Document EvEeM52GoGBmmbz6yQmqMb60V

r / i .... O /- -j WESTINGHOUSE ELECTRIC CORPORATION TRANSFORMER. DIVISIONS MATERIALS ENGINEERING DIVISION ' M a l i s . Eng. Dopt. Report #696 Fluid Insulation Section Proprietary Claas 1 July 25, 1969 INVESTIGATION OF THE GASES PRODUCED BY THE ELECTRICAL ARCING OF INERTEEN I. Synopsis INERTEEN 56201CM, which is AROCLOR 1262 and a small percentage of phenoxypropene oxide as a scavenger, produced sufficient combustible gases when arced to burn when those gases were mixed with air. Under similar tests, INERTEEN 56201KA, which is a mixture of 707. AROCLOR 1256 and 301 trichlorobenzene, produced only about 6X combustible gases, which was below the lower combustible limit of hydrogen-acetylene-air mixtures. The amounts of combustible gases produced when various combinations of the AROCLORS and chlorinated benzenes were mixed was dependent upon the ratio of chlor ine to hydrogen in the molecules. Below a one to one chlorine to hydrogen ratio combustible gases were formed. II. Introduc t ion INERTEEN is the Westinghouse trade name. for a family of chlorinated hydrocarbon liquids which are non-combustible. The generic name for these materials it tfaskarel.M Other manufacturers of electrical apparatus use various other trade names for this family of non-combustible Insulating liquids. However, all of the trade marked brands of askarels contain AROCLOR which is the trade name of Monsanto Chemical Company1^ for their scries of chlorinated biphenyls and chlorinated polyphenyls. These compounds are chemically stable, fire-resistant, heat stable, and have high dielectric strength. Westinghouse has used various blends of the AROCLOR and chlor inated benzenes for many years. The most recent was INERTEEN 56201KA, which consisted of 7OX AROCLOR 1256 (pentochlorodipheny1) and 307, trichlorobenzene. A recent study of the properties and the economics of AROCLOR 1262 (trichlorodiphcnyl),. which is used in capacitors, brought about a change to 100X AROCLOR 1262 for use in transformers. AROCLOR 1262 was assigned the material number 56201CM. Both INERTEEN 56201KA and INERTEEN 562U1CM have been approved by the Underwriters Laboratories^*4 for use in transformers. Matjs. -Eng. Dept.,, Report #694 Fluid Insulation Section Pag 2 Concern arose as to whether the gases formed on arcing INERTEEN 54201CM were similar to the gases formed on arcing INERTEEN 54201KA due to the change in the amounts of chlorine and hydrogen present in the compounds. TnbLc l was prepared to show the hydrogen and chlorine which can be present in chlorinated diphenyl and chLorlnaced benzene. From this it can be seen that In INERTEEN 54201KA the ratio of hydrogen to chLorine la 1 to 1, while in 54201CM it is 7 to 3. On arc ing hydrogen and chlorine atoms are released. As long as there is an amount of chlorine equal to or in excess of the amount of hydrogen, there will be little or no free hydrogen resulting. When the ratio of chlorine to hydrogen falls below one to one, hydrogen will be found in quantities proportionate to the ratio. With INERTEEN 5420LKA the formed gas should approach 100% hydrogen chloride, but in the case of 54201CM an excess of hydrogen approaching 407. should be present. III. Conclusions 1. Rapid burning or explosive mixtures resulted from mixing with air, gases that were produced by electrically arcing ROCLOR 1242 as indicated by Monsanto's test data shown in ''able V. 2. The percentages of combustible gases resulting from the arcing of ROCLOR 1242 varied somewhat among the three locations, (Monsanto, Westinghouse Sharon, West inghouse Canada) but allowing for small differences in techniques the agreement was good. A general average of the arced gases shoved 39 to 46% combustible. (See Ta ble IV). 3. Tests by Monsanto Company showed that the combustible limits of the hydro gen-acetylene-hydrogen chloride mixture were narrower than for hydrogen alone,. For hydrogen and sir the lower combustible limit la 4% hydrogen, and the upper limit is 75% hydrogen. The lower combustible limit of burning was 4.4% hydrogen but the upper limit was less that 20% because of the HC1 present. (See Table VI). 4. Arcing teats made on INERTEEN 54201KA, (70% 1254 AR0CL0R-30% trichlorobcnzeue) did not produce sufficient combustible gases needed to burn under various tests. The total combustible including hydrogen and acetylene were not greater than 4% on any of the tests. (See Table IV). 5. Monsanto's tests of a mixture of 80% AROCLOR 1242 and 20% triretrachlorobe.nzene when arced produced hydrogen and acetylene in the amount of 16 hydrogen and 10 acetylene using a 1500 volt spark. Westinghouse tests made using a carbon arc from a welder showed 16% hydrogen and 4.5% acetylene. Both of these compositions will burn, (Sec Table II for the Westinghouse results and Table V for Monsanto re sults. 6. Canadian Westinghouse tests using a pressure sensing device showed that a spark in the gas space of a transformer with AROCLOR 1242 which had arced would pro duce significant pressures. A 10 pound diaphragm ruptured readily. Similar testings using INERTEEN 34201KA showed only a slight pressure rise with no rupture of a 10 p o u n d diaphragm. (See Figure 3). 7. No measureable amounts of combustible gases wore formed when either INERTEF.N 54 201 CM or 34201KA was subjected to the corona type electrical discharges in the Murrell Test Cell or the Pirelli Test Cell. NPC00032610 905133 Halts. Eng. Dept. Report 0694 Fluid Lnsulotion Section XV, Recommendations Page 3 Due to the combustible gases which may be formed when INERTEEN 54201CM is subjected to an electric arc, it is recommended chat investigations be conducted to determine the potential quantities of such gases which might be generated in a transformer and the probability of a resultant hazard thus occuring. V, Experimental Procedure and Results Samples of INERTEEN 54201KA and 54201CM were subject to an electric discharge in the Merrell Test Ceil described in ASTM D-2298 for Insulating Oil. The sample was admitted under vacuum into the evacuated call at room temperature and 10,000 volts were applied for 1000 minutes. Any evolved gas was collected in an evacuated glass sample container. These samples were then tested by Hass Spectrometer at th Westinghouse R & D Laboratory. The analysis of the gas collected from the discharge in both 54201KA and 54201CM showed approximately 80% nitrogen and 18% oxygen with no combusti ble gases being present. Samples of both materials were then subject to electric stress in the siodified Pirelli' Test Cell. This test cell is so arranged that the electrical stress is applied at the Interface of gas and liquid. The gas used in the test cell was nitrogen. The liquid was degassed by vacuum and heat followed by a period of gas saturation. The gas flow was then stopped and the inter face chamber isolated from the atmosphere. A stress of 100 voits/mil was applied for one hour at the gas liquid interface. The resultant gas gen erated vaa collected in a glass sampling bottle. Both materials were stressed ut 80#C and at 100*C. The collected gas samples were then analyzed at the Westinghouse R & D Laboratories using the Hass Spectrometer. In every case the analysis of the gas showed approximately 817. nitrogen and 18% ox ygen with no combustible gases present. The results of these tests made it appdrent that a ore Intense electrical stress would be required to decompose these materials. Therefore, a setup was made using an arc between two carbon electrodes to breakdown the liquid samples. The carbon electrodes were 5 mm diameter with a gap set as large as possible, but small enough for spontaneous arcing, which was approximately 1/32". The gap could be adjusted by rotating the electrode supports consis ting of a brass fixture supported by 1/4" brass rods mounted in Hicarta spa cers. The power was supplied by a Westinghouse 8N64 D. C. Welder. The voltage was 40 volts under load and Che current was set at 45 amps. The gas generated was collected by displacing liquid in a glass tube on top of an Inver ted. glass funnel located right above the arc. The test set-up is shown in Figure L. Figure 4 shows the fixture ready for arcing and Figure 5 after arcing. Samples of gas thus generated were collected on breaking down INERTEEN 54201KA, 54201GM, and a mixture of 80% 54201CM and 20% tritetrachiorobezune. Samples of the gas from 54201KA and 54201CM were analyzed at Westinghouse R & D Laboratories using the Maas Spectrometer and samples of gas from 54201KA, 5420ICM, and the mixture of 807, 54201CM and 20% tritctrachlorobezcnc were analyzed in our laboratory on the Gas Chromatograph. The Gas Chromatograph is a Burrell Corporation, KromoTog Model K-i. The results of these tests are recorded in Table II. V NPC00032611 905134 Macis. Eng. Dept. Report #694 Fluid Inaulation Section Page 4 Samples of the gas generated from the breakdown of the three liquids were bubbled through a .097 N sodium hydroxide solution and the solution w n then titrated with 0.1N hydrochloric acid solution to determine the amount of H Cl in the gas sample. These results arc recorded in Table H I . Similar Tests have been conducted by Linkert and Gray of Canadian Westinghouse5 and Ur. R. A. Munch of Monsanto Chemical Company. Table IV is e comparison of the results obtain by Westinghouse, Monsanto, and Canadian Wes doghouse. The agreement among the various tests results was good. The Mass Spectrometer identified small amounts of various other gases which could not be identified with the Gas Chrometograph, but the primary interest was in the hydrogen and hydrogen chloride. The amount of hydrogen gener ated when the INERTEEN 5420LCM was broken down quite high In comparison with the 5420LKA end the percentage of hydrogen .chloride waa signifleantly lower. This confirms experimentally what was to be expected theoretically due to the imbalance of hydrogen and chloride in the INERTEEN 54201CM. \ VI. Discussion The experimental resuLts and the theoretical considerations both agreed that in case of an electrical arc sufficient to break the carbon-hydrogen and carbon-chlorine bonds INERTEEN 5420ICM will form gases containing a significant amount of hydrogen. The question then to be considered is the effect this would have on Che use of INERTEEN 54201CM as an Insulating liquid in transformers. It needs to be considered whether we have the ! nonflammability and safety expected. Information in the Bureau of Mines Bulletin on "Limits of Flamraability of Gases and Vapors" indicates that for hydrogen in air the lower explosive limit Is approximately 4.1 percent hydrogen and the upper limit is about 74.2 percent hydrogen, From their information on flammability of hydrogenoxygen ratio, we have plotted Figure 2, which indicates the range of gas mixtures which could cause problem if the INERTEEN were subjected to an electric arc. Table V shows data collected by Monsanto using a modification of the ^ apparatus used by the Bureau of Mlncs& to determine the fiammabiLity of the gases generated by arcing AR0CL0R 1242 and by arcing*a mixture of 807. 1242 and 20Z tritetrachlorobenzene. Figure 3 is a sketch of the apparatus used by Canadian Westinghousfc to compare the effect of a spark on the gases generated by arcing AR0CL0R 1242 and AROCLOR 1254. References 1. Monsanto Chemical Company, Askarel Inspection and Maintenance Guide, September 1963. NPC00032612 905135 Mails. Eng. Dept. Report #694 Fluid Insulation Section Page 5 2. Monsanto Chemical Company, Technical Bulletin No. PL-306, " AROCLOR Plasticyers" December I960. 3. Underwriters Laboratory, Inc, Report on Dielectric Medium, File MH2689, Assignment 67K4825-W February 1, 1968, 4. Underwriters Laboratory, Inc,, Report on Dielectric Medium, File HH2689 application No. 54C627, March 23, 1961. 3. Canadian Westinghouse Company Limited, Research and Development Laborabories, Report No. 02021-350, "Chromatographic Analysis of the Caa Produced by the Electrical Arcing of Aroclor 1242", by D. Gray# and VI. G. Linkert, September 20, 1968. - 6. United States Department of Interior, Bureau of Mines, Bulletin 503 "Limits of Flammability of Gases and Vapors" by H. F. Coward and G. W. Jones, 1952. T. K. Sloat, Manager Fluid Insuistiou Section H. A. Pearce, Jr., Engineer Fluid Insulation Section Mater 'si'cppapa ager Engineering Department /et NPC00032613 905136 Proprietary Class 1 Matla. Eng . Dcpt., Report 694 Fluids Insulation Section COPIES WILL BE SENT TO CANADIAN WESTINGHOUSE, Hamilton, Ontario CANADIAN WESTINGHOUSE, Hamilton, Ontario CANADIAN WESTINGHOUSE, Hamilton, Ontario CANADIAN WESTINGHOUSE, London, Ontario SHARON WORKS, ML-382, Dlstr. Tranair. Division SHARON WORKS, ML-017, Power Transfr. Division SHARON WORKS, ML-384, Power Transfr. Division SHARON WORKS', ML-381, Power Transfr. Division SHARON WORKS. ML-017. Power Transfr. Division SHARON WORKS, ML-017, Power Transfr. Division SHARON WORKS, ML-382, Dlstr. Transfr. Division SOUTH BOSTON WORKS, Power Transfr. Division SOUTH BOSTON WORKS, Power Transfr. Division CANADIAN WESTINGHOUSE, Hamilton, Ontario a) (2) (3) W (5) (6) (7) (8) . (91 (10) (ID (12) (13) (0J L. McEu'en J. Minehiu E. Siliett A. L. Tanton F. J. Brutt If. A. Pearce, Jr. R. B. Pherson R. L. Schwab H. R. Sheppard O'*-). T. K. Sloat Je J. Actleford, Jr. R. Hollister G. C Wilburn J. L. Harbell (6/5/73) Sensitivity Categories* for Technical Documents PROPRIETARY CLASS 1 - Strictly limited. Cannot under any circumstances be dis tributed outside the Company. Inside the Company, recipient must bave a specific need for the information in the conduct of his assigned respon sibilities. All copies of these reports are to be assigned to recipients as shown on the distribution list. PROPRIETARY CLASS 2 - Distribution only within the Company. Copies of such documents cannot be either given or shown to anyone outside the Corpora tion except (1) licensees and (2) the Government tn proposals for contracts. PROPRIETARY CLASS 3 - Unlimited distribution, both inside and outside the Company. PROPRIETARY CLASS 4 - Development Contract Reports. Strictly limited. Copies of such documents can be given only to those persons specifically desig nated by the Contractor or the Westinghouse Contract Supervisor. All copies of these reports are to be numbered and assigned to recipients as shown on the distribution list. Determined by the author with the approval of his Section Manager. For further information, see "Guide for Classification and Distribution of Internal Communications." NPC00032614 905137 Matin. Eng. Dope. Report 0694 Fluid Insulation Section table I Hydrogen - Chlorine Ratios in Chlorinated Diphenyl and Chlorinated Ben2e.ne . AROCLOR 1232 1242 1248 1254 1260 Chlorinated Diphenyl Number of Hydrogen on Ring Percent Chlorine 10 0 9 18.9 8 32*.0 7 42.0 6 48.6 5 54.3 4 59.0 3 63.0 Number of Chlorine on Ring 0 1 2 3 4 5 6 7 Chlorinated Benzene Number of Hydrogen on Ring Percent Chlorine 60 5 .31.6 4 48.2 3 58.5 2 65.8 1 70.8 0 74.8 Number of Chlorine on Ring 0 1 2 3 4 5 6 1NERTEEN 54201K 1NERTEEN 54201CH 807. AROCLOR 1242-20X Tritetrach Lorobeniene Hydrogen 4.4 . 7.0 6.1 Chlorine 4.4 3.0 3,1 Ratio 1-i 7-3 2-1 NPC00032615 905138 Hatls. Eng. Dept. Report #694 Fluid Insulation Section TABLE II Analysis of Gas Generated by an . Electric Arc In INERTEEN 5420LKA and 5*201CM 54201CM Mass Spectrograph Cas Chromatograph 5420IKA Mass Spectrograph Gas Chromatograph 80% 54201CM & 20 X Tritetrachlorobenrene Gas Chromatograph Hydrogen X Gas Content Carbon Dioxide.& Hydrogen Nitrogen Acetylene Carbon Chloride X X Monoxide % X Oxygen X 33 27 2 .3 10 5 <4 13 4 6 7 1 .4 L 1 49 2 60* 71 3 93* 1 16 5' 4 1 73* 1 * By difference TABLE III Percent of UCL in Gas Samples INERTEEN 54201CM INERTEEN 54201KA 80X 5420]CM + Tricetcchlorobeneene * Cl 55X 82% 62t NPC00032616 905139 Metis Eng. Depc. Report #694 Fluid Insulation Section TABLE IV Comparison of Average Percentages of Hydrogen, Acetylene and Hydrogen Chloride Obtained by West1nehouse, Honr.attlo and Canadian Westinghousc_____ _____ 1NERTEEN 54201CM Westinghouse Hass Spectrograph Westinghouse Gas Chromatograph Monsanto Gas Chromatograph Canadian W Gas Chromatograph Gas Content Percent Hydrogen Acetyl en Hydrogen Chloride 33 6 27 7 24 14 35 7 49 60 62 45 INERTEEN 54201KA Westinghouse Mass Spectrograph Westinghouse Gas Chromatograph Monsanto Gas Chromatograph Canadian W Gas Chromatograph 2 .3 2,5 2 1 .4 .3 2 71 93 96 87 NPC00032617 905140 Matla. Eng. Dept., Report #694 Fluid Insulation Section TABLE V Flamiahi11 tv Data From Monsanto Monsanto broke down AROCLOR 1242 and a mixture of 80% AROCLOR 1242 and 20% Tritetrachlorobensene using a 240 volt arc and a 15000 volt spark Gases Formed Sample AROCLOR 1242 AROCLOR 1242 80% AROCLOR 1242 20% rrcB 80% AROCLOR 1242 20% TTCB Voltaae 240 vole AC Arc S amps 15000 volt spark 240 volt AC Arc 5 amps 15000 volt spark Hydrogen 20% 29% 16% 13% Acetylene 13% 17% 10% 4% Hydrogen Chloride 67% 58% 75% 83% Using a modified Bureau of Mines Test using a 5cm tube 5 ft. long synthetic gases were fed into the tube with the balance air to determine minimum and maximum combustible range. Min. - Max. Combustible Ranee Percent Gas Percent Air 2 c 22 KC1 Minimum Maximum 20 13 67 46 78 29 17 16 10 13 4 58 75 83 46 51 42 85 73 58 Ail of these gas mixtures were combustible when mixed with percentages of air in the listed min. max. ranges. NPC00032618 905141 Metis. Eng. Dept., Report #694 Fluid Insulation Section TABLE VI Percentage of Hydrogen, Acetylene, and Oxygen in the Combustible Mixtrues based on Monsanto Tests as per Table V Generated Gas H i c 2% 20 13 20 13 29 17 29 17 1 Composition of Combustible Mix Air 2 C ^ 2 46 10.8 7.0 9 78 4.4 2.9 15 46 15.6 9.2 9 85 4.4 2.6 17 NPC00032619 905142 Fluid Insulation Section FIXTURE FOR ARCING LIQUIDS .T&S&S FIGURE I NPC00032620 905143 Mrttl's-. Dept. Report IfeflA Fluid .Insulation Section BASIS - IOO LITERS ARCED 1242 AROCLOR FIGURE 2 LITERS DILUTION WITH AIR NPC00032621 905144 O M H M U I M i i . I r i L ^ I I H U l i U U o L loL. I U t" I U h ^ U u i U U ^ i u H ILo i ` 1242 Arc created and pressure built up to 3# Gauge. The gas was sampled and found to contain OX hydrogen. Spark created acid pressure sensor rose to 43# in .03 seconds. Returned to 3# in 0.3 seconds. 150# rupture disk was stretched but not broken. 10# rupture disk substituted for 150# disk. Ruptured in 3 milliseconds after sparking. 1256 Arc repeated to 3# Gauge, Sparked and pressure sensor rose to 4#. 150# rupture disk replaced with 10# paper disk. Ho rupture ceured on subsequent trials. FIGURE 3 NPC00032622 905145 Motla. Enf Dope Report #694 Fluid lntulClon Section r *! i I i i l i i. NPC00032623 905146 Mails. Eng. Dept. Report #694 Fluid Insulae Ion Seccin X' NPC00032624 905147