METHODS OF CORRECTING STEAM TURBINE AND ENGINE TESTS TO STANDARD CONDITIONS Standard Conditions for Turbine and Engine Tests. If tests of steam turbines and engines could be always made at some standard vacuum, superheat, and admission pressure, then turbines and engines of the same size and of the same type could be readily compared, and an engineer could determine without any calculations which of two turbines or engines was more economical for at least these standard conditions. But steam turbines and engines even of the same make are not often designed and operated at any standard conditions, so that a direct comparison of steam consumptions has usually no significance. It will be shown now how good comparisons of different tests can be made by a little calculation involving the reducing of the results obtained for varying conditions to assumed standard conditions. The method given here is that generally used by manufacturers for comparing different tests on the same turbine or engine (a " checking " process) or on different types to determine the relative performance. To illustrate the method by an application, a comparatively simple test will first be discussed. Practical Example. Corrections for Full Load Tests. The curve in Fig. 205 shows the steam consumption for varying loads obtained from tests of a 125-kilowatt steam turbine operating at 27.5 inches vacuum, 50 degrees Fahrenheit super- heat, and 175 pounds per square inch absolute admission pres- sure (at the nozzles) . It is desired to find the equivalent steam consumption at 28 inches vacuum, o degrees Fahrenheit super- heat, and 165 pounds per square inch absolute admission pres- sure for comparison with the "guarantee tests" (Fig. 206) of a steam engine of about the same capacity operating at the latter 296 METHODS OF CORRECTING TURBINE AND ENGINE TESTS 297 conditions of vacuum, superheat, and pressure. The manu- facturers of the steam turbine have provided the curves in Figs. 207, 208, and 209, showing the change of economy with 20 10 L60 180 :.'0i) 60 80 100 120 110 Output of Generator in Kilowatts Fig. 205. — Water Rate Curve of a Typical 125-Kilowatt Steam Turbine. (Generator Output.) varying vacuum, superheat, and pressure. With the help of these correction curves, the steam consumption of the tur- bine can be reduced to the conditions of the engine tests. Fig. •2 §35 s«30 \ \ \ ' \ \B » X s s ^ ^c __. .-■ -'" A. Steam Consumption of Engine of g tea nT urb ne. gs* 30 10 60 80 100 120 140 160 180 200 Output of Generator in Kilowatts Fig. 206. — Comparative Water Rate Curves of a Reciprocating Steam Engine and a Steam Turbine. (Both with Standard Generators.) 207 shows that between 27 and 28 inches vacuum a difference of 1 inch changes the steam consumption 1.0 pound. Fig. 208 shows a change of 2.0 pounds per 100 degrees Fahrenheit 298 POWER PLANT TESTING superheat, and from Fig. 209 we observe a change of 5.0 pounds in the steam consumption for 100 pounds difference in admission pressure. Compared with the engine tests the steam turbine was operated at .5 inch lower vacuum, 50 degrees Fahrenheit higher superheat, and 10 pounds higher pressure. At the con- 30 25 20 15 21 Ti 29 Fig. 207. 23 24 25 26 Vacuum Inches of Mercury -Vacuum Correction Curve for a 125-Kilowatt Steam Turbine. ditions of the engine tests, then, the steam consumption of the steam turbine should be reduced . 5 pound to give the equivalent at 28 inches vacuum, but is increased 1.0 pound to correspond to o degrees Fahrenheit superheat, and .5 pound more to bring it to 165 pounds absolute admission pressure. The full load SO 100 120 140 Superheat -Degs. Fahr. Fig. 208. — Superheat Correction Curve for a 125-Kilowatt Steam Turbine. steam consumption for the steam turbine at the conditions required for the comparison is, therefore, 24.5 — .5 + 1.0 +.5, or 25.5 pounds. 1 1 The corrected steam consumption is found to be nearly the same as that which the three correction curves show for the same conditions, that is, about 25.0 pounds. If there had been a difference of more than about 5 per cent between the corrected steam consumption and METHODS OF CORRECTING TURBINE AND ENGINE TESTS 2C9 Persons who are not very familiar with the method of making these corrections will be liable to make mistakes by not knowing whether a correction is to be added or subtracted. A little thinking before writing down the result should, however, pre- vent such errors. When the performance at a given vacuum is to be corrected to a condition of higher vacuum, the correc- tion must be subtracted, because obviously the steam consumption is reduced by operating at a higher vacuum. When the steam consumption with superheated steam is to be determined in its equivalent of dry saturated steam (o degrees superheat) the correction must be added because with lower superheat there is less heat energy in the steam and consequently there is a a 1* .2 5 25 SW20 as,, ft* 10 loo no 120 11)0 200 130 140 150 160 170 180 Steam Pressure -Lbs. Per Sq. In. Abs. Fig. 209. — Pressure Correction Curve for a 125-Kilowatt Steam Turbine. larger consumption. Usual corrections for differences in admission pressure are not large; but it is well established that the economy is improved by increasing the pressure. Corrections for Fractional Loads. It is the general expe- rience of steam turbine manufacturers that full-load correction curves, if used by the following "ratio " or percentage method, can be used for correcting fractional or over loads. This state- ment applies at least without appreciable error from half to one and a half load, and is the only practicable method for quarter load as well. 1 Stated in a few words, it is assumed then that the steam consumption at fractional loads is changed by the same percentage as at full load for an inch of vacuum, that of the correction curves for the same conditions, the "ratio" method as explained on page 300 for fractional loads should have been used also for full load. 1 A very exhaustive investigation of this has been made by T. Stevens and H. M. Hobart, which is reported in Engineering, March 2, 1906. 300 POWER PLANT TESTING a degree of superheat, or a pound pressure. It will now be shown how this method applies to the correction of the steam consumption of the turbine at fractional loads. Now accord- ing to the curve in Fig. 207, the steam consumption at 27.5 inches (25.6 pounds) must obviously be multiplied by the ratio 1 — — . of which the numerator is the steam consumption at 28 25-6 inches and the denominator at 27.5 inches, to get the equivalent consumption at 28 inches vacuum. This reasoning establishes the proper method for making corrections; that is, that the base for the percentage (denominator of the fraction) must be the steam consumption at the condition to which the correc- tion is to be applied. 2 Similarly the correction ratio to change the consumption at 50 degrees Fahrenheit superheat to o degrees 2^0 Fahrenheit is - • ' -, and to correct 175 pounds pressure to 165 pounds the ratio is 24.8 24-3' Data and calculated results obtained by this method may then be tabulated as follows : Conditions of Test. Required Conditions. Correction Ratio. Percentage Correction. 27-5 5°- 175- 28 165 25.0 25.6 25-° 24 . 24.8 24-3 -2.34%* + 4-i7% + 2.06% + 3-89% Superheat, degrees Fahrenheit Admission pressure, pounds absolute * Steps in the calculation are omitted in the table, thus -—- = .9766 or 97.66 per cent 25.6 making the correction 100 — 97.66, or 2.34 per cent. It may seem unreasonable to the reader that these percentages are calculated to three figures when the third figure of the values of steam consumption is doubtful. In practice, however, the ruling of the curve sheets must be much finer and to larger scale so that the curves can be read more accurately. 1 Assuming that this short length of the curve may be taken for a straight line without appreciable error. 2 In nearly all books touching this subject so important to the practical, consulting, or sales engineer, the alternative method of taking the steam consumption at the required conditions as the base for the percentage calculation is implied. By such a method percentage cor- rection curves derived from straight lines like Figs. 208 and 209 would be straight lines and, in application, give absurd results. Actually such percentage corrections will fall on curves. METHODS OF CORRECTING TURBINE AND ENGINE TESTS 301 The signs + and — are used in the percentage column to indicate whether the correction will increase or decrease the steam consumption. " Net correction " is the algebraic sum of the quantities in the last column. The following table gives the results of applying the above " net correction " to fractional loads. \ Load 31.3 kw. i Load 62.5 kw. | Load 93.8 kw. f Load 125 kw. 31.2 + 1.2 3 2 -4 26 .9 4- 1.1 28.0 25.2 + I.O 26 . 2 24-5 + 1.0 2 5-5 S Load 156.3 kw. Steam consumption from test (Fig. 205) Net correction + 3 . 89% Corrected steam consumption 23.6 + 0.9 24-5 Curve B in Fig. 206 shows the corrected curve of steam con- sumption for the steam turbine as plotted from the above table. By thus combining, on the same curve sheet, curves A and B as in this figure, the points of better economy of the turbine are readily understood. Results of economy tests of the various turbines given on the preceding pages are of very little value for comparison when the steam consumptions or "water rates " are given for all sorts of conditions. With the assistance, however, of curves like those shown in Figs. 207, 208 and 209, if they are representative of the type and size of turbine tested, it is possible to make valuable comparisons between two or more different turbines. Some very recent data of Curtis and Westinghouse-Parsons turbines are given below, together with suitable corrections adopted by the manufacturers for similar machines. The following test of a Westinghouse-Parsons turbine, rated at 7500 kilowatts, was taken at Waterside Station No. 2 of the New York Edison Co., and a comparison is made with a test of a five-stage 9000 -kilowatt Curtis turbine at the Fisk Street Station of the Commonwealth Electric Company of Chicago. As no pressure correction is given for the Curtis machine, the New York Edison test is corrected to the pressure at which the other machine was operated (179 pounds per square inch gage). Approximately an average vacuum for the two tests is taken for the standard, and 100 degrees Fahrenheit superheat is used for comparing the superheats. These assumed standard con- 302 POWER PLANT TESTING 7500 KILOWATT WESTINGHOUSE-PARSONS TURBINE, WATER- SIDE STATION NO. 2, NEW YORK EDISON COMPANY Tested September i, 1907 Corrected Correction per cent.* Duration of test, hours Speed revolutions per minute Average steam pressure, pounds gage . . . Average vacuum, ins. (referred to 30 in. barom.) Average superheat, degrees Fahrenheit . . Average load on generator, kilowatts. . . . Steam consumption, pounds per kilo- watt-hour Net correction, per cent Corrected steam consumption, pounds per kilowatt-hour 75° 177-5 27-3 95-7 9830.5 i5-i5t 79 28.5 100 ■3-30 o. 29 •3.80 *The following corrections were given by the manufacturers and accepted by the pur- chaser as representative of this type and size of turbine : Pressure correction .1 per cent for 1 pound. -> Vacuum correction 3.5 per cent for 1 inch. I from Electric Journal, Superheat correction 7.0 per cent for 100 degrees Fahren- | November, 1907, page 658. heit. J t This is -j\ per cent better than the manufacturer's guarantee. 9000 KILOWATT CURTIS TURBINE, FISK STREET STATION, COMMONWEALTH ELECTRIC COMPANY, CHICAGO. Tested in :9°7 Corrected to Correction, per cent.* 7 5° 179 29-55 116 8070. i3-o Average steam pressure, pounds gage . . Average vacuum, inches (referred to 3c 179 28.5 IOO .0 + 12.39 + I.28 Average superheat, degrees F Average load on generator, kilowatts. . . . Steam consumption, pounds per kilo- + 1367 Corrected steam consumption, pounds 14.77 * The following percentage corrections were used': Superheat corrections 8 per cent for 100 degrees Fahrenheit. Vacuum correction 8 per cent for 1 inch from c rve in Fig. 21c Pressure correction not given. E. Bulletin, No. 4531. METHODS OF CORRECTING TURBINE AND ENGINE TESTS 303 ditions make the corrections for each turbine comparatively small. When two tests are to be compared, by far the more intelligent results are obtained if each is corrected to the average conditions of the two tests, rather than correcting one test to the conditions of the other. There is always a chance for various errors when large corrections must be made. These results show a difference of only .20 pound in the cor- rected steam consumption, so that for exactly the same con- ditions these two machines would probably give approximately the same economy. Each turbine is doubtless best for the special conditions for which it was designed. These results are equivalent to respectively 9.58 pounds and 9.72 pounds per indicated horse power, assuming 97 per cent as the efficiency of the generator and 9 1 per cent as the mechani- cal efficiency of a large Corliss engine according to figures given by Stott. 1 From experience with other similar turbines it seems as if the vacuum corrections given are too low for each turbine. The correction for the Curtis turbine was obtained from the curve in Fig. 210, as given between 27 and 28 inches, while it was used between 28.5 and 29.5 inches, where the curve of steam consumption most likely slopes somewhat, as shown by the dotted line in the figure, which was derived from the per- centage change of the theoretical steam consumption calculated from the available energy. The correction of 2.7 per cent per inch of vacuum for the Westinghouse-Parsons turbine is probably too low also, although the percentage correction 1 Electric Journal, July, 1907. It is stated also in this article that the vacuum correction of a Westinghouse-Parsons turbine is 3.5 per cent per inch between 28 and 28.5 inches. Jude states that the vacuum cor- rection for Parsons turbines is 5 to 6 per cent. .2 » M o u -3 18 1G s S N N 1~> \ 10 :« 2± 27 Vacuum Inches of Mercury Fig. 210. — Typical Vacuum Correction Curve for a 5000-Kilowatt Impulse Turbine. 304 POWER PLANT TESTING would not be nearly as large as for the Curtis. If both of these corrections are too low, the effect of increasing them would be to increase the corrected steam consumption of the Curtis turbine and reduce that of the Westinghouse-Parsons. Tests of a 5000 -kilowatt Curtis and a 7 500 -kilowatt Westing- house-Parsons turbine are also recorded here for comparison. The two tests are corrected to the assumed standard conditions of 173.7 pounds gage pressure, 28 inches vacuum, and o degrees Fahrenheit superheat. For the test of the Curtis machine the same percentage corrections were used as for the 9000 -kilowatt turbine; and for the test of the Westinghouse turbine the vacuum correction is that given in the footnote at the middle of page 302 (3.5 per cent per inch), while the other percentage corrections are the same as in the preceding test of a similar machine. The Westinghouse turbine was operated with wet steam. In a test of a reciprocating engine the equivalent economy with dry steam is calculated by merely subtracting the percentage of moisture, but in a turbine test the correction is generally stated as being a little more than twice the percent- age of moisture. In other words, in a turbine test the moisture must be subtracted twice. The reason for this difference in the methods of correcting water rates of engines and turbines is the very large increase in the disk and blade rotation losses in wet steam 5000-KILOWATT FIVE-STAGE CURTIS TURBINE, L STREET STATION, BOSTON EDISON COMPANY Tested January 29, 1907 Corrected to Correction, per cent. 3 720 173-7 28.8 142 5 J 95 *3-5 2 Average steam pressure, pounds gage . . . Average vacuum in. (referred to 30 in. 173-7 28 O O Average superheat, degrees Fahrenheit . . Average load on generators, kilowatts. . . . Steam consumption, pounds per kilo- + 11.36 + 17-76 Corrected steam consumption, pounds !5-9 2 METHODS OF CORRECTING TURBINE AND ENGINE TESTS 305 7500-KILOWATT WESTINGHOUSE-PARSONS TURBINE. INTER- BOROUGH RAPID TRANSIT COMPANY, NEW YORK Tested in 1907 Corrected to. Correction, per cent. 3 Average steam pressure, pounds gage . . . Average vacuum, in. (referred to 30 in. 149.7 27.70 3-° 7135 17.79 173-7 28 O -2.4 -I.05 -6.0 Average moisture, per cent Average load on generator, kilowatts. . . . Steam consumption, pounds per kilo- -9-45 Corrected steam consumption, pounds 16 . 10 It is stated that the steam consumption of the Interborough Company's turbine is 15.87 pounds at full load and 15.54 pounds at 9000 kilowatts when the overload valve opens. The gen- erator connected to this turbine is rated at only 5500 kilowatts. With a generator more nearly the rating of the turbine it is probable still better results would be secured. CHAPTER XIV GAS AND OIL ENGINE AND PRODUCER TESTING The testing of internal combustion engines of the reciprocat- ing type operating with gas, gasoline, kerosene, and alcohol does not differ essentially in the important details from steam engine practice already explained in Chapter XI. Indicator diagrams must be utilized to show the inner workings in the engine cylinder, giving a record of the pressure, " timing " of the valves and ignition for the operation of the engine through a cycle. 1 Brake horse power is measured with a Prony brake or any other type of dynamometer permitting the determination with facility of the power of the engine. If with a Prony brake or similar device (see page 122) then the brake horse power is expressed by the usual formula, B.H.P.= 2 -^ (87) 33,ooo where 1 is the length of the brake-arm in feet, n is the number of revolutions per minute, and w is the net weight indicated by the scales on the brake. Similarly the indicated horse power is given by the Uoual formula for a single-acting steam engine (page 118) except that the number of explosions must be used in calculations instead of the number of revolutions, thus, I.H.P.= Pk*- (88) 33,ooo 1 In what is called usually a four-cycle engine there are four piston strokes — one each for suction, compression, expansion, and exhaust, corresponding to two revolutions of the crank shaft for a complete cycle, while in a so-called two-cycle engine, two strokes of the piston make a complete cycle — corresponding to one revolution of the crank-shaft. In the latter case suction and compression are combined in one stroke and expansion and exhaust in another. 306 GAS AND OIL ENGINE AND PRODUCER TESTING 307 where p is the mean effective pressure in pounds per square inch measured from the indicator diagram, 1 is the length of the stroke in feet, a is the area of the piston in square inches, and n e is the number of explosions per minute, then B H P Mechanical Efficiency = ' ' ' . . . (89) Certain important precautions should be observed when taking indicator diagrams, so that a reasonable degree of accuracy may be expected from the results of tests. In the first place the connections between the indicator and the combustion chamber should be as short as possible. It is much more important that in an internal-combustion engine the volume by which the combustion chamber (clearance) is increased by the indicator connections should be small in comparison to the volume of the engine cylinder than in a steam engine; because by increasing the clearance volume, obviously, the pressure resulting from compression is reduced as well as the pressure due to the explosion. It is in this way that large indicators and indicator connections may cause a considerable reduction in the thermal efficiency of an engine, that is, reducing the efficiency of the transformation of heat energy into work. Tests of gas engines as made commercially have usually three objects in view: (1) Brake horse power. (2) Indicated horse power. (3) Gas or oil consumption per horse power per hour. Many types of gasoline engines, particularly those designed for the automobile service, operate at such high speeds that the indicated horse power cannot be obtained with accuracy. In many other kinds of engines classed in this group the mechan- ical efficiency is very low. It is for these reasons that such engines are rated by the useful or brake horse power instead of by the indicated horse power, as with steam engines. Brake horse power is therefore the prime criterion by which the per- formance of these engines is expressed. Ordinary types of steam engine indicators, moreover, are not very satisfactory for testing gas engines, and many engi- neers prefer to use one of the type shown in the accompanying illustration, Fig. 211. It differs essentially from steam engine 308 POWER PLANT TESTING indicators of the same type in having in the lower part of the main " barrel," a cylinder of smaller diameter than the one just above it, containing the spring. This smaller cylinder takes a piston of only half the area of the standard size. By this means the scale of the spring ordinarily used is doubled and the shock on the small rods and levers of the pencil motion -Crosby Gas Engine Indicator. is only half as great, thus making the liability to breakage and the cost of repairs for such indicators very much less than when a piston of the standard size is used. Measurement of the Fuel Used. For tests of gas engines the gas used is usually measured by means of a gas meter. A so-called " wet " meter (page 143) is always to be preferred, but if a carefully calibrated dry meter is used very good results can be obtained and it is accurate enough for commercial tests. For gasoline, kerosene, and other oil engines the amount of fuel used is preferably determined by direct weighing. The GAS AND OIL ENGINE AND PRODUCER TESTING 309 author has found the automatic indicating scales of the pen- dulum type 1 now generally used by grocers and meat dealers to be most satisfactory. By this means the weight of the oil remaining in the " supply " vessel can be observed regularly throughout a test just as with a gas meter, so that irregularities in operation can be immediately observed. The vessel contain- ing the oil used by the engine when placed on a scales must be connected to the carbureter or the pump, as the case may be, with a very flexible metallic tube made without rubber inser- tion. 2 If an indicating scales is not available a very small platform or grocers beam scales can be used satisfactorily, although if the weight of fuel oil is desired at regular intervals throughout a test some little time is required to balance the poise. Another method very commonly used, however, is to use a cylinder vessel of small diameter provided with a gage glass in which the level of the liquid can be observed. Such a vessel can be calibrated to determine the weight or volume of oil per inch of height measured on the gage glass. Observations taken for a test of a gas or oil engine are in general much more uniform than the corresponding data taken in a steam engine trial. For this reason gas engine tests in particular can be made of much shorter duration than upon a steam engine for the same degree of accuracy. For both gas and oil engines the points plotted on a scale of brake horse power for abscissas and fuel used per unit of time for ordinates will fall along a straight line, similar and resem- bling the Willans line for steam engines and steam turbines (see page 274). A typical set of curves of the results of a test of a gas engine is shown in Fig. 212. Brake horse power is taken for the abscissas, as should always be done for gas and oil engine tests and gas used per hour, to its corresponding scale of ordinates, is given by the top curve. Other curves show the gas used per brake horse power per hour and per indicated 1 Very satisfactory scales for this purpose are made by the Toledo Scale Co. of Toledo, Ohio. -Similar scales of the spring type are not recommended, because necessarily some of the vibrations of the engine will be transmitted to the scales and the indications of the pointer will not be as accurate as they should be. 2 Tubes of this kind are made by the U. S. Metallic Tube Co. of Los Angeles, Cal. 310 POWER PLANT TESTING horse power per hour, the indicated horse power, the mechan- ical efficiency, the number of explosions per minute, the revo- lutions per minute and the thermal efficiency (heat equivalent of the indicated horse power divided by heat supplied). The following paragraphs and tabular forms are taken from the " Rules for Conducting Tests of Gas or Oil Engines " as adopted by the American Society of Mechanical En- gineers. 1 Duration of Test. The duration of a test should depend largely upon its character and the objects in view, and in any case the test should be continued until the successive readings of the rates at which the oil or gas is consumed, taken at, say, half -hourly intervals, become uniform and thus verify each othei. If the object is to determine the working economy, and the period of time during which the engine is usually in motion is some part of twenty-four hours, the dura- tion of the test should be fixed for this number of hours. If the engine is one using coal for generating the gas, the test should cover a long enough period of time to determine with accuracy the coal used in the gas producer; such a test should be of at least twenty-four hours' duration, and in most cases it should extend over several days. Measurement of Fuel. If the fuel used is coal furnished to a gas producer, the same methods apply for determining the con- sumption as are used in steam boiler tests. (See pages 21C-21 7.) 1 Transactions American Society of Mechanical Engineers. Vol. 24, pages 775-79°, Brake Fig. 212. — Typical Economy, Speed, Horse Power and Efficiency Curves of a Five Horse Power Gas Engine. GAS AND OIL ENGINE AND PRODUCER TESTING 311 If the fuel used be gas, the only practical method of measur- ing is the use of a meter through which the gas is passed. Gas bags should be placed between the meter and the engine to diminish the variations of pressure, and these should be of a size proportionate to the quantity used. Where a meter is employed to measure the air used by an engine, a receiver with a flexible diaphragm should be placed between the engine and the meter. The temperature and pressure of the gas should be measured, as also the barometric pressure and temperature of the atmosphere. Data and Results of test of Gas or Oil Engine Standard Form i. Made by of on engine located at to determine 2. Date of trial 3. Type of engine, whether oil or gas 4. Class of engine (mill, marine, motor, for vehicle, pumping or other) 5. Number of revolutions for one cycle and class of cycle 6. Method of ignition 7. Name of builders 8. Gas or oil used (a) Specific gravity (b) Flashing point, deg. F. 1 (open or closed vessel) (c) Burning point, deg. F. . . 9. Dimensions of engine (a) Class of cylinder (working or for compressing the charge) (b) Vertical or horizontal (c) Single or double acting (d) Cylinder dimensions Bore, ins Stroke, ft*. Diameter piston rod, ins Diameter tail rod, ins ( are in- variably provi- ded with an air- tight spiral cas- ing C as shown in Fig. 224, where the suction is at S and the discharge is at D. The fan shown in this figure repre- sents the celebra- ted Rateau designs, which are exten- sively used in Eu- rope. Disk or Propeller Fans are best illus- trated by the so- called " electric " fans so commonly used in offices, shops and dwell- ings. Fans of this type are usually of a very light con- struction with the vanes arranged as in a screw propeller for a ship. In many 1 See Professor Rateau's articles in Revue de M ecanique , vol. 1, pages 629-837. Fig. 225. — Turbine Type (Sirocco) Fan. TESTING OF VENTILATING FANS OR BLOWERS 327 cases fans of this type are not provided with casings, so that it is more difficult to make velocity measurements than with centrifugal fans. Turbine or " Sirocco " fans have an impeller or fan wheel of the " squirrel cage " type, as illustrated in Fig. 225. Fans of this type can be designed to give very high efficiencies. This is due primarily to two characteristic features adopted in these designs. By the use of very short blades a very large intake space for the suction is provided which is practically unobstructed, thus giving a very free " suction." The other important feature of this fan is that the air leaves the blades Fig. 226. — Velocity Diagram for a Turbine Fan. Fig. 227. — Velocity Diagram for a " Standard " Fan. at a higher velocity than that at which the tips of the blades are moving. The importance of this result is shown by a com- parison of Figs. 226 and 227. The former illustrates the type of blading in a turbine or " Sirocco" fan and shows graphically by a velocity diagram, constructed like a parallelogram of forces, the velocity of the tips of the blades V&, the velocity of radial flow in the blades W and the absolute velocity of the discharge V a , which is the velocity of the air with respect to the stationary casing. It will be observed that in Fig. 226, V a is nearly 50 per cent greater than the velocity of the tips Vfc, while in Fig. 227, representing the corresponding velocities for a standard type of fan, the absolute velocity of the discharge V a is actually considerably less than the speed of the tips of the blades. Increased velocity is accomplished in a type of 328 POWER PLANT TESTING fan like Fig. 226, not only by the curvature of the tips of the blades, but also to some extent by making the blades somewhat concave with the inner ends (toward the center) practically radial. By this method of designing the outer edges of the blades have a smaller space between them than the inner edges. This has the effect of reducing the area on the discharge side of the blades and consequently the velocity of the air is increased. Positive Pressure Blowers are used principally for blast furnaces and smelters where a higher pressure is needed than can be efficiently obtained with a centrifugal fan. A section showing the rotors and casing of one of the blowers is shown in Fig. 228. It is often called Root's blower. The efficiency of a blower of this type de- pends on the accuracy of the fitting of the two rotors, A and B, both with respect to each other and to the casing. It is for this reason that when new the efficiency is high, but after being in ser- vice for several years the bearings and the surfaces of the rotors will become worn, so that there is con- siderable leakage and consequent loss of efficiency. Tests of Ventilating Fans or Blowers are made usually by a very simple method ; that is, by determining the necessary data for calculating efficiency by measuring the work done by the fan " on the air " in giving velocity, and the power required to drive the fan alone, excluding bearing friction. The fan is preferably operated by an electric motor, of which the efficiency can be readily determined by a Prony brake test. The power required to overcome the bearing friction of the shaft of the fan may well be first determined by measuring the power input , 1 For normal operation " friction work " is, for machinery in" general, proportional to the speed. Fig. 228. — Typical Positive Pressure Blower. TESTING OF VENTILATING FANS OR BLOWERS 329 (kilowatts) for a series of speeds when the keys fastening the fan to its shaft have been removed and the fan itself has been " blocked " in its casing. The shaft will then revolve in the hub of the fan and in the bearings. After attaching the fan again to the shaft the input to the motor and the work done by the fan " on the air " should be determined for various speeds. Then obviously the ratio of the work done by the fan divided by the power required to drive it after correction for the efficiency i of the motor and bearing friction is the actual efficiency of the fan. In general terms this may be stated as follows : f =input to motor to drive motor and shaft of fan in bear- ings, kilowatts; i =input to motor to drive motor and fan, in kilowatts; e = efficiency of motor for motor input of i kilowatts and at speed of test ; e' =efficiency of motor for motor input of f kilowatts and at speed of test. Then if i n is the net work in horse power to drive fan alone, ei-e'f . , >=^6' • ' • • • • <9 2 > If the fan to be tested is direct-connected to a steam engine, the test is usually made by measuring the indicated horse power of the engine at the various speeds and also with the fan disconnected for no load. If it is possible to do so the fan should be removed from the shaft for the no-load test to determine bearing friction. The work done by the fan " on the air," is most readily calculated by the same method used to calculate the efficiency of hydraulic pumps (see page 357), that is, by multiplying the number of pounds of air delivered per unit of time by the head in feet of air corresponding to the discharge pressure. This product is obviously in terms of work in foot-pounds per unit of time, and dividing by 33,000 the corresponding horse power is obtained. Using the following symbols, the same results may be expressed, however, by the product of " pressure times volume " as follows: 1 Motor efficiency must be necessarily determined for the conditions of each test; that is, for the same kilowatts and speed as for each test. 330 POWER PLANT TESTING v = velocity of air in feet per second; h =head in feet of air necessary to produce a velocity of v feet per second, or, =water pressure p in inches observed with a manometer, produced by the velocity of the air times the ratio. wt. of a cubic foot of water wt. of a cubic foot of air ./— Z — PX62.3 . . v=V 2 gh=^2g wtcuftairfort - . . . ( 93 ) and V /K = velocity in feet per minute is (taking 2g=64.^) Vm ° I096 -W wt.cu.ft P air for test ' ' ' (94) Now the velocity in feet per minute V m multiplied by the area of the section at which the velocity was observed, gives cubic feet C of air discharged per minute, and if P is the total pressure (static + velocity) in pounds per square foot then we have for j the " air horse power " or the work done by the fan " on the CP and efficiency of fan E, is E^-- ^- ( 9 6) In 33,ooo \ n Velocity measurements are usually made with a Pitot tube consisting essentially as shown in Figs. 136 to 138, pages 143- 144, of two tubes with openings at the end, arranged so that one of them " faces " in the direction of flow and the other extends in a radial direction. The former is subjected to the sum of the velocity and static pressures, while the latter receives only the static pressure. 1 Weight of air taken for calculation must be that corresponding to the total pressure in the discharge pipe, the temperature and the humidity. For tables of weight of air see pages 145 and 331, also Kent's "Mechanical Engineers' Pocket-Book," 8th edition, pages 583-588, and "Calculating and Testing Ventilating Systems," issued by U. S. Navy Department, Washington. TESTING OF VENTILATING FANS OR BLOWERS 331 The following table of relative humidity for determinations with a wet- and a dry -bulb thermometer is used by the U. S. Weather Bureau: TABLE OF RELATIVE HUMIDITY, PER CENT Difference between the Dry and Wet Thermometers, Deg. F. Dry Ther- 1 1 1 1 1 1 mometer, Deg. F. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 is 16 17 18 19 20 21 22 23 24 26 28 30 Relative Humidity, Saturation being 100. (Barometer = 3o ins.) 32 89 79 69 S9 49 39 30 20 XI 2 40 92 ■■ 9S 08 60 52 4 5 3 7 29 23 IS 7 So 93 »7 80 74 07 61 SS 49 43 i« 32 27 21 16 n 5 60 94 89 8.1 78 73 68 pz pr 2 z CHAPTER XIX HYDRAULIC MACHINERY Tests of Hydraulic Machinery Tests of Boiler Feed-pumps. In general engineering practice there are two classes of pumps commonly used for " feeding " the water to steam boilers. These are: (1) Motor- or belt-driven pumps ; (2) Steam pumps. Motor-driven feed-pumps are more generally used in Europe than in America, but there are, however, many plants in which feed-pumps operated by direct connection to an electric motor or by belting from a line shaft are used. Fig. 236 shows a good modern example of such a pump. It has three plungers operated from a single shaft with the cranks set 1 20 degrees apart. The valves are accessible for cleaning or for repairs by removing the plates or " covers " C, C, C. The suction pipe is marked S and the discharge pipe D. An air-chamber A is pro- vided to produce, by cush- ioning air in it, a somewhat more steady flow than would be secured without it. A relief valve should be provided on the dis- charge pipe to act as a safety valve in case the pressure in the line should get so high that the pump itself might be broken. 356 Fig. 236. — Belt-driven Feed Pump. HYDRAULIC MACHINERY 357 The power delivered to a belt-driven pump can usually be conveniently measured with a Webber dynamometer or an Emer- son power scales (see pages 136—140), while if it is direct-con- nected to a motor the efficiency of the motor may be obtained by disconnecting it, attaching a Prony brake to the shaft, and measuring the input to the motor with suitable electrical instruments. The equivalent work done " on the water " by the pump is found by multiplying the total head * (suction -f discharge) in feet by the weight of water lifted (foot-pounds) . The quantity of water delivered can be determined by weigh- ing or by calculating the flow over a weir or from an orifice (see pages 156-161). Slip is the difference between the volume swept through by the plunger of the pump ; or, in general, the piston displacement, and the actual volume of the water pumped at the required head. In piston pumps, direct-connected to the steam cylinder without a crank, the length of the stroke is usually variable, and some special method must be adopted for such tests to determine the average length of the stroke. Duty of a pump is usually defined as the number of foot- pounds of work " delivered " by the pump per 1,000,000 B.T.U. supplied. The heat units supplied by the engine are calculated by the A.S.M.E. Rules 2 as the product of the weight of feed- water used by the boiler and the total heat of steam at boiler pressure " reckoned from the temperature of the feed-water." The total heat is to be corrected of course for moisture or super- heat. For a test utilizing a Webber or a similar dynamometer belted to the pump for measuring the power the following form may be used : 1 If the discharge head is measured by a pressure-gage on the dis- charge pipe then the equivalent pressure in pounds per square inch corresponding to the difference in level between the surface of the water- supply and the center of the gage must be added to get the total head. (One foot head of water at about 62 degrees Fahrenheit is equivalent to 0.434 pound per square inch; and conversely, one pound per square inch is equivalent to a head of 2.305 feet of water at the above temperature. See also foot-note page 361. 2 More detailed instructions for steam pumps with steam jackets are given in Transactions American Society of Mechanical Engineers, vol. 12, page 530. 358 POWER PLANT TESTING Test of a Belt-driven Pump (Dynamometer Method) i. Type of pump Made by 2. Diameter of plungers, ins > 3 . Length of stroke, ft 4. Size of suction pipe 5. Size of delivery pipe 6. Speed of dynamometer, r.p.m. . . . 7. Speed of pump, r.p.m 8. Dynamometer reading 9. Delivery pressure, lbs. per sq.in io. Suction pressure, lbs. per sq.in. or inches vacuum. 1 1 . Temperature of water, deg. F Delivery head in feet of water 13. Suction head in feet of water. 14. Total head in feet of water 15. Net weight of water pumped per minute, lbs 16. Work done by pump, ft. -lbs. per min. (i4)X(is) Cubic feet water pumped per minute ■ Plunger displacement, cu.ft. per min Slip, percent [(18) — (17)]^ (18) Net work delivered to pump (by dynamometer) ft. -lbs. per minute, Dynamometer horse power, (20) -f- 33 ,000 ■ Pump horse power, (16)^33,000 Mechanical efficiency, (22) -f- (21) ■ Capacity of pump, gallons delivered per 24 hours A Direct-acting Steam Feed-pump like the one shown in section in Fig. 237 will be tested in a somewhat different manner, and a different set of observations is required. Fig. -Direct-acting Steam Pump. In none of the so-called direct-acting steam pumps has a rotary motion been developed by means of which an eccentric can be made to operate the valve. It is, therefore, necessary to reverse the piston by an impulse derived from itself at the end of HYDRAULIC MACHINERY 359 each stroke. This cannot be effected in an ordinary single- valve engine, as the valve would be moved only to the center of its motion, and then the whole machine would stop. To overcome this difficulty a small steam piston is provided to move the main valve of the engine. In these pumps, the lever A, which is carried by the piston rod, comes in contact with the tappit when near the end of its motion, and, by means of the valve rod R moves the small slide valve which operates the supplemental piston. The sup- plemental piston, carrying with it the main valve V, is thus driven over by steam, and the engine reversed. If, however, the supplemental piston fails accidentally to be moved, or to be moved with sufficient promptness by steam, the lug on the valve Fig. 238. — Outside Packed Plunger Feed-pump. rod engages with it and compels its motion by power derived from the main engine. Outside-packed steam pumps of the plunger type (Fig. 238) are now very commonly used for supplying boiler feed-water, chiefly because at the pump end the only part subjected to ordinary wear is the packing of the plunger stuffing-boxes. Steam is admitted at A and exhausts at E. The suction pipe is at S and the discharge pipe at D. Suitable fittings for the attachment of indicators should' be provided at both the steam and the water cylinders. If the pump is of the ordinary direct-connected type, without a fly- wheel, like the one shown in Fig. 237, some provision must be made to make regular observations of the length of the stroke, 360 POWER PLANT TESTING as it is scarcely ever constant. One method is to attach a suitable arm to the cross-head H, Fig. 237, with a pencil at the end. Strips of tough paper can then be pasted on a board in such a position that the pencil will trace the lengths of the strokes. By shifting the position of the board every minute or two, records will be obtained from which the average length of the stroke can be estimated with considerable accuracy. Another method giving still greater accuracy is to use a counting device designed by Professor Cooley, operated by a mechanism similar to that in a clock (Fig. 240). A cord from the instru- ment is attached to the cross-head of the pump and the clock mechanism moved by this cord integrates or sums the lengths Fig. 240. — Cooley Stroke-measuring Devise. of all the strokes. Conditions should be maintained constant before beginning a test. Form for Steam Pump Test Duration of test Diameter of steam cylinder. . .. Diameter of piston rod Diameter of water plunger Diameter of plunger piston rod. 3 4 5 6. Displacement of plunger, cu. ft 7- f Head end . . [Crank end. Average length of stroke, ft Average number of strokes per minute. . . . : 9. Temperature of water, deg. F 10. Temperature of feed- water to boiler, deg. F. 1 1 . Temperature in steam calorimeter, deg. F . . 12. Feed-water supplied to boiler, lbs. per hour. 1 3 . Quality of steam 14. Dry steam supplied to boiler, lbs. per hour. 15. Boiler pressure, lbs. per sq.in 16. Delivery pressure, lbs. per sq.in HYDRAULIC MACHINERY 361 17. Suction pressure, lbs. per sq.in. or inches vacuum 18. Vertical distance between top of suction pipe where suction gage is attached and center of gage on delivery pipe, ft. 1 19. Total head in feet of water. 20. Weight of water delivered per hour, lbs 21. Plunger displacement, cu.ft. per hour 22. Weight of water by plunger displacement, lbs. per hour 23. Slip of pump, per cent [(21) — (19)]-^ (21) 24. Coal fired per hour, lbs 25. Combustible burned per hour, lbs 26. Steam used per pound of coal, lbs. (actual evaporation) 27. Equivalent evaporation ("from and at 212 deg. F.") per pound of of coal, lbs 28. Actual evaporation per pound of combustible, lbs 29. Equivalent evaporation per pound of combustible, lbs 30. Duty, per 1,000,000 B.T.U 31. Duty, per 100 lbs. coal fired 32. Duty, per 1000 lbs. steam (dry) 33. Capacity, gallons delivered per 24 hours 34. Mean effective pressure, steam cylinders, lbs. per sq.in 35. Mean effective pressure, water cylinders, lbs. per sq.in 36. Indicated horse power, steam cylinders 37. Indicated horse power, water cylinders 38. Dry steam used per indicated horse power per hour (steam cylinders) Cooley Stroke-measuring Device. An apparatus has been developed at the University of Michigan for measuring accu- rately the length of the stroke of the type of pumps in which steam and water cylinders are direct-connected on the same piston rod, such for example as the ordinary steam feed-pumps. In such pumps it scarcely ever happens that there are two strokes in succession that are of the same length and more or less approx- imate methods are usually adopted for obtaining the average length of the stroke during a test. With the stroke-measuring device referred to above each individual stroke is accurately measured and is added by a counting device to the sum of all the other strokes that have preceded. As this apparatus is used in a test of a variable stroke pump, the reading of the counter to the nearest inch can be recorded at the usual times for observations. The difference between two readings is the total length of all the strokes for the interval between observations. If, then, this difference is divided by the total 1 This is usually stated as the vertical distance between the two gages. A vacuum gage, however, on the suction pipe of a pump indicates the vacuum at the level of the top of the suction pipe, and not up to the center of the gage. This was shown by Professor Cooley by attaching a gage by means of a suitable fitting to the suction pipe of a pump so that the gage could be revolved above and below the pipe. It was observed that the reading of the gage remained constant, showing that in a suction pipe of a pump the water does not rise higher than the top of the pipe. 362 POWER PLANT TESTING number of strokes for the same time, the average length of the stroke can be determined accurately. An assembled view of this device is shown in Fig.' 240, and its mechanism is shown in Fig. 241, which it will be observed is the same in principle as the silent ratchet clutches used for the continuous indicator described on page 100. The apparatus is driven by the cord on the wheel W, which moves the ratchet wheels B and C in the same way as Fig. 241. — Mechanism of Stroke-measuring Device. the corresponding parts are moved in the continuous indicator referred to. Numbers on the horizontal plate (Fig. 240) are feet and those on the circular dial are inches. Tests of Centrifugal Pumps. Tests of pumps operating against low heads such as single-stage centrifugal pumps are suitable for, are made in the same way as explained, for the triplex belt-driven feed-pump. It is desired, of course, from the results of the tests to compare the power supplied to the pump with the work done in lifting the water. Power supplied would probably be again measured by some form of transmission dynamometer, and the work done is calculated from the weight of water delivered and the total head against which the pump delivers. 1 1 For more detailed testing of centrifugal pumps see "Centrifugal Pumps," by Lowenstejn and Crissey (D. Van Nostrand Co., 191 1). HYDRAULIC MACHINERY 363 Centrifugal pumps are frequently driven by direct-connected steam turbines.. The horse power required to drive the pump is then determined from a speed-power curve of the turbine (Fig. 203), page 281, obtained usually from a Prony brake test of the turbine. Similarly, if the pump is driven by a variable- speed electric motor, a' speed-power curve of the motor can be used. Usually, however, when a constant speed motor is used it is simpler to determine an efficiency curve of the motor for varying power. Tests of Impulse Water Wheels. Impulse wheels used to operate with water under pressure consist usually of a series of buckets attached to the periphery of a disk or wheel. The btickets are usually divided by a cen- tral rib so that two "pock- ets ' ' are formed (Fig. 242) . The curves for each of the divisions of the bucket are designed to turn the di- rection of the impinging steam without shock. Fig. 243 shows a typical im- pulse wheel. Impulse wheels are designed to operate most efficiently with high heads. It is, therefore, impracticable to measure the head directly in feet, but it is done usually by measuring the pressure near the nozzle N with a gage. When, the center of the gage is at a higher level than the center of the nozzle discharging on the wheel, then this difference in level must be added to the head calculated from the" gage pressure to determine the total head under which the wheel is operating. Power developed is meas- ured usually by a Prony brake connected to the shaft S. In all tests where a large quantity of water is used, the temperature of the water should be recorded and the weight corresponding Fig. -Bucket of an Impulse Water Wheel. 364 POWER PLANT TESTING Fig. 243. — Typical Impulse Water Wheel. Fig. 244. — Water Jet Discharging at High Pressure from the Nozzle of an Impulse Wheel. HYDRAULIC MACHINERY 365 should be used. A view of the jet discharged from one of these nozzles is given in Fig. 244. The type of impulse wheel most in use commercially is called the Pelton, of which typical buckets and the engaging jet of water are shown in Fig. 245. Laboratory tests for a given head are usually run when varying both the load and speed. Make the first test with the load on the Prony brake as light as possible consistent Fig. 245. — Buckets and Jet of a Pelton Wheel. with fairly steady operation of the wheel, and then take a series of tests increasing the load in increments to reduce the speed about ico revolutions per minute in each succeeding test. Duration of tests at each speed should be from twenty to thirty minutes with observations taken every two minutes. The following from may be used for tests : Test of Impulse Wheel General Data: 1. Date 2. Name of wheel and nominal horse power 3. Kind of bucket 4. No. of buckets 5. Angle of buckets 6. Diameter of bucket wheel, inches .' 7. Area of nozzle and delivery pipe 8. Coefficient of discharge for type of nozzle 9. Diameter of brake wheel, inches 10. Length of brake arm, inches 1 1 . Tare of brake, lbs 1 2 . Duration of test 13. Average temperature of water, deg. F 366 POWER PLANT TESTING 14. Average pressure by gage at wheel, lbs. per sq.in 15. Average head at wheel in feet 1 1 6. Quantity of water for total run in pounds 1 7 . Quantity of water in pounds per minute 18. Cubic feet of water per minute 19. Foot-pounds of work per minute calculated from (15) and (17) 20. R.P.M : 2 1 . Net weight on brake, lbs 2 2 . Horse power as measured by brake 23. Over-all efficiency of motor, per cent (22) -=- (19) X33, 000 1 Corrected for vertical distance from the center of the gage to the center of thenozzle. Curves. Plot a curve for each head with speed for abscissas and ''efficiency per cent for ordinates, also curves for the ratio of the velocity of the periphery of the wheel v p to the theoretical velocity due to the head v. t ; that is, v p /v t for abscissas and the maximum horse power developed for ordinates. 1 Tests of Water Turbines. A typical reaction turbine is shown in Fig. 246. The power is transmitted by the main shaft and the smaller shaft is used for controlling the gates regulating the quantity of water passing through the wheel. For testing, a Prony brake can be placed directly on the vertical shaft. In some respects a rope brake is most suitable, as one end can be attached to a spring balance and the other end can be led over a pulley, and will thus support weights on a vertical hanger. It is preferable to have the lower web of the brake- wheel solid, that is, without arms, so that it will retain the cooling water, which should be arranged to flow into it at the rate required. Power supplied is determined by the weight of water used and by the head under which the wheel operates. These quantities are determined in the same general way as for a test of an impulse wheel, already described, except in the case of a reaction turbine, where the housing or casing in which the wheel is placed is always completely filled with water (Fig. 247). ' With this arrangement the turbine receives not only the effect due to the pressure-head, measured from the level in the head- race to the center of the wheel, but also that due to the suction head, measured from the center of the wheel to the level in the tailrace. Data can be recorded in a form similar to that for 1 Plot curves showing effect of head on efficiency if several tests are - run at different heads. Hydraulic machinery 367 tests on impulse wheels. A typical runner for a reaction turbine is shown in Fig. 248. Curves. Plot a curve for each gate opening at a constant head with speed for abscissas and efficiency per cent for ordinates. Tests of Hydraulic Rams. A section of a typical hydraulic ram is shown in Fig. 249. It consists of an air chamber H, Fig. 246. — Typical Reaction Water Turbine. to which is connected the discharge pipe I. There is a check valve G opening into the air chamber from the lower chamber A into which water is brought by the pipe S. There is a waste valve at B. This valve is weighted and opens inward. By means of a nut J on the stem of this valve the lift or amount of opening of the valve can be regulated. When water is supplied 368 POWER PLANT TESTING to the ram, it escapes through the waste valve B with a velocity corresponding approximately to that due to the head under which the water is supplied. The effect of this velocity head is to reduce the pressure on the upper side of the valve so that it becomes unbalanced and closes suddenly. Then the momentum of the column of water in the pipe S becomes sufficient to open Fig. 247. — Reaction Turbine with Submerged Housing. the valve G, and will discharge some water into the discharge pipe I against a considerable head. As soon as the pressures become equalized the valve G closes, the waste valve B opens and water from the supply pipe is again " wasted." This alternate action is produced with regularity, and as a result the water in the supply pipe acquires a certain " backward and forward " wave-motion. As the rule is generally stated the length of the supply pipe leading from the reservoir to the ram HYDRAULIC MACHINERY 369 must be at least five times the head. This length is necessary to secure some resistance to this " backward and forward " wave-motion. A small air chamber shown at P, with a check- valve C opening inward to supply air, is provided in many of these rams, as it improves the efficiency. The rate of opening of the waste valve or the number of pulsations in a given time can be varied by changing the weight on its stem. This appara- FiG. 248. — Typical Runner of a Reaction Turbine, tus is tested usually by measuring the supply and the discharge heads, the weight of water discharged through the delivery pipe Wi and that passing through the waste valve w 2 in pounds per minute. Then the available energy in the water is (Wi +W2)h s , where h s is the supply head; and the useful work is Wihd, where h d is the discharge head, 1 then, Wih d Efficiency (wi +w 2 )h s (ii5) 1 Both the supply and the discharge heads must be measured, of course, from the same datum or " zero " level. 370 POWER PLANT TESTING and the capacity Q in gallons per twenty-four hours is Q = 1440 Wiq, where q is the fraction of a gallon of water in a pound. Satisfactory runs of twenty minutes' duration can usually be made, each run being made with a different lift or " stroke " of the waste valve B. Observations of the heads should be taken every five minutes if they are variable, and weighing as often as necessary, depending on the size of the tanks used. The effect on the efficiency of increasing the lift or " stroke " of the waste valve from one-eighth inch by increments of one- eighth inch is very interesting. Fig. 249. — Section of a Simple Hydraulic Ram. Curves. Plot curves with length of " stroke " as abscissas and take for ordinates: (1) Efficiency; (2) Capacity in gallons per twenty-four hours; (3) Strokes per minute. Fig. 250 shows a slightly different form of ram, as made commercially. The principle of operation is, however, the same as the one in Fig. 249. Letters used for marking the parts are the same in the two figures. Tests of Pulsometers. A type of steam pump called a pulsometer is illustrated in section in Fig. 251. In the form shown here it consists of two chambers AA, joined by tapering necks into which a ball C is fitted so as to move in the direction of the least pressure between seats in these tapering passages. HYDRAULIC MACHINERY 371 The chambers AA, on opposite sides, are connected by means of check or clack valves EE with the " induction " chamber D. Water is delivered through the passage H, which is connected to the chambers through the valves G. Between the chambers is also a vacuum chamber J, connecting them with the ' ' induc- tion " chamber D. Small air valves, moving inward, supply air to the chambers AA by opening when the pressure is less Fig. 250. — Commercial Type of Hydraulic Ram. Fig. 251. — Steam Pulsometer. than atmospheric. Its operation is explained briefly as follows: Starting with the left-hand chamber full of water and with a vacuum in the right-hand chamber, this latter chamber will fill with water which by its momentum due to rushing in suddenly , pushes back the valve C toward the left. Now during this time steam has entered the left-hand chamber to the left of the valve C (before it has shifted) and by exerting a pressure on the surface of the water it forces it through the check valve G, first into the delivery passage H and then into the air chamber J. Then the steam in this left-hand chamber condenses in contact with the cold water and forms a vacuum, permitting the repeti- 372 POWER PLANT TESTING tion of this cycle of events, except that the operations in the two chambers are reversed. Since all the steam used is condensed and discharged with the water lifted, the analysis of the operations in a pulsometer are similar to those in the familiar types of injectors, except that the steam acts in the pulsometer by pressure instead of by impact as in the injector. Using the following symbols : w s = weight of dry steam, pounds; w w = weight of water lifted, pounds; ti =temperature of the water supply, deg. Fahr. ; t 2 =temperature of the water delivered, deg. Fahr. ; r =the latent head of evaporation of the steam in B.T.U.; t s =the temperature of the steam, deg. Fahr. ; hi =the suction head, feet ; h 2 =the delivery head, feet ; hi +h 2 =the total head, feet, then w s (t s -t 2 +r)=w w (t 2 -ti) (i 16) The heat equivalent of the mechanical work done is in B.T.U., — -(w^hx + (w s +w w )h 2 ), 778 and the heat expended is in B.T.U., w s (t s -t 2 +r), and ™. 1 1^.0= • w t „hi + (w s +w w )h 2 . . Thermal Efficiency = —±, — ' ^ \. . . (117) y 778(w s (t s -t 2 +r)) v " And if we neglect the work done in lifting the condensed steam, Efficiency -^^ .... ("8) Curves. Plot with discharge pressures for abscissas curves with both thermal efficiency and capacity (gallons per twenty- four hours) as ordinates. Tests of Injectors. The injector is known particularly in stationary service as the device used for pumping water into the boiler when the feed-pump fails. One of the various forms of injectors sold commercially is shown in Fig. 252. The steam HYDRAULIC MACHINERY 373 supply, the suction or water supply, the delivery or discharge, and the overflow are marked clearly. A double-tube injector, shown similarly in section in Fig. 253, has the parts marked in the same way as in the preceding figure. Method of Operating Injectors. The method to be given, although applicable particularly to the ones described, is, how- ever, more or less generally applicable to all makes. Open wide both the steam- and water-supply (suc- tion) valves. Then close the water-supply (suction) valve slowly till the overflow ceases (for the type shown by Fig. 252) ; or (for the type of Fig. 253) pull the starting lever back a short distance until water appears at the overflow and then continue the movement steadily as far as the lever will go . Regulate the rate of delivery by closing the water-supply (suc- tion) valve. Before testing an injector or indeed even before trying to operate a and particularly the Fig. 252. — Single Tube Steam Injector. new injector, inspect the pipe fittings valves on the water-supply pipe to observe whether they are tight. It is not at all unusual to find that the valve is not air tight, and for this reason it is a very good practice to put always some new wicking in the space for packing around the stem of the valve on the water- supply pipe ; and turn up the cap over the packing tightly. Method of Testing. For the testing of injectors, the arrange- ment of apparatus consists usually of two barrels supported on platform scales, or carefully calibrated tanks fitted with gage glasses. During a test the injector draws water from one barrel 374 POWER PLANT TESTING or tank and discharges it into the other. A test of an injector must be made, of course, with established conditions; that is, with a flying start. This may be accomplished by having the injector draw water from the supply tank, but discharge water through a by-pass connection on the discharge pipe till the test is to begin. For this preliminary operation of the injector the level in the supply tank can be maintained very closely, at any point marked by manipulating a " quick-opening " valve. When the test is to begin, close as quickly as possible this valve on the pipe discharging into the supply tank and turn the dis- charge from the by-pass into the delivery tank. To make this adjustment all the valves to be operated should be of the ' ' quick- opening " type. The pressure against which the injector is to operate is secured by throttling the discharge pipe by means of a globe or an angle valve placed between the injector and the by-pass on the discharge pipe. The quick-opening valve would not be satisfactory. The suction head is measured from the middle of the injector to the average level of the water in the supply tank. The discharge head is obtained by adding to the head in feet corresponding to the pressure indicated by the gage the distance in feet from the center of the gage to a horizon- tal line through the middle of the injector. The temperatures of the water in the supply and delivery pipes must be observed. The injector is stopped at the end of the test by closing the steam valve. The following form, similar to the one used at Purdue Uni- versity, is very complete. Notes explaining the calculations required are given above : Test of an Injector Make of injector Date Number Size of connections: steam in. dia. ; water in. dia. ; discharge in. dia. ; area of discharge ( =a) sq.in. Diameter (minimum) of lifting tube in.; forcing tube in. a. Duration of test b. Steam pressure (average) pounds gage, p s c. Delivery pressure (average) pounds gage, p 2 d. Maximum pressure against which injector will discharge, /> ma x e. Suction-head (average) , feet, li l /. Delivery-head (average) , feet, /z$ g. Temperature of supply (average) t l h. Temperature of delivery (average) t 2 i. Pounds water supplied per hour, w w < HYDRAULIC MACHINERY 375 /. Pounds water and. steam delivered per hour, w m k. Cubic feet of water delivered per hour, Q „ /. Wet steam per hour, w s ( =w m —w w ) ■ ■hi. Dry steam per hour, w' s ( =xw s ) , n. Water delivered per pound wet steam, pounds ( =w w -i-w s ) o. Water delivered per pound dry steam, pounds ( = ic w -t-w' s ) p. Velocity of discharge, feet per second, v{ = 144(3-^36000) q. Energy delivered, raising injection water, B.T.U. per hour r. Energy delivered, heating injection water, B.T.U. per hour s. Energy delivered, velocity of discharge, B.T.U. per hour t. Total energy delivered, B.T.U. per hour u. Energy supplied, B.T.U. per hour v. Thermal efficiency as a boiler-feed apparatus w. Thermal efficiency as a pump x. Horse power y. Dry steam per horse power per hour, pounds The energy of raising injection water = [w w (h 1 + h 2 ) + w s h 2 ]-r- 778, B.T.U. per hour. The energy of heating injection water =w w (q 2 — q 1 ) where q x and q 2 corre- spond to t A and t 2 ., B.T.U. per hour. The energy of discharge =w m v 2 -i- (2gX778), B.T.U. per hour. The total energy delivered = item q + item r + item 5. The energy supplied = w s (xr s +q s — q 2 ) l where r s and q s correspond to p s , and q 2 corresponds to t 2 . x = quality of steam. The thermal efficiency as a boiler feed apparatus = 100 X . item v . item q + item s 1 he thermal efficiency as a pump = 100 X . item v rp, , w w (h l + h 2 )+w s h 2 I he horse power = (120) 60X33,000 The dry steam per horse power, per hour =w s ' -f-item x. 1,000,000 + item p I he pump duty = item t r,ooo,ooo[a' K: .(iii + /i 2 ) +w 8 h 2 ] 778w s (xr s +q s -q 2 ) The weight of steam found by direct weighing may be checked, by calculating (assuming radiation loss negligible) a " heat balance " in which this weight will be the only unknown, thus for this condition, (121) w s (xr s + q s - q 2 ) = -i- w w {h Y + h 2 ) + w s h 2 + (w w + w s ) V — 77 8 L 2 S] w w \h 1 + h+ 7 l8(q 2 ~q 1 ) + 7 ^] +Ww(q2-qi), or, approximately, 71&(xr v + q.-q 2 )-h2 ' ( " 9) CHAPTER XX TESTING THE STRENGTH OF MATERIALS Machines for Testing the Strength of Materials consist, in general, of (i) a power system for producing in the specimen tested the required stresses, and (2) a weighing system to deter- mine the amount of power applied. In the usual form of testing machine the load is applied to the specimen through a train of gears and screws operated either by power or by hand, depend- ing, of course, largely on the capacity. The stress is measured by balancing the force exerted on the specimen by a poise adjusted at the end of a system of levers just as weight is deter- mined with a platform scales. The general principle of most machines for testing materials is illustrated in a simple form in the apparatus for calibrating indicator springs, Fig. 99, page 107. In this case the power is applied to the hand wheel, which exerts two forces equal but opposite in direction, one compressing the spring in the indicator and the other pressing on the platform of the scales. A diagrammatic view of a typical machine for testing materials in tension and compression is shown in Fig. 255. It consists essentially of a table T, to which the upper " head " A is rigidly attached by means of the vertical bars DD. Heavy vertical screws SS, carrying the lower cross-head B are moved up or down by the system of gears GG. Moving the cross-head B downward puts a tensile stress on a test specimen s if it is attached firmly to both the upper " head " A and to the cross- head B. The force applied to the specimen is transmitted by the bars DD to the weighing table T, which rests on the first weighing lever M, having a fulcrum at F. The load on the table T is applied to the lever M at the middle of the table. The long arm of this lever is connected by means of a short link to a second lever N, and this again is connected to the short arm of the lever Q at the other end of which the weighing poise P is 376 TESTING THE STRENGTH OF MATERIALS 377 to be balanced. The position of the poise on this last lever (scale beam) indicates the force applied to the specimen s. Fig. 256 is a remarkably good illustration for showing the parts of a standard testing machine and for explaining its operation. Vertical screws SS, connected by gearing to the power system, move the cross-head B up or down according to the direction of motion. The speed of these screws is controlled and their motion reversed by manipulation of the levers marked Fig. 255. — Diagram of a Simple Machine for Testing the Strength of Materials. li, 1 2 , and 1 3 . The vertical columns supporting the upper " head " A are bolted to the table T, which rests on the system of levers M, N, 0, and Q. The poise P is moved on the lever or scale-beam Q by means of a cord connected to the hand wheel W. The levers are balanced ' ' to zero ' ' by means of the counter- poise C. Adjustment for use with long or short specimens is secured by raising or lowering the upper head A. To prevent sudden jarring of the machine when the load is released by the breaking of a specimen, vertical rods fastened to the base pass up loosely through holes in the table T ? at its four corners, and 37S POWER PLANT TESTING , on their ends large " check-nuts " are screwed. When the machine is in use these nuts must be loose, otherwise they will cause a pressure on the table causing the indication ol the scale- beam to be greater than the weight due to the load on the speci- men. Small testing machines with a capacity not exceeding 50,000 pounds are made to operate by hydraulic pressure. In this Fig. 256. — Standard Testing Machine. type of machine the movable head applying the load to the specimen is moved by the pressure on a piston in a hydraulic cylinder. This hydraulic pressure is produced usually in a small hand-operated pump at the side of the machine. Oil is generally used for the working medium. In order to return the oil to the pump from the cylinder, when the pressure is to be released, a small check valve controlled by a lever or a screw is usually provided. Machines operating hydraulically are not TESTING THE STRENGTH OF MATERIALS 379 satisfactory for large loads, because the leakage from the cylinder is likely to be excessive. When a specimen is to be tested in tension its upper end is fastened into the wedges or " jaws " in the upper " head " and its lower end is similarly gripped in the lower or movable cross- head. For tests in compression the specimen is placed between the movable cross-head and the table. Transverse loads can be applied to long wooden or metal beams with the machine shown in Fig. 256, by placing the beam between the supports or abutments UU', and applying the load by means of the movable cross-head B. Usually a special fitting with a blunt but " definable " edge to local- ize the load is inserted into the cross-head B for such tests. Extensometers. Some of the physical properties of materials are determined by the rate of deformation of the specimen as the stress is applied. To meas- ure the deformation some very accurate instruments have been devised, one of which is shown in Fig. 257. It consists essen- tially of a pair of clamps CC, fitted with sharp-pointed thumb- screws for attaching them to the specimen SS. Two rods B and B', fitted to the upper clamp on opposite sides of the specimen, are provided at their lower ends with adjustable points to be screwed up or down by means of the small milled wheels W and W. Opposite these rods and fastened to the lower clamp are two micrometer screws, usually graduated to ten-thousandths of an inch, for measuring the elongation of the specimen. Electrical connections are made, as shown, with a battery and a bell. As the specimen stretches out the contact points at P and P' are moved apart and the distance the micrometer screws are raised measures the elongation. With the help of the bell Fig. 257. — Extensometer. 380 POWER PLANT TESTING it is possible to make, for all observations, uniformly light contacts. Deflectometer. A very simple device for measuring the deflection of beams is shown in Fig. 258, consisting of a plate P supported upon a steel bar attached to the end supports UU\ Deflections can be measured with this apparatus with s a Fig. 258. — Simple Device for Measuring the Deflection of Beams. the aid of ordinary "inside " calipers, micrometer calipers, or with a special deflectometer. This instrument, illustrated in Fig. 259, is often used to measure the deflection of wooden and metal beams subjected to transverse stress. It can also be used with success to measure the contraction of short specimens in compression. Fig. 259. — Deflectometer. Physical Properties of Materials Defined. The elastic limit is a more or less definite value of the unit stress beyond which, as the stress is increased, the increase in deformation is greater relatively than the increase in stress ; and further, at this point, the deformations produced will not disappear entirely when the stress is removed. Permanent set or " set " is used to represent the lasting deformations produced by stresses greater than the elastic limit. TESTING THE STRENGTH OF MATERIALS 381 Modulus of Elasticity is a term used to express the ratio of the unit stress to the deformation per unit of length 1 accom- panying that stress, within the elastic limit. For example, if f is the stress in pounds per square inch within the elastic limit and s is the accompanying deformation per inch of length in inches, then the modulus of elasticity, in pounds per square inch, is E=f/s (122) The total stress under which a body fails is called its ultimate strength ; and the corresponding unit stress is called the ultimate unit strength, or, for short, simply the ultimate strength. The ratio of the total elongation of a body to its original length is called the percentage of elongation. It is obviously the same as the term unit deformation. For calculations of percentage of elongation, measurements are taken, according to convention, between two gage marks usually 8 inches apart. This percentage of elongation is a measure of the ductility of the material tested. The ratio of the smallest area after rupture to the original area is called the " reduction of cross-section." Resilience, often called the modulus of resilience, is a term used to represent the potential energy stored in a body ; or, from another viewpoint, it is the amount of work that can be done by the body when relieved from a state of stress. More specifically however, it is taken to mean in practice the work, in foot- pounds, done on a cubic inch of a material in stressing it to the elastic limit. For any value of the load, the resilience is equal to the product of half the stress in pounds per square inch at the elastic limit times the distortion (usually the elongation) of the test-piece in feet per inch of length up to the elastic limit, the latter term being the space passed through. Values of resilience calculated as thus defined may be checked by com- paring with the square of the unit stress at the elastic limit divided by 2 4 times the value of the modulus of elasticity (E) . Forms of Test-pieces for Tension Tests. Specimens for testing should be prepared with a great deal of care. Standard forms for round and flat bars 2 for tension tests are shown in Figs. 1 This unit deformation is often called the unit elongation for slender test-pieces, and more generally the strain. 2 See "Materials of Construction," J. B. Johnson. 382 POWER PLANT TESTING 260 and 261. On such test-pieces marks one inch apart are usually made between the limits of the so-called gage length AB, which is generally 8 inches. A standard scale similar to the one in Fig. 262 is of great assistance in marking a test-piece. At the left-hand end a percentage scale is shown from which the -Not lass than Fig. 260. — Standard Round Bar for Tension Tests. percentage of elongation, in a length of 8 inches, can be read directly. The relation between the gage length AB, and the diameter is for round sections, l=8d, and for square sections 1 =gbt. Machine work on specimens for testing should be done carefully, Piece to he same thickness as plata Fig. 261. — Standard Flat Bar for Tension Tests. , so that the material is not torn or weakened in other ways. If there is any flaw, marked irregularity or other defect in the material, the test-piece should be rejected. After, however, a test-piece has been '•' necked " and broken as shown in Fig. 263, the elongation cannot be measured very accurately with such a scale . One method is to measure the elongation from the point of rupture toward each end. In materials in which the ' ' necking ' ' effect is very marked the measured amount of elongation will vary according to the distance of the fracture n ^!r \WiVr\ 4 r\ Wrx pj Fig. 262. — Scale for Marking Test- pieces. TESTING THE STRENGTH OF MATERIALS 383 from the gage marks A and B. If the fracture is midway between these marks then nearly all the elongation will be between these marks ; but if the fracture is near one of the gage marks then a great deal of the elongation will fall outside of the marks, so that the measured elongation is too small. To correct for these discrepancies mentioned, a so-called " equiv- alent elongation " is calculated by the following method: Assume that the standard test-piece, Fig. 260, has been divided originally into 8 equal spaces between the gage marks A and B, and that the nearest number of spaces between the points of fracture (Fig. 263) and the nearer gage mark is 3 ; or in other words, that there are 3 spaces on the shorter portion of the specimen between the point of rupture and the gage mark B. Then the total length to compare with the original is to be measured on the broken test-piece from the point 5 to B on the right (corresponding to 3 spaces), plus from 4 to 5 on the right, 3C Fig. 263. — Test-piece (Round Bar) after Rupture. plus the distance from B to 4 on the left. The sum of these lengths will be the " equivalent " total, length after rupture. In general terms the method may be stated that if the length between gage marks has been divided into x equal spaces and y is the nearest number of these spaces on the shorter portion between the point of rupture and the gage mark, then mark two points M and N on the longer portion, which are y and l/2x spaces, respectively, from the fracture. Place the two portions of the specimen as closely together as possible and measure from the gage mark in the short portion to the mark M. This distance, added to double the distance from M to N, gives the required total length after rupture. In this way the elongation of the " stand- ard " length (x spaces) will be obtained, as if the fracture had occurred midway between the gage marks. Specifications are often made to require that the fracture shall be within the " middle third of the length." Detailed Method for Tension Tests. A standard test-piece to be tested in tension should be without flaws or cracks, and furthermore the material should be monogeneous. Before 384 POWER PLANT TESTING putting it into the testing machine it should be carefully meas- ured. With a scriber or scratch the marks indicating one-inch divisions should be made with the " laying-off " gage. (Fig. 262). Very light punch marks may then be made at each of the division marks accurately along the axis of the bar. 1 At each of these punch marks the diameter of the cross-section should be carefully measured with a micrometer caliper to thousandths of an inch. The outside punch marks, called the "gage marks," are often made a little heavier than the others, so that if an extensometer of the type illustrated in Fig. 257 is used, one of the thumbscrews supporting the clamps at each end can be set accurately but lightly into these marks. In this position the extensometer will measure accurately the elongation of the specimen between the gage marks. The testing machine to be used should then be balanced by adjusting the counterpoise provided for this purpose. It should be observed also that the " check nuts " rest loosely on the table. As the load is increased, however, these nuts should be screwed down a little from time to time so that if the load is suddenly removed when the test-piece breaks, or should happen to slip out of the jaws or wedges holding it in place, the jar on the machine will be very much relieved. After balancing the machine the test-piece should be placed carefully and vertically between the " jaws " or wedges in both the upper and lower heads, and the extensometer should be put in position if one is to be used. Start the machine at a low rate of speed till a load of about 1000 to 2000 pounds is indicated before taking any measurements of elongation. This first load is applied for the purpose of permitting the test-piece to assume a true central position, to allow for some slight slipping of the test-piece in the jaws before it becomes firmly gripped, and also to allow for possible irregularity in the adjustment of the exten- someter or similar auxiliary apparatus. After this light load has been reached, during which time, doubtless for the kind of materials ordinarily tested, there will have been only a very small elongation, the load should be applied continuously and 1 Punch marks can be made accurately along the axis of a round bar by putting them at the " scratch " marks made with a scriber an inch apart along the length of the test-piece on the reflection of a " beam of light " on the bright surface of the test-piece. TESTING THE STRENGTH OF MATERIALS 385 uniformly until the test-piece breaks, stopping only long enough, at the required intervals, to make the necessary observations of elongation and change of shape of the cross-section when " necking " begins. Increments of load are usually determined by taking one-tenth * the product obtained by multiplying the approximate estimated elastic limit of the material in pounds per square inch by the area of the cross-section in square inches. Near the elastic limit it is often found desirable to take observations at intervals of 500 pounds up to the yield point, where suddenly the rate of elongation increases very rapidly. By taking these observations at very close intervals at the elastic limit, data are secured which make the curves to be plotted much more satisfactory. In a standard test the stress, that is, the load, must never be decreased any appreciable amount when it is intended to apply again later an increased load. A stress once applied must be maintained or increased continuously till the end of the test. Extensometers or other apparatus of delicate construction used for measuring the elongation should be removed from the test-piece before the specimen is broken. It is customary to take off such apparatuses just after the elastic limit is reached; although, as a rule, they can be left in place relatively longer for materials that are ductile than for those that are hard and brittle. Micrometers used on extensometers provided with electrical connections are to be set for readings on one side at a time, advancing the screw device until the bell rings, indicating that the contact has been made. Then turn it back just enough to stop the ringing of the bell, and advance the screw on the other side till the bell rings again. After also turning back the micrometer screws just enough to stop the ringing, observa- tions should be taken on both micrometers. Considerable time can be saved if, while these observations are being taken, the attendant having charge of the machine is meanwhile slowly advancing the load. The scale beam should be kept " floating " at all times during a test. Never run the balancing poise out on the scale beam beyond the point necessary to balance the 1 In practice the increments of load in commercial tests are often one- half or one-third of the load at the elastic limit. For laboratory investi- gations it is not unusual to make the increments as small as one-twen- tieth of this load. 386 POWER PLANT TESTING beam. If the scale beam is carefully kept " floating " a point will be observed at from 50 to 7 5 per cent of the maximum load where the scale beam will fall, indicating apparently that it has been advanced too far. This point is called the yield point. It is defined as the stress at which the rate of elongation suddenly and rapidly increases. Beyond the elastic limit, when the extensometer has been removed, the rate of elongation can be measured with consider- able accuracy by means of a large machinist's dividers, with the points set accurately on the " gage marks " at the ends of the standard length. The load when rupture occurs is not usually the maximum. When, therefore, considerable " necking " effect is observed, the poise on the scale beam should be watched closely and when the maximum load has been reached, indicated by the falling of the scale beam, this weight should be quickly observed and recorded and then the poise should be brought back to follow the decreasing load. It will be observed that this part of the work is very interesting if it is done carefully. After the test-piece has been broken, stop the machine, remove the test-piece, clean and return the jaws or wedges to their proper places, and leave the machine in good order. The broken ends of the test-piece should be joined carefully and the length between each of the marks which were originally one inch apart should be measured. Check the sum of these lengths with the overall length between the gage marks. With a micrometer, or preferably a vernier caliper, measure as accurately as possible the diameter of the smallest area at the fracture. The fracture should be carefully examined to observe whether it is fibrous, granular or crystalline; whether coarse, fine or "silky," whether cup-shaped, half-cup, or irregular in shape. Curves and Calculations. Plot a curve with elongation per inch of length (" strain ") as abscissas and stress in pounds per square inch as ordinates. This is the familiar " stress-strain " diagram. Determine from the data obtained the modulus of elasticity (E), 1 the elastic limit, the maximum stress, the ultimate stress, 1 The modulus of elasticity can be determined also from the " stress- strain " diagram by calculating the value of the tangent for the angle between a line drawn through the origin parallel to the straight part of TESTING THE STRENGTH OF MATERIALS 387 per cent elongation in 8 inches, percentage elongation in 2 inches at the fracture, and percentage reduction in area at the fracture. Plot a curve of elongation per inch, using for abscissas 1 the original length in inches and for ordinates the elongations measured for each inch between the gage marks. Report of Tension Tests 1. Date, and names of observers 2. Material to be tested — specification 3 . Makers and brand of material 4. Length between principal punch marks, before test 5. Length between principal punch marks, after test 6. Average width of test-piece, before test (6 readings) : 7. Average width of test-piece, after test (6 readings) 8. Average thickness of test-piece, before test (9 readings) 9. Average thickness of test-piece after test (6 readings) Readings to be taken, for observations 10 and 11 for increments of 2000 lbs. on scale beam until an elongation of 0.04 in. is obtained between readings and then by increments of 1000 lbs. until elongation of 0.03 in. is obtained between readings, and finally by 500 lbs. until piece nears point of maximum strength, when readings should be taken as frequently as possible, keeping beam balanced all the time. 10. Total pull in pounds as shown by the machine 1 1 . Corresponding total elongation in piece in inches 12. Yield point, pounds total (elastic limit) 13. Maximum load, pounds total 14. Breaking load, pounds total 15. Character of fracture, description and sketch The following results are to be computed : 16. Original cross-section, sq.in 17. Final cross-section, sq.in 18. Reduction in area in sq.in., and in per cent of original area 19. Final elongation, total and per cent 20. Stress (pounds per square inch of original area) at elastic limit, maxi- mum load and breaking load. Tests in Compression. When a specimen of which the length is less than five times the smallest dimension is subjected to a load producing compression, it fails usually by crushing. Longer specimens fail usually by bending toward the side of least resist- " stress-strain " curve, reading the scales of co-ordinates, of course, in the units of stress and elongation marked on the diagram. It is obvious that the value of the tangent of the angle in this case is the unit stress divided by the unit elongation, which is, by definition, the modulus of elasticity. 1 If on the curve sheet the " inch marks " on the test-piece are indi- cated by equal divisions on the scale of abscissas, then the points showing elongation for each inch should be plotted midway between the division lines indicating the position of the " inch marks." 388 POWER PLANT TESTING ance. Two general classes of materials are frequently tested in compression: (i) Brittle materials, like brick, stone, wood, cement, cast iron, etc., which fail usually by shearing, and (2) plastic materials, like soft steel, wrought iron, copper, etc., which fail usually by a " flowing " of the metal. Because of the difficulty in measuring the deformations of short speci- mens of the plastic materials, and because the elastic limit in tension is invariably practically the same as in compression, these plastic materials are not often subjected to compressive loads. Methods to be explained here apply, therefore, par- ticularly to materials like wood, brick, stone, and cast iron. Detailed Method for Compression Tests of Short Test-pieces. Specimens of stone, cement, wood, or brick, of which the length is less than five times the smallest dimension, are usually pro- vided in forms approximately cubes, although brick and wood are as often tested in the form of parallelopipeds similar to ordinary commercial bricks. Bearing surfaces of specimens of stone, brick, and cement should be made as nearly flat and parallel as possible, and should then be covered with a thin layer of plaster of Paris. Sized paper in thin sheets should be placed on the bearing surfaces between the specimen and the plaster to prevent the absorption of water from the latter. In order to have the plaster set in a true surface the specimen is placed between the " heads " of the testing machine for about ten minutes after the movable cross-head (B, Fig. 256) has been lowered to press lightly on the plaster. For tests in compression the test-piece is placed on the table T and the load is applied by lowering the movable cross-head B. Dimensions of test-pieces must be carefully measured and recorded before they are put into the testing machine; and, if any of them require the application of plaster of Paris, then the measurements must be made before the plaster is put on. After balancing the testing machine by means of the counterweight with the test-piece on the table, apply the load continuously until the specimen is fractured ; or in the case of plastic materials till the deformation is quite noticeable. In general conduct the test in the same way as for tension, * except that the specimen 1 Measurements of the amount of compression (shortening) of the test-piece cannot be made directly, but must be made between points on the heads of the testing machine. If there is likely to be much TESTING THE STRENGTH OF MATERIALS 389 is compressed instead of being stretched. If the material tested is cast iron or even hard stone or brick, precautions must be taken to protect persons near the machine from flying fragments. If the specimen begins to spall or flake off before it breaks down, the load corresponding as well as similar information should be recorded and placed in the tabulated report under ' ' Remarks. ' ' Usually the specimen breaks down suddenly and the interior cone or pyramid in stone, brick, cement, and cast iron will be plainly seen if the load has been " fairly " applied. In the case of tests of wood, this phenomenon will not be observed, but the lines of cleavage will usually show clearly a constant angle of shearing. Detailed Method of Compression Tests of Long Pieces (Columns). When the length of a specimen to be tested in compression is greater than, at the most, ten times its least dimension, it fails invariably by bending toward the side of least resistance. The condition of the ends of such test-pieces should be as nearly as possible either fixed or perfectly free to turn. Either condition is, however, difficult to obtain. For test-pieces from 1 5 to 20 inches long usually an extensometer may be connected up to read the compression or shortening of the test-piece, if it is desired. 1 The observations will be taken in the same general way as for tension tests except that now the micrometer screws on the extensometer will approach each other, so that these screws must be turned back after taking a measure- ment by an amount greater than the compression that will be produced by the next increment of load. yielding of the parts of the machine, the moving head should be lowered till its steel " compression plate " presses on the corresponding steel block on the table or lower platform with a force of about 1000 pounds. Now measure with micrometers the distance between the points on the two heads used for compression measurements, first with the load of 1000 pounds and then with additional increments of 1000 pounds up to considerably above the breaking load of the material to be tested. From these data a correction curve should be plotted with which to correct the deflections observed when the specimen is tested. When blocks of wood are to be tested in compression, readings of the micrometers should not be taken till a pressure of 500 to 1000 pounds per square inch has been applied. This load will be required to crush the rough fibers. 1 The lateral deflection along the neutral plane is sometimes deter- mined by stretching a fine wire along the length of the specimen parallel to the neutral axis. 390 POWER PLANT TESTING Report on Compression Tests i . Date and names of observers , 2. Kind of material 3. Average thickness of test-piece (4 readings), inches 4. Average width of test-piece (4 readings), inches Machine is to be started and kept running continuously until fracture takes place, the beam being kept balanced carefully all the time. Read- ings to be taken, and calculations made therefrom as follows: In making these tests, wood and brick will be used and two pieces of each kind are to be tested, with each kind of stress. Kind of Wood White Pine. Yellow Pine. Bricks 1 2 1 2 1 Scale reading in pounds at Character of fracture, sketch. Cross-section from items 3 and 4, square inches Breaking stress lbs. per square inch for each piece Average breaking stress for each kind of wood Modulus of elasticity, lbs. per square inches Sketches, Curves, and Calculations. Sketch the character of the fracture for each specimen tested, indicating, for wood, the direction of the grain. Previously, the original shape of the specimen should have been sketched and dimensioned. Calculate the maximum unit stress. If the material was suitable for the measurement of compres- sion, plot "stress-strain" diagrams, and calculate the modulus of elasticity. Transverse Bending Tests. The most common test by transverse or cross-bending is that of a beam usually of either wood or steel, of which the coefficient of elasticity and the " elastic curve " are desired. Deflections of such beams give the data needed. Such tests may be made with a testing machine like the one shown in Fig. 256, which is provided with supporting abutments marked in the figure UU', and by insert- ing into the movable head the attachment for applying the load TESTING THE STRENGTH OF MATERIALS 391 along a line across the beam rather than on a comparatively large area as in the tests already described when the load was applied directly by the flat surface of the cross-head B. Special transverse testing machines are, however, sometimes available. A machine of this kind is illustrated in Fig. 264. In the case of a wooden beam to be tested by loading at the middle, a fine steel wire should be stretched between two pins Fig. 264. — Machine for Transverse Tests. located as accurately as possible above the points of support and on the line of intersection of the neutral plane with the side of the beam. The wire should be fastened to one of these pins and allowed to hang over the other, being kept taut by means of a weight . attached to the free end, Fig. 265. A steel scale, preferably highly polished so that it will show the image of the wire, should be attached in any suitable way to the side of the 392 POWER PLANT TESTING beam, so that the edge along which the scale readings to be observed are marked will be exactly half-way between the two supports. The beam must be protected from indentation by the knife-edges by small bearing plates. The load should be applied centrally in increments to give approximately 7 V inch deflections to the elastic limit, and beyond to give deflections of approximately £j inch. If it can be done successively, the deflections should be read without stopping the test; unless, of course, the permanent set is to be determined, when after each increment, the beam must be released from its load. Fig. 265. — Device for Measuring the Deflection of a Wooden Beam. Curves and Calculations. Plot a curve taking the load applied in pounds for abscissas and deflections in inches for ordinates. Sketch the character of the fracture. Calculate the modulus of elasticity, 1 the modulus of rupture, and the stress in the outer fiber at the elastic limit from the curve. 1 The modulus of elasticity is calculated by the formula - 4 is «> The modulus of rupture from w„lc '«=7P • • (I24) and the stress in the outer fiber at the elastic limit by Wplc f -=rr (,25) "whenw e =load at the elastic limit in pounds per square inch; w M =load at the point of rupture in pounds per square inch; 1 = length of beam (span) in inches ; d = deflection in inches ; c = distance from the neutral axis to the outer fiber in inches; I=moment of inertia, inches. TESTING THE STRENGTH OF MATERIALS 393 Torsion Tests are made to determine the strength of a material to resist twisting forces. A typical machine for such tests is illustrated in Fig. 266. It consists in its essential parts of the frame FF', the " jaw " heads A and B for gripping the rod R to be tested, and the system of weighing levers on which the poise P is balanced. The load is applied to the rod R by power through the gears shown connected to the head B. The power is applied by means of the pulley and gears shown at the right-hand side, or may be applied by hand power by turning hand wheels. With hand power usually more satisfactory results can be obtained than with power applied Fig. 266. — Riehle Torsion Testing Machine. mechanically, because the rate of twisting can be more closely regulated. The amount of twist or the angular deformation is indicated by index-arms connected to opposite ends of the test-piece. An autographic torsion testing machine operated by hand power by means of the crank is sometimes used. The move- ment of the crank tends to rotate the test-piece which at the opposite end of the machine is fastened to a pendulum carrying a heavy bob. The resistance of the pendulum and its weight measure the power applied, which is equal to the length of the lever arm times the sine of the angle of inclination times the constant weight of the bob. 394 POWER PLANT TESTING Tests are made usually by increasing the twisting moment by increments of about 200 inch-pounds, 1 measuring for each increment the torsion angle. Curves and Calculations. Plot a curve, using torsion angle for abscissas and twisting moment for ordinates. Calculate from this curve the' unit stresses 2 (shearing) at the elastic limit, at the point of rupture, and also the maximum value of stress. Determine the torsion angle at the elastic limit and at the point of rupture, the helix angles, 3 and the modulus of elasticity for torsion. 4 Impact Tests. Materials are tested by impact, usually by striking a test-piece with a weight allowed to fall upon it. Metals used in the manufacture of machinery and in railroad construction where it is likely to be subjected to shocks and blows are in many cases tested to determine the effect of the impact due to a blow. Some testing machines for such tests are made like a pile- driver with the weight dropping vertically from a sort of gallows upon the test-piece. A more common form, however, of such machines consists of a pendulum provided with a heavy bob 1 If tests are made of large sections of high-grade material, like, for example, a shaft of nickel steel for a students' class, it is expensive to break many specimens, so that for this reason the twisting moment pro- ducing a maximum stress just inside the elastic limit is computed before making the test, and this value is not to be exceeded. 2 The unit stress is calculated with the formula: M / f s = V c > or f s = Mc/l 2 „ .... (126) where M is the torsional moment in inch-pounds, c is distance in inches from the neutral axis to the extreme fiber, and I p is the polar movement of inertia. When c=r (the radius) as in the case of a cylindrical test- piece, I p = ^tzt*. 3 Torsion produces a peculiar arrangement of the outer fibers in the form of helices, as observed in broken test-pieces. Each one of these fibers makes an angle with its original position equal to its angular dis- tortion a. Any particle on the surface is also moved through an angle /?, having its vertex in the axis and in a plane perpendicular to the axis. Now if we neglect the functions of small angles, we can write approxi- mately la =r/?, where 1 is the effective length of the test-piece and r is the radius. The helix-angle a =r/?/l. * The modulus of elasticity in torsion ("modulus of rigidity"), Es=f s -4- a, as above, then K.--. .......... daw TESTING THE STRENGTH OF MATERIALS 395 intended for delivering a blow on the middle of a test-piece in the shape of a bar, preferably of a rectangular section, held on two knife-edge supports attached to a heavy bedplate. Such machines are particularly designed for comparative tests of cast iron. They are provided with an arc concentric with the move- ment of the bob of the pendulum, graduated to read the verti- cal fall of the bob in feet. A tripping device is attached to the side of the graduated arc for permitting the bob to be supported and then dropped from any height within the limits of the machine. Since the deflection is very small, a device is usually supplied for magnifying it, and by means of a pencil-point traveling over a chart an autographic record is made of the deflections for each blow delivered by the bob. With such instruments the rebound of the test-piece and its permanent set must be carefully excluded from the measured deflection. One way to do this is to draw a " zero line " with the test- piece in place but before a blow is struck. Deflections and permanent set will then be measured on one side of this line and " rebounds " on the other. To determine the center load to be applied that will be equivalent to the impact, the following symbols are used: Let Wi = weight of the bob in pounds; h = the vertical distance it falls, in feet ; w 2 = the equivalent maximum center load, in pounds, and d = the deflection in feet, then Wih = ^w 2 d, 2W 5 h (128) With this value of w 2 the usual properties of the material may be calculated by formulas (123), (124) and (125), page 392. Cement Tests. Cements are tested usually for tensile and crushing strength, for fineness, and for the time required for " setting." Tests for crushing strength (compression) are usually made by crushing cubical blocks in a testing machine designed for general tension and compression tests (see page 378) . For tensile tests, however, special machines, designed particularly for testing cement, are generally used. Because of the nature of the material it is absolutely necessary that the power is 396 POWER PLANT TESTING applied in the line of the axis of the test-piece and also with steadiness and in increments as uniform as possible. There is a standard size and shape for test-pieces of cement, and they must be made in a certain prescribed way in order that different tests may be compared. The standard briquette for testing Fig. 267. — Standard Specimen for Cement. (one square inch section) is shown in Fig. 267, and Fig. 268, shows moulds suitable for making the test-pieces or briquettes, as they are called. Cement-testing machines are invariably provided with moulds for making standard briquettes. These moulds are divided Fig. 268. — Cement Moulds and Briquettes. along a longitudinal center line into two halves which, when placed together, fit closely and are held in place by means of pins, or dowels, preventing endwise movement, and by clamps pressing together the two parts of the mould. The strength TESTING THE STRENGTH OF MATERIALS 397 of the briquettes is affected by the time allowed for hardening, the amount of water used, and by the method of mixing the cement. (See page 400.) Power is applied in the automatic cement-testing machine shown in Fig. 269, by shot dropped from a cylindrical hopper into a pail supported on a scales. The briquette of cement Fig. 269. — Automatic Cement Testing Machine. being tested is held between two shackles or " holders " con- nected to a hand wheel used to regulate the distance between the shackles. When a briquette breaks the scale beam drops and closes automatically a valve, stopping the delivery of shot into the pail. The operation of the machine may be described briefly as follows : Hang the hopper on the hook as shown and put enough shot into it to balance the counterpoise. Now 398 POWER PLANT TESTING put the briquette into the shackles and adjust the hand wheel so that the scale beam will rise nearly to the stop. When the valve is opened shot will begin to fall into the pail. The deliv- ery of the shot into the pail should be slow. When the bri- quette has broken, the scale beam has dropped and the valve Fig. 2' -Hand-operated Cement T has been closed. The weight of shot collected in the pail shows the number of pounds required to break the briquette. A non-automatic type of cement-testing machine is illus- trated in Fig. 270. In this machine the power is applied by moving a hand wheel operating by means of a screw the system of levers transmitting the load to the briquette in the TESTING THE STRENGTH OF MATERIALS 399 shackles. The tension produced by this load may obviously be balanced by weights applied to the scale beam above. In order to secure a very uniform and slow movement of the poise it is carried along the scale beam by a cord moved by a small crank. In applying the load the hand wheel should be moved as slowly and uniformly as possible to avoid a jerking motion. In the figure one of the standard briquette moulds, and a tray used for immersing the briquettes in water after they have set are shown. Fig. 271 shows the proper position of the briquette in the supporting shackles. Test of Cement for Fineness is made by determining the amount by weight of a given sample that will not pass through sieves with meshes of a standard size. The American Society of Civil Engineers recommends the use of sieves of 2500, 5476 and 10,000 meshes per square inch. Sieves with approxi- mately these meshings are known as Nos. 50, 80, and 100. A weighed sample of cement is first passed through No. 50 sieve and the weight of that remaining in the sieve is recorded. That passing through is then put into the next sieve (No. 80) and the residue in this sieve is likewise weighed, while that passing through goes to the finest sieve (No. ico). Results of this test for fineness are expressed by the per- FlG centages that the various residues remaining in the sieves are of the original weight of the sample. Unless cement is ground as fine as flour it has very little " binding power." The coarse particles are nearly as inert for " cementing " as sand. Test of Cement for Time of Setting is made by mixing on a slab of glass circular pats about 3 inches in diameter and one- half inch thick. Then when a blunt needle (one-twelfth inch in diameter as the point) and loaded with a weight of one- quarter pound " ceases to penetrate the entire mass setting is said to have begun." Similarly when a needle one twenty- fourth inch in diameter loaded with a weight of one pound ceases to penetrate at all, setting is said to be ended. Making Briquettes. Neat Cement. Before filling moulds with cement they should be carefully cleaned and rubbed on the 400 POWER PLANT TESTING inside surface with a rag saturated with kerosene. Then clamp the parts together, placing the moulds preferably on a large slab of glass, or a similar material providing a plane surface which will not absorb moisture from the briquettes. A quantity of cement, enough to fill several moulds, should now be mixed with water (usually not as much as 25 per cent by weight) to make a rather stiff but very plastic and homogeneous mixture. It should be " worked " for at least three minutes. Put this mixture into moulds, pressing it down firmly, especially around the sides, preferably with the thumbs, in order that the pressure exerted will not be excessive. Level the surfaces of the briquettes with a trowel, being careful that in this operation too much is not taken from the middle so as to make the briquettes of unequal thickness. Usually cement briquettes are not tested in tension until several days after they have been made, and in that case the moulds with the briquettes in them should be allowed to remain on a slab which will not absorb moisture from them for twenty-four hours, in an atmosphere which is not very dry. After that time the briquettes should be carefully removed from the moulds, put into a suitable tray, like the one shown in Fig. 270, and immersed in a tank containing water. 1 Near one of the ends, each briquette should be marked with significant numbers or letters, so that the maker and the date of making will not be confused. The cement must not, of course, be permitted to begin to set before it is put into the moulds. Mortar Briquettes for testing are made usually of five, two, or three parts of sand to one part of cement. The sand to be used is to be approximately of such fineness that it will all pass through a No. 20 sieve and is all held on a No. 30 sieve. Mortar briquettes for standard tests are made of pure crushed quartz of the kind used in the manufacture of sandpaper. For mortar briquettes less water will be needed than for those of neat (" pure ") cement. The mortar briquettes should be worked with the trowel for at least four minutes. Moulds should be well filled with the mortar, which should then be pressed down to a flat surface with the trowel. The briquettes are to be set aside and later immersed in water as specified for those of neat cement. 1 The water in this tank should be changed once in seven days, and should be kept at normal " room " temperatures. TESTING THE STRENGTH OF MATERIALS 401 For laboratory tests, usually after seven days, or else after twenty-eight days, the briquettes are taken from the water and placed in a cement-testing machine for testing in tension. Apply the load at as uniform a rate as possible, without jerks, and if a non-automatic machine is used the scale beam must be kept " floating " all the time, so that when the briquette breaks the correct load will be indicated by the position of the poise on the scale beam. In some laboratories it is recommended that pieces of thin rubber bands be inserted between the edges of the shack- les or " holders " and the briquette, as in this way it is claimed there is greater certainty of having it break at the middle, that is, in the smallest section where the area is supposedly exactly one square inch. Data regarding tests of neat cement and mortar briquettes should be tabulated in a form similar to the following : Form for Cement Tests i. Date, and names of observers. 2. Name and kind of cement. 3. Makers and location of plant. 4. Distinguishing mark on briquettes. 5. Date of mixing and time. 6. Temperature of room at time of mixing, deg. F. 7. Temperature of water at time of mixing, deg. F. 8. Conditions of setting: as to time briquettes were left in dampened air before immersion, etc. 9. Activity of the cement or time of initial and final setting. 1 10. Fineness of grinding. Kind of Briquette. Neat. Sand. 12 Composition of Briquettes. Per Cent of Cement Per Cent of Water Per Cent of Cement Per Cent of Water Per Cent of San 1. . No. of Briquettes. 7 Day. 28 Day. Remarks 7 Day. 28 Day, 1 — 3 4 S — 7 — 1 2 3 4 S — 7 8 is 16 17 18 Time of test. Breaking Strength, lbs. Appearance of frac- ture — give sketch of each here. Temp, of Room at time of test. deg.F. In reporting the results of these experiments it is important that the effect of different percentages of water, sand, etc., and time of immersion, be fully discussed. 1 The time between the end mixing and the successful resistance to penetration of the needle ■}., inch in diameter with \ pound weight is called the "time of initial setting." Similarly, the time elapsing between the end of mixing and the resistance to penetration of the needle = V inch in diameter with 1 pound weight is called the " time of final setting." APPENDIX Two sets of tables which the author has found useful for "rough and ready" calculations are given on the following pages. Table I is a short table of the more important prop- erties of saturated steam. This table has been taken with permission from Allen and Bursley's Heat Engines. Table II gives for various common substances the specific gravity, density (weight per cubic foot), specific heat and coefficients of expansion per degree Fahrenheit, both linear and volumetric. Coefficient of volumetric expansion is three times the linear coefficient of expansion. TABLE I Properties of Saturated Steam english units go,. h 1! 13 14 1h' B 2s sa]h ill fjfc *Q n ■3*8 Eh° O £± •<* V t h L H V s p .0886 32 1072.6 1072.6 3301.0 .000303 .0886 .2562 60 28.1 1057.4 1085.5 1207.5 .000828 .2562 .5056 80 48.1 1046.6 1094.7 635.4 .001573 .5056 1 101.8 69.8 1034.6 1104.4 333.00 .00300 1 2 126.1 94.1 1021.4 1115.5 173.30 .00577 2 3 141.5 109.5 1012.3 1121.8 118.50 .00845 3 4 153.0 120.9 1005.6 1126.5 90.50 .01106 4 5 162.3 130.2 1000.2 1130.4 73.33 .01364 5 6 170.1 138.0 995.7 1133.7 61.89 .01616 6 7 176.8 144.8 991.6 1136.4 53.58 .01867 7 8 182.9 150.8 988.0 1138.8 47.27 .02115 8 403 404 POWER PLANT TESTING Properties op Saturated Steam english units Continued w °" d to 3 c 3* IS n [So ■p s o a >-s Jo g a "30Q g-s $£ 6 to Ha 0h 3 p p t h L H V 1 p 9 188.3 156.3 984.8 1141.1 42.36 .02361 9 10 193.2 161.2 981.7 1142.9 38.38 .02606 10 11 197.7 165.8 978.9 1144.7 35.10 .02849 11 12 202.0 170.0 976.3 1146.3 32.38 .03089 12 13 205.9 173.9 973.9 1147.8 30.04 .03329 13 14 209.6 177.6 971.6 1149.2 28.02 .03568 14 14.7 212.0 180.1 970.0 1150.1 26.79 .03733 14.7 15 213.0 181.1 969.4 1150.5 26.27 .03806 15 16 216.3 184.5 967.3 1151.8 24.77 .04042 16 17 219.4 187.7 965.3 1153.0 23.38 .04277 17 18 222.4 190.6 963.4 1154.0 22.16 .04512 18 19 225.2 193.5' 961.5 1155.0 21.07 .04746 19 20 228.0 196.2 959.7 1155.9 20.08 .04980 20 21 230.6 198.9 958.0 1156.9 19.18 .05213 21 22 233.1 201.4 956.4 1157.8 18.37 .05445 22 23 235.5 203.9 954.8 1158.7 17.62 .05676 23 24 237.8 206.2 953.2 1159.4 16.93 .05907 24 25 240.1 208.5 951.7 ' 1160.2 16.30 .0614 25 26 242.2 210.7 950.3 1161.0 15.71 .0636 26 27 244.4 212.8 948.9 1161.7 15.18 .0659 27 28 246.4 214.9 947.5 1162.4 14.67 .0682 28 29 248.4 217.0 946.1 1163.1 14.19 .0705 29 30 250.3 218.9 944.8 1163.7 13.74 .0728 30 31 252.2 220.8 943.5 1164.3 13.32 .0751 31 32 254.1 222.7 942.2 1164.9 12.93 .0773 32 33 255.8 224.5 941.0 1165.5 12.57 .0795 33 34 257.6 226.3 939.8 1166.1 12.22 .0818 34 35 259.3 228.0 938.6 1166.6 11.89 .0841 35 36 261.0 229.7 937.4 1167.1 11.58 .0863 36 37 262.6 231.4 936.3 1167.7 11.29 .0886 37 38 264.2 233.0 935.2 1168.2 11.01 .0908 38 39 265.8 234.6 934.1 1168.7 10.74 .0931 39 40 267.3 236.2 933.0 1169.2 10.49 .0953 40 41 268.7 237.7 931.9 1169.6 10.25 .0976 41 42 270.2 239.2 930.9 1170.1 10.02 .0998 42 43 271.7 240.6 929.9 1170.5 9.80 .1020 43 44 273.1 242.1 928.9 1171.0 9.59 .1043 44 45 274.5 243.5 927.9 1171.4 9.39 .1065 45 46 275.8 244.9 926.9 1171.8 9.20 .1087 46 APPENDIX 405 Properties op Saturated Steam — Continued ENGLISH UNITS IX "o 3 +> cr ffl *s o la Kg H ° II s ! m> 6 Sid 1!« P * h L H V 1 P 47 277.2. 246.2 926.0 1172.2 9.02 .1109 47 48 278.5 247.6 925.0 1172.6 8.84 .1131 48 49 279.8 248.9 924.1 1173.0 8.67 .1153 49 50 281.0 250.2 923.2 1173.4 8.51 .1175 50 51 282.3 251.5 922.3 1173.8 8.35 .1197 51 52 283.5 252.8 921.4 1174.2 8.20 .1219 52 53 284.7 254.0 920.5 1174.5 8.05 .1241 53 54 285.9 255.2 919.6 1174.8 7.91 .1263 54 55 287.1 256.4 918.7 1175.1 7.78 .1285 . 55 56 288.2 257.6 917.9 1175.5 7.65 .1307 56 57 289.4 258.8 917.1 1175.9 7.52 .1329 57 58 290.5 259.9 916.2 1176.1 7.40 .1351 58 59 291.6 261.1 915.4 1176.5 7.28 .1373 59 60 292.7 262.2 914.6 1176.8 ' 7.17 .1394 60 61 293.8 263.3 913.8 1177.1 7.06 .1416 61 62 294.9 264.4 913.0 1177.4 6.95 .1438 62 63 295.9 265.5 912.2 1177.7 6.85 .1460 63 64 297.0 266.5 911.5 1178.0 6.75 .1482 64 65 298.0 267.6 910.7 1178.3 6.65 .1503 65 66 299.0 268.6 910.0 1178.6 6.56 .1525 66 67 300.0 269.7 909.2 1178.9 6.47 .1547 67 68 301.0 270.7 908.4 1179.1 6.38 .1569 68 69 302.0 271.7 907.7 1179.4 6.29 .1591 69 70 302.9 272.7 906.9 1179.6 6.20 .1612 70 71 303.9 273.7 906.2 1179.9 6.12 .1634 71 72 304.8 274.6 905.5 1180.1 6.04 .1656 72 73 305.8 275.6 904.8 1180.4 5.96 .1678 73 74 306.7 276.6 904.1 1180.7 5.89 .1699 74 75 307.6 277.5 903.4 1180.9 5.81 .1721 75 76 308.5 278.5 902.7 1181.2 5.74 .1743 76 77 309.4 279.4 902.1 1181.5 5.67 .1764 77 78 310.3 280.3 901.4 1181.7 5.60 .1786 78 79 311.2 281.2 900.7 1181.9 5.54 .1808 79 80 312.0 282.1 900.1 1182.2 5.47 .1829 80 81 312.9 283.0 899.4 1182.4 5.41 .1851 81 82 313.8 283.8 898.8 1182.6 5.34 .1873 82 83 314.6 284.7 898.1 1182.8 5.28 .1894 83 84 315.4 285.6 897.5 1183.1 5.22 .1915 84 85 316.3 286.4 896.9 1183.3 5.16 .1937 85 406 POWER PLANT TESTING Properties of Saturated Steam - english units ■ Continued fig o- ■2P4 (B £.2" fH H c3 oj CD h ^ a ci •p a o a >•$ f 3 o la I" 8 as]h m> 6 p « h L H V V p 86 317.1 287.3 896.2 1183.5 5.10 .1959 86 87 317.9 288.1 895.6 1183.7 5.05 .1980 87 88 318.7 288.9 895.0 1183.9 5.00 .2002 88 89 319.5 289.8 894.3 1184.1 4.94 .2024 89 90 320.3 290.6 893.7 1184.3 4.89 .2045 90 91 321.1 291.4 893.1 1184.5 4.84 .2066 91 92 321.8 292.2 892.5 1184.7 4.79 .2088 92 93 322.6 293.0 891.9 1184.9 4.74 .2110 93 94 323.4 293.8 891.3 1185.1 4.69 .2131 94 95 324.1 294.5 890.7 1185.2 4.65 .2152 95 96 324.9 295.3 890.1 1185.4 4.60 .2173 96 97 325.6 296.1 889.5 1185.6 4.56 .2194 97 98 326.4 296.8 889.0 1185.8 4.51 .2215 98 99 327.1 297.6 " 888.4 1186.0 4.47 .2237 99 100 327.8 298.4 887.8 1186.2 4.430 .2257 100 101 328.6 299.1 887.2 1186.3 4.389 .2278 101 102 329.3 299.8 886.7 1186.5 4.349 .2299 102 103 330.0 300.6 886.1 1186.7 4.309 .2321 103 104 330.7 301.3 885.6 1186.9 4.270 .2342 104 105 331.4 302.0 885.0 1187.0 4.231' .2364 105 106 332.0 302.7 884.5 1187.2 4.193 .2385 106 107 332.7 303.4 883.9 1187.3 4.156 .2407 107 108 333.4 304.1 883.4 1187.5 4.119 .2428 108 109 334.1 304.8 882.8 1187.6 4.082 .2450 109 110 334.8 305.5 882.3 1187.8 4.047 .2472 110 111 335.4 306.2 881.8 1188.0 4.012 .2493 111 112 336.1 306.9 881.2 1188.1 3.977 .2514 112 113 336.8 307.6 880.7 1188.3 3.944 .2535 113 114 337.4 308.3 880.2 1188.5 3.911 .2557 114 114.7 337.9 308.8 879.8 1188.6 3.888 .2572 114.7 115 338.1 309.0 879.7 1188.7 3 878 .2578 115 116 338.7 309.6 879.2 1188.8 3.846 .2600 116 117 339.4 310.3 878.7 1189.0 3.815 .2621 117 118 340.0 311.0 878.2 1189.2 3.784 .2642 118 119 340.6 311.7 877.6 1189.3 3.754 .2663 119 120 341.3 312.3 877.1 1189.4 3.725 .2684 120 121 341.9 313.0 876.6 1189.6 3.696 .2706 121 122 342.5 313.6 876.1 1189.7 3.667 .2727 122 123 343.2 314.3 875.6 1189.9 3.638 .2749 123 APPENDIX 407 Properties of Saturated Steam english units Continued h § CO ^ -1 §3 « 03 tfi ° - " a a la r°"o QgO ^ I s K £ o H o Pi ^ V t h L H V D V 124 343.8 314.9 875.1 1190.0 3.610 .2770 124 125 344.4 315.5 874.6 1190.1 3.582 .2792 125 126 345.0 316.2 874.1 1190.3 3.555 .2813 126 127 345.6 316.8 873.7 1190.5 3.529 .2834 127 128 346.2 317.4 873.2 1190.6 3.503 .2855 128 129 346.8 318.0 872.7 1190.7 3.477 .2876 129 130 347.4 318.6 872.2 1190.8 3.452 .2897 130 131 348.0 319.3 871.7 1191.0 3.427 .2918 131 132 348.5 319.9 871.2 1191.1 3.402 .2939 132 133 349.1 320.5 870.8 1191.3 3.378 .2960 133 134 349.7 321.0 870.4 1191.4 3.354 .2981 134 135 350.3 321.6 869.9 1191.5 3.331 .3002 135 136 350.8 322.2 869.4 1191.6 3.308 .3023 136 137 351.4 322.8 868.9 1191.7 3.285 .3044 137 138 352.0 323.4 868.4 1191.8 3.263 .3065 138 139 352.5 324.0 868.0 1192.0 3.241 .3086 139 140 353.1 324.5 867.6 1192.1 3.219 .3107 140 141 353.6 325.1 867.1 1192.2 3.198 .3128 141 142 354.2 325.7 866.6 1192.3 3.176 .3149 142 143 354.7 326.3 866.2 1192.5 3.155 .3170 143 144 355.3 326.8 865.8 1192.6 3.134 .3191 144 145 355.8 327.4 865.3 1192.7 3.113 .3212 145 146 356.3 327.9 864.9 1192.8 3.093 .3233 146 147 356.9 328.5 864.4 1192.9 3.073 .3254 147 148 357.4 329.0 864.0 1193.0 3.053 .3275 148 149 357.9 329.6 863.5 1193.1 3.033 .3297 149 150 358.5 330.1 863.1 1193.2 3.013 .3319 150 152 359.5 331.2 862.3 1193.5 2.975 .3361 152 154 360.5 332.3 861.4 1193.7 2.939 .3403 154 156 361.6 333.4 860.5 1193.9 2.903 .3445 156 158 362.6 334.4 859.7 1194.1 2.868 .3487 158 160 363.6 335.5 858.8 1194.3 2.834 .3529 160 162 364.6 336.6 858.0 1194.6 2.801 .3570 162 164 365.6 337.6 857.2 1194.8 2.768 .3613 164 166 366.5 338.6 856.4 1195.0 2.736 .3655 166 168 367.5 339.6 855.5 1195.1 • 2.705 .3697 168 170 368.5 340.6 854.7 1195.3 2.674 .3739 170 172 369.4 341.6 . 853.9 1195.5 2.644 .3782 172 174 370.4 342.5 853.1 1195.6 2.615 .3824 174 176 371.3 343.5 852.3 1195.8 2.587 .3865 176 i08 POWER PLANT TESTING Properties of Saturated Steam english units Concluded a> § ■ Kg r <5 ^"d "§3i£ 2 pgd $* J 5 K ►3 o H ° o p-i 3* V t h L H V i V 178 372.2 344.5 851.5 1196.0 2.560 .3907 178 180 373.1 345.4 850.8 1196.2 2.532 .3949 180 182 374.0 346.4 850.0 1196.4 2.506 .3990 182 184 374.9 347.4 849.3 1196.7 2.480 .4032 184 186 375.8 348.3 848.5 1196.8 2.455 .4074 186 188 376.7 349.2 847.7 1196.9 2.430 .4115 188 190 377.6 350.1 847.0 1197.1 2.406 .4157 190 192 378.5 351.0 846.2 1197.2 2.381 .4200 192 194 379.3 351.9 845.5 1197.4 2.358 .4242 194 196 380.2 352.8 844.8 1197.6 2.335 .4284 196 198 381.0 353.7 844.0 1197.7 2.312 .4326 - 198 200 381.9 354.6 843.3 1197.9 2.289 .4370 200 202 382.7 355.5 842.6 1198.1 2.268 .4411 202 204 383.5 356.4 841.9 1198.3 2.246 .4452 204 206 384.4 357.2 841.2 1198.4 2.226 .4493 206 208 385.2 358.1 840.5 1198.6 2.206 .4534 208 210 386.0 358.9 839.8 1198.7 2.186 .4575 210 212 386.8 359.8 839.1 1198.9 2.166 .4618 212 214 387.6 360.6 838.4 1199.0 2.147 .4660 214 216 388.4 361.4 837.7 1199.1 2.127 .4700 216 218 389.1 362.3 837.0 1199.3 2.108 .4744 218 220 389.9 363.1 836.4 1199.5 2.090 .4787 220 222 390.7 363.9 835.7 1199.6 2.072 .4829 222 224 391.5 364.7 835.0 1199.7 2.054 .4870 224 226 392.2 365.5 834.3 1199.8 2.037 .4910 226 228 393.0 366.3 833.7 1200.0 2.020 .4950 228 230 393.8 367.1 833.0 1200.1 2.003 .4992 230 232 394.5 367.9 832.3 1200.2 1.987 .503 232 234 395.2 368.6 831.7 1200.3 1.970 .507 234 236 396.0 369.4 831.0 1200.4 1.954 .511 236 238 396.7 370.2 830.4 1200.6 1.938 .516 238 240 397.4 371.0 829.8 1200.8 1.923 .520 240 242 398.2 371.7 829.2 1200.9 1.907 .524 242 244 398.9 372.5 828.5 1201.0 1.892 .528 244 246 399.6 373.3 827.8 1201.1 1.877 .532 246 248 400.3 374.0 827.2 1201.2 1.862 .537 248 250 401.1 374.7 " 826.6 1201.3 1.848 .541 250 275 409.6 383.7 819.0 1202.7 1.684 .594 275 300 417.5 392.0 811.8 1203.8 1.547 .647 300 350 431.9 407.4 798.5 1205.9 1.330 .750 350 APPENDIX 409 TABLE II Properties of Common Substances Specific Gravity. Weight per Cubic Foot, Lbs. Weight of One Cubic Inch, Lb. Specific Heat. Coefficient of Expan- sion per Deg. F. Volu- metric. Aluminum Bismuth Brass Copper Coal (anthracite) . . Coke Gasoline Glass Gold Ice (at. 32° F.) . . . . Iron (cast) Iron (wrought) . . . Lead Limestone Mercury (at 32° F.) Cement Nickel Platinum Pine (white) Silver Steel Tin Zinc 2.60 9.82 8.10 8.79 1.43 1.00 .68 2.89 19.26 .92 7.5 7.74 11.35 3.16 13.60 2.24 8.90 21.5 .55 10.47 7.83 7.29 7.19 161. 613. 503. 545. 88.7 62.4 42.4 180.7 1200. 57.5 465. 582. 708. 197. 849. 140. 547. 1342. 34. 653. 486. 452. 445. .095 .353 .293 .318 .058 .037 .105 .697 .033 .271 .280 .411 .114 .492 .083 .321 .779 .020 .379 .292 .264 .260 .212 .031 .094 .092 .241 .203 .198 .032 .504 .130 .114 .031 .217 .033 .20 .109 .032 .65 .056 .116 .056 .095 .000011 .000008 .00001 .000009 .000005 .000008 .000006 .000007 .000016 .000033 .000008 .000007 .000005 .0000025 .000011 .000007 .000012 .000016 .000033 .000024 .00003 .000028 .000014 .000024 .000018 .000021 .000048 .000100 .000024 .000020 .000015 .000008 .000033 .000020 .000035 .000048 INDEX PAGE Absorption Refrigerating Machines 345 Air Compressors, Testing of 334 _ 336 Air Engines 350 Air, Flow of 142 Air " Horse Power. 33o Air, Velocity and Volume of 143, 330 Air Supplied to Furnace, Determination of, from Flue Gases 194 Alden Brake Dynamometer I2 8 Allen-Moyer Gas Apparatus i 9I Alternating Current, Measurement of 282, 294 Ammonia, Leakage of 342 Ammonia, Properties of, Tables 341-342 Ammonia Refrigerating Plants 339 Amsler Planimeter 66 Analysis of Coal, Proximate 1 78, 2 1 1 Analysis of Flue Gases 181, 212 Analysis of Flue Gases, Typical Examples of 187 Anemometers !47 Areas, Measurement of 66, 120 Ash in Fuel, Determination of ! 80 Aspirator ! 83 Atwater's Fuel Calorimeter ^7 Available Energy 266 Averager (Planimeter) , Amsler 66 Averager (Planimeter), Coffin ^5 Bachelder Indicator n c Back-firing Indicator Diagrams ^7 Balance-sheet of Gas Engine -, I . Balance-sheet of Boiler 214, 223 Balance-sheet of Steam Engine 257 Barraclough & Mark's Engine Tests 274 Belts, Testing Tension in ■, r . Bending (Transverse) Tests 300 Blowers, Testing of , 2 s Boiler Balance-sheet 214, 223 411 412 INDEX PAGE Boiler Efficiency ' 213, 221 Boiler Feed-pumps, Testing of 356, 358 Boiler Feed- water, Measurement of, in Tests 205 Boiler Heat Balance 214, 223 Boiler Horse Power 201 Boiler Summary Sheets 215 Boiler Test Log Sheet ." 202 Boiler Testing : Datum Lines of Water and Fire in Tests 207 Boiler Testing, Principal Objects of 200 Boiler Testing, Rules for (A. S. M. E.) 202 Bourdon Pressure Gage, Theory of 5,6 Bourdon Pressure Gage with Steel Tube 7 Brake Dynamometers 122-129 Brake Horse Power 126, 306 Brake Pulley, Design for 127 Briquettes of Neat Cement and Mortar for Testing 399-401 Bristol-Durand Integrator for Circular Diagrams 11, 80 Brumbo's Pulley 115 Burning Point of Oils, How Tested 311 Calibration of Pressure Gages 12 " Vacuum " 21 " Indicator Springs 104 Thermo-Electric Thermometers and Pyrometers 38 " Mercury Thermometers 26 Calorific Value of Fuel, Determination of 162, 177 Calorific Value of Gas . 1 74 Calorimeter, Barrel (for Steam) 62 ' ' , Bomb (for Fuels) ,. 163 ' ' , Combined Separating and Throttling 57~6o ' ' , Condensing (Steam) 46, 62 ' ' , Calibration of *. 65 ' ' , Electric 61 ' ' , Fuel 162 ' ' , Junkers 1 74 ' ' , Separating (Steam) 56 Throttling 46 ' ' , Wire-drawing (or Throttling) 46 ' ' Charts for Determining Moisture in Steam 50, 54 Nipples 48 Capillary Corrections for Mercury 4 Carbonic Acid (C0 2 ) Apparatus for Determining 188, 191, 196 Carbonic Acid Refrigerating Machine 340- Cement Tests 395 Centrifugal Fans, Testing of 325 ' ' Pumps, Testing of 362 Chemical Analysis of Fuels to Determine Calorific Value 177; INDEX 413 PAGE Chimney Gases, Weight of, Calculated from Analysis 225 Clearance of an Engine Cylinder, Determination of 240 Coal Analysis 178 Coal, Calorific Value of, from Analysis 177 Coals Recommended for Standard Boiler Trials (A. S. M. E.) 204 Coal Testing: Proximate Analysis 178 Coal Calorimeters 162 Coefficient of Dilution 192 Discharge for Orifices 158 Weirs 161 Expansion of Mercury 19 Coefficients of Expansion of Various Substances Appendix " of Friction of Friction Wheels 355 Coffin Planimeter 75 Columns, Testing of 389 Combined Separating and Throttling Calorimeters 57 Commercial Steam Engine Testing 244 ' ' Turbine Testing 281 Compression Tests of Materials 387 Compressors, Air, Testing of 334 ' ' , Ammonia, Testing of 346 Condensers, Testing of 243 Continuous Indicator 99 Conversion of Pressures 3 ' ' Temperatures and Heat Units 26 Cooley-Hill Continuous Indicator 99 Cooley Indicator Spring Tester 105 Cooley Stroke Measuring Counter 361 Corliss Engine Diagrams, Normal and Abnormal 235, 251 Corliss Engine Valve-setting 234 Correcting Steam Engine and Steam Turbine Tests to Standard Con- ditions 296 Correction Curves for a Steam Turbine 297 Correction for Stem Exposure of Mercury Thermometers 31 Crosby Indicator 88 Curve, Hyperbolic as Applied to Indicator Diagram 250 ' ' , Typical ' ' Error " 21 Curves for Determinations of Moisture in Steam 50, 54 Cut-off, How Determined 250 Dead-center to Set Engine On 231 Dead-weight Gage Testers 13 Deflectometers 380 Density of Water at Different Temperatures 4 ' ' Ammonia (Liquid) 34 1 ' ' Substances, Table of Appendix Diagram Factor of Engines : 252 414 INDEX PAGE Differential Dynamometer, Calibration of 137 Direct Current, Measurement of 293 Draft Gages 23 Ducts, Loss of Velocity in 333 " , Measuring Velocity of Air in 143, 330 Durand-Bristol Integrator 1 1 , 80 Duty of a Steam Pump 357 Dynamometer, Alden 128 , Differential 134 , Dynamo 131 , Emerson "Scales" 133 , Flather 140 , Hints on Management 124-125 , Rope and Strap 126 , Water Brake : 128 , Webber Transmission 136 Eccentric, Setting of, Effect on Indicator Diagram 235 Econometer 198 Economy of Steam Engine Compared with Ideal 253 Efficiency of Steam Boiler 213, 221 ' ' Fans or Blowers 330 Gas Engine 310, 314 ' ' Gas Producer 324 Refrigerating Machines 343 Steam Engines 259 " " Turbines 286 ' ' Compared with Rankine Cycle 287 Ejector for Flue Gases 183 Elastic Limit Defined 380 Electric Dynamometers 131 Electrical Measurement of Power 293 ' ' Instruments, Precautions to be Observed 294 ' ' Pyrometers 36 Emerson Fuel Calorimeter 168 ' ' Power Scales 138 Engine and Boiler Tests 245 Engine Lubricators 276 Engine Test, Balance-sheet 259 Entropy Defined 262 Entropy Diagram of Rankine Cycle 267 Entropy-temperature Diagrams 262-268 Extensometer 379 Fans, Ventilating 325 Feed- water, Measurement of 242, 246 ' ' , Thermometer and Gage 44 INDEX 415 PAGE Feed Pumps, Testing of 356, 358 Flashing Point of Oil, How Determined : 311 Flather's Dynamometer 140 Fliegner's Formula 334 Flow of Air 143-148, 334 ' ' Steam 148 " Water 150 Flue Gas Analysis 181 Flue Gas, Determination of Air Supply from . 194 Flue Gases, Loss of Heat in 214, 225 ' ' , Weight of 225 Form for Report of Boiler Test 215 Gas Engine Test 311 Steam Engine Test 254 " Pump Test 360 Francis Formula for Weirs 160 Friction Brakes and Dynamometers 122 ' ' Horse Power , 229 ' ' Wheels, Tests of 355 Fuel: Calculation of Heating Value 166, 177 Fuel Calorimeters 163 Fuel for Gas and Oil Engines, Measurement of 308, 310 Fuels, Calorific Value of t 162 Fuel Testing 162, 178 Gages, Bourdon 6 , Calibration of 12 , Diaphragm , 8 , Pressure 6 , Recording 10 , Vacuum 2,9 Gage Notch 1 59 Gage Testers 13 Gas, Calorific Value of 1 74 ' ' , Measurement of 142, 308 Gas Engine Balance Sheet 314, 324 " , Efficiency of 310, 314 ' ' Fuel, Measurement 310, 312 " Indicator Diagrams: Normal, "Suction," etc. 316 showing "Timing" of Ignition 318 Gas Fuels 318 ' ' Meters 142 ' ' Producers 310 Gasoline, Measurement of 308, 310 Goss Dynamometer 133 " Guarantee " Tests 286 416 INDEX Head at a Pump (Suction, Discharge, Total), Denned 357 Heat Balance of Boiler 214, 223 Gas Engine 314 Steam Engine 257 Refrigerating Plant 347 Heat Units, Conversion of 26 Heating Value of Fuels Calculated from Analysis 177 by Experiment 162 Heat Unit Basis of Engine Testing 246, 260 Hoists, Efficiency of 353 Hook Gage , 1 59 Horse Power, Boiler 201 , Brake 126 , Indicated 118, 247, 306 Hot-air Engine, Testing of . : 351 Humidity of Air 331 Hydraulic Machinery, Testing of 356 Hydraulic Motors, Testing of 363 ' ' Rams, Testing of 367 Hyperbolic Curve Applied to Engine Indicator Diagrams 250 Impellers of Fans 327 Impulse Water Wheels .• 363 Indicated Horse Power, Calculation of 118, 247, 306 Engine Constant 121 Indicator, Bachelder 95 Care of 97 Continuous 99 Crosby 88 Crosby Gas Engine 308 High-pressure (Ordnance) , 335 Optical 101 " Star Brass " 93 Tabor 93 Thompson 85 Watt 85 Indicator Diagrams, Analysis of 249 , Calculation of Steam Consumption from 272 from Flather's Dynanometer 141 of Gas and Oil Engines 315 showing Back Firing 317 of Suction Stroke 316 Taken with Light Spring Attachment 316 Indicator Testing Apparatus 112 Spring Testing 105 ' ' , Calibration of 104 Injector, Method of Operating • 373 INDEX 417 TAGE Injector, Testing 372 Test, Form for Report on 374 " Used in Boiler Testing, Correction Applied to Feed-water 205 Integating Instruments, Durand-Bristol 80 , Planimeters . .' 66-78 "Internal" Horse Power 286 Junkers Gas Calorimeter 1 74 Latent Heat of Ammonia 342 Steam, Table of Appendix Leakage Test of a Boiler 242 ' ' of Steam in Tests 241 Light Spring Indicator Diagrams of Suction Stroke of Gas Engine 316 Log Form for Indicator Spring Test 1 1 1 ' ' for Mechanical Efficiency Test of Engine 230 ' ' for Pressure Gage Test 20 ' ' for Thermometer Calibration 28, 31 Losses of Head in Ducts 333 Low Heat Value of Gas 176 Lubricators, Engine 276 Mahler Bomb Calorimeter 164 Machines for Testing Strength of Materials 376-398 Manometers 1-4 Mean Ordinate, Determination of 78, 120 ' ' Effective Pressure by Coffin Planimeter 78 Mechanical Pyrometer 38 Efficiency '. 229, 307, 330, 335 Mercury Columns, Cleaning of 4 ' ' , Corrections for 4 ' ' and Equivalent Pressure per Unit Area 3 ' ' Column for Calibrating Gages 18, 22 ' ' , Expansion of 19 and Appendix Meter, Gas 143 ' ' , Venturi 155 Modulus of Elasticity 381 Moisture in Coal 178 Steam (by Charts) 50, 54 ' ' , Determination of 46 Mortar, Testing of . 400 Moulds for Cement Briquettes 396 Napier's Formula ; . 148 Oil Engines, Measurement of Fuel for 308 "Orsat" Apparatus 188 418 INDEX PACE Optical Pyrometers 39 Pantograph Reducing Motion for Indicators _ 115 Parallel Rule for Dividing Diagrams 120 Parr Calorimeters 169 Pendulum Reducing Motions 113 Permanent "Set" Denned 380 Perry Optical Indicator 102 Pitot Tubes 143, 330 Planimeter, Amsler 66 Calibration of 79 ' ' , Coffin 75 ' ' , Polar, Theory of 67 , Roller 78 Positive Pressure Blowers 328 Power, Measurement of 122 Power Scales, Emerson 138 Pressure (lbs. per square inch) and Equivalent Head of Water or of Air . 3 Pressure and Temperature of Steam, Table of Append'x Pressure Type of Gas Producer 319 Pressure Gages > 5 j> " . , for Measuring Draft 23 " , Calibration of . . . 12, 18 ' , Recording 10 Pressure Gage Tester, Dead-weight 13 Pressure Scales, Crosby 17 Prony Brake ...... • - 122 Proximate Analysis of Coal !.............. 1 78 Pulsometer, Testing of 370 Pump, Centrifugal, Testing of 359 Pumping Engine Trials 357, 360 Pumps, Effective Head at 357 ' ' , Testing of Feed 356, 358 Pyrometer, Calibration of 38 ' ' , Calorimetric 41 Cones 43 ' ' , Electric Resistance 37 ' ' , Mechanical 38 ' ' , Mercury 38 ' ' , Optical . . . ! 39 5 ' , Radiation 40 " , Recording 38 ' ' , Thermo-electric 36 Quality of Steam, How Calculated 49 ' ' Determined from Charts 50, 54 INDEX 419 PAGE Radiation Loss in Calorimeters 55, 162, 166, 173 Rankine Cycle Steam Engines and Turbines 253, 261, 287 Ratio of Expansion 252 Reaction Water Turbines 366 Recording CO2 Apparatus. 196 ' ' Gages 10 ' ' Thermometers 34 ' ' Pyrometers 38 Reducing Motions for Indicators 1 13-1 18 Refrigerating Plants 337 Capacity 344 Report of Boiler Test, Forms for 215 ' ' Gas Engine Test, Forms for 311 ' ' Steam Engine Test, Forms for 254 Resilience 381 Rider Hot Air Engine 350 Rope Brake 126 ' ' , Hints on Management of (footnote) 125 Rope Drives, Tension in 354 Rotation Losses in Turbines 289 Rules for Boiler Testing (A. S. M. E.) 202 Gas Engine Testing (A. S. M. E.) 310 ' ' Steam Engine Testing (A. S. M. E.) 241 Sampling Bottle for Flue Gases 182 Sampling Coal . ., 210 Tubes for Flue Gas (A. S. M. E.) 184 Scales for Weighing Fuel . 206, 309 Seger Pyrometer Cones 43 Separating Calorimeter 56 " Set " (Permanent) Defined 380 Siphons for Steam Gages 7 Sirocco Fans 326 Slip in Pumps 357 Smoke Observations 212 Specific Gravities of Various Substances, Table of Appendix Specific Heat of Ammonia 342 " Various Substances, Table of Appendix " Superheated Steam 270 ' ' Volume of Steam Appendix Speed-output Curves 281, 288 Spring Tester, Indicator 104 Standard Conditions for Ventilating Fans (U. S. Navy) 333 Engine and Turbine Tests 296 ' ' Gases 176, 324 Steam,. Flow of 148 Steam Calorimeters 4 6_6 5 420 INDEX PAGE Steam Consumption Calculated from Indicator Diagram 272 Determined from Feed- water 242 when Using Surface Condenser 24^ Calculated from Heat Balance 285 Steam Engine Balance-sheet 257 Lubricators 276 Testing, Rules for 241 Thermal Efficiency of 259 Steam Measurement 242 " Tables of Properties of 29, Appendix Stem Exposure of -Thermometers, Correction for 31 Stroke-measuring Counter 361 Suction Gas Producer, Testing of 319 " Stroke Diagrams of a Gas Engine 316 " Head of Pump, Measurement of with Gage (footnote) 360 Superheated Ammonia 341 Steam, Flow of 149 " Specific Heat of 270 Tabor Indicator , 93 Temperature-entropy Diagrams 253 Temperature, Measurement of 25 Scales, Conversion of 26 Tension Tests of Materials 379, 382 Testing Boilers 200 " Gas Engines 306 " Hydraulic Motors 356 " Impulse Water Wheels _ 363 " Refrigerating Machines 337, 341 " Steam Engines 229 " " Pumps 360 " " Turbines 279 " Strength of Materials 376 " Ventilating Fans and Blowers ^. . . . 328 " Water Turbines 366 Test-pieces, Standard Shapes and Sizes for 382, 396 Thermal Efficiency of a Boiler 221 " Gas Engine , 310, 314 " " Steam Engine 259 " " " Turbine 286 Thermo-electric Pyrometers and Thermometers 36 Thermometer, Alcohol 26 " and Pressure Gage Combined , 44 " Calibration of 26, 51 " Correction for Stem Exposure 31 " for Flue Gases 36 " High Temperatures (footnote), ... 29, 38 INDEX 421 PAGE Thermometer, Mercury 25 ' ' , Recording 34 with Mercury Well for Steam Pipes 44 " , Regraduating of (footnote) 25 " , Standard 26 " , Thermo-electric 36 , Wet and Dry Bulb 331 Thompson Indicator 85 Throttling Calorimeters 46 Timing of Ignition 318 Torque (footnote) 122 " of Steam Turbine 289 Total Heat of Saturated Steam, Table of Appendix " " Superheated Steam 270 Trammels, Method of, for Setting Engine on Dead Center 231 Transmission Dynamometers 135-140 Transverse Bending Tests 390 Turbine Dynamometer, Westinghouse 130 Ultimate Strength Defined 381 Vacuum Gages 2-3, 9 on Suction Pipes of Pumps (footnote) 360 Valve Setting, D-slide and Piston Types . 230 , Corliss Type 234 Velocity of Air 143, 330 Ventilating Fans 325 ' ' Systems, Testing of 332 Venturi Water Meter . . . . _ 155 Volatile Matter in Coal 179 Volume of Air Discharged by a Blower 330 " of a Pound of Steam, Table of Appendix Volumetric Efficiency of Refrigerating Machine 343 Water Brake 128 ' ' Cooled Brake Pulley 127 ' ' Equivalent of Calorimeters 64, 162 ' ' Flow through Circular Orifice or Nozzle 156 ' ' Friction Dynamometer, Westinghouse 130 ' ' Measurement of by Weir 159 ' ' Measuring Tank, Continuous 152 ' ' Meters 150 ' ' Meter, Venturi 155 * ' Rate Curve 281 * ' Turbines 366 " ' , Weight of at Different Temperatures 4 ' ' Wheels, Testing of . . 363 422 INDEX PAGE Watt's Indicator. 85 Weak Spring Indicator Diagrams 316 Webber's Transmission Dynamometer 136 Weighing Machine, for Water 152. Weight of Air Required to Burn a Pound of Fuel 194, 225, 228 Weight of Air, Table of 145 ' ' Chimney Gases 225 ' ' Flue Gases 225 ' ' a Cubic Foot of Steam, Table of Appendix ' ' Various Substances, Table of Appendix Water at Different Temperatures 4 Westinghouse Water Brake 130 Wet and Dry Bulb Thermometer for Humidity 331 Willan's Law 274 Lines 274, 276, 283, 309 Wire-drawing Calorimeters 46 SEP 14 1911