SHRINKAGE Of WOOD IN SHIP CONSTRUCTION December 1942 y. S. DEPT. OF AGRfCULft LIBRARY ATLANTA SRAhtfcM way em ATLANTA, QEOHOiA UNJV. OF FL LIB DOCUMENTS DEPT. U.S. DEPOSITORY NC. 11424 UNITED STATES DEPARTMENT OF AGRICULTURE FOREST SERVICE FOREST PRODUCTS LABORATORY Madison, Wisconsin In Cooperation with the University of Wisconsin SHRINKAGE OE -JOOD III SHIP CONSTRUCTION By L. V. TEESEALE, Senior Engineer Wood, like many other materials, shrinks as it loses moisture and swells as it absorbs moisture. While green, freshly cut logs may contain water ranging in quantity from 30 to 300 percent, "based on the weight of the oven-dry wood, the removal of only the last 25 or 30 percent of this moisture content has the effect of shrinking the ^ood upon drying; and since wood in service is never totally dry, the possible shrinkage effect falls within a relatively narrow range of moisture content . Water is held in the wood in two distinct ways — as imbibed water in the walls of the wood cells, and as free water in the cell canities. When wood "begins to dry, the free water leaves first, followed "by the imbibed water. The fiber-saturation point is that stage in the drying or wetting of wood at which the cell walls are saturated and the cell cavities are free from water. Eor most woods, this point is between 25 and 30 percent moisture content. The dimensions of wood change only when the moisture content drops below the fiber-saturation point. Since in seasoning green wood, the surface dries more rapidly than the interior and reaches the fiber-saturation point first, shrinkage may start near the surface of a piece while the average moisture content of the board or timber is still considerably above the fiber-satura- tion point. Wood shrinks most in the direction of the annual growth rings (tangentially) , about one-half to two-thirds as much across these rings (radially), and very little, as a rule, along the grain (longitudinally). The combined effects of radial and tangential shrinkage on the shape of various sections in drying from the green condition are illustrated in fig- ure 1. When a board or portion of a board is cross-grained, the lengthwise shrinkage resulting from a combination of crosswise and longitudinal shrink- age is greater than that in a straight-grained piece. Shrinkage is usually expressed as a percentage of the green dimensions, which represent the natural size of the piece. Table 1 gives the range in shrinkage in different directions for most of the commercially important native species. Mimeo. R1U2U -1- Table 1. — Range in average shrinkage of a number of native species of wood . Direction of shrinkage : From green : to oven-dry : condition From green to air-dry con- dition (12 to 13 percent moisture content) : Percent of Percent of : green size green size : 1+.3 to 1^ 2.1 to 7 Radial : 2 to g.5 1 to 1+.2 : .1 to .2 .05 to .1 : 7 to 21 3.5 to 10.5 Shrinkage in drying is proportional to the moisture lost below the fiber- saturation point. Approximately one-quarter of the total shrinkage possible has occurred in wood seasoned to a moisture content of 18 to 20 percent, about one-half when the moisture content is 12 to 13 percent, and about three-fourths in lumber kiln dried to a moisture content of about 6 to 7 percent. Vood is a hygroscopic substance; that is, it lias • the.' property of absorbing or giving off moisture according to the conditions of the surround- ing atmosphere. T iJhen wood is subjected to a constant temperature and rela- tive humidity, it will in time come to a definite moisture content determined by the prevailing humidity, which is called the equilibrium moisture content. This relationship between the moisture content of wood and the surrounding atmospheric conditions is shown on figure 2. Atmospheric temperatures are constantly changing, both during a single day and with the seasons. The relative humidity also fluctuates. The rate of exchange of moisture betxireen wood and atmosphere is, however comparatively slow, and the equilibrium moisture content is based on the average humidity prevailing over an ex- tended period rather than on changes during very short periods, even though fluctuations may then be through a wide range. Thickness of the wood is also a factor, the moisture content of thin material reacting to atmospheric changes more rapidly than thick material. Surface coatings retard the rate of exchange of moisture between the wood and the atmosphere. All wood used in ships above the water line, as well as interior parts not exposed to water, null in time attain a moisture content in equilibrium with the atmospheric conditions surrounding • it . This moisture content may vary more or less according to changes in atmospheric conditions and exposure to wetting, but will in general remain consistently below the fiber-saturation point. Sxterior planking below the water line that is exposed to the vessel's interior atmosphere will have a moisture content on the inner face that is below the fiber-saturation point, while its outer face when in contact with the water will be at or above the f iber-saturation point. Deck boats and other types of small water craft that are kept out of the water most of the time will attain a moisture content consistently below the fiber-saturation point, probably as low as 10 percent at times. Mimeo. R1U2U -2- Hence, if wood intended for use in "boats is adequately seasoned and installed at a moisture content in accord with its service conditions, there is every prospect of satisfactory performance without serious changes in size cr dis- tortion of section. On the other hand, if green or partially seasoned material is used under conditions such that further drying will take place after installation, some shrinkage may he expected. In general, the heavier snecies of wood shrink more across the grain than the lighter ones. Heavier wood of the same species also shrinks in this direc- tion more than lighter wood. T ifhen freedom from shrinkage is a more important requirement than hardness or strength, as, for example, in the planking of small boats, a lightweight species should be chosen. When it is important to combine hardness or strength with low shrinkage, such as in the dase of treenails, some species like black locust is to be preferred. Average tan- gential, radial, and volumetric shrinkages for individual domestic species when dried from the green condition to various moisture-content values are given in table 2. The average tangential, radial, and volumetric shrinkage for a limited number of tropical species when dried from the green to an oven-dry condition is given in table 3. This table is included for compara- tive purposes only, since very few tropical species are available to United States beat builders due to wartime conditions. Theoretically, and for practical purposes, the normal moisture content- shrinkage relation may be considered a direct one, from zero shrinkage at the fiber-saturation point to maximum shrinkage at zero moisture content. Actually, however, the relationship is similar in boards to the curves in figure 3« The curves represent average values, and the shrinkage of an individual board may, of course, be above or below the average amount indi- cated. Changes in moisture content in seasoned wood, such as those caused by seasonal variations in relative humidity, produce changes in dimension that are proportional to the moisture-content changes. For example, assume that a piece of flat-sawed southern yellow pine sheathing at 12 percent moisture content loses 5 percent of its moisture. The shrinkage curve (marked "tangential") indicates that the shrinkage in width from the green con- dition to 7 percent moisture content would be 5 percent, and that from the green condition to 12 percent moisture content would be 3-l/ 2 percent. The difference of 1-1/2 percent is the shrinkage in the width of the board due to the 5-"°e r cent loss in moisture. These curves represent average values, and the shrinkage of an individual board may be somewhat below or above the indicated amount, Moisture pickup below the fiber-saturation point causes expansion or swelling just as moisture loss causes shrinkage. For instance, reversing the above example, a pickup of 5 percent in moisture content from 7 to 12 percent would cause a flat -sawed board to swell about 1-1/2 percent of the green width. Where swelling is restrained during a cycle of moisture pickup, such as in the case of tightly fitted planking or decking, the surfaces tend to buckle or become deformed or slightly crushed at the edges. Upon redrying to the Mimeo. RlU2*+ -3- 1 Z 3 4 5 6 7 8 9 10 SHRINKAGE (PER CENT OF GREEN DIMENSION) Tigure 3. — Typical moisture-shrinkage curves. These curves are for Douglas-fir and southern yellow pine and may "be used for estimating the amount of change in dimension that will take place with change in the moisture content of the wood. ZM-22048-F figure U. — Changes in curred wood aaabar caused by shrinkage and swelling. (Hot drawn to scale.) original moisture content, such pieces will shrink and assume a dimension somewhat less than before swelling started, and the joint will open. Re- peated cycles of swelling and shrinking under pressure in the course of time cause the piece to "become narrower. Tight recaulking when the joints are opened "by shrinkage will cause more rapid loss of dimension than would occur ' if the caulking is tapped in lightly. Moreover, seam composition should "be of a type that will remain soft in service so that it can expand and contract as the wood shrinks and swells. Since the swelling and shrinkage are propor- tional to the width of the piece, it follows that the narrower the piece is the less the joint will open or close. With proper caulking, the joint can take up some of the expansion "before the forces developed tend to deform the contact surfaces of the itfood; hence, on the basis of shrinkage effects, narrow material gives better service than wide material. Exposed decking, particularly if unpainted, is subject to very severe exposure and rapid moisture changes, hence, the width used is generally not more than one-third greater than the thickness, and often less. Planking contributes materially to the strength of the vessel and, within certain limits, the wider the plank is the greater are' its strength properties. On the other hand, to prevent trouble from expansion, particularly above the water line, ■ the width must be limited. Sir.ce the exposure is less severe than for decking and the material is invariably well protected with paint to minimize moisture changes, such planking can safely be wider than decking, and generally, the width is equal to three or four times its thickness. Lifeboats, deck boats, utility boats, or similar types that are out of the water the greater part of the time should be constructed of material having . a moisture content that is at or slightly below that which they will attain" in service. This mean? a moisture content generally not less than 10 percent and net more than 18 percent. To prevent rapid moisture changes, such boats should be well painted, inside and out. Any further protection from sun and rain, such as boat covers, will minimize moisture changes and lengthen the useful life of the boat. When not in service, such boats should be supported at least a foot above the ground and should preferably be stored in open, well ventilated, unheatea sheds, or covered to protect them from the sun. Effects of Change of Moisture Content on Bent '»*ood Members The checks that develop in bent members, such as ribs, frames, and similar parts, are largely the result of unequal drying but the stresses set up because of tne bending accentuate the checks to some degree. Heat in com- bination with moisture tends to plasticize wood, hence stock intended for bending is heated with hot water or steam immediately before bending. The contained heat causes rapid evaporation from the surface, thus setting up the first factor favorable for checking. As the drying progresses after the stock has cooled, any checks present may.be extended. Checking may be min- imized by using stock having a moisture content of 15 to 20 percent. If the moisture content is 12 percent or less, the breakage during bending is likely to be more frequent than if more moisture were present. "imeo. R1U2U Seme change of shape of "bent mefbers that are not restrained .nay be expected with a change of moisture content. Wh.eth.er restrained or not, such parts develop complex interior stresses because of moisture changes and resulting shrinkage. IPor example, assure the frame to be semicircular. "'.Tien loss of moisture occurs the radial dimension, being across the grain, tends to de- crease, which in turn leads to decreases in the length of the inner or outer arc, or both. The upset fibers in the inner arc also tend to shrink while there is little or no tendency for the fibers in the outer arc to shrink. Consequently, if the piece is not restrained the curvature tends tc increase or "close the bend" (fig. ty). If it is restrained so that no change in curvature can take place, a tension stress is developed across the grain (radially) with compression stress in the outer arc and tension stress in the inner arc. . • . Conversely, an increase in moisture content results in radial swelling- and welling of the upset fibers of the inner arc, tending to "open" or decrease the curvature. If the piece is restrained from movement, the stresses de- veloped are the reverse of those- caused by shrinkage. Breakage of bent members after the "bending operation ! — often after they have been installed in the boat — is in some cases caused by shrinkage stresses or a. combination of such stresses with some weakness in the piece, such as localized diagonal grain, a concealed defect, check or split, or a partial undetected break that occurred in "bending. Shrinkage of Pl ywood Forroal straight-grained wood has relatively high strength and sli ht shrink- age properties in the direction of the grain.. Across the grain, however, the shrinkage is relatively high and the strength property is low. In plywood with the grain of adjacent plies perpendicular, the lateral shrinkage of ad- jacent plies is almost completely restrained, and the length and width of ply-root 3 panels .--re hence only slightly affected by moisture content changes. The shrinkage in the thickness of the plies, however, is unopposed, hence the panel will shrink in thickness just as does normal wood. The shrinkage of a plywood panel in the t^o lateral directions will be the sum of the longitudinal shrinkage and the longitudinal compression assumed, by the plies. m his shrinkage will vary with the species, the ratio of ply thicknesses, the number of plies, the character of the grain, and the combi- nation of species. The average shrinkage obtained from several hundred tests on a variety of combinations of species and. thicknesses in bringing 3-ply wood from the soaked to the oven-dry condition was about 0.^5 percent paral- lel to the face grain and O.67 percent perpendicular to the face grain. The species included in the tests were basswood, birch, black '-'a.lnut , chestnut, elm, mahogany, Spanish cedar, spruce, sugar maple, sweetgum, tupelo, and yellowpoplar . Mimeo. Rlkzh -5- Prom this it is seen that the shrinkage of plywood is only about one-tenth as great as that across the grain of an ordinary board. The total lateral shrinkage of a l-l/2-inch southern yellow pine board with two l/l2-inch sweet gum face veneers was only 1 percent, or about one-seventh of the norma shrinkage. The values given for shrinkage are based on a moisture content ranging from a green or soaked condition to an oven-dry condition. In serv ice the change in moisture content will be much less, generally not more than enough to cause a dimensional change of one-fourth to one-half of that represented in the tests. Mimeo. Rl^ -6- Table 2.-- Shrinkage values for commercially Important woods grown In the United States Shrinkage (percent of dlmensl green) t Dm green Oven dried to percent molstur (test values) Percent : Percent i Percent : Percent I Percent : Percent : Percent : Percent : Percent : Percent : Percent : Percent Beeot.Amei BlrohS. . . rloan. Birch, paper Butternut Cedar: Alaska. yellow. Eastern red. ■ . Inosnsa Northern white- . Port Orford white— Atlantic white Western red- Cherry, blaok Chestnut Cottonwood: Eaetern Northern blaok Cypress, southern. *.■ Douglas-fir: Coast region •Inland Empire* region Rooky Mountain region Eln: Blaok Sweet Tupelo Haokberry Hemlock: Eastern Western Hlokory: Peoanl Trues Honey loouet Larch, western Locust, blaok Magnolia: Cuoumbertrae Southern.... Mahogany, West Indies ■ Maple: Blgleaf Black Red. Silver. Loblolly Lodgepole Longleaf Eastern white- Red Ponderoea Bhortleaf Sugar Western white Poplar, yellow Redwood Spruce: Eaeternii Engelmann Sitka Sycamore , American Tamarack Walnut, blaok : 1.2 : 2.0 1-9 i l!o : 2.0 : 1.6 : l'l : 1.3 : 2.8 : 1.7 2.2 : l.i : 2.2 .8 : .7 : 1-5 , - : !s : .8 1.2 i'l 1.2 : lJ ! 1,7 : .7 : .6 : i'l 1.2 : .0 : .8 1.8 1-7 : 1.0 : 2-3 1:1 • 1*0 • :' 1.2 1 2.0 : 1.0 : i'i I .9 : 1.6 2.1* : 1.? : z'.o : 1.? i 2.2 ' 7 1 1.6 : '.& : lis 2.1 : 1.1 : 1-9 ! 1:1 : 2.5 : 1.2 : 2^2 .8 1.7 1 1-1 : 2.0 : 1.2 : 2.2 : 1.8 : 1:1 : 1.0 : 2.0 1.1 : l.y : l.J : 2.2 : 1.5 : 1.6 : - 9 1 1.2 ! .9 i l.g : 1.2 : I'l : 1.0 : 2.0 .8 : 1.8 2. •) : 1.1 : 2.2 : l."» : 2-? : 1.2 : 1.8 : 1.1 : 1.7 ! M ! : 1.2 : : 1.0 : ; \i \ 1:1 : 1.0 : 1.8 1.8 ; M | 1.1 : 1.1 .8 : a : 1.1 : 1-9 u i3i l.g been taken a£ [ 1:1 n 1:1 2.6 n n Vc 5-7 hi l.i 3.8 5.8 3- u 6.6 3-1 li 7-9 5.0 8.2 3-8 2:1 8.2 tl ?:? 5-1 2-5 5.6 5.6 2.1 1.5 3.5 •5 3.8 n • 9 3 5 • 5 1:1 • 2 2.1 • 9 3.8 1.8 .« b 2.8 •3 2.6 .0 5-2 2.9 2.7 2.8 1:2 5-3 3-S n I:i Li 3-7 6.7 lis 2.1 2.5 5.0 u 7-0 u 6.9 1:1 1:1 n l:] : 5.8 6.0 2.2 3-2 1:1 6.8 B *-9 3-7 n 3-2 3-3 tl 1:1 5-2 6.8 6.2 3.8 1:1 2.6 6.6 5-0 3.6 5.8 6.o 7.5 2.8 3.6 3.0 2.2 3-7 u n 7> 8.0 1:1 6.8 7.0 11.8 12.2 12.2 12.2 7.6 6.9 5.8 5-7 i; 1:1 8.7 10.6 9-3 7-9 io. u 10. k u 9 :f 10.4 1:1 10.2 l li W 10.2 9-2 5.a 8.7 1:i 7.8 u i?: 8.9 11.5 6.6 8.1 6.9 ilh... shrinkage v&lu laet 3 columns of thle table. 2Th..e shrinkage values have been taken laet 3 crOumns of thle table. ^Theee shrinkage values have been taken the last 3 columns of this table. "Average of Blltmore white ash, blue ash, green -fourth the shrinkage to the oven-dry condition ae given one-half the shrinkage to the oven-dry condition ce given in three-fourths the shrinkage to the oven-dry condition as glv ■•Average of sweet bl "Average of lowland yello ■Average of blgl hlokory, f shagbark h utmeg hlokory, water hickory, and pecan, kory, mockemut hickory, pignut hickory, k, pin oak, red oak, scarlet oak, souths Average of black oak, laurel oak, pin oak, r and willow oak. 10 Average of bur oak, chsstnut oak, post oak, swamp cheetnut Average of black sprues, red spruce, and whits apruoe. Z U 411732 F , swamp red T ab 1 e 3 • — Tropical woods — percent directional and volume shrinkage from green to oven-dry condition . (Excerpt from "Tropical "'. T oodc , " Volume 7L Ellwood S. Harrar, Buke University.) September U 19^2, by Spec ies : Radial : Tangential : Volume Aboudikro (Ivory Coast) Entanc 1 rciVTa- 0, la c Tr lin.cricum : 5.61 : 9. ^ : 15.69 Allacede (Phil. Is.) llac ec dendron cel^bicum i u.6s : 6.97 : 12. Us Almon (Fhil. In.) Shcrea eximia • • 6 99 7 f)9 l*s . 5U Amaranth (Trop. Aner.) Pelt o^'vie "^anici 1 " 1 ata . . . «. • ~*> 7S i 5. SO 10. 17 Amarellc (Brazil) PI a t hvmeni a reticulata • £ 07 6 71 • i qLl Andiroba ("rop. Aner.) TlQ P M 1 SkY PpCl Q • s hU £ ?7 6. c:_5 Araca/ (Brazil) Terminalia af 4 *. januarensis : 7 01 U. SQ 10 . 6U Avcdire' ( T,r . Africa} : k. 07 19 10. 6U Ayous (W. Africa) Triplochiton sclerox^lon • 2.U9 5. 11 7 sh Blackbean, Australian Castanos'oermum australe s 2.76 7.0U 10. 07 Bosse' ("West Africa) Guar e a ced.rata : ^ SO s 96 9 37 3ox> f ood, Indian Buxus s em ~ ervir ens . . . : ^.0^ IO.7S 16. Us Bubinga (*■!. Africa) Copaifera aff. m essmanii ...i U.13 : 9.56 ! 15.3S Capcmo (Trop. Amer.) Brosimum Alicastrum 5.12 : 9. k : 15.36 I'imeo. RlU2U Table 3. — Tropical woods — percent directional and volume shrinkage from green to oven-dry condition , (continued) S"^ © c i e s Volume Cherry, African 7. so • 13.72 Coccobolo (Cent. Amer.) ■ 2.65 H.26 : 7.20 3"bony, Macassar (Dutch -&» I.) 9.1U : lU.89 Franerie ( T tf. Afric,a) : 6.12 : lU.lg Gaboon (V. Africa) : 5.63 6.10 : 12.62 G-arapa (Brazil) ; U.53 S.ll 13.95 rip. "n pq] a A"l Troc ( T't* a Attiot* 1 U^UllUcxX U ix± vco \ X I U O* AillcJ x • / S.33 lU.70 >JX CCIlilcd,! 1/ • urU J. clxl a. ^ ; ! 3^1 U.22 g.00 Guapinol (Tro~o. Amer. ) : 3-00 5.22 g.65 1 3.HU ^•77 S.U9 Eoa (Hawaii) 5.^7 6.19 12.39 Koko (Andaman Is.) 6.62 9.7^ Lacewood (Australia) 7.20 11. U7 Lauaan, Red. (Phil. Is.) 3.27 s.oU 11. g6 Laurel, East Indian S.9S 15.^3 I-'imeo. R1U2H -"b- Table 3. — Tropical woods — percent directional and volume shrinkage green to oven-dry condition , (continued) from Snecies . xvao. lai Tangential Volume T.-inVho ('"J 4 f V i r» q ^ ±j ± IILU d ^ v » A-L I XLd J S.06 lU.37 Macacauba (Brazil) i " : 6.^2 11.51 Mahoganv , African ! u.96 8.36 16.88 Mahogany, Columbian :' 2.U6 3.80 . 6.53 i'la^iogany, Cuban 2.U3 U.Uy Mahogany, St. Jago : 3.19 ^•13 7.89 Mahogany (Peru) Swietenia macrophylla 3.15 ^.39 8.07 Mahogany (San Domingo) : 2.06 2.91 5.22 Manconia ( ,,r . Africa) 6. 1+2 11.51 Maple, Australian I 3-7^ 8.36 15.^7 Movingui ('f. Africa) : 3.08 5.18 10.66 Ibarra \, Jrxi 1 J. . is.; ! 2.5^ 3.63 • 6.81 Orientalwood (Australia) \ U.5U S.55 13.7^ Padouk, African 8.52 Padouk, Andaman k.oS 7.9^ Mimeo. R1U2U Table 3. — Tropical woods — percent directional and volume shrinkage from green to oven-dry condition . ( c ont i nued J >J|/C01CO ' Rpri i nT • T'q "OP* p^n t, i q1 • X Ct--X^ G- !. U X CLX. • 8.55 : 13.31 P?al nc QTii q 1 T^Vi i 1 To 1 JTdlUbdUlo ^^rliJ.J-» io. J i : 7.S1 ! 13.37 Pearv/ood (Europe) : !+.30 : lU.65 ! 19.79 Pero"ba, White (3razil) < 1 fcfe 6.20 : 9.82 P^imavera (Cent. Arner.) : U. 23 5.05 9.62 H.0 S SWO Oct , -'^1*3.2 1 1 lcill ; 3M 7.70 12.31 x\U CWUUU-, — 'do t X liC 1 . X rill 5.71 7.18 Rosewood, Prench. (Madagascar) ; 3.25 5.3S 9.17 CI. 1 J CL G ^ * 9 1 luc, / 1 5.91 7.U2 13.99 d "f" 1 vi i,rA n/^ P oTr"l ■Jciij jlziwu ou. , ueyion 5.71 8.51 Sfl ti n lnrn W T ri i » n ! 6.12 9. 18 15.18 Sof, ^ «tt "R Pr^ 1 AnotrQl i q 1 u a U Xil^y | XVoU. \_ Alio I a.X Id / 7.96 16.93 l ?ft"h»c« ^ ^-rn' A^p-r ^ -i- CX