TN295 
 
 ■ 
 
 No. 9132 
 
 
 I 
 
 m m 
 
 
 
 
 ■ 
 
 ■ < 
 
 
 ■ 
 
 ^H 
 
 Era 
 
 
 'H*.>. 
 
 WAS 
 
 l&>-^< m 
 

 ' V t *' 
 
 
 
 . 
 
 IV 
 
 *P "^ 
 
 
 *. y ,-»X >°* •£&%«* , 
 
 "ov* 
 
 *,• *°-v 'X2SH&: ^°^ *.e sv »p^.. 
 
 J> o • " • . <S> 
 
 ^ ■ • i 1 - 
 
 \.? 
 
 / V^V v-^v v^v 
 
 
 
 
 
 
 V* , 
 
 ** 
 
 *bv D 
 
 ." ; V *:«K^ > 0v ^ V 
 
 
 
 
 ^0< 
 
 
 ^o< 
 
 
 \ 
 
 ..\ n ^ 
 
 
 
 
 .v .... V 
 
 
 
 ^> v 
 
 /■^fc\ <?'*&>» /sJmkX <?£&S /*s£ikS 
 
 » ^ v 
 
 0^ -' 
 
 
 
 
 
 
 £ 
 
 ^ 
 
 
 r%, iik- j>\_ -m?.-- ^\^ --m; y\ wa ; --^ ; y > 
 
 . * 
 
 ^ % ^,c^ 
 
 
 ^^ °%,. 
 
 
 > ^ V 
 
 
 
 
 1^ .'•<., * 
 
 
 
 
 
 ^^ 
 
 
 •c> ' . . 5 . G 
 
 ;° £^ •WW' A v -^ '^« c,^ o V 
 a * «p u ^., o V J§sV^ * Ay ^*, • ©WIS » «? u ^, «x 
 
 
°* 
 
 <*- A * 
 
 *P ^ 
 
 ^ v" .;•»- cv aP* .l^L:* > 
 
 V *'*•' C* 
 
 
 
 w v 
 
 A*° >:*fc *-. « 
 
 
 iVA. 
 
 ^^ 
 
 ■> A'*\> oV/^Wv\\V" <!•>",■. 
 
 ,. a* *« o -5f!' ./ ^ 
 
 ^ v ..'••. ** 
 
 <, 
 
 '% S.-&&* /^MrS, /.-'mkS ,/.:££. V <■"*, 
 
 ^o T 
 
 <?> * o „ ° <J, 
 
 
 "oV 
 
 x -^. 
 
 ;- "'W •* 
 
 L* *> 
 
 V* 
 
 
 v*--*V v*^*/ c v3- / / v^^v * 
 
 
 
 
 » A? *« 
 
 
 
 4 o 
 
 
 
 V «^ A 
 
 
 ^d **^r * <\ 
 
 
 J- ^ 
 
 'q.. *•.'*' a° 
 
 
 
 
 "of 
 
 
 
 
 V 
 
 
 
 r* a 
 
 ,/%/' 
 
 
 v^ 
 
 
 • ^ A* .^ 
 
 
 
 ;• . ^ 
 
 q, *'V%» 5 .v^- 
 
 V 
 
 y»! 
 
 
 O d 
 
 
 
 °o ,** ,C^ 
 
 ^ 
 
 VJ. A* *j 
 
 v^ 
 
 ^ 
 
 1 *' 
 
 *, v 
 
 
 o. *'..«* A 
 
 ;* * v <* -».' 
 
 V •*'£* o^ 1? SIX*.* * V % 
 
 * >$• A V • rf^Sr A c . ^n .<? * S» • *^. A V * r 
 
 '* *•*■ 0^ °^ 
 
 
 
 V^ 1 
 
 % A 
 
 -,. # ^ -.y^w. i? -> ..«»,. ^v ^ 
 
 : a> 
 
 V-* 
 
 ■//^\ C°*.jSii>o /\>^\ / 
 
 ^. *'.. s 4 A 
 
 'oV 
 
 ;. ^o^ •* 
 
 x0 ^ ; 
 
 
 4>°* 
 
 4 ?* 
 
 
 
 x/" .•'«&"•. %.** .•« 
 
 
 
 o* .•••• w *o " 
 
 J? 
 
 
 * v . 
 
 rat* •?■ «. 
 
 
 
 *b>* 
 
 s* y^^L*. ^ 
 
 
 ">* % 
 
 ^o< 
 
 
 A" .i'., % o* ..... 
 
 a^ ♦!«iZS&' % ° 
 
 r oV" 
 
 
 »«V-. q*. a^> .-wiL'* > v v .1*°- ^ a5> >*VL'* "> 
 
 
 <?>. "o." 3 A 
 
 o. * , 
 
 
 : ^^ 
 
 
 • ^^ ,'isSfei-. \^* ^^afe'. ^ ^ .: 
 
 V «'°J 
 
 u jv9' 
 
 ^ V ^ 
 
 
 
 % 
 
 ■^ 
 
 
 ;• * v ^ WW .^ ^ ^^. ; ^ ^ 
 
 
 *bv° 
 
 
 
 
 ,4q 
 
 
 ^/(• n *ca a* i /rafV' «* «. »a^^«* "fv a* * 
 
 . s * A <. -..»* «G V V ♦*7^T» A <. -o.»" A 
 
 
 
 ♦<\. 
 
 ^ ^. 
 
 aP v . «i^L'*. ^ . v % . . i v- %. a, 6 -^'* ^ 
 
 <*■ «. ' *'/ 1 ^^.* "£j. a"* * 
 
 •#>» < 
 V-0^ 
 
 *feV 
 
 
 •^d« 
 
 q, *'T^i s A <.'»•- • .g v 
 
 
 '.' «> 
 
 L% V 
 
 
 • °» o 
 
 
 
 
 aV *^ „!! 
 
 ^x;> 
 
 
 
 
 
 
IC 9132 
 
 Bureau of Mines Information Circular/1987 
 
 A Statistical Analysis of Metal and Nonmetal 
 Mine Fire Incidents in the United States 
 From 1950 to 1984 
 
 By Shail J. Butani and William H. Pomroy 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
Information Circular 9132 
 
 A Statistical Analysis of Metal and Nonmetal 
 Mine Fire Incidents in the United States 
 From 1950 to 1984 
 
 By Shail J. Butani and William H. Pomroy 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
 Donald Paul Hodel, Secretary 
 
 BUREAU OF MINES 
 Robert C. Horton, Director 
 

 
 Library of Congress Cataloging in Publication Data: 
 
 Butani, Shail J. 
 
 A statistical analysis of metal and nonmetal mine fire incidents in 
 the United States from 1950 to 1984. 
 
 (Information circular/United States Department of the Interior, Bureau of Mines; 9132) 
 
 Includes bibliographical references. 
 
 Supt. of Docs, no.: I 28.27: 9132. 
 
 1. Mine fires — United States — Statistics. 2. Mine fires — United States — Statistical 
 methods. I. Pomroy, William H. II. Title. III. Series: Information circular (United States. 
 Bureau of Mines); 9132. 
 
 T«29SrrJ4— [TN3151 363.379 
 
 86-600299 
 
CONTENTS 
 
 Page 
 
 Abstract 1 
 
 Introduction 2 
 
 Data analysis methods 2 
 
 Chi-square 3 
 
 t-test 4 
 
 Spearman rank order correlation 6 
 
 Levels of significance and rounding 7 
 
 Incidence rate 7 
 
 Reported fires 7 
 
 Underground fires 8 
 
 Time trends 8 
 
 Ore type 11 
 
 Ignition source 12 
 
 Burning substance 13 
 
 Location 15 
 
 Equipment involved 17 
 
 Means of detection 18 
 
 Duration 18 
 
 Number of injuries 19 
 
 Number of deaths 21 
 
 Mining method 21 
 
 Successful extinguishing agent 22 
 
 Surface fires at surface and underground mines 23 
 
 Time trends 23 
 
 Ore type 26 
 
 Ignition source 27 
 
 Burning substance 27 
 
 Location 28 
 
 Equipment involved 28 
 
 Means of detection 29 
 
 Duration 29 
 
 Number of injuries 30 
 
 Number of deaths 30 
 
 Successful extinguishing agent 31 
 
 Comparison of underground and surface fires 32 
 
 Time trends 32 
 
 Ore type 32 
 
 Ignition source 32 
 
 Burning substance 32 
 
 Location 32 
 
 Equipment involved 32 
 
 Means of detection 32 
 
 Duration 33 
 
 Number of injuries 33 
 
 Number of deaths 33 
 
 Successful extinguishing agent 34 
 
 Nonreportable fires 34 
 
 Underground fires 36 
 
 Ingition source 36 
 
 Burning substance 36 
 
 Location 36 
 
ii 
 
 CONTENTS— Continued 
 
 Page 
 
 Equipment involved 36 
 
 Successful extinguishing agent 36 
 
 Surface fires 36 
 
 Ignition source 36 
 
 Burning substance 36 
 
 Location 36 
 
 Equipment involved 36 
 
 Successful extinguishing agent 36 
 
 Opinion data from mine safety directors 36 
 
 Underground mine fires 37 
 
 Surface mine fires 37 
 
 Summary of opinion data 37 
 
 Summary 38 
 
 Ignition source 38 
 
 Burning substance 39 
 
 Location 39 
 
 Equipment involved 39 
 
 Successful extinguishing agent 39 
 
 Conclusions 40 
 
 Appendix 41 
 
 ILLUSTRATIONS 
 
 1. Underground fire incidents by year, 1950-84 8 
 
 2. Average number of fire incidents per year, underground versus surface, 
 
 during three time periods 10 
 
 3. Fire incidence rates by year, underground versus surface, 1978-84 11 
 
 4. Average underground fire incidence rates by ore type, 1978-84 13 
 
 5. Surface fire incidents by year, 1950-84 24 
 
 6. Average surface fire incidence rates by ore type, 1978-84 27 
 
 7. Percentage of mobile equipment fires, underground versus surface, during 
 
 three time periods 33 
 
 8. Percentage of fires detected immediately, underground versus surface, 
 
 during three time periods 34 
 
 TABLES 
 
 1. Reported underground fire incidents by year 6 
 
 2. Ranking of reported underground fire incidents by year 7 
 
 3. Number of reportd underground fire incidents by three time periods 9 
 
 4. Reported underground fire incidence rates, 1978-84 10 
 
 5. Reported underground fires by principal ore, three time periods 12 
 
 6. Average reported underground fire incidence rates by ore, 1978-84 12 
 
 7. Reported underground fires by ignition source, three time periods 14 
 
 8. Reported underground fires by burning substance, three time periods 14 
 
 9. Reported underground fires by location, three time periods 15 
 
 10. Reported underground electrical fires by location, three time periods 16 
 
TABLE S — Con t inued 
 
 iii 
 
 Page 
 
 11. Reported underground welding or cutting fires by location, three time 
 
 periods 16 
 
 12. Reported underground fires by equipment involved, three time periods 17 
 
 13. Reported underground electrical fires by equipment, three time periods.... 18 
 
 14. Reported underground fires by means of detection, three time periods 18 
 
 15. Reported underground fires by duration, three time periods 19 
 
 16. Reported underground fires by ignition source and duration, two time 
 
 periods 20 
 
 17. Reported underground fires by number of injuries, three time periods 20 
 
 18. Reported underground injury fires by ignition source, three time periods.. 20 
 
 19. Reported underground injury fires by location, three time periods 21 
 
 20. Reported underground injury fires by equipment, three time periods 21 
 
 21. Reported underground fires by number of deaths, three time periods 21 
 
 22. Reported underground fires by mining method, three time periods 22 
 
 23. Reported underground fires by successful extinguishing agent, three time 
 
 periods 22 
 
 24. Reported underground fires, joint distribution of successful extinguishing 
 
 agents , by three time periods 23 
 
 25. Reported surface fire incidents by year 23 
 
 26. Ranking of reported surface fire incidents by year 25 
 
 27. Number of reported surface fire incidents by three time periods 25 
 
 28. Reported surface fire incidence rates, 1978-84 25 
 
 29. Reported surface fires by principal ore, three time periods 26 
 
 30. Average reported surface fire incidence rates by ore, 1978-84 26 
 
 31. Reported surface fires by ignition source, three time periods 28 
 
 32. Reported surface fires by burning substance, three time periods 28 
 
 33. Reported surface fires by location, three time periods 29 
 
 34. Reported surface fires by equipment involved, three time periods 29 
 
 35. Reported surface fires by means of detection, three time periods 30 
 
 36. Reported surface fires by duration, three time periods 30 
 
 37. Reported surface fires by number of injuries, three time periods 31 
 
 38. Reported surface fires by number of deaths, three time periods 31 
 
 39. Reported surface fires by successful extinguishing agent, three time 
 
 periods 31 
 
 40. Nonreportable underground fires 35 
 
 41. Nonreportable surface fires 35 
 
 42. Average rankings of opinion data for ignition source, burning substance, 
 
 and successful extinguishing agent, underground fires 37 
 
 43. Average rankings of opinion data for ignition source, burning substance, 
 
 and successful extinguishing agent, surface fires 37 
 
 44. Major study findings of reported fires 38 
 
 45. Major study findings of nonreportable fires 39 
 
UNIT OF MEASURE ABBREVIATIONS USED IN THIS 
 
 REPORT 
 
 h hour pet 
 
 percent 
 
 min minute yr 
 
 year 
 
A Statistical Analysis of Metal and Nonmetal 
 
 Mine Fire Incidents in the United States 
 
 From 1950 to 1984 
 
 By Shail J. Butani' and William H. Pomroy 2 
 
 ABSTRACT 
 
 This Bureau of Mines publication presents a statistical analysis of 
 official U.S. Mine Safety and Health Administration (MSHA) metal and 
 nonmetal mine fire reports from 1950 through 1984, plus accounts of se- 
 lected nonreportable fires (less than 30 min and no injury) and opinion 
 data on fire hazards from mine safety directors. Fires were analyzed by 
 time trends, ore type, ignition source, burning substance, location in 
 mine, equipment involved, means of detection, duration, number of injur- 
 ies and fatalities, mining method, and successful extinguishing agent. 
 
 The leading ignition sources were electricity in underground fires and 
 engine heat in reported surface fires. The most frequent burning sub- 
 stance was combustible liquids for all nonreportable fires, reported 
 surface fires, and reported underground fires from 1978 to 1984. For 
 underground fires reported prior to 1978, timber was the leading burning 
 substance. Mobile equipment was the type most frequently involved in 
 both underground and surface fires. Underground fires occurred most of- 
 ten in haulageway or drift areas, and reported surface fires occurred 
 most often in plant and mill buildings, while most nonreportable surface 
 fires occurred in other areas. The most common methods of extinguish- 
 ment were water hose lines for reported fires and dry chemical hand- 
 portable extinguishers for nonreportable fires. 
 
 ' Mathematical statistician. 
 ■^ Supervisory mining engineer. 
 Twin Cities Research Center, Bureau of Mines, Minneapolis, MN. 
 
INTRODUCTION 
 
 In support of the Bureau of Mines re- 
 search on mine fire protection, two sep- 
 arate baseline studies on mine fire in- 
 cidents have been prepared. The first 
 study addressed coal mine fires. The re- 
 sults of that study are summarized in the 
 Bureau's Information Circular 8830, "A 
 Statistical Analysis of Coal Mine Fire 
 Incidents in the United States From 1950 
 to 1977." This second report, address- 
 ing metal and nonmetal mine fires, is the 
 companion to IC 8830. Together, these 
 two reports provide a comprehensive, fac- 
 tual summary of the mining industry's 
 fire experience. 
 
 This report, like IC 8830, covers 
 (1) MSHA fire incident reports, (2) rec- 
 ords of nonreportable fires (less than 30 
 min and no injury), and (3) mine safety 
 directors' opinions on fire hazards. The 
 MSHA fire reports provide the most reli- 
 able, objective, and accurate historcial 
 record of the major noncoal mine fires 
 that occurred during the study period 
 (1950 through 1984). However, MSHA fire 
 reports alone understate the true magni- 
 tude of the fire problem because metal 
 and nonmetal mines have been legally re- 
 quired to report fires to MSHA only since 
 1968. Although some fires were reported 
 prior to 1968, doubtless a great many 
 were not. Also, the reporting regu- 
 lations that took effect in 1968 spe- 
 cify that only fires lasting 30 min or 
 longer or involving an injury need to be 
 
 reported. MSHA fire reports are thus 
 limited in scope by MSHA's legal author- 
 ity. In order to provide a more compre- 
 hensive data base, it was also necessary 
 to gather and analyze mine company rec- 
 ords of nonreportable fires. 
 
 Finally, since mine fires are relative- 
 ly rare events, it is desirable to ana- 
 lyze not only the fires themselves, but 
 also the "near-misses" (which occur much 
 more frequently) and the unsafe condi- 
 tions that could give rise to future 
 fires. Thus, opinion data from mine 
 safety directors were collected and sep- 
 arately analyzed in an effort to broaden 
 the discussion and to characterize and 
 rank mine fire hazards in general. The 
 1950 to 1977 data were collected via a 
 research and development contract with 
 Allen Corp. of America. These data have 
 been previously published in the form of 
 a Bureau Open File Report and are avail- 
 able from the National Technical Informa- 
 tion Service (NTIS). 3 The 1978 through 
 1984 data have not been previously pub- 
 lished. The reported mine fire data have 
 also been formatted and entered into a 
 computer data base using the Lotus 1-2-3 
 software 4 for IBM-compatible personal 
 computers. Use of this computer data 
 base would greatly simplify the analysis 
 of metal and nonmetal mine fires for the 
 specialized purposes of individual users. 
 Microdata are also available from the au- 
 thors on hard copy. 
 
 DATA ANALYSIS METHODS 
 
 Three statistical techniques were em- 
 ployed to analyze the data collected: 
 the chi-square test, the t-test, and 
 the Spearman rank order correlation. A 
 
 3 Baker, R. M., J 
 and J. Wishmyer. 
 raphy of Metal and 
 ports (contract JO 
 America) . Volume 
 81 , 1980, 64 pp. 
 Volume II. BuMine 
 284 pp.; NTIS PB 
 BuMines OFR 68(3)- 
 PB 81-223745. 
 
 . Nagy, L. B. McDonald, 
 
 An Annotated Bibliog- 
 
 Nonmetal Mine Fire Re- 
 
 295035, Allen Corp. of 
 
 I. BuMines OFR 68(1)- 
 
 ; NTIS PB 81-223729. 
 
 s OFR 68(2)-81 , 1980, 
 
 81-223737. Appendix. 
 
 1 , 1980, 390 pp.; NTIS 
 
 detailed discussion of these methods can 
 be found in basic statistical books. 5 
 Additionally, incidence rates were com- 
 puted where possible. Each of these 
 methods of analysis is discussed below. 
 
 ^Reference to specific products does 
 not imply endorsement by the Bureau of 
 Mines . 
 
 5 Mendenhall, W. Introduction to Sta- 
 tistics. Wadsworth Publ. Co., Inc., 
 1965, 305 pp. 
 
 Siegel, S. Nonparametric Statistics 
 for the Behavioral Sciences. McGraw- 
 Hill, 1956, 312 pp. 
 
CHI- SQUARE 
 
 Chi-square (x 2 )» as an analysis tech- 
 nique, is used in situations requiring 
 comparison of an observed number of re- 
 sponses in each of certain chosen cate- 
 gories with an expected number, which is 
 based on the null hypothesis (see discus- 
 sion of null hypothesis below). The ob- 
 served values or frequencies are those 
 obtained by direct observation or by tab- 
 ulation of the various types of data. 
 The expected or theoretical frequencies 
 are generated on the basis of some under- 
 lying assumption (the null hypothesis), 
 independent of the observed data. The 
 computational formula for chi-square val- 
 ues is 
 
 ,2 _ 
 
 = I 
 
 (0, - E,) 2 
 
 i=l 
 
 where K = number of categories , 
 
 1 = observed number of cases in 
 i* h category, 
 
 difference is said to be significant. 
 Thus, the null hypothesis concerning the 
 theoretical frequencies is rejected; that 
 is, the differences between the observed 
 and expected frequencies are greater than 
 would be expected to exist by chance. 
 
 Throughout this report, the null hy- 
 pothesis used for analyzing data within a 
 time period is that an equal number of 
 fires are expected for all categories 
 within each factor studied — ore type, 
 ignition source, location, etc. As an 
 example, refer to the tabulation below, 
 which shows reported underground fires by 
 ignition source, 1968-77: 6 
 
 Electrical 
 
 Welding sparks or hot slag 
 
 Engine heat 
 
 Spontaneous combustion. . . . 
 
 Friction 
 
 Explosives 
 
 Other 
 
 Total 
 
 Number 
 
 Pet 
 
 39 
 
 46.4 
 
 16 
 
 19.0 
 
 7 
 
 8.3 
 
 13 
 
 15.5 
 
 6 
 
 7.1 
 
 1 
 
 1.2 
 
 2 
 
 2.4 
 
 84 
 
 100.0 
 
 and 
 
 E| 
 
 K 
 
 I 
 i=l 
 
 = expected number of cases in 
 i +h category under the null 
 hypothesis, 
 
 directs one to sum over all 
 categories. 
 
 That is, chi-square equals the summa- 
 tion of the squared value of the differ- 
 ence of observed and expected values 
 divided by the corresponding expected 
 value. 
 
 Tf the agreement between the observed 
 and expected frequencies is close, the 
 differences (Oi - Ej) will be small, and 
 consequently, chi-square will be small. 
 On the other hand, if the differences are 
 large, the value of chi-square will also 
 be large. The larger chi-square is, the 
 more likely it is that the observed 
 values did not come from the population 
 on which the null hypothesis is based. 
 If the computed value of chi-square is 
 greater than the tabled value of chi- 
 square for a specified significance level 
 (oc; usually a = 0.05 or 0.01) with K — \ 
 degrees of freedom (d.f.), then the 
 
 According to the above-stated hypothesis, 
 all types of ignition sources would be 
 expected to have caused an equal or near- 
 ly equal number of fires. Thus, the ex- 
 pected frequency for each ignition source 
 type would be the average frequency of 
 
 all ignition source types specified , 
 this case, the expected value is 
 
 In 
 
 39+16+7+13+6+1+2 
 K = 7 
 
 = 12.0, 
 
 where K = number of different types of 
 ignition sources. 
 
 Thus, the computation of chi-square is as 
 follows: 
 
 2 _ (39 - 12. 0) 2 (16 - 12.0) 
 * 12 12 
 
 + • • . + -r-z — 85.67. 
 
 6 Data from table 7, which appears later 
 in this report. 
 
The tabled value of chl-square with 6 
 (K - 1) d.f. at a = 0.01 is 16.81. Since 
 the computed value of chi-square (85.67) 
 is greater than the tabled value of chi- 
 square (16.81), the null hypothesis is 
 rejected. This implies that all ignition 
 source types are not equally involved in 
 metal and nonmetal underground fires for 
 the period 1968-77. 
 
 When the expected frequencies are small 
 (less than five), the chi-square test is 
 generally not valid. Also, the chi- 
 square test is not appropriate for the 
 actual number of fires across time peri- 
 ods, because the reporting requirements 
 before 1968 were less stringent and, con- 
 sequently, a number of fires were not 
 reported pre-1968. Hence, the underlying 
 assumption of an equal number of fires 
 per year is not met. Therefore, the chi- 
 square test is not applied across all 
 situations; instead, the t-test, con- 
 cerning proportions (probabilities), is 
 used. 
 
 t-TEST 
 
 The first type of t-test is for the 
 one-sample case; it is concerned with de- 
 termining whether the proportion of ob- 
 served fires in a category within a time 
 period is more than what is expected un- 
 der the null hypothesis (that all catego- 
 ries have the same proportion of fires). 
 Only probabilities that are greater than 
 what is expected are analyzed, because 
 the concern is for those categories that 
 are hazardous. The computational formula 
 for the t-test is 
 
 Po 
 
 Pod - Po) 
 
 t = 
 
 where p = observed proportion of acci- 
 dents falling in the cate- 
 gory of interest, 
 
 p = proportion of accidents ex- 
 pected under the null hy- 
 pothesis (in this report, p 
 = 1.00 divided by the number 
 of categories) , 
 
 and n = total number of specified fires 
 within a time period. 
 
 A comparison is made between the calcu- 
 lated value of t and the tabled value of 
 t; the tabled value of t with its associ- 
 ated d.f. (n - 1) is such that the area 
 under the distribution curve to the right 
 of t a is equal to n, the significance 
 level. If the calculated value of t is 
 greater than the tabled value of t, the 
 difference is said to be significant and 
 not merely due to chance. Consequently, 
 the null hypothesis is rejected in favor 
 of the alternative hypothesis; namely, 
 the category has proportionately more 
 fires than what is expected. As an ex- 
 ample, refer to the tabulation below, 
 which shows reported underground electri- 
 cal fires by location, 1950-67: 7 
 
 Number 
 
 Haulageway-drif t 
 
 Substation-shop-storage- 
 pump 
 
 Shaft-raise-winze 
 
 Working face 
 
 Mined-out waste 
 
 Other 
 
 Total 
 
 Pet 
 
 31.8 
 
 4 
 
 18.2 
 
 7 
 
 31.8 
 
 2 
 
 9.1 
 
 I 
 
 4.5 
 
 1 
 
 4.5 
 
 22 
 
 100.0 
 
 The null hypothesis is tested, that 
 the probability of electrical fires in 
 haulageway-drif t locations is equal to 
 0.167 (1.00 divided by the number of 
 categories) . 
 
 t = P ~ Po 
 
 J P( 
 
 >0 - 
 
 - Po) 
 
 V 
 
 0. 
 
 n 
 
 ,318 
 
 - 0.167 
 
 10. 
 
 167 
 
 (0.833) 
 
 22 
 
 = 1.91. 
 
 The tabled value of t with 21 d.f. at 
 the 0.05 significance level is 1.72. 
 Since the computed value is greater than 
 
 ^Data from table 10, which appears lat- 
 er in this report. 
 
the tabled value, the null hypothesis is 
 rejected; that is, the relative frequency 
 of the haulageway-drif t location as the 
 site of electrical fires is significantly 
 more than its share of one out of every 
 six fires. 
 
 The absolute t-values ( 1 1 1 ) for the 
 one-sample case, where applicable, are 
 shown at the bottom of the tables, under 
 the appropriate time periods. 
 
 The second type of t-test concerns the 
 difference between two proportions. This 
 test determines whether the observed dif- 
 ference between proportions from the two 
 samples is a real difference or is due to 
 random variation. The null hypothesis in 
 this test is Pj = P2, where Pj, and P2 
 are the proportion of cases in population 
 1 and population 2, respectively, pos- 
 sessing a certain attribute that is of 
 interest. The computational formula for 
 this test is 
 
 t = 
 
 Pi ~ P2 
 
 ^-»(k + k) 
 
 where pi = number of cases in sample 1 
 possessing a certain attri- 
 bute, divided by ni, 
 
 P2 = number of cases in sample 2 
 possessing a certain attri- 
 bute, divided by n2, 
 
 rij = sample size drawn from popu- 
 lation 1, 
 
 n2 = sample size drawn from popu- 
 lation 2, 
 
 and 
 
 n lPl + n 2P2 
 ni + n2 
 
 A comparison is then made between the 
 absolute calculated value of t and the 
 tabled value of t. As before, the tabled 
 value of t with its associated d.f. (nj^ 
 + n2 - 2) is such that the area under the 
 distribution curve to the right of t a is 
 equal to a, the significance level. The 
 absolute value of t is compared to the 
 
 tabled value because the alternative 
 hypothesis is that either Pj is greater 
 than P2 or P2 is greater than Pj. Again, 
 if the absolute calculated value of t is 
 greater than the tabled value of t, then 
 the difference between the two propor- 
 tions is significant and not merely due 
 to chance. Consequently, the null hy- 
 pothesis that the two probabilities are 
 equal is rejected in favor of one proba- 
 bility being greater than the other. 
 
 The following data can be used to test 
 the hypothesis that the proportion of un- 
 derground electrical fires located in 
 haulageway-drif t locations in 1950 to 
 1967 is equal to the proportion in 1968 
 to 1977. Reported underground electrical 
 fires by location, two time periods: 8 
 
 Period I Period II 
 
 1950-67 
 
 1968-77 
 
 No. 
 
 Pet 
 
 No. 
 
 Haulageway-drif t. 7 31.8 18 
 Substation-shop- 
 storage-pump. .. . 4 18.2 13 
 Shaft-raise-winze 7 31.8 3 
 
 Working face 2 9.1 4 
 
 Mined-out waste.. 1 4.5 
 
 Other _J_ 4.5 1 
 
 Total 22 100.0 39 
 
 Pet 
 
 46.2 
 
 33, 
 
 .3 
 
 7, 
 
 .7 
 
 10, 
 
 .3 
 
 
 .0 
 
 2, 
 
 .6 
 
 100.0 
 
 t = 
 
 Pi - P2 
 
 >>-»>(^) 
 
 where pi = y~- = 0.318, 
 
 1 q 
 P2 =3! = 0.462, 
 
 ni = 22, 
 
 n 2 = 39, 
 
 22 (0.318) + 39 (0.462) 
 
 and 
 
 22 + 39 
 
 = 0.410. 
 
 8 Data from table 10. 
 
0.318 - 0.461 
 
 J 
 
 (0.410) (0.590) 
 
 _T 
 
 22 
 
 1 
 
 39 
 
 |t| = 1.09. 
 
 The absolute calculated value of t = 1.09 
 is less than the tabled value of t 
 = 1.645 (with 59 d.f. and a = 0.05), so 
 the null hypothesis cannot be rejected. 
 That is, the difference between the two 
 proportions could be due to random 
 variation. 
 
 The absolute t-values for the two- 
 sample case, where applicable, are given 
 in the last two columns of the tables. 
 
 SPEARMAN RANK ORDER CORRELATION 
 
 The Spearman rank order correlation 
 (r s ) is a measure of concordance or 
 agreement, the tendency of two ranks to 
 be similar. This test was chosen as the 
 measure of correlation because it is non- 
 parametric — it makes no assumption about 
 the shape of the distribution curve of 
 the underlying population. It was also 
 chosen because both variables can be mea- 
 sured in at least an ordinal scale so 
 that the objects or individuals under 
 study may be ranked in two ordered 
 
 series. For example, using the data in 
 table 1, the number of fires, denoted as 
 Xj, X2, ... Xn, and the year, denoted as 
 Yj, Y2, ... Yn, may be ranked as shown 
 in table 2. Rank correlation may then be 
 used to determine the relation between 
 the X's and Y's. The formula used in 
 calculating the Spearman rank order 
 correlation is as follows: 
 
 = 1 - 
 
 N 
 6 I d 
 
 N(N 2 - 1) 
 
 where d| = Xj - Y| 
 
 and N = total number of observations. 
 
 The value of r s is always equal to or 
 greater than -1.0 and equal to or less 
 than 1.0; that is, -1.0 < r s < 1.0. If 
 the computed value of r s exceeds the ta- 
 bled value based on N and a (significance 
 level) , then the rank order correlation 
 is said to be significant; that is, the 
 agreement between the two sets of ranks 
 is greater than would be expected to oc- 
 cur by chance. 
 
 Using the ranked tabulated data in ta- 
 ble 2 as an example, the Spearman rank 
 order correlation would be 
 
 = 1-6 
 
 (3 
 
 l) 2 + (18.5 - 2) 2 + ... + (23.0 - 34) 2 + (20.5 - 35) 2 ] 
 
 35 [(35) 2 - 1] 
 
 = 0.569 
 
 For N = 35 and a = 0.01, the tabled value 
 of r s is 0.432. Since the computed value 
 of r s is greater than the tabled value, 
 
 the rank order correlation 
 significant. 
 
 is said to be 
 
 TABLE 1. - Reported underground fire incidents by year 
 
 Year 
 
 1950.. 
 
 1951.. 
 
 1952.. 
 
 1953.. 
 
 1954.. 
 
 1955.. 
 
 1956.. 
 
 1957.. 
 
 1958.. 
 
 1959.. 
 
 I960.. 
 
 1961.. 
 
 Number of 
 incidents 
 
 1 
 
 5 
 3 
 3 
 2 
 4 
 6 
 10 
 2 
 7 
 3 
 3 
 
 Year 
 
 1962. 
 1963. 
 1964. 
 1965. 
 1966. 
 1967. 
 1968. 
 1969. 
 1970. 
 1971. 
 1972. 
 1973. 
 
 Number of 
 incidents 
 
 
 
 2 
 
 4 
 
 2 
 
 2 
 
 2 
 
 4 
 
 
 
 4 
 
 5 
 12 
 21 
 
 Year 
 
 1974.. 
 1975.. 
 1976.. 
 1977.. 
 1978.. 
 1979.. 
 1980.. 
 1981.. 
 1982.. 
 1983.. 
 1984.. 
 Total 
 
 Number of 
 incidents 
 
 15 
 10 
 
 8 
 13 
 23 
 10 
 
 7 
 11 
 12 
 
 7 
 
 6 
 229 
 
TABLE 2. - Ranking of reported underground fire incidents by year 
 
 Year 
 
 Number of 
 incidents 
 
 Rank 
 
 Year 
 
 Number of 
 incidents 
 
 Rank 
 
 
 Year 
 
 Incidents 
 
 Year 
 
 Incidents 
 
 1950 
 
 1 
 
 1 
 
 3.0 
 
 1968 
 
 4 
 
 19 
 
 15.5 
 
 1951.... 
 
 5 
 
 2 
 
 18.5 
 
 1969 
 
 
 
 20 
 
 1.5 
 
 1952 
 
 3 
 
 3 
 
 11.5 
 
 1970 
 
 4 
 
 21 
 
 15.5 
 
 1943 
 
 3 
 
 4 
 
 11.5 
 
 1971 
 
 5 
 
 22 
 
 18.5 
 
 1954 
 
 2 
 
 5 
 
 6.5 
 
 1972 
 
 12 
 
 23 
 
 30.5 
 
 1955 
 
 4 
 
 6 
 
 15.5 
 
 1973 
 
 21 
 
 24 
 
 34.0 
 
 1956.... 
 
 6 
 
 7 
 
 20.5 
 
 1974 
 
 15 
 
 25 
 
 33.0 
 
 1957 
 
 10 
 
 8 
 
 27.0 
 
 1975 
 
 10 
 
 26 
 
 27.0 
 
 1958 
 
 2 
 
 9 
 
 6.5 
 
 1976 
 
 8 
 
 27 
 
 25.0 
 
 1959 
 
 7 
 
 10 
 
 23.0 
 
 1977 
 
 13 
 
 28 
 
 32.0 
 
 1960 
 
 3 
 
 11 
 
 11.5 
 
 1978 
 
 23 
 
 29 
 
 35.0 
 
 1961 
 
 3 
 
 12 
 
 11.5 
 
 1979 
 
 10 
 
 30 
 
 27.0 
 
 1962 
 
 
 
 13 
 
 1.5 
 
 1980.... 
 
 7 
 
 31 
 
 23.0 
 
 1963 
 
 2 
 
 14 
 
 6.5 
 
 1981 
 
 11 
 
 32 
 
 29.0 
 
 1964 
 
 4 
 
 15 
 
 15.5 
 
 1982 
 
 12 
 
 33 
 
 30.5 
 
 1965 
 
 2 
 
 16 
 
 6.5 
 
 1983 
 
 7 
 
 34 
 
 23.0 
 
 1966 
 
 2 
 2 
 
 17 
 18 
 
 6.5 
 6.5 
 
 1984 
 
 Total. 
 
 6 
 
 35 
 
 20.5 
 
 1967 
 
 229 
 
 NAp 
 
 NAp 
 
 NAp Not applicable. 
 
 0.569 (significant at 1-pct level) 
 
 LEVELS OF SIGNIFICANCE AND ROUNDING 
 
 Levels of significance are provided in 
 tbe tables for all statistical analyses 
 reported. Significant results are iden- 
 tified by eitber one or two asterisks, 
 denoting probabilities of 0.05 and 0.01, 
 respectively. Lack of an 
 cates tbat tbe results 
 cant; tbat is, tbere 
 0.05 probability tbat tbe results could 
 have occurred by chance. In some cases, 
 no test was made, for any one of several 
 reasons. These cases are denoted as "not 
 determined" (ND). 
 
 The sum of the percentage components in 
 the tables may not exactly equal 100.0 
 because of rounding. 
 
 INCIDENCE RATE 
 
 asterisk indi- 
 were not signifi- 
 is a greater than 
 
 incidence rates (IR). Incidence rates 
 are used to standardize or normalize the 
 number of fires on the basis of exposure 
 hours so that meaningful comparisons can 
 be made. The incidence rates in this re- 
 port represent the number of fires per 
 10,000 full-time workers and are calcu- 
 lated as 
 
 IR = 
 
 number of fires 
 total hours worked 
 
 x 20 million h, 
 
 where 20 million h is equivalent to 
 10,000 full-time workers working 40 h per 
 week and 50 weeks per year. 
 
 Data from 1980 9 may be used as an exam- 
 ple. The incidence rate for 1980 is 
 
 IR 
 
 80 54,200,000 
 
 x 20,000,000; 
 
 The last type of analysis is con- 
 cerned with the relative measure based on 
 
 IR 80 " 2 ' 6 ' 
 
 REPORTED FIRES 
 
 For the period 1950-77, MSHA fire re- 
 ports were acquired through searches of 
 the official files of all MSHA metal- 
 nonmetal inspection offices and the head- 
 quarters office at Arlington, VA. The 
 
 inspection offices included in the search 
 are listed in the appendix. 
 
 ^Data from table 4, which appears later 
 in the report. 
 
At each MSHA office, project personnel 
 searched through all accident report 
 files on fires, explosives, and igni- 
 tions. Only reports on fires were uti- 
 lized, but other reports were reviewed in 
 order to locate any incorrectly filed or 
 titled reports. In the review process, 
 the fire reports that were considered to 
 be minor burn accidents were eliminated. 
 For the period 1977-84, MSHA's Health and 
 Safety Analysis Center (HSAC) computer 
 data base of mine accidents was searched. 
 Several fire reports not contained in 
 HSAC's files were also provided by MSHA's 
 Denver Mining Technology Center, Ventila- 
 tion Branch. 
 
 MSHA fire reports were analyzed under 
 two headings: underground fires and sur- 
 face fires at underground and surface 
 mines. 
 
 UNDERGROUND FIRES 
 
 Time Trends 
 
 The first analysis determines time 
 trend effects on the number of fires in 
 each year. Table 1 lists 229 fire inci- 
 dents in underground mines by year of oc- 
 currence. The Spearman rank order corre- 
 lation, based on table 2, between years 
 and number of fires is 0.569, compared 
 with the tabled value of 0.432 at the a 
 = 0.01 level. This indicates the rank 
 order correlation is significant; that 
 is, on the average, the number of fires 
 reported has increased with time. This 
 can be seen in figure 1, which shows 
 there were a higher number of fires for 
 which a report could be located for the 
 period beginning in 1972. 
 
 W W W 
 
 Lfl 
 
 W W i W h 
 
 
 WWW 
 
 / / 
 
 www 
 
 y < 
 
 W- 
 
 y / y 
 
 ', y 
 
 Rl 1 ^ Wi 
 
 y 
 y 
 
 s 
 
 ' ft M 
 
 
 Figure 1. — Underground fire incidents by year, 1950-84. 
 
The analysis is geared toward a break- 
 down by three time periods of mine law: 
 pre-1967 Metal and Nonmetal Act, post- 
 1967 Metal and Nonmetal Act, and post- 
 1977 Mine Health and Safety Act. Table 3 
 displays the numbers of fire incidents by 
 the three time periods. The chi-square 
 for this table is computed on the assump- 
 tion of an equal number of fires each 
 year. This highly significant value in- 
 dicates that differences exist in the 
 frequency of occurrence of reported fires 
 during the three time periods. This 
 could be because prior to 1968 record- 
 keeping was not required for accidents 
 (including fires) that occurred in metal 
 and nonmetal mines. Although the Bureau 
 was the designated authority over metal 
 and nonmetal mining, investigative re- 
 porting of accidents was not undertaken 
 on a regular basis, and no specific 
 agency was charged with this activity. 
 Usually the technical relationship of the 
 mine operator with the nearest Bureau 
 representative or the seriousness of the 
 occurrence were factors determining 
 whether or not a fire incident was 
 reported. 
 
 Based on discussions with personnel 
 from each of the MSHA offices visited 
 during this project, it is also apparent 
 that a significant percentage of reports 
 of fire incidents for this period (pre- 
 1968) have been misplaced over the years 
 of shift and reorganization of personnel 
 and offices in the Bureau, the Mine and 
 Environmental Safety Association (MESA), 
 and MSHA. If the conservative assumption 
 is made that all fire incidents occurring 
 after passage of the Metal and Nonmetal 
 Mining Act of 1967 have been located, a 
 
 TABLE 3. - Number of reported underground 
 fire incidents by three time periods 
 
 Period 
 
 Ob- 
 served 
 
 Ex- 
 pected 
 
 Average 
 per year 
 
 I: 1950-67 
 
 II: 1968-77... 
 Ill: 1978-84.. 
 
 61 
 92 
 76 
 
 118 
 65 
 46 
 
 3.4 
 
 9.2 
 
 10.9 
 
 Total or av. 
 
 229 
 
 229 
 
 6.5 
 
 level) 
 
 58.32 
 
 (significant at 1-pct 
 
 simple calculation using average fires 
 per year shows that possibly 65 pet of 
 all fires that occurred before passage of 
 the act have not been accounted for. As 
 will be shown later in this report, there 
 has been a dramatic reduction in the 
 average duration of fires reported fol- 
 lowing passage of the act and an increase 
 in reporting of fire incidents, mandated 
 by the creation of MESA and, ultimately, 
 MSHA. This supports the hypothesis that 
 proportionately more fires less than 24 h 
 in length went unreported during the 
 early years under consideration. If, on 
 the other hand, the assumption of an 
 average yearly fire incidence is in 
 error, the fire hazardousness of the un- 
 derground mining environment must be 
 examined. 
 
 Over the years following the mid- 
 1960' s, underground metal and nonmetal 
 mining equiment has developed an in- 
 creased dependency on diesel power and 
 electricity as prime movers. This could 
 represent a significant source of in- 
 crease in fire hazard. Furthermore, the 
 growth of the industry during the 1970' s 
 in size of operations and number of 
 miners employed also represents an over- 
 all potential increase in the underground 
 fire hazard. 
 
 Significantly then, the trend seen 
 in figures 1 and 2 is either toward in- 
 creased awareness (via reporting) of 
 fires in underground metal and nonmetal 
 mines or toward increased hazardousness 
 of the environment. 
 
 To fully explain the fire hazard by 
 year, the analysis should be based on 
 relative measures such as incident rates, 
 that is, the number of fires occurring 
 per 20 million h worked, the equivalent 
 of 10,000 full-time workers per year. 
 (See "Data Analysis Methods.") This type 
 of analysis was performed for the period 
 1978-84, since this was the most current 
 period and the only period for which 
 the hours could be easily obtained from 
 MSHA's data base (table 4). The expected 
 number of incidents for each year is com- 
 puted on the assumption that the number 
 of fires would be in proportion to the 
 relative exposure (for example, for 1980 
 
10 
 
 IB 
 
 16 -- 
 
 tr 
 
 
 < 
 
 UJ 
 
 14 
 
 y 
 
 
 IX 
 
 
 LI 
 
 12 
 
 1 
 
 
 01 
 
 
 h 
 
 
 Z 
 
 10 
 
 LU 
 
 
 D 
 
 
 H 
 
 
 u 
 
 H 
 
 z 
 
 
 H 
 
 
 111 
 
 
 CD 
 
 6 
 
 < 
 
 
 II 
 
 
 UJ 
 
 
 > 
 
 
 < 
 
 4 
 
 2- 
 
 KEY 
 
 ^j Underoround 
 
 1950-67 1968-77 1978-84 
 
 Figure 2. — Average number of fire incidents per year, underground versus surface, during three time periods. 
 
 TABLE 4. - Reported underground fire incidence 
 rates, 1978-84 
 
 Year 
 
 Observed 
 number of 
 incidents 
 
 Hours 
 
 worked, 
 
 10 3 h 
 
 Incidence 
 rate 1 
 
 Expected 
 number of 
 incidents 
 
 1978 
 
 23 
 
 10 
 
 7 
 
 11 
 
 12 
 
 7 
 
 6 
 
 51,900 
 55,000 
 54,200 
 54,500 
 37,700 
 27,050 
 27,400 
 
 8.9 
 3.6 
 2.6 
 4.0 
 6.4 
 5.2 
 4.4 
 
 13 
 
 1979 
 
 14 
 
 1980 
 
 13 
 
 1981 
 
 13 
 
 1982 
 
 9 
 
 1983 
 
 7 
 
 1984 
 
 7 
 
 Total or IR 
 
 76 
 
 307,750 
 
 4.9 
 
 76 
 
 X 2 = 13.05* (significant at 5-pct level). 
 ! Per 20 million h worked. 
 
 it would be (54,200/307,750) x 76). The 
 chi-square value computed from the ob- 
 served and expected number of fires indi- 
 cates that not all years were equally 
 
 hazardous. This fact can be easily seen 
 from figure 3, which also shows that 
 there is no apparent time trend on inci- 
 dence rates for this period. 
 
11 
 
 9-- 
 
 7 -■ 
 
 h 6 
 < 
 
 H 
 
 ID 5 
 U 
 
 z 
 111 
 
 Q 
 
 H 4 
 
 u 
 
 z 
 
 H 
 
 3-- 
 
 1-- 
 
 197B 
 
 1982 
 
 KEY 
 
 Underground 
 
 Surface 
 
 11 
 
 1963 19B4 
 
 Figure 3. — Fire incidence rates (based on 20 million h worked), by year, underground versus surface, 1978-84. 
 
 Ore Type 
 
 Table 5 shows underground fires by ore 
 type for the three time periods. Analy- 
 sis within any given time period indi- 
 cates that the differences among differ- 
 ent ore types are highly significant; all 
 three overall chi-square values are sig- 
 nificant at the 0.01 level. The most no- 
 ticeable comparison is between ore types 
 iron and salt in the period 1950-67. 
 These data also show that the distribu- 
 tion of fires by ore type has changed 
 from period to period. 
 
 According to the t-values, the relative 
 frequency of copper fires decreased from 
 1968-77 to 1978-84, lead-zinc fires in- 
 creased from 1950-67 to 1968-77, iron 
 fires decreased from 1950-67 to 1968-77 
 and again from 1968-77 to 1978-84, and 
 salt fires increased from 1950-67 to 
 1968-77. 
 
 Since an estimate of the hours worked 
 in each ore type was readily available 
 from MSHA's data base for the period 
 1978-84, a detailed analysis by ore type 
 for this period was performed (table 6). 
 The relative measure of rate is based on 
 the number of fire incidences per 20 mil- 
 lion h worked, which is equivalent to 
 10,000 full-time workers per year. (See 
 "Data Analysis Methods.") The expected 
 number of fires for each type of ore 
 mined was computed on the assumption that 
 all mines were equally hazardous. Hence, 
 the expected number of fires for each ore 
 mined would be in proportion to the rela- 
 tive exposure (for example, for salt it 
 would be (15,400/307,750) * 76). Since 
 the chi-square value of 26.55 is highly 
 significant, it means all ore types were 
 not equally hazardous. The relative mea- 
 sure of incidence rates as depicted in 
 
12 
 
 TABLE 5. - Reported underground fires by principal ore, three time periods 
 
 Ore 
 
 Period I 
 1950-67 
 
 No. 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 Ill 
 
 Copper 
 
 Lead-zinc 
 
 Iron 
 
 Salt 
 
 Silver 
 
 Other 
 
 Total specified. 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 X 2 : 
 
 Specified 
 
 Without "Other' 
 ND Not determined. 
 
 10 
 4 
 
 25 
 
 3 
 
 15 
 
 17.5 
 7.0 
 
 43.9 
 
 .0 
 
 5.3 
 
 26.3 
 
 21 
 
 15 
 8 
 
 11 
 9 
 
 25 
 
 23.6 
 16.9 
 9.0 
 12.4 
 10.1 
 28.1 
 
 9 
 18 
 
 1 
 12 
 
 6 
 30 
 
 11.8 
 23.7 
 
 1.3 
 15.8 
 
 7.9 
 39.5 
 
 40 
 37 
 34 
 23 
 18 
 70 
 
 18.0 
 16.7 
 15.3 
 10.4 
 8.1 
 31.5 
 
 0.87 
 1.72' 
 
 4.91* 
 
 2. 76' 
 1.04 
 ND 
 
 1.95 
 
 1.09 
 
 2.17' 
 
 ND 
 
 ND 
 
 ND 
 
 57 
 
 100.0 
 
 89 
 
 100.0 
 
 76 
 
 100.0 
 
 222 
 
 100.0 
 
 57 
 
 4 
 
 93.4 
 6.6 
 
 89 
 3 
 
 96.7 
 3.3 
 
 76 
 
 
 
 100.0 
 .0 
 
 222 
 
 7 
 
 96.9 
 3.1 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 45.63 
 47.29* 
 
 15.97 
 8.81 
 
 41.32 
 17. 70 1 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 TABLE 6. - Average reported underground fire 
 incidence rates by ore, 1978-84 
 
 Ore 
 
 Salt 
 
 Lead-zinc 
 
 Silver 
 
 Copper 
 
 Other 2 
 
 Total or IR 
 
 Observed 
 number of 
 incidents 
 
 12 
 
 18 
 
 6 
 
 9 
 
 31 
 
 76 
 
 Hours 
 
 worked, 
 
 10 3 h 
 
 15,400 
 44,050 
 15,850 
 48.650 
 183,800 
 
 307,750 
 
 Incidence 
 
 rate 
 
 15.6 
 8.2 
 7.6 
 3.7 
 3.4 
 
 5.0 
 
 Expected 
 number of 
 incidents 
 
 4 
 11 
 
 4 
 12 
 45 
 
 76 
 
 X = 26.55 (signif icanct at 1 pet level) 
 'Per 20 million h worked. 
 2 Includes iron (1 incident). 
 
 figure 4 shows salt to be the most haz- 
 ardous of all ore types, followed by 
 lead-zinc and silver. 
 
 Ignition Source 
 
 Table 7 shows fires by ignition source 
 for the three time periods. The overall 
 chi-square values indicate that the dif- 
 ferent ignition source totals are signif- 
 icantly different for each of the three 
 time periods. The overwhelming ignition 
 source for all the time periods was elec- 
 trical. The distribution of specified 
 fires among the seven ignition source 
 types was very similar for the first two 
 
 time periods but changed for the third 
 period. One such category is "Spontane- 
 ous combustion," where the proportion of 
 fires was significantly down from 15.5 
 pet in 1968-77 to 1.4 pet in 1978-84. 
 According to the t-test values, the pro- 
 portion of electrical fires was also sig- 
 nificantly down between 1968-77 and 1978- 
 84, but the proportion of engine heat 
 fires had increased for the same time 
 period. With electrical fires removed, 
 the differences among the remaining igni- 
 tion sources are significant at the 1-pct 
 level for the periods 1968-77 and 1978- 
 84, but are not significant for the 1950- 
 67 period. 
 
13 
 
 20 
 
 19-- 
 
 18-- 
 
 17 
 
 16-- 
 
 15-- 
 
 14 -- 
 
 13-- 
 
 12-- 
 
 11- 
 
 UJ 10-- 
 
 u 
 
 9-- 
 B-- 
 
 7-- 
 
 6-- 
 
 5- 
 
 4- 
 
 3-- 
 
 2- 
 
 1 -- 
 
 
 
 Average (all types)> 
 
 Salt Lead-zinc Silver Copper Other 
 
 Figure 4. — Average underground fire incidence rates (based on 20 million h worked), by ore type, 1978-84. 
 
 Since 1970, there has been some effort 
 by persons in the Bureau and MSHA to ban 
 smoking in underground metal and nonmetal 
 mines. Smoking is, at present, banned in 
 certain areas of a mine, such as fueling 
 areas. There have been about a half- 
 dozen reported underground mine fires di- 
 rectly attributed to smoking during the 
 study period. These fires involved the 
 ignition of mine timbers and lasted 
 longer than 24 h, but resulted in no in- 
 juries or fatalities. In fires where the 
 cause was unspecified, smoking was iden- 
 tified as a possible or probable cause in 
 at least four cases. 
 
 A survey of MSHA district managers re- 
 vealed a cross section of opinions relat- 
 ing to the hazardousness of smoking in 
 underground metal and nonmetal mines. 
 
 However, most acknowledged that smoking 
 materials, where evident of a fire's 
 cause, would be the first substance to be 
 consumed by the fire and, thus, might be 
 involved in more fires than is generally 
 known. 
 
 Burning Substance 
 
 Table 8 shows underground fires by 
 burning substance for the three time pe- 
 riods. One of the underlying assumptions 
 in utilizing either the chi-square test 
 or the t-test is that of independence. 
 This means among other things that any 
 given observation falls into and only one 
 category. In the case of fires catego- 
 rized by burning substances, it would 
 mean that each fir* 1 is classified into 
 
14 
 
 TABLE 7. - Reported underground fires by ignition source, three time periods 
 
 Ignition source 
 
 Period I 
 1950-67 
 
 Period II 
 1968-77 
 
 Period III 
 1978-84 
 
 Total 
 
 1 
 
 t| 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 22 
 
 42.3 
 
 39 
 
 46.4 
 
 21 
 
 29.6 
 
 82 
 
 39.6 
 
 ND 
 
 2.15* 
 
 10 
 
 19.2 
 
 16 
 
 19.0 
 
 7 
 
 9.9 
 
 33 
 
 15.9 
 
 ND 
 
 1.60 
 
 2 
 
 3.8 
 
 7 
 
 8.3 
 
 18 
 
 25.4 
 
 27 
 
 13.0 
 
 1.02 
 
 2.87** 
 
 8 
 
 15.4 
 
 13 
 
 15.5 
 
 1 
 
 1.4 
 
 22 
 
 10.6 
 
 ND 
 
 3.05** 
 
 3 
 
 5.8 
 
 6 
 
 7.1 
 
 9 
 
 12.7 
 
 18 
 
 8.7 
 
 ND 
 
 1.16 
 
 3 
 
 5.8 
 
 1 
 
 1.2 
 
 2 
 
 2.8 
 
 6 
 
 2.9 
 
 ND 
 
 ND 
 
 4 
 
 7.7 
 
 2 
 
 2.4 
 
 13 
 
 18.3 
 
 19 
 
 9.2 
 
 ND 
 
 ND 
 
 52 
 
 100.0 
 
 84 
 
 100.0 
 
 71 
 
 100.0 
 
 207 
 
 100.0 
 
 
 
 52 
 
 85.2 
 
 84 
 
 91.3 
 
 71 
 
 93.4 
 
 207 
 
 90.4 
 
 
 9 
 
 14.8 
 
 8 
 
 8.7 
 
 5 
 
 6.6 
 
 22 
 
 9.6 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 40.35** 
 
 85.67** 
 
 122.61** 
 
 
 
 
 10. 
 
 40 
 
 23. 
 
 67** 
 
 20 
 
 .10 
 
 
 
 
 
 Electrical 
 
 Welding sparks or hot 
 
 slag 
 
 Engine heat 
 
 Spontaneous combustion 
 
 Friction 
 
 Explosives 
 
 Other 
 
 Total specified. . . . 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 X 2 : 
 
 Specified 
 
 Without electrical. 
 
 ND Not determined. Significant at 5-pct level. Significant at 1-pct level. 
 TABLE 8. - Reported underground fires by burning substance, 1 three time periods 
 
 
 Per 
 
 iod I 
 
 Period II 
 1968-77 
 
 Peri 
 
 od III 
 
 Total 
 
 It| 
 
 Burning substance 
 
 1950-67 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet . 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 48 
 6 
 9 
 3 
 
 18 
 
 57.1 
 7.1 
 
 10.7 
 3.6 
 
 21.4 
 
 40 
 21 
 29 
 18 
 26 
 
 29.9 
 15.7 
 21.6 
 13.4 
 19.4 
 
 19 
 31 
 
 16 
 
 9 
 
 12 
 
 21.8 
 35.6 
 18.4 
 10.3 
 13.8 
 
 107 
 58 
 54 
 30 
 56 
 
 35.1 
 19.0 
 17.7 
 9.8 
 18.4 
 
 4.00** 
 1.86* 
 2.07* 
 2.40** 
 ND 
 
 1.32 
 
 Combustible liquids... 
 
 3.42** 
 ND 
 
 
 ND 
 ND 
 
 Total specified.... 
 
 84 
 
 100.0 
 
 134 
 
 100.0 
 
 87 
 
 100.0 
 
 305 
 
 100.0 
 
 
 
 Specified 
 
 84 
 3 
 
 96.6 
 3.4 
 
 134 
 1 
 
 99.3 
 .7 
 
 87 
 3 
 
 96.7 
 3.3 
 
 7 
 312 
 
 97.8 
 2.2 
 
 
 
 
 
 
 
 87 
 
 100.0 
 
 135 
 
 100.0 
 
 90 
 
 100.0 
 
 
 100.0 
 
 
 
 
 X 2 : 
 
 ¥■ * 
 
 79. QV 
 
 10 
 3 
 
 .85* 
 .11 
 
 16.*?** 
 
 
 
 
 Without timber. . . . 
 
 14. 
 
 00** 
 
 16. 
 
 82** 
 
 
 ND Not determined. 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 In many fires, more than 1 substance was burning. 
 
 only one burning substance type. The 
 data on fires by burning substances, of 
 course, violates this assumption. How- 
 ever, in analyzing the data of this ta- 
 ble, it was found that fire in one burn- 
 ing substance type was independent of 
 fires in other burning substance types. 
 If this assumption is true, then the re- 
 sults of the analysis given below are 
 valid for the intended purpose. The 
 large chi-square values for each of the 
 
 three time periods indicate that the dif- 
 ferences among the different burning sub- 
 stance are significant. These data show 
 that timber was the most frequent sub- 
 stance burning, while wiring insulation 
 and combustible liquids were also fre- 
 quently involved in fires, especially 
 since 1968. 
 
 The t-test values indicate that the 
 distribution of fires by burning sub- 
 stance has changed from one period to 
 
15 
 
 another. There was a significant de- 
 crease in the proportion of fires attrib- 
 uted to timer from 1950-67 to 1968-77 and 
 1978-84. It should be noted, however, 
 that the decrease was not significant be- 
 tween the second and third time periods. 
 The fires due to combustible liquids, on 
 the other hand, have increased from one 
 period to the next. There was a definite 
 shift from timber as the most frequent 
 burning substance to combustible liquids, 
 insulation, and rubber between 1950-67 
 and 1968-77. 
 
 Location 
 
 The locations of fires in underground 
 metal and nonmetal mines appear in table 
 9. The large chi-square values for the 
 three time periods indicate that signif- 
 icant differences exist in the frequency 
 of fires occurring at various loca- 
 tions. Fires occurred predominantly in 
 haulageway-drif t entry areas for the pe- 
 riods 1968-77 and 1978-84; for the period 
 1950-67, they occurred in shaft-raise- 
 winze areas. This result also means 
 there was a change in the distribution 
 of fire location from one period to 
 another. The number of fires occurring 
 in mined-out wastes and shaft-raise-winze 
 
 areas declined while those in haulageway- 
 drift areas increased. To determine the 
 sources of fires at the various loca- 
 tions, the data on the two most common 
 sources were analyzed separately. 
 
 Table 10 shows electrical fires by lo- 
 cation for the three time periods. Since 
 there are too many categories for the 
 number of observations, the expected fre- 
 quencies are small (less than five). 
 Hence, the chi-square test is not valid 
 for this situation. The data in this ta- 
 ble and in table 11, therefore, were ana- 
 lyzed by the t-test. The t-values for 
 all three time periods indicate that the 
 relative frequency of electrical fires 
 along haulageway-drif t areas, where lit- 
 erally miles of cable are installed, was 
 more than what would be expected by 
 chance. Electrical fires in the moving 
 equipment, such as load-haul-dumps, oc- 
 curred frequently here. The t-values 
 also show the relatively high frequencies 
 of electrical fires at substation-shop- 
 storage-pump areas for the 1968-77 period 
 and at shaft-raise-winze areas for the 
 1950-67 period to be significant. In 
 comparing the distribution of fires 
 across time periods, the t-test revealed 
 no significant difference in the pro- 
 portion of fires between 1950-67 and 
 
 TABLE 9. - Reported underground fires by location, three time periods 
 
 Location 
 
 Period I 
 1950-67 
 
 No. 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. 
 
 Pet 
 
 I vs. 
 
 II 
 
 $1 
 
 ¥TF 
 
 II VS. 
 Ill 
 
 Haulageway-drif t. . . 
 Shaft-raise-winze. . 
 Mined-out waste. . . , 
 
 Working face , 
 
 Substation-shop- 
 storage-pump i 
 
 Other 
 
 Total specified. 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 X 2 : 
 
 Specified 
 
 Without haul- 
 
 ageway-drif t. 
 
 ND Not determined. 
 
 12 
 
 20 
 
 13 
 
 5 
 
 8 
 3 
 
 19.7 
 
 32.8 
 
 21.3 
 
 8.2 
 
 13.1 
 4.9 
 
 61 
 
 100.0 
 
 61 
 
 
 
 100.0 
 .0 
 
 61 
 
 100.0 
 
 18.77 
 
 19.06 
 
 42 
 
 11 
 
 10 
 
 9 
 
 16 
 3 
 
 46.2 
 
 12.1 
 
 11.0 
 
 9.9 
 
 17.6 
 3.3 
 
 42 
 9 
 2 
 6 
 
 6 
 6 
 
 59.2 
 
 12.7 
 
 2.8 
 
 8.5 
 
 8.5 
 8.5 
 
 96 
 40 
 25 
 20 
 
 30 
 12 
 
 43.0 
 
 17.9 
 
 11.2 
 
 9.0 
 
 13.5 
 5.4 
 
 3.34 
 3.10 1 
 1.74* 
 ND 
 
 ND 
 ND 
 
 1.64 
 ND 
 
 1.97 s 
 ND 
 
 ND 
 ND 
 
 91 
 
 100.0 
 
 71 
 
 100.0 
 
 223 
 
 100.0 
 
 91 
 
 1 
 
 98.9 
 1.1 
 
 71 
 
 5 
 
 93.4 
 6.6 
 
 223 
 6 
 
 97.4 
 2.6 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 61.11 
 6.41 
 
 94.38 
 4.28 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
16 
 
 TABLE 10. - Reported underground electrical fires by location, three time periods 
 
 Location 
 
 Period I 
 1950-67 
 
 No. 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 Ill 
 
 Haulageway-drif t. . 
 Subs tat ion-shop- 
 
 storage-pump 
 
 Shaft-shaft-winze. 
 
 Working face 
 
 Mined-out waste... 
 Other 
 
 Total specified 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 Haulageway-drif t. . 
 Subs tat ion-shop- 
 
 storage-pump. 
 Shaft-raise-winze. 
 
 31.8 
 
 18.2 
 
 31.8 
 
 9.1 
 
 4.5 
 4.5 
 
 13 
 3 
 
 4 
 
 1 
 
 46.2 
 
 33.3 
 
 7.7 
 
 10.3 
 
 .0 
 
 2.6 
 
 11 
 
 3 
 2 
 
 1 
 3 
 
 55.0 
 
 15.0 
 
 10.0 
 
 .0 
 
 5.0 
 
 15.0 
 
 36 
 
 20 
 
 12 
 
 6 
 
 2 
 
 5 
 
 44.4 
 
 24.7 
 
 14.8 
 
 7.4 
 
 2.5 
 
 6.2 
 
 1.09 
 
 1.27 
 2.44' 
 
 ND 
 ND 
 ND 
 
 0.64 
 
 1.50 
 ND 
 ND 
 
 ND 
 ND 
 
 22 
 
 100.0 
 
 39 
 
 100.0 
 
 20 
 
 100.0 
 
 100.0 
 
 22 
 
 
 
 100.0 
 .0 
 
 39 
 
 
 
 100.0 
 .0 
 
 20 
 1 
 
 95.2 
 4.8 
 
 98.8 
 1.2 
 
 22 
 
 100.0 
 
 39 
 
 100.0 
 
 21 
 
 100.0 
 
 82 
 
 100.0 
 
 1.91' 
 
 0.19 
 1.91' 
 
 4.94 
 
 2.79* 
 
 ND 
 
 4.60 
 
 ND 
 ND 
 
 ND Not determined. 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 TABLE 11. - Reported underground welding or cutting fires 
 by location, three time periods 
 
 Location 
 
 Shaft-raise-winze. 
 Haulageway-drif t. . 
 Substation-shop- 
 storage-pump. 
 Mined-out waste. . . 
 
 Working face , 
 
 Other , 
 
 Total , 
 
 1950-67 
 
 No. 
 
 10 
 
 Pet 
 
 60.0 
 20.0 
 
 20.0 
 .0 
 .0 
 .0 
 
 100.0 
 
 1968-77 
 
 No. 
 
 16 
 
 Pet 
 
 37.5 
 43.8 
 
 .0 
 
 6.3 
 
 .0 
 
 12.5 
 
 100.0 
 
 1978-84 
 
 No. 
 
 Pet 
 
 42.9 
 28.6 
 
 28.6 
 .0 
 .0 
 .0 
 
 100.0 
 
 Total 
 
 No. 
 
 15 
 11 
 
 4 
 1 
 
 2 
 
 33 
 
 Pet 
 
 45.5 
 33.3 
 
 12.1 
 
 3.0 
 
 .0 
 
 6.1 
 
 100.0 
 
 Shaft-raise-winze. 
 Haulageway-drif t. . 
 
 4.44 
 2.57 1 
 
 Significant at 1-pct level. 
 
 1968-77, and between 1968-77 and 1978-84 
 for the haulageway-drif t entry areas and 
 substation-shop-storage-pump areas. The 
 same analysis for the shaft-raise-winze 
 area, on the other hand, showed a signif- 
 icant decline in the proportion of fires 
 from 1950-67 to 1968-77. 
 
 Table 11 contains the analysis for 
 welding-cutting fires by location for two 
 of the time periods. Owing to the small 
 
 sample size in each period, which leads 
 to large fluctuations, the total number 
 of fires was analyzed so that statisti- 
 cally meaningful conclusions could be 
 drawn from these data. The t-tests for a 
 one-sample case showed that fires from 
 welding or cutting operations occurred 
 more frequently in the mine shaft-raise- 
 winze and in haulageway-drif t areas than 
 in other areas of the mine. 
 
17 
 
 Equipment Involved 
 
 Table 12 categorizes fires by equipment 
 involved, for the three time periods. 
 During the tabulation of these data, it 
 became obvious that some changes in the 
 reporting mechanism have occurred over 
 time for the "Unspecified," "None," and 
 "Unknown" categories. The data for these 
 categories, therefore, are such that it 
 is not possible to obtain a separate fire 
 count for each category. It should be 
 noted that the combination of the three 
 categories comprises a major portion of 
 all the fires by equipment type. Where 
 equipment is specified, the chi-square 
 values for the last two time periods are 
 significant, which indicates that some 
 equipment types were more involved in 
 fires than others. The most prone pieces 
 of equipment were either mobile, such as 
 load-haul-dumps, or electrical (that is, 
 stationary electrical equipment) . Two 
 t-tests were performed for the mobile and 
 electrical equipment. The first compared 
 the 1950-67 and 1968-77 periods; the sec- 
 ond compared the 1968-77 and 1978-84 
 periods. All of the t-values were not 
 significant. 
 
 To fully explain the relative fire haz- 
 ard of the various types of equipment 
 used in underground mines, fire incidents 
 
 should be normalized based on actual ex- 
 posure hours or the number of equipment 
 pieces. Although some data to support 
 such an analysis were located, it was 
 found that the data available were inade- 
 quate to normalize the equipment most 
 frequently associated with underground 
 fires. The authors believe that esti- 
 mates of equipment populations beyond the 
 scope of presently available data would 
 lead to highly questionable results; 
 therefore, no attempt was made to compute 
 incidence rates by equipment for under- 
 ground fires at this time. 
 
 An analysis of the leading ignition 
 source of fires where equipment is in- 
 volved is presented in table 13. As in 
 table 12, the equipment category "Unspe- 
 cified-unknown -none" comprises a major 
 portion of all the electrical fires. 
 Also, as explained before, the small num- 
 ber of fires precludes the use of the 
 chi-square test for this table. The 
 t-tests show that the relative frequency 
 of electrical equipment in electrical ig- 
 nition fires was much higher than the 
 expected (0.25) for all time periods. 
 The same type of analysis also shows that 
 the relative frequency of mobile equip- 
 ment was significantly more than expected 
 for the period 1968-77. 
 
 TABLE 12. - Reported underground fires by equipment involved, three time periods 
 
 
 Per 
 
 iod I 
 
 Period II 
 
 Period III 
 
 Total 
 
 It| 
 
 Equipment 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 6 
 
 10 
 1 
 6 
 
 26.1 
 
 43.5 
 
 4.3 
 
 26.1 
 
 24 
 
 16 
 
 5 
 
 7 
 
 46.2 
 
 30.8 
 
 9.6 
 
 13.5 
 
 31 
 9 
 6 
 6 
 
 59.6 
 17.3 
 11.5 
 11.5 
 
 61 
 35 
 12 
 19 
 
 48.0 
 
 27.6 
 
 9.4 
 
 15.0 
 
 1.64 
 
 1.07 
 
 ND 
 
 ND 
 
 1.38 
 
 
 1.61 
 
 ND 
 
 Other 
 
 ND 
 
 
 
 Total specified.... 
 
 23 
 
 100.0 
 
 52 
 
 100.0 
 
 52 
 
 100.0 
 
 127 
 
 100.0 
 
 
 
 
 23 
 38 
 
 37.7 
 62.3 
 
 52 
 40 
 
 56.5 
 43.5 
 
 52 
 24 
 
 68.4 
 31.6 
 
 127 
 102 
 
 55.5 
 45.5 
 
 
 Unspecif ied-unknown- 
 
 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 
 
 
 7.09 
 
 17.69** 
 
 33.69** 
 
 
 
 
 ND Not determined. 
 # * 
 
 Significant at 1-pct level. 
 
 1 The reporting mechanism did not permit the classification of fires into 3 separate 
 
 categories. 
 
18 
 
 TABLE 13. - Reported underground electrical fires by equipment, three time periods 
 
 El 
 
 Equipment 
 
 Period I 
 1950-67 
 
 No. 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 Ill 
 
 Electrical 
 
 Mobile , 
 
 Conveyor 
 
 Other 
 
 Total specified.. 
 
 Specified 
 
 Unspecif ied-unknown- 
 none 
 
 Total 
 
 54.5 
 
 36.4 
 
 .0 
 
 9.1 
 
 12 
 
 15 
 
 I 
 
 42.9 
 
 53.6 
 
 .0 
 
 3.6 
 
 47.1 
 
 41.2 
 
 5.9 
 
 5.9 
 
 26 
 26 
 
 1 
 3 
 
 46.4 
 
 46.4 
 
 1.8 
 
 5.4 
 
 0.66 
 
 .97 
 
 ND 
 
 ND 
 
 ND 
 ND 
 ND 
 ND 
 
 11 
 
 100.0 
 
 28 
 
 100.0 
 
 17 
 
 100.0 
 
 56 
 
 100.0 
 
 11 
 
 11 
 
 50.0 
 
 50.0 
 
 28 
 
 11 
 
 71.8 
 
 28.2 
 
 17 
 
 81.0 
 
 19.0 
 
 56 
 
 26 
 
 68.3 
 
 31.7 
 
 22 
 
 100.0 
 
 39 
 
 100.0 
 
 21 
 
 100.0 
 
 82 
 
 100.0 
 
 t| 
 
 Electrical. 
 Mobile 
 
 2.26 
 .87 
 
 2.19 
 3.49 J 
 
 2.10 
 
 1.54 
 
 ND Not determined. Significant at 5-pct level. Significant at 1-pct level. 
 TABLE 14. - Reported underground fires by means of detection, three time periods 
 
 
 Period I 
 
 Period II 
 1968-77 
 
 Period III 
 
 Total 
 
 It| 
 
 Means of detection 
 
 1950-67 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pct 
 
 No. 
 
 Pet 
 
 Ill 
 
 Operator-worker 
 
 9 
 
 20 
 3 
 8 
 7 
 7 
 
 16.7 
 
 37.0 
 5.6 
 14.8 
 13.0 
 13.0 
 
 28 
 
 23 
 
 10 
 
 6 
 
 8 
 
 14 
 
 31.5 
 
 25.8 
 
 11.2 
 
 6.7 
 
 9.0 
 
 15.7 
 
 29 
 
 21 
 7 
 6 
 
 3 
 
 43.9 
 
 31.8 
 10.6 
 
 9.1 
 .0 
 
 4.5 
 
 66 
 
 64 
 20 
 20 
 15 
 24 
 
 31.6 
 
 30.6 
 9.6 
 9.6 
 7.2 
 
 11.5 
 
 1.96* 
 
 1.42 
 1.15 
 
 1.57 
 
 ND 
 ND 
 
 1.59 
 
 Workers (not 
 
 .82 
 
 Shift boss-foreman.... 
 
 ND 
 ND 
 ND 
 
 
 ND 
 
 
 
 Total specified. . . . 
 
 54 
 
 100.0 
 
 89 
 
 100.0 
 
 66 
 
 100.0 
 
 209 
 
 100.0 
 
 
 
 54 
 7 
 
 88.5 
 11.5 
 
 89 
 3 
 
 96.7 
 3.3 
 
 66 
 10 
 
 86.8 
 13.2 
 
 209 
 20 
 
 91.3 
 8.7 
 
 
 
 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100. 0_, 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 
 
 
 18.44** 
 
 26. 
 
 21** 
 
 59.09** 
 
 
 
 ND Not determined. 
 
 Means of Det 
 
 Signi 
 ectio 
 
 f icant 
 n 
 
 at 5- 
 
 pet lev 
 The 
 
 el. 
 
 t-te 
 
 **Sign 
 st show 
 
 ifica 
 s the 
 
 nt at 1 
 re was 
 
 -pet level, 
 a significant 
 
 The person(s) discovering a fire are 
 listed in table 14. The large chi-square 
 values for each of the three time periods 
 indicate that the frequencies for the six 
 different categories are statistically 
 different. The most frequent detection 
 of fire occurs Immediately or after a 
 short period of time by workers In the 
 area. Most metal and nonmetal mine fire 
 reports do not identify the specific job 
 title of the person discovering a fire. 
 
 increase in the relative frequency of the 
 category "Operator-worker (immediate)" 
 between the periods 1950-67 and 1968-77. 
 The same type of analysis, however, shows 
 the changes to be nonsignificant for the 
 next three categories. 
 
 Duration 
 
 Duration of fires appears in table 15. 
 Note that for the periods 1950-67 and 
 1978-84, the duration is not specified 
 
19 
 
 TABLE 15. - Reported underground fires by duration, three time periods 
 
 
 I 
 1950-67 
 
 II 
 1968-77 
 
 III 
 
 Total 
 
 |t| 
 
 Duration, h 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 5 
 2 
 4 
 8 
 23 
 
 11.9 
 4.8 
 9.5 
 
 19.0 
 54.8 
 
 11 
 5 
 31 
 20 
 17 
 
 13.1 
 
 6.0 
 
 36.9 
 
 23.8 
 
 20.2 
 
 8 
 6 
 
 21 
 8 
 
 14 
 
 14.0 
 10.5 
 36.8 
 14.0 
 24.6 
 
 24 
 13 
 56 
 36 
 54 
 
 13.1 
 
 7.1 
 
 30.6 
 
 19.7 
 
 29.5 
 
 ND 
 
 ND 
 3.23** 
 
 ND 
 3.92** 
 
 ND 
 
 
 ND 
 
 
 ND 
 
 
 1.43 
 
 24+ 
 
 ND 
 
 
 
 Total specified.... 
 
 42 
 
 100.0 
 
 84 
 
 100.0 
 
 57 
 
 100.0 
 
 183 
 
 100.0 
 
 
 
 
 42 
 19 
 
 68.9 
 31.1 
 
 84 
 8 
 
 91.3 
 8.7 
 
 57 
 19 
 
 75.0 
 25.0 
 
 183 
 46 
 
 79.9 
 20.1 
 
 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 
 33. 
 
 95** 
 
 22. 
 
 90** 
 
 13.26* 
 
 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 for an estimated 30 and 25 pet, respec- 
 tively, of all the reported fires. It 
 should also be noted that many of the 
 fires lasting less than 1/2 h were not 
 reportable based on duration. Most of 
 these were reportable because of an asso- 
 ciated injury; some were reported as a 
 courtesy; and others were investigated by 
 MSHA inspectors after hearing about the 
 fire from a third party. As a general 
 rule, reports of these fires were found 
 at MSHA subdistrict offices in the form 
 of internal memos or letter reports. 
 
 Of those fires that were reported and 
 where the duration is specified, the chi- 
 square values are significant for all the 
 periods. This means the number of fires 
 are not evenly distributed across the 
 five duration categories. In the 1950-67 
 time period, the most frequent duration 
 was 24 h or more. The t-test value be- 
 tween the 1950-67 and 1968-77 periods in- 
 dicates there was definitely a decline in 
 the relative frequency of fires for this 
 category. Based on these results and on 
 discussions with the concerned officials, 
 it can be concluded that prior to 1968 
 fires lasting 24 h or longer were more 
 frequently reported than others. 
 
 Further analysis of ignition sources by 
 duration was performed for two time peri- 
 ods, 1968-77 and 1978-84, and the results 
 are presented in table 16. It should be 
 noted that for period 1978-84, the data 
 on duration of fires are less than com- 
 plete. In about 25 pet of the reports, 
 
 the duration of the fire was unspecified. 
 Of the specified durations in both time 
 periods, all the spontaneous combustion 
 fires lasted 24 h or longer, and about 50 
 pet of the electrical ignition fires 
 lasted 1 to 4 h. 
 
 Number of Injuries 
 
 Fire injuries for the three time peri- 
 ods appear in table 17. In all the time 
 periods, the fires where no injuries oc- 
 curred comprise an overwhelming propor- 
 tion of all the fires. Once the fires 
 with no injuries are eliminated, the sam- 
 ple size for each of the time periods 
 becomes too small to draw any statisti- 
 cally meaningful conclusions. A cross- 
 tabulation of injury fires by ignition 
 source appears in table 18. As shown, 
 the source of most injury fires was elec- 
 trical, followed by welding and engine 
 heat. Prior to 1978, there were no in- 
 juries involved with fires occurring from 
 engine heat. Yet, during the 1978-84 
 time period, engine heat was the largest 
 source of all injury fires. Cross-tabu- 
 lations of injury fires by location and 
 equipment appear in tables 19 and 20. 
 Injury fires occurred predominantly along 
 haulageway-drif ts, followed by working 
 face areas. Prior to 1968, it appears 
 that more injury fires started at the 
 shaft or in a mined-out area. Since 
 1968, they seem to have occurred more in 
 haulageway-drif ts and working face areas. 
 
20 
 
 TABLE 16. - Reported underground fires by ignition source and duration, 
 two time periods 
 
 Ignition source 
 
 to 0.5 h 
 
 1968-77 
 
 1978-84 
 
 0.5 to 1.0 h 
 
 1968-77 
 
 1978-84 
 
 1+ to 4 h 
 
 1968-77 
 
 1978-84 
 
 4+ to 24 h 
 
 1968-77 
 
 1978-84 
 
 Electrical , 
 
 Welding sparks or 
 
 hot slag 
 
 Engine heat 
 
 Spontaneous 
 
 combustion 
 
 Friction 
 
 Explosives 
 
 Other 
 
 Unspecified 
 
 Total 
 
 Electrical 
 
 Welding sparks or 
 
 hot slag 
 
 Engine heat 
 
 Spontaneous 
 
 combustion 
 
 Friction 
 
 Explosives 
 
 Other 
 
 Unspecified 
 
 Total 
 
 19 
 
 3 
 2 
 
 
 3 
 
 2 
 2 
 
 11 
 
 
 3 
 
 
 3 
 
 3 
 
 1 
 
 11 
 
 31 
 
 21 
 
 20 
 
 24+ h 
 
 1968-77 
 
 1 
 
 2 
 1 
 
 12 
 1 
 
 
 
 
 
 17 
 
 Unspecified 
 
 Total 
 
 1978-84 
 
 1968-77 
 
 14 
 
 8 
 
 1978-84 
 
 1968-77 
 
 39 
 
 16 
 7 
 
 13 
 6 
 1 
 2 
 
 19 
 
 92 
 
 1978-84 
 
 21 
 
 1 
 9 
 2 
 13 
 5 
 
 76 
 
 TABLE 17. - Reported underground fires by number of injuries, 
 three time periods 
 
 Number of injuries 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 
 
 54 
 4 
 2 
 1 
 
 
 88.5 
 
 6.6 
 
 3.3 
 
 1.6 
 
 .0 
 
 85 
 3 
 3 
 1 
 
 
 92.4 
 
 3.3 
 
 3.3 
 
 1.1 
 
 .0 
 
 63 
 
 11 
 
 2 
 
 
 
 
 
 82.9 
 
 14.5 
 
 2.6 
 
 .0 
 
 .0 
 
 202 
 
 18 
 
 7 
 
 2 
 
 
 
 88.2 
 
 1 
 
 7.9 
 
 2 to 5 
 
 10 + 
 
 3.1 
 .9 
 .0 
 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 
 TABLE 18. - Reported underground injury fires 
 by ignition source, three time periods 
 
 Ignition source 
 
 Electrical 
 
 Welding sparks or hot slag 
 
 Engine heat 
 
 Spontaneous combustion. . . . 
 
 Friction 
 
 Other 
 
 Total 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 13 
 
 27 
 
21 
 
 The type of equipment most involved in 
 injury fires was mobile-type equipment. 
 In fact, in the 1978-84 period, all of 
 the injury fires with a known equipment 
 type involved mobile equipment. 
 
 Number of Deaths 
 
 Table 21 lists fires by number of 
 deaths. The only significant result is 
 the overall reporting frequency of non- 
 fatal fires. Once the nonfatal fires are 
 removed, the sample sizes in each of the 
 three periods become too small to draw 
 any statistically meaningful conclusions. 
 
 Mining Method 
 
 Table 22 shows fires by mining method 
 for the three time periods. For a large 
 
 portion of the fire reports, mining meth- 
 od was not specified for the 1950-67 and 
 1978-84 periods. The chi-square values 
 for all time periods indicate significant 
 differences among the fire frequencies 
 associated with the five specified mining 
 methods. The t-test values show the fire 
 incidences occurred with a higher than 
 expected relative frequency in the caving 
 method for the period 1950-67 and in the 
 room-and-pillar method for the periods 
 1968-77 and 1978-84. The t-values show 
 that the relative frequency of fires in- 
 creased from one period to the next for 
 the room-and-pillar method and decreased 
 for the caving method. These results in- 
 dicate that the distribution of the fire 
 incidents for the mining methods has 
 changed from period to period. For the 
 open stoping method, however, the change 
 
 TABLE 19. - Reported underground injury fires by location, 
 three time periods 
 
 Location 
 
 Haulageway-drif t 
 
 Working face 
 
 Shaft-raise-winze 
 
 Substation-shop-storage-pump 
 
 Mined-out area 
 
 Other 
 
 Unspecified 
 
 Total 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 10 
 6 
 3 
 3 
 2 
 2 
 1 
 
 13 
 
 27 
 
 TABLE 20. - Reported underground injury fires by equipment, 
 three time periods 
 
 Equipment 
 
 Mobile 
 
 Electrical 
 
 Other 
 
 Unspecified-unknown -none. 
 Total 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 11 
 
 
 2 
 
 13 
 
 14 
 3 
 1 
 9 
 
 27 
 
 TABLE 21. - Reported underground fires by number of deaths, 
 three time periods 
 
 Number of deaths 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 
 
 57 
 2 
 
 2 
 
 93.4 
 
 3.3 
 
 .0 
 
 3.3 
 
 86 
 2 
 2 
 2 
 
 93.5 
 2.2 
 2.2 
 2.2 
 
 76 
 
 
 
 
 100.0 
 .0 
 .0 
 .0 
 
 219 
 4 
 2 
 
 4 
 
 95.6 
 
 1 
 
 1.7 
 
 6+ 
 
 .9 
 1.7 
 
 
 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 
 
22 
 
 TABLE 22. - Reported underground fires by mining method, three time periods 
 
 M 
 
 Mining method 
 
 Period I 
 1950-67 
 
 No, 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 Ill 
 
 Roora-and-pillar. . . 
 
 Caving 
 
 Open stoping 
 
 Cut-and-fill 
 
 Other 
 
 Total specified 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 X 2 : Specified.... 
 
 |t|: 
 
 Room and pillar 
 Caving 
 
 ND Not determined 
 
 10 
 
 21 
 
 13 
 
 1 
 
 6 
 
 19.6 
 41.2 
 25.5 
 2.0 
 11.8 
 
 31 
 18 
 
 16 
 
 14 
 
 7 
 
 36.0 
 20.9 
 18.6 
 16.3 
 8.1 
 
 34 
 5 
 5 
 
 10 
 5 
 
 57.6 
 8.5 
 8.5 
 
 16.9 
 8.5 
 
 75 
 44 
 34 
 25 
 18 
 
 38.3 
 22.4 
 17.3 
 12.8 
 9.2 
 
 2.03 
 2.54' 
 1.02 
 2.59 1 
 ND 
 
 2.57' 
 2.02* 
 1.70* 
 ND 
 ND 
 
 51 
 
 100.0 
 
 86 
 
 100.0 
 
 59 
 
 100.0 
 
 196 
 
 100.0 
 
 51 
 10 
 
 83.6 
 16.4 
 
 86 
 6 
 
 93.5 
 6.5 
 
 59 
 17 
 
 77.6 
 22.4 
 
 196 
 33 
 
 85.6 
 14.4 
 
 61 
 
 100.0 
 
 92 
 
 100.0 
 
 76 
 
 100.0 
 
 229 
 
 100.0 
 
 22.23 
 
 ND 
 3.78* 
 
 17.84 
 
 3.72 
 
 53.80 
 
 7.23 
 
 ND 
 
 ND 
 
 * 3Flf 
 
 Significant at 5-pct level. Significant at 1-pct level. 
 
 TABLE 23. - Reported underground fires by successful extinguishing 
 agent, 1 three time periods 
 
 Extinguishing agent 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 
 26 
 
 5 
 
 9 
 
 18 
 
 10 
 
 38.2 
 7.4 
 13.2 
 26.5 
 14.7 
 
 43 
 27 
 18 
 10 
 24 
 
 35.2 
 22.1 
 14.8 
 8.2 
 19.7 
 
 20 
 
 16 
 
 7 
 
 3 
 
 5 
 
 39.2 
 
 31.4 
 
 13.7 
 
 5.9 
 
 9.8 
 
 89 
 48 
 34 
 31 
 39 
 
 36.9 
 
 
 19.9 
 
 
 14.1 
 
 
 12.9 
 16.2 
 
 
 
 Total specified.... 
 
 68 
 
 100.0 
 
 122 
 
 100.0 
 
 51 
 
 100.0 
 
 241 
 
 100.0 
 
 
 68 
 3 
 
 95.8 
 4.2 
 
 122 
 3 
 
 97.6 
 2.4 
 
 51 
 34 
 
 60.0 
 40.0 
 
 241 
 40 
 
 85.8 
 
 
 14.2 
 
 Total 1 
 
 71 
 61 
 
 100.0 
 NAp 
 
 125 
 92 
 
 100.0 
 NAp 
 
 85 
 76 
 
 100.0 
 
 NAp 
 
 281 
 229 
 
 100.0 
 
 
 NAp 
 
 NAp Not applicable. 
 More than one successful extinguishing agent was used in some fires. 
 
 is significant only from 1968-77 to 1978- 
 84. The relative frequency of the cut- 
 and-fill also increased from 1950-67 to 
 1968-77. 
 
 To better explain the relative fire 
 hazard of the various types of mining 
 methods, it would be more appropriate to 
 relate the fires to the average number of 
 mines associated with each mining method 
 and time period. As no reliable data on 
 the average number of mines by mining 
 methods were available, this type of 
 analysis was not performed. 
 
 Successful Extinguishing Agent 
 
 Tables 23 and 24 delineate fires by 
 successful extinguishing agent for the 
 three time periods. Table 23 gives the 
 frequency with which an extinguishing 
 agent was successfully used. The fires 
 in which two agents were successfully 
 used were entered under both categories 
 in this table. Table 24 gives the bivar- 
 iate distribution of the fires, showing 
 which extinguishing agents were used 
 together. Note that in 1978-84, the 
 
23 
 
 TABLE 24. - Reported underground fires, joint distribution of successful 
 extinguishing agents, by three time periods 
 
 Period 1 
 
 Water 
 
 Dry chemical. . . 
 
 Burned out 
 
 Sealing 
 
 Other 
 
 Unspecified. . . . 
 l I: 1955-67; II 
 
 Water 
 
 II 
 
 III 
 
 Dry chemical 
 
 II 
 
 III 
 
 Burned out 
 
 II 
 
 III 
 
 Sealing 
 
 II 
 
 III 
 
 Other 
 
 II 
 
 III 
 
 Unspecified 
 
 II III 
 
 19 
 
 14 
 
 13 
 
 3 
 
 10 
 
 2 
 
 4 
 
 10 
 
 34 
 
 1968-77; III: 1978-84. 
 
 TABLE 25. - Reported surface fire incidents by year 
 
 
 Number of 
 
 
 Number of 
 
 
 Number of 
 
 Year 
 
 incidents 
 
 Year 
 
 incidents 
 
 Year 
 
 incidents 
 
 1950... 
 
 1 
 
 1962... 
 
 4 
 
 1974... 
 
 10 
 
 1951... 
 
 
 
 1963... 
 
 3 
 
 1975... 
 
 5 
 
 1952... 
 
 
 
 1964... 
 
 2 
 
 1976... 
 
 12 
 
 1953... 
 
 1 
 
 1965... 
 
 
 
 1977... 
 
 13 
 
 1954... 
 
 
 
 1966... 
 
 
 
 1978... 
 
 29 
 
 1955... 
 
 
 
 1967... 
 
 1 
 
 1979... 
 
 31 
 
 1956... 
 
 3 
 
 1968... 
 
 4 
 
 1980... 
 
 16 
 
 1957... 
 
 1 
 
 1969... 
 
 4 
 
 1981... 
 
 21 
 
 1958... 
 
 1 
 
 1970 
 
 4 
 
 1982 1 .. 
 
 5 
 
 1959... 
 
 1 
 
 1971 
 
 2 
 
 1983... 
 
 10 
 
 1960... 
 
 
 3 
 
 1972... 
 1973 
 
 6 
 5 
 
 1984... 
 Total 
 
 14 
 
 1961... 
 
 212 
 
 Statistics for 1982 in surface mining of stone, clay, col- 
 loidal phosphate, and sand and gravel are based on July through 
 December only. These operations were excluded from MSHA juris- 
 diction by House Joint Resolution 370, December 15, 1981. On 
 July 15, 1982, the Congress approved House Resolution 6685, 
 which restored MSHA's jurisdiction to all operations except 
 those of States or political subdivisions. 
 
 successful extinguishing agent was not 
 specified in about 45 pet (34/76) of the 
 fire reports. Where the extinguishing 
 agents were specified, it was found that 
 none of the standard statistical tests 
 were applicable, as the basic assumption 
 of independence was violated. That is, 
 water and dry chemicals were used more 
 frequently than other agents in fires in- 
 volving two agents. Hence, the following 
 conclusions are based on simple examina- 
 tion of the data: (1) in 1968-77, the 
 frequency with which two successful ex- 
 tinguishing agents were used is much 
 higher than in the other two periods, 
 (2) water was the predominant agent in 
 
 all three periods, (3) there was a defi- 
 nite shift from sealing to dry chemical, 
 and (4) the relative frequency of "Burned 
 out" was approximately the same through- 
 out the three periods. 
 
 SURFACE FIRES AT SURFACE 
 AND UNDERGROUND MINES 
 
 Time Trends 
 
 As with underground fires, the first 
 analysis looked at the time trends. Ta- 
 ble 25 lists 212 fire incidents by year 
 for surface fires. The number of fires 
 for 1982 is based only on July through 
 
24 
 
 December for stone, clay, colloidal 
 phosphate, and sand and gravel, because 
 these operations were excluded from 
 MSHA's jurisdiction from January through 
 June of that year. 
 
 It can be seen from table 25 and fig- 
 ure 5 that there were a higher number of 
 fires for which a report could be located 
 for the period beginning in 1972. This 
 fact is further confirmed by the Spearman 
 rank order correlation test based on ta- 
 ble 26. The highly significant r s value 
 of 0.855 indicates that, on the average, 
 the number of fires reported increased 
 with time. 
 
 Table 27 displays the number of fire 
 incidents by three time periods. The 
 chi-square for this table is computed on 
 the assumption of an equal number of 
 fires for each year. This highly signif- 
 icant value indicates, as shown in figure 
 
 2, that differences exist in the fre- 
 quency with which surface fires were re- 
 ported during the three time periods. As 
 with underground fires, this could be be- 
 cause prior to 1968 there was no record- 
 keeping requirement for accidents (in- 
 cluding fires) that occurred in metal and 
 nonmetal mines. Also, a significant per- 
 centage of reports of fire incidents dur- 
 ing this period (pre-1968) have been 
 misplaced over the years of shift and re- 
 organization of personnel and offices in 
 the Bureau and MSHA. 
 
 Significantly then, the trend as seen 
 in figures 2 and 5 is either toward in- 
 creased awareness (via reporting) of 
 fires in surface metal and nonmetal mines 
 or toward an increasingly hazardous envi- 
 ronment, or both. 
 
 Table 28 gives the surface incidence 
 rates for the period 1978-84. These 
 
 n , , n 
 
 $$$ 
 
 W-H 
 
 $_ 
 
 UM 
 
 IV 
 
 14 
 
 1 1> 
 
 / s / 
 
 W WW 
 
 / 
 
 / 
 / 
 / 
 
 ', 
 
 111 
 
 in 
 
 n\ 
 
 *** 
 
 
 www 
 
 1 1 
 
 J4LJ4 
 
 Figure 5. — Surface fire incidents by year, 1950-84. 
 
25 
 
 TABLE 26. - Ranking of reported surface fire incidents by year 
 
 Year 
 
 Number of 
 incidents 
 
 Rank 
 
 Year Incidents 
 
 Year 
 
 Number of 
 incidents 
 
 Rank 
 
 Year Incidents 
 
 1950. 
 1951. 
 1952. 
 1953. 
 1954. 
 1955. 
 1956. 
 1957. 
 1958. 
 1959. 
 1960. 
 1961. 
 1962. 
 1963. 
 1964. 
 1965. 
 1966. 
 1967. 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 10 
 
 11 
 
 12 
 
 13 
 
 14 
 
 15 
 
 16 
 
 17 
 
 18 
 
 10.5 
 
 4.0 
 
 4.0 
 
 10.5 
 
 4.0 
 
 4.0 
 
 17.0 
 
 10.5 
 
 10.5 
 
 10.5 
 
 4.0 
 
 17.0 
 
 20.5 
 
 17.0 
 
 14.5 
 
 4.0 
 
 4.0 
 
 10.5 
 
 1968 
 
 1969 
 
 1970 
 
 1971 
 
 1972 
 
 1973.... 
 
 1974 
 
 1975 
 
 1976 
 
 1977 
 
 1978 
 
 1979 
 
 1980 
 
 1981 
 
 1982 1 . ... 
 
 1983 
 
 1984 
 
 Total. 
 
 4 
 
 4 
 
 4 
 
 2 
 
 6 
 
 5 
 
 10 
 
 5 
 
 12 
 
 13 
 
 29 
 
 31 
 
 16 
 
 21 
 
 5 
 
 10 
 
 14 
 
 19 
 20 
 21 
 22 
 23 
 24 
 25 
 26 
 27 
 28 
 29 
 30 
 31 
 32 
 33 
 34 
 35 
 
 20.5 
 20.5 
 20.5 
 14.5 
 26.0 
 24.0 
 27.5 
 24.0 
 29.0 
 30.0 
 34.0 
 35.0 
 32.0 
 33.0 
 24.0 
 27.5 
 31.0 
 
 212 
 
 NAp 
 
 NAp 
 
 5FTT 
 
 = 0.855 (significant at 1-pct level), 
 surface mining of stone, clay, colloidal phosphate, and 
 on July through December only. These operations were ex- 
 ion by House Joint Resolution 370, December 15, 1981. On 
 ss approved House Resolution 6685, which restored MSHA's 
 ions except those of States or political subdivisions. 
 
 NAp Not applicable. r 
 Statistics for 1982 in 
 sand and gravel are based 
 eluded from MSHA jurisdict 
 July 15, 1982, the Congre 
 jurisdiction to all operat 
 
 TABLE 27. - Number of surface fire incidents by three time periods 
 
 Period 
 
 Ob- 
 served 
 
 Ex- 
 pected 
 
 Average 
 per year 
 
 Period 
 
 Ob- 
 served 
 
 Ex- 
 pected 
 
 Average 
 per year 
 
 I: 
 II: 
 
 1950-67 
 
 1968-77... 
 
 21 
 65 
 
 109 
 61 
 
 1.2 
 6.5 
 
 III: 1978-84. 
 Total or av. 
 
 126 
 
 42 
 
 18.0 
 
 212 
 
 212 
 
 6.1 
 
 239.31 (significant at 1-pct level). 
 
 TABLE 28. 
 1978-84 
 
 Reported surface fire incidence rates, 
 
 Year 
 
 Observed 
 number of 
 incidents 
 
 Hours 
 
 worked, 
 
 10 3 h 
 
 Incidence 
 rate 1 
 
 Expected 
 number of 
 incidents 
 
 1978 
 
 29 
 31 
 16 
 21 
 5 
 10 
 14 
 
 418,650 
 440,200 
 417,100 
 400,350 
 221,550 
 289,150 
 299,450 
 
 1.4 
 1.4 
 .8 
 1.0 
 .4 
 .7 
 .9 
 
 21 
 
 1979 
 
 22 
 
 1980 
 
 21 
 
 1981 
 
 20 
 
 1982 
 
 H 
 
 1983 
 
 15 
 
 1984 
 
 15 
 
 
 
 Total or IR 
 
 126 
 
 2,486,450 
 
 1.0 
 
 126 
 
 X 2 = 13.16 (significant at 5-pct level) 
 1 Per 20 million h worked. 
 
26 
 
 incidence rates are equivalent to the 
 number of fires per 10,000 full-time 
 workers (See "Data Analysis Methods.") 
 The chi-square values computed from the 
 observed and expected number of fires in- 
 dicate that not all years were equally 
 hazardous. Figure 3 shows years 1978 and 
 1979 to have the highest incidence rate. 
 
 Ore Type 
 
 Table 29 shows surface fires by ore 
 type for the three time periods. Fires 
 occurred most often at surface limestone 
 and iron mines , especially for the period 
 1978-84. The chi-square test for the pe- 
 riod 1968-77 indicates that the differ- 
 ences in ore totals, after the "Other" 
 category is eliminated are not signifi- 
 cant. This test was not performed for 
 
 the period 1950-67 because 11 fires among 
 4 different categories yield an expected 
 frequency of less than 3 fires per cate- 
 gory. An examination of the data for 
 this period shows that the frequency 
 of fires was the highest in iron ore. 
 According to the t-values, the only sig- 
 nificant change in the relative fre- 
 quency was in iron ore between 1950-67 
 and 1968-77. 
 
 A detailed analysis by ore type was 
 also performed for the most current peri- 
 od, 1978-84 (table 30). The chi-square 
 value of 12.71 is significant at the 
 5-pct level, indicating that all ore 
 types were not equally hazardous. The 
 relative measure of incidence rates as 
 depicted in figure 6 shows iron to be the 
 most hazardous and sand and gravel to be 
 the least hazardous. 
 
 TABLE 29. - Reported surface fires by principal ore, three time periods 
 
 
 Period I 
 1950-67 
 
 Peri 
 
 od II 
 
 Period III 
 
 Total 
 
 It| 
 
 Ore 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 1 
 7 
 3 
 
 
 
 7 
 
 5.6 
 
 38.9 
 
 16.7 
 
 .0 
 
 38.9 
 
 9 
 5 
 5 
 5 
 31 
 
 16.4 
 9.1 
 9.1 
 9.1 
 
 56.4 
 
 24 
 
 18 
 
 9 
 
 9 
 
 64 
 
 19.4 
 
 14.5 
 
 7.3 
 
 7.3 
 
 51.6 
 
 34 
 
 30 
 
 17 
 
 14 
 
 102 
 
 17.3 
 
 15.2 
 
 8.6 
 
 7.1 
 
 51.8 
 
 1.16 
 2.96** 
 
 .89 
 
 1.33 
 
 ND 
 
 0.48 
 1.00 
 
 
 .42 
 
 
 .42 
 ND 
 
 Total specified. . . . 
 
 18 
 
 100.0 
 
 55 
 
 100.0 
 
 124 
 
 100.0 
 
 197 
 
 100.0 
 
 
 
 
 18 
 3 
 
 85.7 
 14.3 
 
 55 
 
 10 
 
 84.6 
 15.4 
 
 124 
 2 
 
 98.4 
 1.6 
 
 197 
 15 
 
 92.9 
 7.1 
 
 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 
 
 
 X 2 : Specified without 
 
 ND 
 
 2.00 
 
 10.80* 
 
 
 
 
 ND Not determined. 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 TABLE 30. - Average reported surface fire incidence rates 
 by ore, 1978-84 
 
 Ore 
 
 Observed 
 number of 
 incidents 
 
 Hours 
 
 worked, 
 
 10 3 h 
 
 Incidence 
 rate 1 
 
 Expected 
 number of 
 incidents 
 
 
 18 
 
 24 
 
 9 
 
 9 
 
 66 
 
 198,350 
 458,400 
 236,850 
 362,400 
 1,230,450 
 
 1.8 
 
 1.0 
 
 .8 
 
 .5 
 
 1.1 
 
 10 
 
 Crushed limestone 
 Sand and gravel. . 
 
 23 
 12 
 19 
 62 
 
 
 
 Total or IR 
 
 26 
 
 2,486,450 
 
 1.0 
 
 126 
 
 ,2 _ 
 
 = 12.71 (significant at 5-pct level). . 
 : Per 20 million h worked. 
 2 Includes 2 incidents in unspecified ores. 
 
27 
 
 1.B-- 
 
 1.6-- 
 
 < 
 
 DC 
 
 111 i 
 
 U 
 
 z 
 111 
 
 Q 
 
 H -B + 
 
 u 
 
 Z 
 
 H 
 
 .6- 
 
 .4-- 
 
 Average (all types)- 
 
 Iron Crushed limestone Copper Sand and gravel Other 
 
 Figure 6. — Average surface fire incidence rates (based on 20 million h worked), by ore type, 1978-84. 
 
 Ignition Source 
 
 Table 31 shows fires by ignition source 
 for the three time periods. The number 
 of unspecified fires comprises a major 
 portion of the total fires, especially 
 for the 1978-84 period, thus raising some 
 questions on the completeness of the 
 data. The chi-square values with and 
 without the "Other" category indicate 
 that the different ignition source to- 
 tals are significantly different for the 
 time periods 1968-77 and 1978-84. No 
 chi-square test was performed for the 
 first period as there were too few obser- 
 vations. It is apparent, however, that 
 the leading ignition source for this pe- 
 riod was electrical. The t-tests also 
 show that the proportion of engine heat 
 
 fires increased from 1968-77 to 1978-84 
 while the proportion of electrical fires 
 declined over the same time. 
 
 Burning Substance 
 
 The data on surface fires by burning 
 substance for the three time periods are 
 shown in table 32. In analyzing the data 
 of this table, the same basic assumption 
 was made on independence of burning sub- 
 stances as outlined for its counterpart 
 table 8 on underground fires. The chi- 
 square test shows that the differences in 
 the frequency with which the various sub- 
 stances were involved in fires were sig- 
 nificant for the periods 1968-77 and 
 1978-84 but not for the period 1950-67. 
 The relative frequency of combustible 
 
28 
 
 TABLE 31. - Reported surface fires by ignition source, three time periods 
 
 
 Period I 
 1950-67 
 
 Peri 
 
 od II 
 
 Period III 
 
 Total 
 
 It| 
 
 Ignition source 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 1 
 6 
 3 
 1 
 8 
 
 5.3 
 31.6 
 15.8 
 
 5.3 
 
 42.1 
 
 9 
 15 
 11 
 
 1 
 20 
 
 16.1 
 
 26.8 
 
 19.6 
 
 1.8 
 
 35.7 
 
 38 
 16 
 22 
 8 
 18 
 
 37.3 
 15.7 
 21.6 
 
 7.8 
 17.6 
 
 48 
 37 
 36 
 10 
 46 
 
 27.1 
 20.9 
 20.3 
 5.6 
 26.0 
 
 1.20 
 
 .40 
 
 .37 
 
 ND 
 
 ND 
 
 2.79** 
 
 
 1.68* 
 
 
 ND 
 
 
 ND 
 
 
 ND 
 
 
 
 Total specified. . . . 
 
 19 
 
 100.0 
 
 56 
 
 100.0 
 
 102 
 
 100.0 
 
 177 
 
 100.0 
 
 
 
 
 19 
 
 2 
 
 90.5 
 9.5 
 
 56 
 9 
 
 86.2 
 13.8 
 
 102 
 24 
 
 81.0 
 19.0 
 
 177 
 35 
 
 83.5 
 16.5 
 
 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 
 X 2 : 
 
 ND 
 
 ND 
 
 17. QT** 
 
 24.fiR** 
 
 
 
 
 11. 
 
 56** 
 
 23. 
 
 05** 
 
 
 ND Not determined. Significant at 5-pct level. Significant at 1-pct level. 
 TABLE 32. - Reported surface fires by burning substance, 1 three time periods 
 
 
 Period I 
 1950-67 
 
 Period II 
 
 Period III 
 
 Total 
 
 It| 
 
 Burning substance 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 Combustible liquids... 
 Construction material. 
 
 4 
 5 
 5 
 4 
 11 
 7 
 
 11.1 
 13.9 
 13.9 
 11.1 
 30.6 
 19.4 
 
 24 
 19 
 11 
 9 
 6 
 16 
 
 28.2 
 22.4 
 12.9 
 10.6 
 7.1 
 18.8 
 
 70 
 
 25 
 
 8 
 
 11 
 
 3 
 
 9 
 
 55.6 
 19.8 
 6.3 
 8.7 
 2.4 
 7.1 
 
 98 
 49 
 24 
 24 
 20 
 32 
 
 39.7 
 
 19.8 
 
 9.7 
 
 9.7 
 
 8.1 
 
 13.0 
 
 2.04* 
 1.07 
 
 ND 
 
 ND 
 3.40** 
 
 ND 
 
 3.92** 
 
 ND 
 1.64 
 
 
 ND 
 
 
 1.65* 
 
 
 ND 
 
 
 
 Total specified.... 
 
 36 
 
 100.0 
 
 85 
 
 100.0 
 
 126 
 
 100.0 
 
 247 
 
 100.0 
 
 
 
 
 36 
 
 
 100.0 
 .0 
 
 85 
 1 
 
 98.8 
 1.2 
 
 126 
 9 
 
 93.3 
 6.7 
 
 247 
 10 
 
 96.1 
 3.9 
 
 
 
 
 
 
 
 36 
 
 100.0 
 
 86 
 
 100.0 
 
 135 
 
 100.0 
 
 257 
 
 100.0 
 
 
 
 
 X 2 : 
 
 6. 
 
 00NS 
 ND 
 
 16. 1 * l 
 
 150-1Q** 
 
 
 
 
 Without combus- 
 tible liquids. . . . 
 
 h 
 
 ID 
 
 24 
 
 .36* 
 
 
 ND Not determined. 
 
 Significant at 5-pct level. **Signif icant at 1-pct level. 
 1 In many fires, more than 1 substance was burning. 
 
 liquids fires definitely increased across 
 time periods, while that of timber 
 decreased. 
 
 Location 
 
 The locations of fires in surface metal 
 and nonmetal mines appear in table 33. 
 For the first two time periods, the over- 
 whelming majority of fires occurred in 
 
 the surface building location, while for 
 the last period, the split was about SO- 
 SO between the surface building and a 
 surface location other than the building. 
 
 Equipment Involved 
 
 Table 34 shows surface fires by equip- 
 ment for the three time periods. As with 
 its counterpart table 12 on underground 
 
29 
 
 TABLE 33. - Reported surface fires by location, three time periods 
 
 E 
 
 Location 
 
 Period I 
 1950-67 
 
 No. 
 
 Pet 
 
 Period II 
 1968-77 
 
 No. 
 
 Pet 
 
 Period III 
 1978-84 
 
 No. 
 
 Pet 
 
 Total 
 
 No. 
 
 Pet 
 
 I vs. 
 
 II 
 
 II vs. 
 
 Ill 
 
 Surface building. . 
 
 Surface 
 
 Other 
 
 Total specified 
 
 Specified 
 
 Unspecified 
 
 Total 
 
 ND Not determined 
 
 16 
 
 
 4 
 
 80.0 
 
 .0 
 
 20.0 
 
 44 
 
 
 
 21 
 
 67.7 
 
 .0 
 
 32.3 
 
 54 
 
 67 
 
 
 
 44.6 
 
 55.4 
 
 .0 
 
 114 
 67 
 25 
 
 53.3 
 32.5 
 12.1 
 
 1.06 
 ND 
 
 ND 
 
 3.00 
 7.50* 
 ND 
 
 20 
 
 100.0 
 
 65 
 
 100.0 
 
 121 
 
 100.0 
 
 206 
 
 100.0 
 
 20 
 1 
 
 95.2 
 4.8 
 
 65 
 
 
 100.0 
 .0 
 
 121 
 5 
 
 96.0 
 4.0 
 
 206 
 6 
 
 97.2 
 2.8 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 Significant at 1-pct level. 
 
 TABLE 34. 
 
 Reported surface fires by equipment involved, three time periods 
 
 
 Period I 
 1950-67 
 
 Period II 
 
 Period III 
 
 Total 
 
 It| 
 
 Equipment 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 
 
 1 
 
 2 
 
 14 
 
 0.0 
 
 5.9 
 
 11.8 
 
 82.4 
 
 8 
 
 4 
 
 6 
 
 18 
 
 22.2 
 11.1 
 16.7 
 50.0 
 
 51 
 9 
 
 4 
 17 
 
 63.0 
 
 11.1 
 
 4.9 
 
 21.0 
 
 59 
 14 
 12 
 49 
 
 44.0 
 
 10.4 
 
 9.0 
 
 36.6 
 
 2.11* 
 
 .61 
 
 .47 
 
 ND 
 
 4.07** 
 
 
 ND 
 2.09* 
 
 
 ND 
 
 Total specified.... 
 
 17 
 
 100.0 
 
 36 
 
 [ 100.0 
 
 81 
 
 100.0 
 
 134 
 
 100.0 
 
 
 
 Unspecif ied-unknown- 
 
 17 
 
 4 
 
 81.0 
 19.0 
 
 36 
 29 
 
 55.4 
 44.6 
 
 81 
 45 
 
 64.3 
 35.7 
 
 134 
 78 
 
 63.2 
 36.8 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 
 
 
 ND Not determined. 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 The reporting mechanism did not permit the classification of fires into three sep- 
 
 arate categories. 
 
 fires, changes in the reporting mechanism 
 for the "Unspecified," "None," and "Un- 
 known" categories make it impossible to 
 obtain a separate fire count for each 
 category. The combination of these three 
 categories comprises a major portion of 
 the total fires for all the time periods. 
 Where equipment is involved, in recent 
 time periods, more fires were associated 
 with the mobile equipment type. 
 
 Means of Detection 
 
 Means of fire detection appear in table 
 35. In the most recent time period, the 
 means of detection was not specified for 
 a large portion (21.4 pet) of the fires, 
 thus raising some questions as to the 
 completeness of the data. Of the spe- 
 cified means of detection, almost all 
 were operators and/or workers. The 
 
 relative frequency of fires discovered 
 immediately increased while that of fires 
 not immediately discovered decreased. 
 This is further substantiated by the next 
 table, which classifies fires by dura- 
 tion. Hence, either operators and work- 
 ers were becoming incresingly aware of 
 fires and discovering them at an early 
 stage, or for some reason there were more 
 fires under 30 min being reported in the 
 most recent time period. 
 
 Duration 
 
 Duration of surface fires appears in 
 table 36. Duration is not specified for 
 a large portion of fires for all the pe- 
 riods; the analysis of this table is thus 
 limited to those fires where the duration 
 is specified. As there were too few spe- 
 cified fires in the period 1950-67, a 
 
30 
 
 TABLE 35. - Reported surface fires by means of detection, three time periods 
 
 
 Per 
 
 iod I 
 
 Period II 
 
 Period III 
 
 Total 
 
 It| 
 
 Means of detection 
 
 1950- 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 Operator-worker 
 Workers (not 
 
 7 
 
 8 
 
 4 
 
 36.8 
 
 42.1 
 21.1 
 
 29 
 
 26 
 
 7 
 
 46.8 
 
 41.9 
 11.3 
 
 84 
 
 12 
 3 
 
 84.8 
 
 12.1 
 3.0 
 
 120 
 
 46 
 14 
 
 66.7 
 
 25.6 
 7.8 
 
 0.76 
 
 ND 
 
 ND 
 
 5.14** 
 4.33** 
 
 
 ND 
 
 
 
 Total specified.... 
 
 19 
 
 100.0 
 
 62 
 
 100.0 
 
 99 
 
 100.0 
 
 180 
 
 100.0 
 
 
 
 
 19 
 2 
 
 90.5 
 9.5 
 
 62 
 3 
 
 95.4 
 4.6 
 
 99 
 27 
 
 78.6 
 21.4 
 
 180 
 32 
 
 84.9 
 15.1 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 
 .. ..... — ... . ... _.. . — j 
 
 ND Not determined. 
 
 *Sign 
 
 ificant 
 
 at 1 
 
 -pet level. 
 
 
 
 
 
 
 TABLE 36. - Reported surface fires by duration, three time periods 
 
 
 Per 
 
 iod I 
 
 Period II 
 
 Period III 
 
 Total 
 
 It| 
 
 Duration, h 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 to 0.5 
 
 3 
 
 1 
 3 
 4 
 
 
 27.3 
 
 9.1 
 
 27.3 
 
 36.4 
 
 .0 
 
 16 
 
 10 
 8 
 2 
 3 
 
 41.0 
 
 25.6 
 
 20.5 
 
 5.1 
 
 7.7 
 
 29 
 
 10 
 
 6 
 
 2 
 
 2 
 
 59.2 
 
 20.4 
 
 12.2 
 
 4.1 
 
 4.1 
 
 48 
 
 21 
 
 17 
 
 8 
 
 5 
 
 48.5 
 
 21.2 
 
 17.2 
 
 8.1 
 
 5.1 
 
 0.83 
 1.17 
 
 .84 
 2.81** 
 
 .95 
 
 1.69* 
 
 0.5 to 1 
 
 .58 
 
 1+ to 4 
 
 1.05 
 
 4+ to 24 
 
 ND 
 
 24 + 
 
 .73 
 
 
 
 Total specified. ... 
 
 11 
 
 100.0 
 
 39 
 
 100.0 
 
 49 
 
 100.0 
 
 99 
 
 100.0 
 
 
 
 Specified 
 
 11 
 
 10 
 
 52.4 
 47.6 
 
 39 
 26 
 
 60.0 
 40.0 
 
 49 
 77 
 
 38.9 
 61.1 
 
 99 
 113 
 
 46.7 
 53.3 
 
 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 
 
 
 
 ND 
 
 * * 
 
 16.51 
 
 51.51** 
 
 
 
 Significant at 5-pct level. 
 
 Significant at 1-pct level. 
 
 chi-square test was not performed. For 
 the remaining two time periods, the num- 
 ber of fires is not evenly distributed 
 across the fire duration categories. As 
 mentioned previously, the relative fre- 
 quency of fires under 30 min has in- 
 creased in recent times. Also, the pro- 
 portion of fires was significantly down 
 in the 4- to 24-h category between 1950- 
 67 and 1968-77. 
 
 Number of Injuries 
 
 Table 37 shows surface fires by number 
 of injuries for the three time periods. 
 In the first two time periods, approxi- 
 mately 85 pet of the fires involved no 
 injuries, whereas in the most recent time 
 period, this was true for only about 45 
 pet of the cases. In recent times there 
 
 has definitely been a substantial in- 
 crease in the proportion of fires being 
 reported with one injury. This means ei- 
 ther more workers have been sustaining 
 fire injuries or changes in law and/or 
 attitude have occurred that are affecting 
 the reporting; that is, injuries that 
 were previously not reported were being 
 reported in the most recent time period. 
 
 Number of Deaths 
 
 Table 38 contains fires by number of 
 deaths. The only significant result, 
 fortunately, is the overall reporting 
 frequency of nonfatal fires. Once the 
 nonfatal fires are removed, the only re- 
 maining category is fires involving one 
 death. 
 
31 
 
 TABLE 37. - Reported surface fires by number of injuries, three time periods 
 
 
 Per 
 
 iod I 
 
 Period II 
 
 Peri 
 
 od III 
 
 Total 
 
 It| 
 
 Number of injuries 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pct 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 
 18 
 2 
 
 1 
 
 
 
 85.7 
 9.5 
 
 .0 
 4.8 
 
 .0 
 
 54 
 7 
 4 
 
 
 
 83.1 
 
 10.8 
 
 6.2 
 
 .0 
 
 .0 
 
 56 
 
 65 
 
 4 
 
 1 
 
 
 
 44.4 
 
 51.6 
 
 3.2 
 
 .8 
 
 .0 
 
 128 
 
 74 
 
 8 
 
 2 
 
 
 
 60.4 
 
 34.9 
 
 3.8 
 
 .9 
 
 .0 
 
 ND 
 ND 
 ND 
 ND 
 ND 
 
 5.12** 
 
 1 
 
 5.51 ** 
 
 2 to 5 
 
 ND 
 
 ND 
 ND 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 NAp 
 
 NAp 
 
 NAp Not applicable. 
 
 ND Not determined. 
 
 Significant 
 
 at 1- 
 
 pet level. 
 
 
 TABLE 38. - Reported surface fires by number of deaths, 
 three time periods 
 
 Number o 
 
 f deaths 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 Total 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 
 
 18 
 3 
 
 
 85.7 
 
 14.3 
 
 .0 
 
 59 
 6 
 
 
 90.8 
 
 9.2 
 
 .0 
 
 123 
 3 
 
 
 97.6 
 
 2.4 
 
 .0 
 
 200 
 
 12 
 
 
 
 94.3 
 
 1 
 
 5.7 
 
 2 + 
 
 .0 
 
 
 
 
 
 21 
 
 100.0 
 
 65 
 
 100.0 
 
 126 
 
 100.0 
 
 212 
 
 100.0 
 
 TABLE 39. - Reported surface fires by successful extinguishing agent, 1 
 three time periods 
 
 
 Period I 
 
 Peri 
 
 od II 
 
 Period III 
 
 Total 
 
 It| 
 
 Extinguishing agent 
 
 1950-67 
 
 1968-77 
 
 1978-84 
 
 No. 
 
 Pet 
 
 I vs. 
 II 
 
 II vs. 
 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 No. 
 
 Pet 
 
 Ill 
 
 
 10 
 9 
 1 
 2 
 
 45.5 
 
 40.9 
 
 4.5 
 
 9.1 
 
 29 
 
 11 
 
 10 
 
 9 
 
 49.2 
 18.6 
 16.9 
 15.3 
 
 14 
 4 
 6 
 4 
 
 50.0 
 14.3 
 21.4 
 14.3 
 
 53 
 
 24 
 17 
 15 
 
 48.6 
 22.0 
 15.6 
 13.8 
 
 ND 
 2.07* 
 1.45 
 
 ND 
 
 ND 
 
 
 ND 
 
 
 0.50 
 
 
 ND 
 
 Total specified. . . . 
 
 22 
 
 100.0 
 
 lJ9 
 
 100.0 
 
 28 
 
 100.0 
 
 109 
 
 100.0 
 
 
 
 
 22 
 2 
 
 91.7 
 8.3 
 
 59 
 11 
 
 84.3 
 15.7 
 
 28 
 100 
 
 21.9 
 78.1 
 
 109 
 113 
 
 49.1 
 50.9 
 
 
 Total 1 
 
 24 
 21 
 
 100.0 
 NAp 
 
 70 
 65 
 
 100.0 
 NAp 
 
 128 
 126 
 
 100.0 
 NAp 
 
 222 
 212 
 
 100.0 
 NAp 
 
 
 
 
 
 11.82 
 
 18.49** 
 
 9.71* 
 
 
 
 
 ND Not determined. 
 Significant at 5-pct level. 
 
 Significant at 1-pct level, 
 ^ore than 1 successful extinguishing agent used in some fires. 
 
 Successful Extinguishing Agent agent 
 
 (100/ 
 
 Table 39 gives the frequency with which speci 
 
 an extinguishing agent was successfully five, 
 
 used in each of the three time periods. perio 
 
 The fires in which two agents were sue- tingu 
 
 cessfully used were entered under both these 
 
 categories in this table. In period table 
 
 1978-84, the successful extinguishing manne 
 
 was not specified in about 80 pet 
 
 126) of the reported fires. Of the 
 
 fied fires, there were only three, 
 
 and two fires, respectively, in 
 
 ds I, II, and III that used two ex- 
 
 ishing agents successfully. Since 
 
 members are relatively small, this 
 
 is not analyzed in the same 
 
 r as its counterpart table 23 on 
 
32 
 
 underground fires. Instead, the standard 
 tests were performed. The data show wa- 
 ter to be consistently successful in 
 about 50 pet of the specified fires for 
 all three periods. The burned-out method 
 has been used less frequently since 1968. 
 
 COMPARISON OF UNDERGROUND 
 AND SURFACE FIRES 
 
 Time Trends 
 
 Ignition Source 
 
 Electrical equipment was the leading 
 ignition source of undergropund fires. 
 For surface fires, engine heat was the 
 leading ignition source, followed by 
 electrical equipment (tables 7 and 31). 
 For both underground and surface, the 
 proportion of electrical fires decreased 
 while the proportion of engine heat fires 
 increased between 1968-77 and 1978-84. 
 
 Although the total number of fire inci- 
 dents was approximately the same for un- 
 derground and surface, the distribution 
 of these incidents across the three time 
 periods was more variable for surface 
 than for underground (fig. 2). The aver- 
 age number of surface fires per year for 
 which a report could be located ranged 
 from 1.2 in 1950-67 to 18.0 in 1978-84. 
 For underground fires the range was from 
 3.4 to 10.9. Figure 2 also shows that 
 for the most recent time period there 
 were, on the average, 18.0 surface fires 
 reported per year versus 10.9 for under- 
 ground. Hence, it would appear that in 
 the recent time period the surface area 
 is more hazardous than the underground 
 area. However, this is not the case, as 
 can be seen from figure 3. This is be- 
 cause many more hours were worked in the 
 surface area than in the underground 
 area. Figure 3 also shows 1978 to be the 
 most hazardous year. 
 
 Burning Substance 
 
 As indicated in tables 8 and 32, more 
 than one substance was burning in many 
 fires. Combustible liquids were the pre- 
 dominant burning substance for both un- 
 derground and surface fires in the period 
 1978-84. Also, the relative frequency of 
 fires involving combustible liquids in- 
 creased from one period to the next for 
 both underground and surface fires. The 
 proportion of fires involving timber, on 
 the other hand, dropped sharply between 
 1950-67 and 1968-77. 
 
 Location 
 
 Since the major categories for location 
 (tables 9 and 33) differ vastly for un- 
 derground and surface fires, they are not 
 compared. 
 
 Equipment Involved 
 
 Ore Type 
 
 Of the major ore types, copper and iron 
 were the only ones involved in both un- 
 derground fires and surface fires. Over 
 the entire period, copper was the princi- 
 pal ore involved in underground fires and 
 was third in surface fires. Of the spe- 
 cified ore types, copper accounted for 
 about 18.0 pet of the underground fires 
 and 8.6 pet of the surface fires (tables 
 5 and 29) . In terms of incidence rates 
 for the 1978-84 period, it ranks low for 
 both underground and surface fires (fig- 
 ures 4 and 6). 
 
 The predominant equipment involved in 
 both underground and surface fires was of 
 the mobile type. Figure 7 shows a com- 
 parison of mobile equipment fires in 
 surface and underground mines over the 
 three time periods. The relative fre- 
 quency of fires involving this equipment 
 type increased from one period to the 
 next for both surface and underground 
 fires. 
 
 Means of Detection 
 
 Most fires incidents were detected by 
 the operators and/or workers, both in 
 
33 
 
 u 
 
 Q 
 
 20 - 
 
 KEY 
 
 Underground 
 
 1950-67 1968-77 1978-84 
 
 Figure 7. — Percentage of mobile equipment fires, underground versus surface, during three time periods. 
 
 surface and underground. Also, data in 
 tables 14 and 35 show that, on the whole, 
 the proportion of fires discovered 
 immediately increased with each time 
 period (fig. 8). 
 
 Duration 
 
 For the period 1950-67, about 55 pet of 
 the reported underground fires lasted 
 longer than 24 h. There were no surface 
 fires reported in this category for the 
 same period (tables 15 and 36). The per- 
 centage of underground fires reported in 
 the 0- to 0.5-h category was approximate- 
 ly the same across the three time peri- 
 ods, while for surface fires there was 
 definitely an increase between 1968-77 
 and 1978-84. 
 
 Number of Injuries 
 
 As indicated by tables 17 and 37, the 
 majority of underground and surface fires 
 did not cause injuries. However, for the 
 most recent time period, this was not 
 true in the case of surface fires. In 
 this case, only 44 pet of the fires had 
 no injuries, 52 pet had one injury, and 
 4 pet had two or more injuries. Also, 
 for both underground and surface, there 
 were no fires resulting in 10 or more 
 injuries. 
 
 Number of Deaths 
 
 Both in underground and surface, ap- 
 proximately 95 pet of the fires were non- 
 fatal (tables 21 and 38). 
 
34 
 
 r-. 90 
 
 
 
 o 
 
 ** 80 
 
 u 
 
 Q 
 
 50 - 
 
 30 
 
 20-- 
 
 KEY 
 
 Undaroround 
 
 Surf aca 
 
 1950-67 19BB-77 1978-84 
 
 Figure 8. — Percentage of fires detected immediately, underground versus surface, during three time periods. 
 
 Successful Extinguishing Agent 
 
 As can be seen in tables 23 and 39, wa- 
 ter was the predominant successful agent 
 used to extinguish both underground and 
 
 surface fires. Also, in both underground 
 and surface fires, two extinguishing 
 agents were successfully used in some of 
 the fires. 
 
 NONREPORTABLE FIRES 
 
 Since 1968, MSHA regulations have re- 
 quired that fires lasting 1/2 h or more 
 or involving an injury must be reported. 
 Each reported fire is investigated by an 
 MSHA inspector who then prepares a fire 
 report. Noninjury fires under 1/2-h dur- 
 ation are not reportable. 
 
 To develop a data base that represented 
 mine fires in general, 16 local and re- 
 gional safety directors in charge of 12 
 
 surface mines and 12 underground mines 
 were interviewed to obtain descriptions 
 of nonreportable fires at their mines 
 over the last 5 yr. Internal company 
 memos on these fires were also provided. 
 Nonreportable data were analyzed in two 
 sets: under the headings of underground 
 fires and surface fires. A total of 20 
 underground (table 40) and 22 surface 
 (table 41) fire reports were obtained. 
 
35 
 
 TABLE 40. - Nonrepor table underground fires 
 (Based on sample of 20 reports) 
 
 Ignition source: 
 
 Electrical 
 
 Welding 
 
 Friction 
 
 Engine heat 
 
 Spontaneous combustion. 
 
 Unspecified 
 
 By burning substance: 1 
 
 Combustible liquids.... 
 
 Wiring insulation 
 
 Timber 
 
 Rubber hose 
 
 By location: 
 
 Haulageway-drif t 
 
 Substation < 
 
 Shaft 
 
 Mined-out area 
 
 Number 
 
 10 
 4 
 3 
 1 
 1 
 1 
 
 By location — Con. 
 
 Working face 
 
 Other 
 
 Unspecified 
 
 By equipment involved: 
 
 Mobile 
 
 Electrical 
 
 Maintenance-shop 
 
 None 
 
 Unspecified-unknown 
 
 By successful extinguishing 
 agent: 2 
 
 Dry chemical 
 
 Cut off electrical power. 
 
 Water 
 
 Seals 
 
 Other 
 
 *More than 1 burning substance involved in some fires. 
 
 2 More than 1 successful extinguishing agent used in some fires. 
 
 TABLE 41. - Nonreportable surface fires 
 (Based on sample of 22 reports) 
 
 Ignition source: 
 
 Welding 
 
 Electrical 
 
 Engine heat , 
 
 Friction , 
 
 Other , 
 
 Unspecified 
 
 By burning substance: 1 
 Combustible liquids.. 
 
 Insulation , 
 
 Rubber , 
 
 Other , 
 
 Unspecified 
 
 By location: 
 
 Surface 
 
 Surface building. . . . 
 Unspecified 
 
 Number 
 
 5 
 4 
 3 
 2 
 2 
 6 
 
 12 
 3 
 2 
 4 
 2 
 
 14 
 5 
 3 
 
 By equipment involved: 
 
 Mobile 
 
 Conveyor 
 
 Electrical 
 
 None 
 
 Unspecified 
 
 By successful extinguishing 
 agent: 2 
 
 Dry chemical 
 
 Water 
 
 Burned out 
 
 Other 
 
 Unspecified 
 
 Number 
 
 1 
 1 
 4 
 
 15 
 
 5 
 4 
 1 
 2 
 
 Number 
 
 14 
 2 
 1 
 1 
 4 
 
 13 
 1 
 1 
 5 
 3 
 
 More than 1 burning substance involved in some fires. 
 2 More than 1 successful extinguishing agent used in some fires. 
 
36 
 
 UNDERGROUND FIRES 
 
 Ignition Source 
 
 Most of the nonreportable underground 
 fires were electrical in origin, pri- 
 marily from wiring shorts. The sec- 
 ond highest ignition sources were weld- 
 ing sparks or hot slag and friction, 
 primarily from welding near combustibles 
 and overheating brakes, respectively. 
 
 Burning Substance 
 
 Combustible liquids, wiring insulation, 
 and timber were nearly equal in frequency 
 of involvement. 
 
 Location 
 
 As with reported fires, the most fre- 
 quent location of nonreportable fires was 
 along haulageways and in drift entries 
 where activity is high. Nonreportable 
 fires also occurred with a relative high 
 frequency at substations or switchboxes. 
 
 Equipment Involved 
 
 Unlike the reported fires, it was pos- 
 sible to separate the fires which in- 
 volved no equipment from those that were 
 unspecified or unknown. Fires where no 
 equipment was involved usually resulted 
 from welding or cutting, where combusti- 
 bles such as timber or grease and oil 
 were ignited. Where equipment was in- 
 volved, the most frequent type was mo- 
 bile, as was the case with reported 
 fires. 
 
 Successful Extinguishing Agent 
 
 Unlike the reported fires, the most 
 frequently used agent here was one or 
 more hand-portable dry chemical extin- 
 guishers. Also, water and the local 
 
 removal of electric power were used with 
 relative frequency. 
 
 SURFACE FIRES 
 
 Ignition Source 
 
 Welding, electrical, and engine heat 
 were frequently cited as ignition 
 sources. Welding sparks can ignite 
 grease or oil that has accumulated on 
 vehicles. Electrical fires are generally 
 from wiring shorts that ignite diesel 
 fuel or hydraulic fluid. Engine heat can 
 ignite hydraulic fluid or diesel fuel 
 from a ruptured line. 
 
 Burning Substances 
 
 As with reported fires, combustible 
 liquids were most frequently involved. 
 
 Location 
 
 Most surface fires occurred along haul- 
 age roads or in the pit area on mobile 
 machinery. 
 
 Equipment Involved 
 
 An overwhelming majority of nonreport- 
 able surface fires involved mobile ma- 
 chinery such as haulage trucks, front-end 
 loaders, and flatbed welding trucks. 
 These findings are consistent with those 
 from fires that were reported. 
 
 Successful Extinguishing Agent 
 
 Most of these fires were extinguished 
 by hand-portable dry chemical extin- 
 guishers and one automatic fire suppres- 
 sion system. This finding is quite dif- 
 ferent from that for reported fires where 
 water is consistently the most success- 
 ful extinguishing agent used. 
 
 OPINION DATA FROM MINE SAFETY DIRECTORS 
 
 When gathering data on rare events such 
 as fires, it is often advisable to ask 
 knowledgeable individuals to give hazard 
 opinion data. Consequently, after the 
 
 mine safety directors related their non- 
 reportable fire data, they were asked to 
 rank various ignition sources, burning 
 substances, successful extinguishing 
 
37 
 
 agents, and equipment on their relative 
 frequencies of occurrence in nonreport- 
 able fires. Because of the size of the 
 opinion data sample, it is difficult to 
 specify with any certainty the reliabil- 
 ity. These data are presented as a sam- 
 ple of the metal and nonmetal fire ex- 
 perience of the group of mine safety 
 personnel. 
 
 UNDERGROUND MINE FIRES 
 
 Table 42 contains ranked opinions of 
 ignition sources, burning substances, and 
 successful extinguishing agents for un- 
 derground nonreportable fires. Each item 
 is listed according to the average rank 
 assigned to it by the safety direc- 
 tors interviewed. The top three ignition 
 sources were welding sparks or hot slag, 
 electrical arcing, and friction. The top 
 two burning substances were combustible 
 liquids and insulation. The top two suc- 
 cessful extinguishing agents were dry 
 chemicals and water. 
 
 SURFACE MINE FIRES 
 
 Table 43 contains ignition sources, 
 burning substances, and successful extin- 
 guishing agents for surface nonreportable 
 fires. Each item is listed according to 
 the average rank assigned to it by the 
 safety directors interviewed. The top 
 three ignition sources were welding 
 sparks or slag, engine heat, and elec- 
 trical arcing. The top two burning sub- 
 stances were combustible liquids and 
 rubber. The top two successful extin- 
 guishing agents again were dry chemicals 
 and water. 
 
 SUMMARY OF OPINION DATA 
 
 After the safety directors were inter- 
 viewed concerning nonreportable fires, a 
 consensus of general opinions became ap- 
 parent. The primary causes of nonreport- 
 able underground fires were poor mainte- 
 nance of electrical equipment and poor 
 housekeeping, coupled with a lack of 
 
 TABLE 42. - Average rankings of opinion data for ignition source, 
 burning substance, and successful extinguishing agent, 
 underground fires 
 
 Rank 
 
 Ignition source 
 
 Rank 
 
 Burning substance 
 
 Rank 
 
 Extinguishing agent 
 
 1... 
 
 Welding sparks or 
 
 1... 
 
 Combustible liquids. 
 
 1... 
 
 Dry chemical. 
 
 
 slag. 
 
 2... 
 
 Wiring insulation. 
 
 2... 
 
 Water. 
 
 2... 
 
 Electrical. 
 
 3... 
 
 Rubber (hose or 
 
 3... 
 
 Other. 
 
 3... 
 
 Friction. 
 
 
 belt). 
 
 
 
 4... 
 
 Engine heat. 
 
 4... 
 
 Timber, lagging, 
 
 
 
 5... 
 
 Spontaneous 
 
 
 etc. 
 
 
 
 
 combustion. 
 
 5... 
 
 Other. 
 
 
 
 6... 
 
 Other. 
 
 
 
 
 
 TABLE 43. - Average rankings of opinion data for ignition source, 
 burning substance, and successful extinguishing agent, 
 surface fires 
 
 Rank 
 
 Ignition source 
 
 Rank 
 
 Burning substance 
 
 Rank 
 
 Extinguishing agent 
 
 1... 
 
 Welding sparks or 
 
 1... 
 
 Combustible liquids. 
 
 1... 
 
 Dry chemical. 
 
 
 slag. 
 
 2... 
 
 Rubber (hose or 
 
 2... 
 
 Water. 
 
 2... 
 
 Engine heat. 
 
 
 belt). 
 
 3... 
 
 Other. 
 
 3... 
 
 Electrical. 
 
 3... 
 
 Insulation. 
 
 
 
 4... 
 
 Spontaneous 
 
 4... 
 
 Other. 
 
 
 
 
 combustion. 
 
 5... 
 
 Timber, lagging, 
 
 
 
 5... 
 
 Other. 
 
 
 etc. 
 
 
 
38 
 
 caution on the part of welders when work- 
 ing near combustibles. Welding on under- 
 ground mining machinery frequently in- 
 volves the spot ignition of combustibles 
 that accumulate on this equipment. The 
 primary causes of nonreportable surface 
 fires were poor equipment design, poor 
 maintenance, and poor housekeeping. The 
 routing of fluid lines on surface mining 
 equipment was frequently cited as con- 
 tributing to an increased fire hazard. 
 
 Several safety directors mentioned ef- 
 forts undertaken at their mines to re- 
 route or shield these lines from the ig- 
 nition sources of electrical arcing and 
 engine heat. Leaks of hydraulic fluid, 
 oil, and other lubricants frequently col- 
 lect in hard-to-reach locations. These 
 substances eventually are ignited by a 
 cutting torch, electrical arc, or engine 
 heat. 
 
 SUMMARY 
 
 Major findings of the study appear 
 in tables 44 and 45. The most frequent 
 ignition sources, burning substances, 
 eqiupment types, locations, and success- 
 ful extinguishing agents of reported and 
 nonreportable fires are discussed below. 
 
 IGNITION SOURCE 
 
 The majority of underground mine fires, 
 both reported and nonreportable, were 
 
 electrical. Electrical equipment was al- 
 so the primary cause of underground fires 
 resulting in injuries during the period 
 1950-84. In recent times, however, en- 
 gine heat has become the leading cause of 
 fires resulting in injuries. Engine heat 
 was the leading ignition source for re- 
 ported surface fires, and welding for 
 nonreportable surface fires. 
 
 TABLE 44. - Major study findings of reported fires 
 (Factors listed in sequence of significance) 
 
 Category 
 
 1950-77 
 
 1978-84 
 
 Total 
 
 
 UNDERGROUND 
 
 
 Ignition source. . 
 
 Electrical, welding. . 
 
 Electrical, engine 
 
 Electrical, welding, 
 
 
 
 heat. 
 
 engine heat. 
 
 Burning substance 
 
 Timber, insulation, 
 
 Combustible liquids, 
 
 Timber, combustible 
 
 
 combustible liquids. 
 
 timber, insulation. 
 
 liquids, insulation. 
 
 
 Haulageway-drif t , 
 shaft-raise-winze. 
 
 Haulageway-drif t, 
 shaft-raise-winze. 
 
 Haulageway-drif t, 
 shaft-raise-winze. 
 
 
 Equipment 
 
 Mobile, electrical.. 
 
 Mobile, electrical.. 
 
 Mobile, electrical. 
 
 involved. 
 
 
 
 
 Successful extin- 
 
 Water, dry chemical. 
 
 Water, dry chemical. 
 
 Water, dry chemical. 
 
 guishing agent. 
 
 
 
 
 SURFACE 
 
 Ignition source. . 
 
 Electrical, welding, 
 
 Engine heat, weld- 
 
 Engine heat, electri- 
 
 
 engine heat. 
 
 ing, electrical. 
 
 cal, welding. 
 
 Burning substance 
 
 Combustible liquids, 
 
 Combustible liquids, 
 
 Combustible liquids, 
 
 
 construction 
 
 construction 
 
 Construction 
 
 
 material. 
 
 material. 
 
 materials. 
 
 
 
 
 Surface building, 
 surface. 
 
 
 building. 
 
 Equipment 
 
 Mobile, electrical... 
 
 Mobile, conveyor.... 
 
 Mobile, conveyor, 
 
 involved. 
 
 
 
 electrical. 
 
 Successful extin- 
 
 Water, burned out.... 
 
 Water, dry chemical. 
 
 Water, burned out. 
 
 guishing agent. 
 
 
 
 
39 
 
 TABLE 45. - Major study findings of nonreportable fires 
 (Factors listed in sequence of significance) 
 
 Category 
 
 Underground 
 
 Surface 
 
 
 NONREPORTABLE FIRES 
 
 
 
 Combustible liquids, wiring insula- 
 tion, timber. 
 
 Mobile, electrical, maintenance-shop 
 Dry chemical, cut off electrical 
 power, water. 
 
 Welding, electrical, 
 
 engine heat. 
 Combustibles liquids, 
 
 insulation. 
 
 
 Successful extinguishing 
 agent. 
 
 building. 
 Mobile. 
 Dry chemical. 
 
 OPINION DATA 
 
 
 
 Welding, engine heat. 
 Combustible liquids, 
 
 
 Combustible liquids, wiring 
 
 
 insulation. 
 
 rubber (hose or 
 belt). 
 
 Successful extinguishing 
 agent. 
 
 
 Dry chemical, water. 
 
 
 BURNING SUBSTANCE 
 
 The most frequent burning substance in 
 reported underground fires was timber, 
 followed by combustible liquids and in- 
 sulation. In nonreportable underground 
 fires, combustible liquids, wiring insu- 
 lation, and timber were involved with 
 about equal frequency. For reported sur- 
 face fires, the most frequent burning 
 substance was combustible liquids, fol- 
 lowed by construction material. In non- 
 reportable surface fires, combustible 
 liquids were most frequently involved. 
 
 LOCATION 
 
 Reported underground fires occurred 
 along haulageways or in drift entries 
 where electrical and diesel equipment are 
 concentrated. Nonreportable fires also 
 occurred at these locations more fre- 
 quently than at any other. 
 
 Reported surface fires occurred primar- 
 ily in mill buildings, and nonreportable 
 surface fires occurred primarily on mo- 
 bile equipment along haulage roads or in 
 the pit area. In recent times, however, 
 the reported surface fire locations were 
 about evenly split between surface build- 
 ing and surface area other than building. 
 
 EQUIPMENT INVOLVED 
 
 The equipment most frequently involved 
 in reported and nonreportable underground 
 and surface fires is the mobile type such 
 as load-haul-dumps. These vehicle fires 
 are more quickly detected and are extin- 
 guished with hand-portable extinguishers 
 since there is less structure to collect 
 smoke or otherwise conceal the fire. 
 
 SUCCESSFUL EXTINGUISHING AGENT 
 
 The most frequently successful extin- 
 guishing agent for reported fires was wa- 
 ter. For nonreportable fires, dry chemi- 
 cal hand-portable fire extinguishers were 
 used most often. This Is consistent with 
 the duration of reportable fires. The 
 first attack on a fire is generally with 
 a hand-portable extinguisher. If the at- 
 tempt is successful, then the fire is 
 most likely extinguished at the nonre- 
 portable stage. If the fire has grown in 
 size, or initial extinguishing attempts 
 prove unsuccessful, then the fire will 
 probably become reportable while water is 
 brought to the area and applied to the 
 fire. 
 
40 
 
 CONCLUSIONS 
 
 Fire incidence rates in both surface 
 and underground mines are not declining. 
 Despite the considerable efforts of mine 
 safety personnel, regulatory action, and 
 fire safety-related research, little 
 progress toward reducing the incidence of 
 fire in recent years (1978-84) is appar- 
 ent. One possible explanation, which is 
 supported to some extent by the data on 
 ignition sources, burning substances, and 
 equipment involved, is that fire hazards 
 are changing as mining methods, mate- 
 rials, and equipment evolve. As specific 
 fire hazards are recognized and cor- 
 rected, new mining technology introduces 
 other hazards into the workplace. This 
 explanation suggests that the present 
 level of fire safety effort may not suc- 
 ceed in reducing fire incidence rates, 
 and that an accelerated pace of activity 
 with particular focus on newly emerging 
 mining technologies is required if inci- 
 dence rates are to be reduced. 
 
 Conclusions regarding the relative im- 
 portance of a given fire hazard for one 
 time period do not necessarily hold for 
 subsequent periods, suggesting the value 
 of regular updates to the fire incident 
 data base. Timely collection, analysis, 
 and publication of such data will help 
 ensure that fire safety efforts will ad- 
 dress the greatest needs. 
 
 Finally, the depth and breadth of the 
 data contained in this report is signifi- 
 cant in itself. The level of detail pro- 
 vided will enable users to concentrate 
 their efforts on the research instead of 
 on collecting data, thereby maximizing 
 effectiveness. The relative hazardous- 
 ness of various equipment and procedures 
 has been presented in clear and concise 
 fashion. The data were put on a floppy 
 disk in the Lotus 1-2-3 format to facili- 
 tate the analysis. Microdata are also 
 available from the authors on hard copy. 
 
41 
 
 MSHA fire reports for 1950 
 spection offices: 
 
 Northeastern District 
 
 Northeastern District Office 
 Pittsburgh, PA 15213 
 
 Pittsburgh Subdistrict Office 
 Pittsburgh, PA 15213 
 
 Albany Subdistrict Office 
 Albany, NY 12201 
 
 Southeastern District 
 
 Southeastern District Office 
 Birmingham, AL 35209 
 
 Birmingham Subdistrict Office 
 Birmingham, AL 35209 
 
 Knoxville Subdistrict Office 
 Knoxville, TN 37902 
 
 North Central District 
 
 North Central District Office 
 Duluth, MN 55802 
 
 Duluth Subdistrict Office 
 Duluth, MN 55802 
 
 Vincennes Subdistrict Office 
 Vincennes, IN 47591 
 
 APPENDIX 
 through 1977 were acquired from the following MSHA in- 
 South Central District 
 
 South Central District Office 
 Dallas, XX 75209 
 
 Dallas Subdistrict Office 
 Dallas, TX 75209 
 
 Rolla Subdistrict Office 
 Rolla, MO 65401 
 
 Rocky Mountain District 
 
 Rocky Mountain District Office 
 Denver, CO 80225 
 
 Denver Subdistrict Office 
 Denver, CO 80225 
 
 Salt Lake City Subdistrict Office 
 Salt Lake City, UT 84115 
 
 Western District 
 
 Western District Office 
 Alameda, CA 94501 
 
 Bellevue Subdistrict Office 
 Bellevue, WA 98004 
 
 Phoenix Subdistrict Office 
 Phoenix, AZ 85004 
 
 INT.-BU.0F MINES,PGH.,PA. 28451 
 
 U.S. GOVERNMENT PRINTING OFFICE: 1987 - 605-017/60007 
 
W84 
 
U.S. Department of the Interior 
 Bureau of Mines— Prod, end Dtstr. 
 Cochrens Mill Road 
 P.O. Box 18070 
 Pittsburgh. Pe. 15236 
 
 OFFICIAL BUSINESS 
 PENALTY FOR PRIVATE USE, S300 
 
 ] Do not wi sh to recei ve thi s 
 material, please remove 
 from your mailing list. 
 
 ^ Address change* Please 
 correct as indicated. 
 
 AN EQUAL OPPORTUNITY EMPLOYER 
 
 LIBRARY OF 
 
 CAT.' 
 WASHINGTON 
 
 CONGRESS RC'.'-24 
 
 IN PUBLICATIONS DIV 
 DC 20540 
 

 A-* 
 
 -Jv A v *J*/W * r v * „ Z. ' "*> A v 
 
 
 ^v \¥§if °' ***** 
 
 
 o »**A'* > 
 
 VV 
 
 jV 
 
 r ^ ->3 
 
 
 
 . aV ** 
 
 
 
 
 
 •P. 4 
 
 
 .^ .."'. *** 
 
 
 
 »• «o 
 
 *bV 
 
 
 
 
 mm?° ** v< * 
 
 o I* 
 
 
 * V •#• "CNrd » A 
 
 
 
 
 >°^ 
 
 
 
 V '. . l tT'' ,^ 
 
 
 '. o 
 
 
 :X 
 
 
 ^0* 
 ■A o 
 
 .* ,0 
 
 ^ 
 
 
 A <-, 'o . . • o > 
 
 
 j.0 * i^(>»! * ^ 
 
 .^ *v *W . ♦ 
 
 ^ A^ '.LyOk.* ^ 
 
 ^ ' .>Va'- ^ c^ .* 
 
 ^ V 
 
 <*."*••"»* a &v "%.*'••«* ,<y <*. '•.»* ^c> 
 
 * „v 
 
 . + 
 
 ^o* 
 
 
 
 ^"^ 
 
 
 
 JP> 
 
 
 
 
 
 
 <{5 x. 
 
 
 
 
 
 V^ % 
 
 
 vV^ 
 
 X-iHt-X y.-afc^ y,-i&-X y.-afc^. / 
 
 •\"-"^-> %-^'/ v^-y %-^y .. v^;/.. v ; 
 
 5- .^V/Ji'- >« c*' 
 
 
 
 ^^ v .* 
 
 **# 
 
 -$• A v " 
 
 A ^ 
 
 
 
 
 "V 
 
 •• ^ 
 
 
 
 
 K' 1 ^ " * n ^y^fs .j* \ 
 
 - ^O ' A*" t .^« ^ 
 
 
 q,. •Tr.'* aQ 
 
 
 
 iP": 
 
 ^. * o » o ' ■$ 
 
 o,. *•.■'• A V, *» 
 
 
 
 
 •^ 
 
 .^■J 
 
 o V 
 
 
 
 
 ^ ^. "•' 
 
 ;^l-. ^> 
 
 /^ "v^V ^' ; ^'\^' % % - T V .. V 
 
 
 "oV 
 
 » ^ 
 
 
 '^ A v *^rh»K * -^ V » 1 Z.' ^V. A- 
 
 ♦ A* %*. 
 
 -i- A v * 
 
 
 
 ,. ^v J . 
 
 * A* 
 
 
- -ov* 
 
 .«* / 
 
 
 ^ 
 
 
 * o°"°. <* 
 
 
 "by 
 
 
 o»« 
 
 *" * 
 
 
 
 *■ 
 
 .•lo. 
 
 
 ** .<^%** 
 
 '- J\ : -0m': }\ 'lit* y% 
 
 :> c^c^> .<-«^\ <?*ii£k » **^'*+ /*X&S> 
 
 
 
 ^%ik^^^ 
 
 
 
 *0 'o, 
 
 
 bt 
 
 
 * A V ^ - v 
 
 s*. ^.<» .^ir^o «*...«? „*i &\ \/ ; 
 
 
 
 A V "^ -' 
 
 
 
 ^d« 
 
 
 »* "yu 
 •^o? 
 
 " -y 
 
 I* » ' • 
 
 ^°- 
 
 
 
 
 
 
 
 
 i* 
 
 ^0* 
 
 
 
 
 -^ * 
 
 ,0- 0° 
 c,"^ '"W/W^' A V "^ 
 
 
 
 
 »°^, 
 
 ^ 
 
 ^ "^j. 
 
 .•1°, 
 
 A« ^ *'.-• ^ O ►.TV,' 0' ^ ♦.,,•' ^' 
 
 
 
 
 
 
 v .{.' 
 
 
 
 
 o V 
 
 » ^ 
 
L 'B«**Y OF CO N G Ress 
 
 Si 
 
 
 p I 
 
 '*.