Document 065173nyRndG1xBDQ5N295Zvd

IEEE Symposium on E le c tric a l In su la tio n , Montreal, Quebec, Canada June 1976 a flammability test fob dielectric fluids L. Maivdlcom and R. L. Millar West Inghouse Research Laboratories Pittsburgh, Pennsylvania 15235 Suaseary A nod If Leaden of the Cleveland Open Cup Test, ASTM D92, was developed for dlelectrte fluids in nonfire hazard power capacitors and tranaloneera. This was on the basis of considering flee situations in these apparatus. Two criteria for non-flaemability were deduced and applied in tbia teat: that burning nay only be initiated at the boiling point, and that burning cease on cooling when separated iron the heat source. A number ot fluids wore tested. Effects of soot are noted. Introduction la undertaking to develop alternate dielectric fluids for trl- and penta- chloroblphenyls. for non flammable or non-lire hazard power capacitors and distribution transforstri. It was realized that there La a need eo determine the eeeential fluid (non-) flammability properties for these applications. Possi ble problems of ecological pollution by chlorinated biphenyls, that nay halt their continued use, have necessitated searches for alternates. It la the fluid component la impregnated power equipment that causes particular consideration for lira-hazard. Attention was directed to understand why power equipment has bean satisfactorily non-fleasable with these chlori nated biphenyls - as shown by experience of AO years so that the same could be achieved with new fluids. This was deemed necessary because It was realized that the Cleveland Open Cup Teat, ASl>l D92, which is extensively used to determine flammability of dielectric fluids, and according to which trl- and pence- chlorinated biphenyls aro classified es non flammable. nay not always be sufficiently reliable, "bon-flanmablc" has really only relative significance, rather implying difficultly flammable, es probably almoet anything will burn if sufficient hast and oxygen are made available to It. L. J. Frisco and E. N. Mat heal caused polytetraf luoroethylcne (Teflon) to bu m . In a simple test where the oxygen concentration and temperature were high. Teats for (non-) flammability of a liquid generally give an intensity factor chat will cause the liquid to lgnits or burn continuously, such as temperature In the Cleveland Open Cup Tast and oxygen concentration in the Critical Oxygao Index Test.^ But the flammability or fire hazard problcn due to a fluid in a power capacitor or transformer Is caused by an arc, which can causa chlorinated biphenyls to burn in Its vicinity. The fact that non-flammability is achieved with them. In these applications, is thsrefore not uaequlvlcally associated with their high flame points in the Cleveland Open Cup Teat. Only when the latter occure can there be extensive burning of the fluid, with oxygen available froa the aablent, since there Le no gas space In a power capaci tor and the aoounc of gas above the fluid In a trans former le very limited. Then, the fire could spread. Since the fluid la ignited by an arc, which due to its very high temperature could Ignite almost any fluid, the magnitude of ice ignition temperature has little effect at this point. The situation la helpless within the unit while arcing persista. Capacitors Impregnated with trichloroblphenyl have been consumed almost com pletely when the feule currant euatelning Che arc waa not interrupted for a long time, which could occur only rarely whan the circuit breaker or fuse felled to operate. However, burning within a unit should be ex tinguished very soon after the current is interrupted. But moreover. It le reasonable to Insist Chet the burn ing fluid be extinguished immediately when such fluid escapes from the unit, and hence from the source of heat. This would prevent spreading of the fire. It should be noted that, under these circumstances, the tesq>ereture in the mess of a burning fluid cannot exceed its atmospheric preaeure boiling point. This la because the fluid Is wnconflned, due to the rupture of ice container. These consideraClone lead to two assentisi flamma bility characteristics of a suitable fluid for non flammable or non-fire hazard power equipment. It sheuld not burn freely In air below ite boiling point. And, If it does burn at thst tempersture, the flame la extinguished as soon as it atarte to cool when removed free the eource of heat. The two ere inter related, etnee, if the ignition temperature of a fluid is below the boiling point, it is possible that it * could continuously burn at iaaat in that temperatura interval, either during cooling from the boiling tem perature, or with the flame increasing Its temperature If ignited Below the boiling temperature. An additional problem it presented when fluid is heated to the atmospheric pressure boiling point end higher temperatures. Some of it will be vaporized Into the ambient, which could constitute an explosive mixture with the oxygen. This hazard can ba amplified or even dominated by the presence of flammable volatile gaeee, such as hydrogen, generated by erelng in the fluid. Such e situation is analagous to releasing compressed flammable gas, from e tank, Into an area where there are flanea or sparks. This matter of a potencial explosion arising from a fluid exposed to a failure arc should be dealt with separately, and not from the point of view of flammability, although there is some relationship. Background - F l i n b l l l t y Considerations __________ for Dielectric Fluids_________ This is a simple analysis of a bad condition of a fire In a fluid tilled power capacitor or transferner, as a basis of thu required fluid flammability proper ties to avoid spreading of the fire. The events lead ing to a fire In the unit are insulation breakdown (due to deterloracion or overvoltage), then high current arcing, trtiich is followed by rupture of the container. Experimental and Results A flammability teat was accordingly devised to determine ehe ignition temperature for continuous burn ing of a fluid, whether It wee boiling then, and whether it would cease to burn after it waa separated from the source of the heat. The equipment consisted of e Cleveland Opeo Cup Test cell, or cup, a high temperature hot plat (Corning PC-33), an aluminum block 6 in square by 2 in, 170 NPC00026369 753706 n thermocouple and bridge, and a Runscn burner. The tlirrnoccuplc lend tiros vre f a s t e n e d to th cup handle, and lea sensing bead was located Inside the cup, about 1/3J in fro the bottom. A snail piece of Teflon vaa attached to the bead to weigh it down and to keep It fren touching the heated metal. The hot plate and the alunlnuo block were next to each other with thetr surfaces at about the same level, so that the cup could be easily moved froe the hot pete onto ihe aluminum block. Thsy were located In a draft-free enclosure. For any tost, the hot plate was always at its hignest tampers ture, and tha aluminum block was Initially at room temperature. In a typical tase, a sample of fluid In the cup was Ignited with the burner to continuous burning while heated on the hot plate, and then moved onto the sur face ef the aluminum block and coolod until Che flasM extinguished, temperature being recorded throughout, and also time during tha cooling. At tha time of igni tion the fluid was checked visually If it was boiling. The cup was claaned carefully before each test, removing an* pool, and was filiad with fluid to the designated level. A specimen was assumed to be burning continuously when the flame persisted for 10 seconds. When a fluid was tasted for the first time, the burner flsoe was appllad for about one second every 5-10 C rise. I n e subsequent test the flame might be kept on the surface for several seconds at a lowar bulk fluid temperature to affect forced ignition. Or, the sample was ignited at a higher temperature than in*the first test. This was the most meaningful cast, particularly when the speciseti urns ignited at ot near boiling. The following fluida were tested; pencechiorobiphenyl (Aroclor 125*), trlehloroblphenyl (MCS 1016), mineral oil, various solutions of trlehloroblphenyl and mineral oil, cricresyland tributyl phosphate, dilsooonyl phthalace, and polydimethyl end polymechylphenyl silicone (Dow Corning 200, 20 cs, end DC 350). The trlehloroblphenyl fluid end mineral oil served ee re spective "markers" of non-fire hazardous and fire haz ardous fluids In this teat. Solutions of thesa two vers checked to observe the effects of a controlled chenge in flaemablllty. This le also of practical interest because both fluids ere often used at the seme manufacturing sits, and may be mixed inadvertantly. Tricreeyl phosphate, the silicones, and tha phthalate ester were of particular interest because they ere oc casionally considered as non-flammabls. about 35 C. both silicone oils required relatively high teeprrature* to be Ipnited to continuous burning, Figs. 2 and 3. but once this occurred, they continue-, to burn during cooling. In fact, polvdimethvl silicone continued to burn frtmi about 290 C to less than 80 C without its flume extinguishing. It should be noted that the ignition temperatures of the silicone oils and tricreayl phosphate ero about as high or higher than that of trlehloroblphenyl, but they era more flesmabla by the criteria presented here. It was observed Chat Che chlorinated biphenyl fluids produced considerable soot end flashes of fla-e at considerable distances from the specimen in this test, as well as in the Cleveland Open Cup Text, whan exposed to a flame at or naar boiling. The soot and flashes of flame aeem to be associated. Sore fluid esy be sorbed on the soot, and, perhapa by catalysis, may burn even at a relatively low temperature. Tha possibility ef a fire spreading this way should be con sidered. But it is not as dangerous as when a fluid like mineral oil is ignited, where the bulk fluid burns continuously, while here only vary avail amounts are involved. Dlxrusslon The results of this tost art quite reasonable. It ia significant that positive flammability results, indicating fire hazard, ware obtained with polynethyl phenyl silicons (DC 300), pclydlmethyl silicone and tricreeyl phosphate whose Cleveland Open Cup Test fire paints are either about the same as.or higher than trlchloroblphenyl. These fluids burn freely, after being ignited at atmospheric conditions, because of their relatively high hydrogen contents. However, the aagnitude of the Cleveland Open Cup test fire point, or the ignition temperature in its modification here, ie significant when considered in conjunction with ehe non-fLanaabllity criteria that Separate fluids into two distinct catcgorlss. Firs Point magnitude say then be considered with respect tc each category to further subdivide different fluids. This ia especially Important for working with fluids, especially ones Chat are a fire hazard in power equip ment, like mineral oil, where e fire through e shop mishap or auto combustion has a finite, hopefully very email, probability. Results of these tests are shown In Figs. 1, 2 and 3 whera the temperature of the fluid is plotted as a function of Cine, from the ignition to e continuous, 10 second, flema to extinction of the flame Avile cooling. The criteria for a fluid to be eon-fire hazardous In power equipment ere satisfied by only pentachloroblphenyl, sod by trlehloroblphenyl end its solutions of up to 10 w 2 mineral oil, rig. 1. In fact, pentaehloroblphenyl did not even burn for 10 seconds when ignited at Its boiling point. A solution of 20 w Z mineral oil In trlehloroblphenyl could be ignited well below Its boil ing point, and then continued to burn uotll its temper ature decreased by about *0 C. Hineral oil Itself con tinued te burn for some time daring cooling until the (lame estinguiehed, its temperature, in fact, increasing during the ignition period and during the initial cool ing period. Fig. 2. This effect was observed In some of the other teats hers. Tributyl phosphate. Fig. 2. showed this effect of continuing to burn during cooling even when the sample wms "forced" to ignite. On the other hand, tricreayl phosphate burned continuously, for about 10 escanda, end then the flame extinguished Immediately, when It was ignited at 355 C end 366 C. Fig. 3. When It was ignited at 420 C, still below boiling, its flame extinguished only after it cooled Acknowledgement The suchore wish to thank Dr. T. V. Dakin, Westlngbouse Research Laboratories, for his suggestions to work with the Cleveland Open Cup Test procedure end to follow Che temperature of the burning fluid, la this investigation. . Rsfarencas 1. L. J. Frisco and K. K. hath, m a s a Report. Evalua tion of Thin Wall Spacecraft Wiring, HAS 9-4549, September 28. 1965. 2. ERA Symposium - Flammability and Smoke Testing Techniques, Tha Institution of Electrical Ln* inser, February 6, 1974, ERA 74-22. Fspers by E. 1. Chars ley end R. A. Schult, end by A. C. Day where an adaption of the Critical Oxygen Index Test to liquids it presentad. 171 NPC00026370 753707 <00 -- 1-- -- 1---- -- r-- -- i-- -- i-- -- i-- b 6 360 1 P4 320 3 p . s - b 1 5 u m - b lo llin g Temperature^, 1 Flame Extinguished >- - 9 E \ \\ - 1 TrlchlorflMphenyf _ 2 TrkhloroWphenyl 200 2 n * Mineral 00 -- - 3 TrtcmoroMphtnyl _ 5 w* Mineral 08 160 4 Trlchlorebiphenyl -- 10w * Mineral OU _ - 5 Trlehloroblphenyi 120 - 20w% Mineral O il-- A fwlaniuitcifnhvlonrroiJoiiinphnafmnydi _ l __ 1__ __ i__ i 1 i 1 0 12 34 h H Coding Time, rains/tes toe Ignition Hy. 1-Flam nubfiity temperature reiaMonsMss tor dielectric fluids In a n o d ifM Cleveland Open Cup Ia si Sufttalpts refer to different lasts an the same Quid 400 -- i-- -- i-- -- i-- -- 1-- - 1 i 360 - 320 a 280 - -240 1c 200 - 160 L 120 - 10 i ___ i uH sac v*lgnRlon (Flame [tinquiihel xForced Ignlion 1 Mineral Oil 2 Tributyl Phosphite 3 Mydlmathyt Silicone - - 2 ,% - __ 1__ 23 4 Cooling Time, minutes __ i__ Fig. 2-Flim m aM Ry temperature relationships lor dietedrlc Quids In a meddled Cleveland Open Cup lest, Subscripts refer to different tests an the same fluid. i F Y )-fle a N M M ity temperature relationships ferdlvfectflc Quids In a modified Cleveland Open Cup tost. Sufcseriptt refer to different tests an the same fluid m NPC00026371 753708 APPROVED: J. "Siri;'Division Manager Chemical Sciences POOR QUALITY ORIGINAL * - r " * -r-K ' .** * .. * A FLAMMABILITY TEST FOR DIELECTRIC FLUIDS L. Maendelcoran end R. L. Miller *- ? * * ? ! * ~ ` ^ * * ' ' 1 il :.*v vr **i. ^ . f'Ti PI December 31, 1975 753709