TN 295 
 .U4 
 
 H 
 
 ■ I 
 
 No. 9117 
 
 
 ■ 
 
 ■ 
 
 ■ 
 
 ■ 
 
 I 
 
 
 ■ 
 
 I 
 

 «fev* 
 
 3' > 0v t, -SIBIJv /°- 
 
 ++# 
 
 
 
 i • .?* . q* .4? v . , -v- v *> a . : 
 
 W 
 
 * *>v^- >. .*.** •ygfef- ^ ^ *>w^ >. ^ stm**. ^ ^ .: 
 
 ^.^ 
 
 . , « ,G V \5 *< v. • * A 
 
 
 • T * A v * 
 
 V V • 
 
 • A^ 
 
 
 u "%% 
 
 •oV 
 
 iP^k 
 
 ^o* 
 
 
 .G r "q, ***^T» a 
 
 "^* 
 
 • \>« 
 ,^^ j ^ii^ a>"3 
 
 %q„ *<VVV» A 
 
 vv • 
 
 
 
 / *>' 
 
 r ^0 
 
 
 of 
 
 *^A A^ 
 
 TvT' .A 
 
 
 <*-_ 
 
 ^ ^ .wa, v/ /jflter-, v,^ •jsk^-. v> 
 
 : V* - 
 
 ^■=» J 
 
 •of 
 
 
 
 u a>^ 
 
 
 ^■0^ 
 
 
 o V 
 * <K "J 
 
 .' *>' 
 
 
 
 ^ ^ a'^». ** 
 
 ^oV* 
 
 
 ^0* 
 
 4 o 
 
 
 
 :• >. ^ »*^a^'. ^ a^ .: 
 
 
 
 
 "o>" 
 
 
 
 -xV n » • . ** 
 
 
 
 .A* . • M" « ^o 
 
 
 ° A 4 ' 
 
 ^°^ 
 
 
 G^ V *7K^ t » A 
 
 ° a^ 
 
 
 V »!.»# 
 
 
 
 < . . s \> V> ° • » <> 
 
 
 '• A V ^ 
 
 
 ;• . ^ ^ -y^^- ^ ^ --?aB?»* . * v ^ -y^^* ,^ 
 
 ** & 
 
 **<? 
 
 **%> 
 
 *° *^v*. > 
 
 v| 
 
 
 *OK 
 
 
 
 '•» ^O 
 
 4 o " i 
 
 
 
 ^ 
 
 
 ,0* »•"• 
 
 
 
 
 
 ^n^ 
 vT 7 
 
 „ A „ » . . "fe 
 
 ^* ,A V 
 
 
 
 
 
 %S " 
 
 
 v v^ 
 
 
 
 ^ 
 
 ^ 
 
 
 
 4 ^ 
 
 > ^ 
 
 
 .V ,A 
 
 
 v^ v 
 
 
 0^ 
 
 
 
 ^ *»;o» ^- o^ *.",,•• ^o- T^_ *.- B o' ^ 
 
 • % ^ 
 
 
 
 a*° *j&fc' ^* « v »*.i"i"' ^ a* v *>^4:' ^ « v »^"i"» ^ A*' * 
 
 
 1 'i.^'* > 
 
 v^ % 
 
 4? ^ • -, 
 
 "• A V ^ 
 V ^ % 
 
 ^<F 
 
 
 '^ .0* .•^..••. ^> 
 
 
 „4 o 
 > 4- * 
 
 ° A^ 1 
 *b>* 
 
 q,. *'^*' a 
 
 
 
 
 ^ f 
 
 
 
 
 'of 
 
 * A v 
 
 O" A W # ' 
 
 .«> ..v'..^. 
 
 <V , o » 
 
 
 vv 'raM. ^^ °^!5^' vv 'M^x%° ^^ °^!15^- vv 
 
 
 • if o 
 
 ^4 
 
 v v 
 
 
 
 at: *bv* :ffl^'- ^ < • 
 
 ^ V 
 
 q. ♦*V.T» ,/V 
 
 «• A V ^ 
 
 ^.O 
 
 ^. -„^^^: ^q. 
 
 
 4-^ Q-, 
 
 * o 
 
 
 .o- - 
 
 
 
 ,G V ^r> *<vvV* A 
 
 
 \s 
 
 ' (P^yJ^S " /V* 
 
 <P. 4 
 
 f»* ^ 
 
 
cw« 
 
 y **>. - *c<. 
 
 
 
 ^%. 
 
 
 V * b, 
 
 
 
 v^ 
 
 ** ** ' /sk*. >* **' *w 
 
 l> " A V «^ - 
 
 vv 
 
 
 ^* 
 
 •f ^ •-«?,♦ «v ^ •TWv * v "> 
 
 
 "W 
 
 * A V *^ : 
 
 
 
 
 
 » I • •* > 
 
 +>. "' <V 
 
 
 *by 
 
 t : 'Mf\ ****** °M& % ^ ** 
 
 r » « » < 
 
 
 '«> » < 
 
 <, *rr: 
 
 I**** 
 
 
 
 »* A 
 
 
 ^■ v 
 
 
 O o %. 'jay 1 *** j 
 
 F , " • o, *> V N s 
 
 h* O 
 
 ft? 1V> 
 
 
 
 A ° 
 
 
 
 <> ♦*T7i* A 
 
 
 
 
 »P^ v 
 
 c 0> / 
 
 •>-d* 
 
 V ..iilf. "«> 
 
 k * smfef- \. .<? Wa\ ^ a* .*^fef. >« ^* .: 
 
 
 
 ^^ 
 
 
 ♦^ 
 
 
 
 
 
 
 » aV ^ . 
 
 
 
 IV . i • < 
 
 -^ * 
 
 ^O 5 
 
 * ^* 
 
 .v q^ * o » o g ,0 
 
 b# 
 
 r o '."".I* A 
 
 P V > 
 
 
 vv !^c?^: ^^ 
 
 Q~ . » * « 
 
 ^6* 
 
 ,^ .••-".-. ^ 
 
 "ov* 
 
 S' » 
 
 
 ?-^-$W ; y ,! 
 
 
 
 
 
 
 
 
 
 %A^ .-S 
 
 
 
 t ^ /fflfe'v W* .-isK«i-. ^.^ /Jfe\ V./ .^M» *«^ /^K\ ^ 
 
 
 
 
 
 
 VJ A 
 
 ". T * a 
 
 ,. ^ v 
 
 
 
 
 O.. 'o . » » ^V 
 
 <, *' ... 
 
 
 "*bv^ 
 
 s s^y* 
 
 
 •/ \-^?\y.. v 1 ^-/ \-^^v v^-/ \-^\/ *. 
 
 
 
 
 
 
 
 o, »o . 
 
 
 / + c^r^°o 
 
 :- ^o/ 
 
 
 <> *'7Vi* ,0 
 
 
 ,v ^ *" • » * A <. *^TT" ' A v "o, 
 
 *bv" 
 
 *p^ 
 
 
 
IC 9117 
 
 Bureau of Mines Information Circular/1986 
 
 Transformer Fluid Fires 
 in a Ventilated Tunnel 
 
 By Margaret R. Egan 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 
Information Circular 9117 
 
 Transformer Fluid Fires 
 in a Ventilated Tunnel 
 
 By Margaret R. Egan 
 
 UNITED STATES DEPARTMENT OF THE INTERIOR 
 Donald Paul Hodel, Secretary 
 
 BUREAU OF MINES 
 Robert C. Horton, Director 
 
.at 
 
 Library of Congress Cataloging in Publication Data: 
 
 Egan, Margaret R 
 
 Transformer fluid fires in a ventilated tunnel. 
 
 (Information circular ; 9117) 
 
 Bibliography. 
 
 Supt. of Docs, no.: I 28.27: 9117. 
 
 1. Mine fires. 2. Mine ventilation. 3. Electric transformers. I. Title. II. Series: Information 
 circular (United States. Bureau of Mines) ; 9117. 
 
 TN295.U4 
 
 [TN315] 
 
 622 s 
 
 [622'.8] 
 
 86-600288 
 
CONTENTS 
 
 Page 
 
 Abstract 1 
 
 Introduction 2 
 
 Instrumentation 2 
 
 Fire tunnel 2 
 
 Thermocouples 2 
 
 Flow probes and pressure transducers 2 
 
 Gas monitors 3 
 
 Smoke monitors 3 
 
 Fuel-consumption monitor 4 
 
 Typical test procedure 4 
 
 Calculations 4 
 
 Product generation rates 4 
 
 Combustion yields 4 
 
 Heat-release rates 5 
 
 Production constants 5 
 
 Smoke particle diameters 5 
 
 Burning rate 6 
 
 Brand comparison results. 6 
 
 Gas concentrations and heat production 6 
 
 Smoke characteristics 8 
 
 Combustion yields 8 
 
 Production constants 8 
 
 Discussion of results 8 
 
 Conclusions 10 
 
 Fuel comparison results and discussion 10 
 
 Scaling results and discussion 11 
 
 Appendix. — List of symbols 13 
 
 ILLUSTRATIONS 
 
 1. Schematic of intermediate-scale tunnel 3 
 
 2. CO concentrations (A), CO2 concentrations (B), heat-release rates (C)» and 
 
 heats of combustion (D) for three brands of transformer fluid 7 
 
 3. Particle mass concentrations (A), number concentrations (5), and mass mean 
 diameters (C) for three brands of transformer fluid 9 
 
 4. Burning rates for gasoline and transformer fluid 12 
 
 TABLES 
 
 1. Average gas concentrations, heat-release rates, and heats of combustion for 
 
 transformer fluid 7 
 
 2. Toxic-gas concentrations for transformer fluid 8 
 
 3. Smoke characteristics for transformer fluid 8 
 
 4. Mean particle sizes and obscuration rates for transformer fluid 8 
 
 5. Combustion yields for transformer fluid 8 
 
 6. Production constants for transformer fluid 9 
 
 7. Gas, heat, and smoke concentrations for the three fuels tested 10 
 
 8. Normalized gas and smoke concentrations for the three fuels tested 10 
 
 9. Particle size and obscuration rates for the three fuels tested 11 
 
 10. Production constants for the three fuels tested 11 
 

 UNIT OF MEASURE ABBREVIATIONS USED IN 
 
 THIS REPORT 
 
 cm 
 
 centimeter 
 
 mg/m 3 
 
 milligram per cubic meter 
 
 cm/min 
 
 centimeter per minute 
 
 pg/m 3 
 
 microgram per cubic meter 
 
 °F 
 
 degree Fahrenheit 
 
 min 
 
 minute 
 
 g 
 
 gram 
 
 ym 
 
 micrometer 
 
 g/cm 3 
 
 gram per cubic centimeter 
 
 m 3 /s 
 
 cubic meter per second 
 
 g/(m 3, ppm) 
 
 gram per (cubic meter 
 times part per million) 
 
 P 
 
 particle 
 
 
 
 p/cm 3 
 
 particle per cubic i 
 
 g/g 
 
 gram per gram 
 
 
 centimeter 
 
 g/kJ 
 
 gram per kilo joule 
 
 pet 
 
 percent 
 
 g/s 
 
 gram per second 
 
 p/g 
 
 particle per gram 
 
 kg 
 
 kilogram 
 
 p/kJ 
 
 particle per kilojoule 
 
 kJ/g 
 
 kilo joule per gram 
 
 ppm 
 
 part per million 
 
 kW 
 
 kilowatt 
 
 ppm/min 
 
 part per million 
 per minute 
 
 In 
 
 logarithm, natural 
 
 
 
 
 
 psi 
 
 pound per square inch 
 
 m 
 
 meter 
 
 
 
TRANSFORMER FLUID FIRES IN A VENTILATED TUNNEL 
 
 By Margaret R. Egan 1 
 
 ABSTRACT 
 
 The Bureau of Mines subjected three commercially available brands of 
 transformer fluid to a series of combustion studies. The experiments 
 were conducted in the intermediate-scale fire tunnel, which was designed 
 to simulate environmental conditions in underground mines. The work was 
 divided into three phases. In phase one, the brands were compared for 
 gas production, smoke characteristics, and combustion yields. In phase 
 two, the production constants of transformer fluids and other fuels were 
 compared for the rate of formation of gas and smoke as a function of 
 fire size. In phase three, several diameters of liquid-pool fires were 
 compared in terms of fire size and burning rate. 
 
 These transformer fluid measurements will be added to the existing 
 coal and wood data that can be used as a basis of comparison for future 
 studies of other mine combustibles. Further research into the combus- 
 tion product emissions from combustible materials found in underground 
 mines will lead to improved and realistic fire detection and suppression 
 systems. 
 
 'Chemist, Pittsburgh Research Center, Bureau of Mines, Pittsburgh, PA. 
 
INTRODUCTION 
 
 The Bureau of Mines conducts research 
 to ensure that mines are safe and health- 
 ful places to work. Exceptional circum- 
 stances, such as an underground fire, 
 pose additional problems involving health 
 and safety. Among these dangers are 
 those affecting escape and rescue. Early 
 fire detection increases the opportunity 
 for human escape. Rescue efforts are 
 often hampered by reduced visibility due 
 to smoke and the potential toxicity of 
 the combustion products. More efficient 
 detection devices and rescue equipment 
 can be designed once the hazardous pro- 
 ducts of combustible materials are known. 
 Therefore, the smoke characteristics of 
 fires in ventilated mine passageways are 
 investigated by the Bureau. 
 
 These experiments are part of a series 
 of continuing projects in which combus- 
 tible materials are burned in a simulated 
 
 mine environment. Previous studies^ have 
 shown that the intermediate-scale fire 
 tunnel used for the current studies can 
 successfully predict full-scale fire 
 conditions. 
 
 Transformers are used to produce elec- 
 trical power needed to operate mine mach- 
 inery. Heat, which is an inherent by- 
 product of this process, is removed by 
 transformer fluid. Fires are potential 
 risks whenever petroleum oil, the basic 
 component of transformer fluid, is used 
 in any electrical equipment. 
 
 The objectives of this study are (1) to 
 compare three brands of transformer fluid 
 for gas production and smoke characteris- 
 tics, (2) to compare these evaluations 
 with similar data for wood and coal, and 
 (3) to compare fire sizes and burn- 
 ing rates of several pool sizes of 
 transformer fluid. 
 
 INSTRUMENTATION 
 
 FIRE TUNNEL 
 
 THERMOCOUPLES 
 
 The transformer fluid fires were con- 
 ducted in an intermediate-scale fire tun- 
 nel located at the Bureau's Pittsburgh 
 Research Center. A diagram of the tunnel 
 with its data-acquisition system is shown 
 in figure 1. The tunnel measures 0.8 m 
 wide by 0.8 m high by 10 m long and is 
 divided into three sections. The first 
 horizontal section begins at the air- 
 intake cone, which gradually enlarges to 
 a hinged portion that can be lifted to 
 allow entrance for the placement of oil 
 pan. Next is the fire zone in which is 
 located the gas burner followed by the 
 oil pan. The fire zone and the remaining 
 horizontal section are lined with fire- 
 brick and instrumented with thermo- 
 couples, flow probes, and sampling ports. 
 The diffusing grid begins the vertical 
 section of the tunnel. Located in this 
 section is an orifice plate that can be 
 manually adjusted to attain the desired 
 airflow. The final section is horizontal 
 and ends at an exterior exhaust fan. 
 
 The thermocouple arrays were located 
 1.57, 2.36, 3.15, 4.72, 6.30, and 7.87 m 
 from the gas burner. Additional thermo- 
 couples are located on the air-intake 
 cone and at the exhaust. In all, a total 
 of 28 thermocouples were used to measure 
 the temperature distributions resulting 
 from the fires. Their locations are 
 shown in figure 1 . 
 
 FLOW PROBES AND PRESSURE TRANSDUCERS 
 
 The air pressure produced by the ex- 
 haust ventilation is detected by the 
 transducer. As pressure increases, the 
 capacitance decreases. This change is 
 then converted to a linear electric sig- 
 nal. Nonlinearily is described as <±0.1 
 
 2 Lee, C. K. , R. F. Chaiken, J. M. Sin- 
 ger, and M. E. Harris. Behavior of Wood 
 Fires in Model Tunnels Under Forced 
 Ventilation Flow. BuMines RI 8450, 1980, 
 58 pp. 
 
INTERMEDIATE-SCALE FIRE TUNNEL 
 10- m length 
 
 I2 _ m length 
 0.6l _ m diam duct 
 
 1.22 m 
 
 0.8 -m square duct 
 
 Fire zone 
 
 intake 
 
 _*-?_B_ 
 
 u Load cell 
 
 Air 
 
 Manually 
 adjustable 
 orifice plate 
 
 7 exhaust 
 Ventilation 
 
 fan 
 ( 2-speed ) 
 
 1 — Diffusing grid 
 
 -Q305 - rTrdiam 
 entrance duct 
 ( hinged and movable) 
 
 CALOUT/DECNET- 
 
 Tapered-element 
 
 oscillating 
 
 microbalance 
 
 Condensation 
 nuclei monitor 
 
 CO meter 
 
 CO2 meter 
 
 Pressure transducers 
 
 MM 
 
 48" 
 channel 
 
 data- 
 acqui 
 sition 
 system 
 
 3X 
 = detector 
 
 t...28— ' 
 thermocouples 
 
 i_ -Load cell 
 
 - Digital input 
 
 for CNM 
 
 range 
 
 PDP 
 I /44 
 
 Control 
 terminal 
 
 VAX 
 11/780 
 
 Printer 
 
 VAX 
 terminal 
 
 CALCOMP 
 plotter 
 
 KEY 
 
 Pressure transducer (flow probe) • Thermocouples 
 
 Differential pressure transducer ■ Sampling ports 
 
 3X detector 
 
 FIGURE 1.— Intermediate-scale fire tunnel (top) and data-acquisition system (bottom). 
 
 pet full range of the output or ±0.00001 
 psi differential. The locations of the 
 pressure transducers are also shown in 
 figure 1. 
 
 GAS MONITORS 
 
 The carbon monoxide (CO) 
 sures accurately within 1 
 range or ±5 ppm. 
 
 The carbon dioxide (C0 2 ) 
 sures accurately within 1 
 range or ±250 ppm. 
 
 SMOKE MONITORS 
 
 The number concentration (N Q ) was 
 obtained with a condensation nuclei 
 
 analyzer mea- 
 pct of full 
 
 analyzer mea- 
 pct of full 
 
 monitor, manufactured by Environment One 
 Corp. ^ The monitor measures the con- 
 centration of submicrometer airborne par- 
 ticles (p) using a cloud chamber. The 
 particulate cloud attenuates a light beam 
 that ultimately produces a measurable 
 electrical signal. The accuracy is 
 stated as ±20 pet of a point above 30 pet 
 of scale on the linear ranges, 3,000 to 
 3000,000 p/cm3. 
 
 The mass concentration (M Q ) was ob- 
 tained by a TEOM tapered-element os- 
 cillating microbalance, developed by 
 Rupprecht & Patashnick Co. , Inc. It 
 
 ^Reference to specific equipment does 
 not imply endorsement by the Bureau of 
 Mines. 
 
measures the mass directly by depositing 
 the particles on a filter attached to an 
 oscillating tapered element. The change 
 in the oscillating frequency of the tap- 
 ered element is directly proportional to 
 the change in mass. The apparatus is 
 capable of measuring dust concentrations 
 with a better than 10-pct accuracy at the 
 250-ug/m 3 level. 
 
 A three-wavelength light transmission 
 technique was also used to measure smoke 
 characteristics and obscuration. White 
 light was transmitted through a smoke 
 cloud to the detector. The beam was 
 split into three parts, and each passed 
 
 through an interference filter centered 
 at wavelengths of 0.45, 0.63, or 1.00 pm. 
 Each photodiode output was amplified and 
 recorded as a linear electric signal. 
 This technique was developed by Bureau 
 personnel. 4 
 
 FUEL-CONSUMPTION MONITOR 
 
 The weight-loss data were obtained by a 
 strain-gauge conditioner in conjunction 
 with a load cell, which has a range up to 
 22.68 kg. The accuracy of the strain 
 gauge is stated as 0.05 pet of full scale 
 or ±11.3 g. 
 
 TYPICAL TEST PROCEDURE 
 
 The transformer fluid was poured into a 
 stainless steel pan inside the tunnel. 
 The shaft of the pan extended through the 
 tunnel floor and was supported on the 
 load cell so that continuous weight loss 
 could be recorded. 
 
 Prior to each experiment, background 
 readings were obtained after the cone was 
 closed and the exhaust fan was started. 
 All instruments were continuously scanned 
 and recorded throughout the experiment. 
 
 A gas jet, located immediately upstream 
 from the pan, was the igition source. 
 The burner flow rate was adjusted so 
 that the flame licked the side of the 
 
 oil-filled pan. The flash point of the 
 transformer fluid, as stated by the man- 
 ufacturers, is about 300° F. At a ven- 
 tilation rate of approximately 0.65 m 3 /s, 
 the transformer fluid took about 1 min to 
 ignite. The ventilation rate was then 
 lowered to approximately 0.47 m 3 /s during 
 the steady-state burning. 
 
 The gas burner was turned off once 
 the transformer fluid had ignited. The 
 flames spread quickly, engulfing the en- 
 tire pan within 1 min. The flames began 
 to die down as the transformer fluid was 
 consumed. The experiment was concluded 
 when the flames were no longer visible. 
 
 CALCULATIONS 
 
 It is necessary to measure certain pa- 
 rameters in order to compare the steady- 
 state combustion products and utilmately 
 the hazards of various fuels. Among 
 these values are gas concentrations, mass 
 and number concentrations of smoke par- 
 ticles, ventilation rate, and mass-loss 
 rate. Other combustion properties can be 
 calculated once these are known. 
 
 PRODUCT GENERATION RATES 
 
 In a ventilated system, the generation 
 rates (Gx)'of CO2 and CO are related to 
 the bulk average concentration increase 
 above ambient, ACO2 and ACO, by the 
 expressions 
 
 and 
 
 Geo = M C0 V A ACO, (2) 
 
 Geo, = M C0 , V A AC0 2 
 
 (1) 
 
 where Mco 2 = l- 97 x 10_3 g/(m 3 *ppni) ; 
 
 M C o = 1.25 x 10 -3 g/m 3 «ppm); 
 
 and V A = incoming cold gas flow, 
 m 3 /s. 
 
 COMBUSTION YIELDS 
 
 Once the generated rates are known and 
 the mass-loss rate of the fuel (Mf) is 
 
 4 Cashdollar, K. L., C. K. Lee, and 
 J. M. Singer. Three-Wavelength Light 
 Transmission Technique To Measure Smoke 
 Particle Size and Concentration. Appl. 
 Opt., v. 18, No. 11, 1979, pp. 1763-1769. 
 
calculated using the load-cell assembly, 
 the true yield of the combustion product 
 (Yx) can be calculated by the expression 
 
 Y X = Gx/Mf. 
 
 (3) 
 
 The yields for mass (M ) and number 
 (N ) concentrations are calculated in a 
 similar manner by the expression 
 
 Yx = 
 
 AX C X V A 
 
 • ! 
 
 Mf 
 
 (4) 
 
 where 
 
 Cx = appropriate units conversion factor: 
 1.00 x 10" 3 when M Q is in mg/m 3 or 
 1.00 x 10 6 when N is in p/cm 3 ; 
 
 and 
 
 AX = smoke concentration increase above 
 ambient (when M is measured as mg/ 
 in 3 and N is measured as p/cra 3 ). 
 
 HEAT-RELEASE RATES 
 
 It has been shown^ that the actual 
 heat-release rate realized during a fire 
 can be calculated from the expression 
 
 Qa = 
 
 He 
 Kco. 
 
 Geo. 
 
 He - Heo (Keo) 
 Kco 
 
 Geo. 
 
 (5) 
 
 where Qa = actual heat release, kW; 
 
 He = net heat of complete combus- 
 tion of the fuel (40.7 kj/g 
 for transformer fluid); 
 
 Substituting the values in equations 1, 
 2, and 5 yields 
 
 Qa = V A [0.0251 (AC0 2 ) 
 
 + 7.07 x 10~ 3 (ACO)]. (6) 
 
 Since measurements of V A , ACO2, and 
 ACO were made continuously, the actual 
 heat-release rates could be calculated 
 using equation 6. 
 
 A typical fire rarely realizes the 
 state of complete combustion. For this 
 reason, the actual heat of combustion 
 (Ha) during a fire is usually less than 
 the total heat of combustion (He). By 
 measuring both the actual heat-release 
 rate, equation 6 above, and the fuel 
 mass-loss rate (Mf), the actual heat of 
 combustion can be calculated from the 
 expression 
 
 H A = QA/Mf. 
 
 PRODUCTION CONSTANTS 
 
 (7) 
 
 In an actual mine fire, it is often 
 difficult, if not impossible, to calcu- 
 late the actual heat of combustion. 
 Moreover, since the true yield of a com- 
 bustion product depends upon this infor- 
 mation, significant errors can result in 
 predicting the resultant concentration 
 increases. For flaming fires, the rel- 
 ative hazards tend to increase with the 
 actual heat-release rate that results. 
 
 For this reason, production constants 
 or beta values (Bx) can be calculated for 
 a given product by the expression 
 
 3x = Gx/Qa« 
 
 (8) 
 
 Kco 2 = stoichiometric yield of CO2 
 = 3.19 g/g; 
 
 Heo = heat of combustion of CO 
 =10.1 kJ/g; 
 
 and Kco = stoichiometric yield of CO 
 = 2.58 g/g. 
 
 5 Tewarson, A. Heat Release Rate in 
 Fires. Fire and Mater., v. 4, No. 4, 
 1980, pp. 185-191. 
 
 Using the rate of formation of gas or 
 smoke as a function of the fire size 
 is also beneficial in comparing the 
 combustion hazards of different fuels. 
 
 SMOKE PARTICLE DIAMETERS 
 
 Measurements of both number and mass 
 concentrations of the smoke provide 
 important information relative to the 
 yields, equation 4, and production con- 
 stants, equation 7. They can also be 
 
used to calculate the average size of the 
 smoke particles, with the expression 
 
 In T (XI. 00) In T (XI. 00) 
 
 TTd m 3 
 
 p p N = 1 x 10 3 M , (9) 
 
 where p p = individual particle den- 
 sity, g/cm 3 ; 
 
 particle, ym; 
 
 and 1 x 10 3 = the appropriate units con- 
 version factor. 
 
 Assuming a value of p p = 1.4 g/cm 3 , 
 then the mass mean diameter of the 
 particles can be calculated from 
 
 d m = 11 
 
 ■" Ct) 
 
 1/3 
 
 (10) 
 
 where the particle diameter is expressed 
 in micrometers. 
 
 Using the three-wavelength smoke detec- 
 tor, the transmittance (T) of the light 
 through the smoke can be calculated for 
 each wavelength. The extinction-coeffi- 
 cient ratio can be calculated for each 
 pair of wavelengths by the following log- 
 transmission ratios: 
 
 or 
 
 In T (X0.63) In T (X0.45) 
 
 In T (X0.63) . 
 In T (X0.45) 
 
 (11) 
 
 Using these extinction coefficients and 
 the curve found by Cashdollar, *> the mean 
 particle size (d 32 ) can be determined. 
 
 The smoke-obscuration rate is the per- 
 centage of the ratio of the light de- 
 tected after passing through the smoke 
 compared with the background light. 
 
 BURNING RATE 
 
 It is difficult to compare the hazards 
 from liquid-fuel fires with those of 
 fires involving solid fuels such as coal 
 and wood. The major reason for this dif- 
 ficulty is the speed with which liquid- 
 fuel fires develop compared with that for 
 solid-fuel fires. However, liquid-pool 
 fires can be compared by using the total 
 burning rate, which is calculated by di- 
 viding the depth of the fuel (in centi- 
 meters) by the burning time (in minutes). 
 This information can then be used to 
 compare various pan diameters. 
 
 BRAND COMPARISON RESULTS 
 
 All the values listed in this report 
 are an average of steady-state burning 
 stage that was arbitrarily selected to be 
 the 20th to the 30th minute after igni- 
 tion. Ten experiments were completed us- 
 ing three brands of transformer fluid: 
 four with Texaco 7 fluid and three each 
 with Shell and Gulf fluids. All three 
 brands showed similar results for gas 
 production, heat release, and heat of 
 combustion. However, the tested brands 
 showed somewhat different results for 
 smoke characteristics. 
 
 GAS CONCENTRATIONS AND HEAT PRODUCTION 
 
 The CO and CO2 concentrations, heat- 
 release rates, and heats of combustion 
 
 6 Figure 9 of work cited in footnote 4. 
 
 graphs for an average test of each brand 
 of transformer fluid are found in figure 
 2. The initial spike found on the CO2, 
 heat-release rate, and heat of combustion 
 was caused by the natural gas burner. 
 
 The CO and CO2 productions were contin- 
 uously monitored throughout the experi- 
 ment. The CO concentration remained 
 fairly constant, rising slightly as the 
 flames died. The CO 2 concentration grad- 
 ually rose at an average rate of 22 ppm/ 
 min as the fuel was consumed. 
 
 The heat-release rate remained fairly 
 constant for most of the experiments, in- 
 creasing slightly just before the flames 
 died. The average mass loss was 2.026 g, 
 at a rate of 0.866 g/s . The heat of 
 
 'Reference to specific brands does not 
 imply endorsement by the Bureau of Mines. 
 

 i 
 
 i i 
 
 1 
 
 B 
 
 " 
 
 
 
 K\ 
 
 ll 
 
 
 r / 
 
 \\ 
 
 
 I 
 
 ti 
 
 ^-jpjfcr^ 
 
 1 rp 
 
 
 
 \\ 
 
 "1 \M 
 
 
 
 \\ 
 
 "I 
 
 
 
 l\ - 
 
 1 
 
 
 
 1\ . 
 
 1 
 
 - 
 
 
 
 1 
 
 - 
 
 
 
 \ - 
 
 
 
 I 1 
 
 
 FIGURE 2.— CO concentrations (A), C0 2 concentrations (S), heat-release rates (C), and heats of combustion (0) 
 for three brands of transformer fluid. 
 
 combustion also remained fairly constant 
 throughout the steady-state burning but 
 rose sharply just before the fuel was 
 completely consumed. The average values 
 for each brand and the grand average 
 (which includes all experiments) are 
 listed in table 1. 
 
 During a 30-min period after igni- 
 tion, six gas samples were taken, each 5 
 min apart. All brands showed a slight 
 decrease in oxygen and a slight increase 
 
 TABLE 1. - Average gas concentrations, 
 heat-release rates, and heats of 
 combustion for transformer fluid 
 
 Brand 
 
 CO, 
 ppm 
 
 C0 2 , 
 ppm 
 
 Qa» 
 
 kW 
 
 kJ/g 
 
 
 120 
 
 97 
 
 120 
 
 1,682 
 1,871 
 1,784 
 
 20.3 
 22.2 
 21.0 
 
 22.9 
 
 
 25.3 
 
 24.5 
 
 
 113 
 
 1,769 
 
 21.1 
 
 24.1 
 
in argon and nitrogen concentrations. 
 These samples were also tested for hydro- 
 carbons C1-C3. The brands showed very 
 little variation. The concentrations in- 
 creased slightly as combustion proce- 
 eded. Table 2 lists the averages for 
 each brand. 
 
 SMOKE CHARACTERISTICS 
 
 The number concentration slowly in- 
 creased until the fuel was almost 
 completely consumed, then it started to 
 drop. The mass concentration varied 
 throughout the experiments. An average 
 was taken during the steady-state burning 
 when it was the most stable. Using these 
 values, the mass mean diameter was 
 calculated. The average values for each 
 brand are listed in table 3. The mass 
 number concentrations and the mass mean 
 diameters for an average test of each 
 brand are found in figure 3. 
 
 Using the three-wavelength smoke de- 
 tector, the average mean particle size 
 (d32) was calculated for each wavelength. 
 These averages and the obscuration rates 
 for each brand are listed in table 4. 
 
 COMBUSTION YIELDS 
 
 The gas concentration yields, as ex- 
 pected, showed little variation. How- 
 ever, the mass and number concentration 
 yields of smoke particles showed a wide 
 variation between brands. The average 
 yields are listed in table 5. 
 
 PRODUCTION CONSTANTS 
 
 The production constants or beta values 
 were calculated as a function of the fire 
 size. For the tested brands of trans- 
 former fluid, the fire sizes were very 
 similar. Therefore, it is expected that 
 the beta values reflect the same vari- 
 ability as the gas and smoke concen- 
 trations. Table 6 lists the average pro- 
 duction constants for each brand. 
 
 DISCUSSION OF RESULTS 
 
 The Shell transformer fluid generated a 
 slightly larger CO2 concentration, thus 
 increasing the corresponding fire size 
 
 TABLE 2. - Toxic-gas concentrations 
 for transformer fluid, ppm 
 
 Brand 
 
 Methane 
 
 Ethane 
 
 Ethy- 
 lene 
 
 Acety- 
 lene 
 
 Texaco. . . 
 Shell.... 
 Gulf 
 
 22.1 
 15.3 
 22.5 
 
 4.5 
 
 1.7 
 
 .9 
 
 19.0 
 11.5 
 16.9 
 
 9.8 
 16.6 
 16.2 
 
 TABLE 3. - Smoke characteristics for 
 transformer fluid 
 
 Brand 
 
 Texaco. . . . 
 
 Shell 
 
 Gulf 
 
 Average, 
 
 No, 
 p/cm 3 
 
 1,567,250 
 
 1,126,900 
 
 416,533 
 
 1,089,930 
 
 Mo, 
 mg/m 3 
 
 40.6 
 
 9.0 
 
 58.1 
 
 35.3 
 
 dm, 
 ym 
 
 0.309 
 .223 
 .601 
 
 .386 
 
 TABLE 4. - Mean particle sizes and obscuration rates for transformer fluid 
 
 Brand 
 
 In T (X0.63) 
 
 In T (X0.45)' 
 
 ym 
 
 In T (XI. 00), 
 
 In T (X0.45)' 
 
 ym 
 
 In T (XI. 00) 
 In T (X0.63)' 
 
 ym 
 
 Average 
 
 d 32, 
 ym 
 
 Obscuration 
 
 rate, 
 
 pet 
 
 
 0.302 
 .339 
 .336 
 
 0.351 
 .395 
 .431 
 
 0.391 
 .440 
 .538 
 
 0.348 
 .391 
 .435 
 
 39.9 
 
 Shell 
 
 40.8 
 
 
 59.6 
 
 
 .323 
 
 .388 
 
 .450 
 
 .387 
 
 46.1 
 
 TABLE 5. - Combustion yields for transformer fluid 
 
 Brand 
 
 Yco, 
 g/g 
 
 V C0 2 , 
 g/g 
 
 10^°p/g 
 
 g/g 
 
 Brand 
 
 Yco, 
 g/g 
 
 Yco 2 » 
 g/g 
 
 Yn , 
 10H°p/g 
 
 g/g 
 
 
 0.083 
 .064 
 
 1.829 
 1.952 
 
 8.85 
 5.92 
 
 0.021 
 .005 
 
 Average. . 
 
 0.081 
 
 1.894 
 
 2.23 
 
 0.031 
 
 
 .077 
 
 1.885 
 
 5.99 
 
 .019 
 
2.0 
 
 I 1 I i ■ 
 A A S 
 
 1.8 
 
 l\l\ 
 
 1.6 
 
 A j \ 
 
 » 1.4 
 
 1 VI \ A 
 
 E 
 
 i \ / \ 
 
 o 10 
 
 z 
 o 
 
 
 o 
 
 cc 0.8 
 
 -ikf / v v \ 
 
 UJ 
 CD 
 
 W'-ajV J \ 
 
 § 0.6 
 
 z 
 
 0.4 
 
 
 
 s. 
 
 0.2 
 
 "S - 
 
 
 V 
 
 1 1 1 1 
 
 10 20 30 40 50 
 
 TIME, min 
 
 FIGURE 3.— Particle mass concentrations (A), number 
 concentrations (B), and mass mean diameters (C) for 
 three brands of transformer fluid. 
 
 TABLE 6. - Production constants for transformer fluid 
 
 Brand 
 
 Bco> 
 10~ 3 g/kJ 
 
 Bco ? » 
 10~ 2 g/kJ 
 
 &N rt > 
 
 10 10 p/kj 
 
 lO" 1 * g/kJ 
 
 
 3.47 
 2.53 
 3.30 
 
 7.69 
 7.74 
 7.70 
 
 3.77 
 
 2.34 
 
 .91 
 
 9.17 
 
 
 1.95 
 12.60 
 
 
 3.14 
 
 7.71 
 
 2.48 
 
 7.75 
 
 (Q/\) and the heat of combustion (Ha). It 
 also produced less CO, methane, and ethy- 
 lene. However, the Texaco fluid pro- 
 duced more ethane but less acetylene than 
 
 the other tested brands. These analyses 
 were based on relatively few experiments 
 and may only reflect the range of gas 
 production. Considering this, the gas 
 
10 
 
 concentrations generated by the tested 
 brands of transformer fluid were similar. 
 
 The differences between the brands were 
 more evident in their smoke character- 
 istics. The most noticeable variations 
 were the Shell fluid's 55 pet lower mass 
 concentration and the Gulf fluid's 82 pet 
 lower number concentration. These low 
 values were also reflected in the re- 
 duced combustion yields and production 
 constants. 
 
 Since the particle size can be calcu- 
 lated by two independent methods, the di- 
 ameters obtained by one method should 
 confirm those obtained by the other. The 
 calculations indicate good agreement be- 
 tween the average d m and d 32 . 
 
 However, differences were apparent in 
 comparing the particle sizes of each 
 brand. The small mass concentration of 
 the Shell fluid has lowered the d m , while 
 
 the d 32 approximates the average. The 
 low number and high mass concentrations 
 for the Gulf fluid resulted in the 
 largest calculated d m . By either method, 
 the Gulf fluid produced the largest par- 
 ticles. This was corroborated by its 
 high obscuration rate. 
 
 CONCLUSIONS 
 
 The results of these experiments showed 
 little variation between the transformer 
 fluid brands for CO and CO production, 
 heat release, and heat of combustion. 
 However, the smoke-characteristic calcu- 
 lations indicate that the Gulf fluid 
 produced the heaviest and thickest 
 smoke, while the Shell fluid gen- 
 erally produced the lowest toxic-gas 
 concentrations . 
 
 FUEL COMPARISON RESULTS AND DISCUSSION 
 
 Earlier wood and coal experiments were 
 conducted in the same intermediate-scale 
 fire tunnel. Since the same instrumenta- 
 tion was used in the collection of 
 all the data, it was possible to com- 
 pare them. The gas, heat, and smoke 
 concentrations for the three fuels stud- 
 ied are found in table 7. 
 
 In these experiments, wood produced the 
 most heat relative to the fuel consumed. 
 For better comparison, the other fuels 
 were normalized to this fire intensity. 
 Table 8 has these normalized values. 
 Burning coal produced the most hazardous 
 smoke and gas concentrations. 
 
 TABLE 7. - Gas, heat, and smoke concentrations for 
 the three fuels tested 
 
 Fuel 
 
 CO, 
 ppm 
 
 co 2 , 
 
 ppm 
 
 kW 
 
 No. 
 10 6 p/cm 3 
 
 mg/m 3 
 
 
 145 
 
 76 
 
 113 
 
 6,759 
 
 909 
 
 1,769 
 
 110.2 
 
 5.4 
 
 21.1 
 
 6.04 
 1.68 
 1.09 
 
 49.1 
 
 
 7.7 
 
 
 35.3 
 
 TABLE 8. - Normalized gas and smoke concentrations for 
 the three fuels tested 
 
 Fuel 
 
 CO, 
 PPm 
 
 C0 2 , 
 PPm 
 
 N , 
 10 6 p/cm 3 
 
 mg/m 3 
 
 Wood 
 
 Coal 
 
 Transformer fluid. 
 
 145 
 
 1,551 
 
 590 
 
 6,759 
 
 18,550 
 
 9,239 
 
 6.04 
 
 34.28 
 
 5.69 
 
 49.1 
 157.1 
 184.4 
 
11 
 
 TABLE 9. - Particle size and obscuration rates for 
 the three fuels tested 
 
 Fuel 
 
 d m » 
 
 d32» 
 
 Obs 
 
 >curation rate, 
 
 ym 
 
 urn 
 
 
 pet 
 
 0.223 
 
 ND 
 
 
 8.6 
 
 .177 
 
 0.272 
 
 
 18.2 
 
 .386 
 
 .387 
 
 
 46.1 
 
 Wood 
 
 Coal 
 
 Transformer fluid. . 
 ND Not determined. 
 
 TABLE 10. - Production constants for the three fuels tested 
 
 Fuel 
 
 Wood , 
 
 Coal 
 
 Transformer fluid. 
 
 6co> 
 10- 3 g/kj 
 
 1.58 
 5.26 
 3.14 
 
 Bco 2 > 
 10~ 2 g/kj 
 
 10.44 
 8.89 
 7.71 
 
 10 10 p/kj 
 
 5.79 
 
 11.29 
 
 2.48 
 
 1Q- 1 * g/kj 
 
 4.93 
 4.08 
 7.75 
 
 Transformer fluid produced a thick, 
 dense smoke. It surpassed coal in the 
 mass of smoke particles. It also pro- 
 duced the largest particle size and ob- 
 scuration rate. The values can be 
 found in table 9. 
 
 Note the discrepancy between the d m and 
 the d32 for coal. The d32 may be less 
 reliable because it is determined by the 
 complex refractive index of the par- 
 ticles, which is not precisely known. If 
 the refractive index used in the calcu- 
 lations is incorrect, then the diameter 
 will be inaccurate. However, the obscur- 
 ation rates do indicate that the particle 
 diameter for coal should be smaller than 
 that calculated for transformer fluid. 
 
 Coal generated the most CO and smoke 
 particles, using the formation rate of 
 gas and smoke as a function of the heat 
 produced. Transformer fluid produced the 
 largest mass concentration. These values 
 are confirmed by the normalized concen- 
 trations found in table 8 in which wood 
 was found to generate the most CO 2 rela- 
 tive to the fire size. The production 
 constants are found in table 10. 
 
 Based on these experiments, coal was 
 the most hazardous of the fuels studied. 
 However, transformer fluid produced 
 the thickest smoke with the largest 
 particles. 
 
 SCALING RESULTS AND DISCUSSION 
 
 In liquid-fuel fires, the surface area 
 of the pool determines the intensity of 
 the fire. In order to study this phenom- 
 enon, three round pans (25, 50, and 71 cm 
 diam), filled with transformer fluid, 
 were used. These experiments followed 
 the same test procedure as described 
 above. Texaco transformer fluid was used 
 for all the scaling studies. 
 
 Only one experiment was completed using 
 the 71-cm pan because the flame intensity 
 was so great that the range of most of 
 the smoke and gas detectors was exceeded. 
 A 50-cm pan was used in two experiments 
 but, again, the instruments were reaching 
 
 their upper limits. In order to repeat 
 the experiments using the larger pans, 
 the gas and smoke monitors would have to 
 be disconnected. The remaining four ex- 
 periments were the ones previously 
 reported in the portion on comparison of 
 brands. 
 
 Gasoline scaling studies were reported 
 by Hertzberg. 8 In those experiments, the 
 burning time was plotted as a function 
 of the pool diameter. In figure 4, the 
 
 8 Hertzberg, M. The Theory of Free 
 Ambient Fires. Combust. and Flame, v. 
 21, 1973, p. 202. 
 
12 
 
 0.5 
 
 i 1 r 
 
 i i i i i 
 
 20 
 
 30 40 50 60 
 POOL DIAMETER, cm 
 
 70 80 90 
 
 FIGURE 4.— Burning rates for gasoline and transformer 
 fluid. 
 
 transformer fluid results were super- 
 imposed on his figure 3C The two 
 smaller pan sizes showed good agreement. 
 But the 71-cm pan result was incon- 
 sistent with the gasoline data. The 
 size of the pan approximated the entry 
 width of the tunnel, which could have 
 limited the oxygen supply to the fire. 
 This would have reduced the combustion 
 rate, thus increasing the burning time. 
 To test this theory, higher ventilation 
 rates could be used to increase the 
 available oxygen supply. Future scal- 
 ing experiments should be conducted 
 without the delicate gas and smoke 
 instrumentation. 
 
 C 244 ! 
 
13 
 
 dm 
 
 <*32 
 
 Geo 
 
 GC0 2 
 
 Gx 
 
 H A 
 
 H C 
 
 HCO 
 Keu 
 Kuu 2 
 M co 
 
 Mco, 
 
 conversion factor of a combustion 
 product 
 
 mass mean diameter, ym 
 
 mean particle size, ym 
 
 generated rate of CO, g/s 
 
 generated rate of CO2, g/s 
 
 generated rate of a combustion 
 product, g/s 
 
 actual heat of combustion, kj/g 
 
 net heat of combustion of the 
 fuel, kJ/g 
 
 heat of combustion of CO, kj/g 
 
 stoichiometric yield of CO, g/g 
 
 stoichiometric yield of CO2, g/g 
 
 density of CO, g/(m 3 *ppm) 
 
 density of CO2, g/(m 3 *ppm) 
 
 APPENDIX.— LIST OF SYMBOLS 
 
 Mf fuel mass loss rate, g/s 
 
 M particle mass concentration, 
 mg/ra 3 
 
 N particle number concentration, 
 p/cm 3 
 
 Qa actual heat of combustion, kW 
 
 T transmission of light, volts 
 
 V A ventilation rate, m 3 /s 
 
 Yx 
 
 3x 
 
 AX 
 
 X 
 Pp 
 
 yield of a combustion product, 
 g/g or p/g 
 
 grams of product per unit kilo- 
 joule of heat release 
 
 measured change in a given 
 quantity 
 
 wavelength, ym 
 
 individual particle density, 
 g/cm 3 
 
 A U.S. GOVERNMENT PRINTING OFFICE: 1986-605-017/40.100 
 
 INT.-BU.0F MINES,PGH.,PA. 28377 
 

U.S. Department of the Interior 
 Bureau of Mines— Prod, and Distr. 
 Cochrans Mill Road 
 P.O. Box 18070 
 Pittsburgh, Pa. 15236 
 
 OFFICIAL BUSINESS 
 PENALTY FOR PRIVATE USE. MOO 
 
 | J Do not wish to receive this 
 material, please remove 
 from your mailing list* 
 
 I | Address change* Please 
 correct as indicated* 
 
 AN EQUAL OPPORTUNITY EMPLOYER 
 
♦ v ii v ^ •-** ^^^^^^^^ 
 
 9* 
 
 
 * M> A 9 
 
 
 <, •*ttj« ,& v ^ -?.>- a - - <. -*7?rr«' ,G^ V * 
 
 ^6« 
 
 O ,1-0 . 
 
 
 
 
 ,/.. %*^v v^V v^v V^V % 
 
 \ /fe % ASfeX Xtf*-X .^*te-V ^ 
 
 
 iP^K 
 
 i* A 
 
 
 ~*¥ 
 
 >.,* v 
 
 ^^' 
 
 
 -.- y\ --SK- /\ 
 
 "W 
 
 .♦ViSK'. ^ ^.i 
 
 W 
 
 
 
 *bv' 
 
 /.<J^\ c *.C^*°o /.^% J 
 
 «-.v. s /,^:A y,:^sfe.x ..•♦..sas.-.v. y\-a;t\. .<?^-% * 
 
 ■-.■- 
 
 9 °^ 
 
 i^ 
 
 ^ "^ 
 
 
 -of 
 
 *bv' 
 
 ^°^ 
 
 
 
 
 • -e* 
 
 C^ .^ 
 
 * i. _T' 
 
 ^••"rr,.'"^- 
 
 
 
 
 s 1 ' a v "^ - « 
 
 ^^ V 
 
 
 
 
 
 
 r ^6* 
 
 .♦ M .... 
 
 5, *^«*;* .r 
 
 
 
 ^ v »: 
 
 %- 
 
 
 of . V^- ^. .* •*■ 
 
 v v .»«M% ^ 
 
 
 
 4* *.^«5. 
 
 »■> 
 

 ^0* 
 
 
 J^l'-. V 
 
 
 * 4- 1 
 
 
 
 
 
 «U* V • 
 
 
 
 ^yJKi*% /VB--\ /••!»&.*' 
 
 
 
 Key 
 
 ft* ''ft 
 
 -* 1 ^ ' 
 
 
 
 
 
 .* *K 
 
 • or •, 
 
 O V . ftji 
 
 ^ ... 
 
 0' .'tr^'* V 
 
 "^d* 
 
 .S> <*l* • i i ■AT <>. 
 
 
 jPV 
 
 V '/ 
 
 
 > ^ .** . l ^^'. ^ ^ .Wa\ +* ^ /fife\ ^ ^ .^Va\ V,^ 
 
 
 
 ^^ 
 
 
 ^^' 
 
 
 
 o > . 
 
 » ^ : . 
 
 <. *••»• ^ V ^^^^'a' V < A ' , -f^^.G^ ^^TTT.'.A 
 
 ' .^' ....,^. "•* ^ 
 
 "oV' 
 
 
 ^d* 
 
 V-^'V* 'V*^^/ ^' ; ^"'\^' % % -"-"f° 
 
 
 •bv' 
 
 ^°^ 
 
 
 
 v^ v 
 
 
 
 v \.^ .*»&•. \,^ .*' 
 
 vv 
 
 
 
 
 * v 
 
 
 ,%^ 
 
 . (if O »^« 
 
 .- 4 ^ ^o -*W' A^ 
 
 4 o 
 
 
 
 
 6»" ^ *^f.T* A 
 
 n \ _ O « * - ^^-\ «v . 
 
 1 • * o 
 
 % * 
 
 
 *bV 
 
 
 
 
 T?r.- a 
 
 
 O 
 
 ^ '••"»• ^ 
 
 ^% ^t, 
 
 :- ^o« :^K' ^> ^^^** A0< ?■ 
 
 :< 
 
 V^\/ %-^V .V^\/ %- 
 
 v »!• 
 
 '■ ' %/ ' :'Mk: \<f' :£&; V* ' 
 
 o OV cp 
 
 
 ^ ** »^ 
 
 
 *" /v ^ 
 
LIBRARY OF CONGRESS 
 
 002 955 954 5 
 
 
 ■