Document NGJaq5Y6Zj5Zxa08nn4dKMEyp

COMMENTS CF THE UTILITY SOLID WASTE ACTIVITIES GROUP, THE EDISON ELECTRIC INSTITUTE, THE AMERICAN PUBLIC POWER ASSOCIATION AND THE NATIONAL RURAL ELECTRIC COOPERATIVE ASSOCIATION IN RESPONSE TO ADVANCE NOTICE OF PROPOSED RULEMARINC CONCERNING RISKS POSED BY FIRES INVOLVING PC3-CONTAINING TRANSFORMERS 49 Fed. Re?. 11070 (March 23, 1994) Docket No. OPTS-92035 of Counseli WALD, HARJULADER ROSS 1300 Nineteenth Street, N.W. Washington, D.C. 20039 Submitted to The United Sta Environmental June 19, 1994 MQNS 213857 U99 ABLE OF CONTENTS INTRODUCTION AND SUMMARY OF POSITION ........................................................ 2 I. DATA REGARDING USE OF ASKAREL TRANSFORMERS AND FIRE-RELATED INCIDENTS .................................................................................... * A, Inventory of Utility-Ovnad A*karal Transformara . . . , - 1. Current Invantory : 2. Voluntary Phasa-Out Program* , , , . ........................ a S. Fraquancy and Natura of Fira Incidant*................................ _; 1. Fraquancy of tha Incidant* . , ..................................... . 33 2. Exposura Laval* ..................................... .... _; 3. Claan-Up Coat* ........ .......................................... i; 4. Factors Dacraasmg Fira Incidant*................................. 12 C. Ratrofilling Expananca .................................................................. 13 D. Impact of RPC Data on EPA Initial Estiratas ... .; 1. Fraquancy of Fira Incidant* ..................................................... li 2. Claan-up Cost* ........................................................................... 1: II. REPRESENTATIVES OF THE BINGHAMTON, SAN FRANCISCO AND CHICAGO INCIDENTS ........................................................................... 15 A, Fira* Associatad with Askaral Transformers Are Rare . .i B. Potantial Risks Ara Ovarstatad ................................................... 1. Causa* of Firas May Ba pravantabla . ............................. 13 2. Laval* of Contaminants Formad IsInconsistant . . .3 3. Confounding Sourca* of PCCFs and PCDDs ......................... 4. Formation of Raactlon Products ..................................... 2. 22 5. Exposura and Haaith Effacts of PCSs, PCDFs and PCDDs...................................................................... 2; C. Conclusions Ragarding tha Binghamton, San Francisco and Chicago Incidant* ................... ...... 2: MGNS 213858 III, MECHANISMS OF PCDf AND PC3D FORMATION IV, OTHER RELEVANT FACTORS ............................ A. Substitute Fluids for PCSs .... B. Risk from Contaminated Water , , , C. Tbe Consequences of Building Fires CCNCLCSICN ...................................................................... HONS 213859 The following comments are submitted or behalf of the Utility Solid Waste Activities Croup (USWAG), the Edison Electric Institute (EEI), the American Public Power Association IAPPA), end the National Rural Electric Cooperative Association (NRECA) in response to an Advance Notice of Proposed Rulemaking (ANPR) issued by the U.S, Environmental Protection Agency (EPA). 49 Fed. Reg. 110": (March 23, 1994). specifically, EPA sought data regard..-g t-.e risxs posed by fires involving electrical transformers that contain PCBs and the mechanisms for eliminating or mitigating these risks. '.'SWAG is an informal consortium of ESI, APPA and NRECA and approximately 65 electric utility operating companies, ESI is the principal national association of investor-owned electric light and power companies. APPA is the nations, associat.on of publicly-owned electric utilities. NRECA .3 the national association of rural electric cooperatives. Together, USWAG members represent more than 95 percent of the total electric generating capacity of the United States, and service more than 95 percent of the nation's consumers o electricity.^/ A list of USWAG members participating m these comme*ts is attached hereto as Exhibit 1. HONS 213860 *% 2- tmtbopucticn and suy-MABY ct ?cs:t:cn In August 1992. EPA Authorized the use of a poculac.c- of transferrers containing or contaminated with PC9s for the remainder of their useful lives. This authorization was based on EPA's determination that the continued use of such equipment did not pose an unreasonable risk to human health cr the environment. Underlying this ultimate conclusion were t-e following findings {see 49 fed. Beg. at 11270): 1. The resulting reduction m risk, in the absence cf such authorization, was insufficient to outweigh the substantia costs (billions of dollars) involved in such an aetien. ,< it 2. Servicing conditions, maintenance activities, an inspection programs reduced the exposure nsxs associated w. the use of PCBs in transformers. 3. Beleases of PC3s to t.ne environment and expesure to humans were minimal. ` 4. Accelerated phase-out and other risk reduction programs were not reasonable when compared to the potential reduction m release of PCBs achieved. Prompted by a challenge to these findings filed => EOF/NRDC and reports of three fire incidents involving PC3 Transformers, EPA issued the ANPB in March 1994 seeking further information on risks to the public health and environ-? posed by fire-related events. Notwithstanding the fact-fird.-g mission intended by the ANPB, EPA concluded, based only c.n very limited data there presented, that "PCB-Transforme- fires pose relatively high risks, occur with unknown frequency and can reault in relatively high clean-up costs." Id. at 1127'.. These comments and the analyses, studies and materials on which they are based demonstrate that EPA's initial conelus.c-s cannot be supported. MONS 213861 -3- Specifically, in response to the AN"?R, USwag conducted en extensive survey of electric utility companies regarding the current inventory of high-concentration, PC3 Transformers (referred to fleremefter as "askarel transformers*): fire incidents that have involved such equipment; ensuing clean-up costs and other matters.V These data and other information discussed below regarding the matters on which EPA solicited data, show that: 1. The risks posed to human health and the environment in the event of a fire-related incident involving an askarel transformer, including any effects from exposure to the products of PCB combustion, are de minimis if not non existent. 2. The probability of fire-related events occurring is extremely low. 3. The costs involved m cleaning-up the two fire :r.c.dents identified by EPA as "relatively n;gh"^V are not representative. 4. The costs associated with the available options for mitigating or eliminating the minimal risks posed by fires involving askarel transformers, are not reasonable when compare; to the potential reduction in exposure risk; or put another way, the resulting reduction in risk clearly would not outweigh the substantial costs involved. As EPA is aware, the electric utility industry nas been deeply and responsibly involved in the management of PCBs for more than a decade. See USWAG/EEI/NRECA, Comments and Studies on the Use of Polychlorinated Biphenyls m Response to An Order of the United States Court of Appeals for the V A report on the survey conducted by Resource Planning Corporation is attached as Exhibit 2 ("RPC 1994 Report"). 49 Ped. Reg. at 11071. MONS 213862 -4- Distriet of Columbia Court, Feb, 12, 1982, Vol. I at 1-2 ("1982 Comments"). The industry continues to be support.ve control measures that can ba justified by the body of scient evidence.*/ But however Chat evidence is weighed, there nas been no acknowledgement of the effectiveness of the PCS control measures already in place. Thus there is growing industry concern that this failure to recognize the effect c the controls in substantially eliminating release of PCSs in the environment and exposure of humans to PCSs has distorted the public's view regarding the continued use of PCSs. Current control measures, regardless of whether one new views them to be justified by the scientific evidence, a effective. The need for additional and mere restrictive --i is highly questionable. Even in the case of fire incidents involving askarel transformers, where EPA believes that certain by-products of pyrolysis or combustion -- more so tr. the PCSs themselves -- pose some sort of risk to human heal-, the perceived human risk described by EPA appears to be base on the presence of the chemical (whether PCSs, PCDFs or PC" rather than on the actual exposure levels and the effect of that actual exposure. Overall, the existing evidence shows that fires involving askarel transformers are rare, that u:. / It is USWAG's overall view that this evidence, notwiths ~ prior Congressional and regulatory action, does not warrant many of the restrictions in place today. MONS 213863 -5- have program* underway to eliminate any perceived potential risks resulting from even these rare incidents, and t.nat further regulation of this industry's equipment is unnecessary, while the focus of these comments is that no further regulation is warranted because of fires involving askarel transformers, OSWAG wants to emphasize that the arguments against further regulation of mineral oil transformers are even stronger. I. DATA REGARDING USE Of ASKAREL TRANSFORMERS AND FIRE-RELATED INCIDENTS________________________________________________________________ The electric utility industry has used equipment containing RGBs for over 50 years. Some equipment, because of its location or electrical properties, was designed to use PC3s as a dielectric fluid. This equipment included transferors that would be used in locations where fire prevention was a significant factor. Thus, as shown in CSWAG's 1981 survey, of the 39,640 askarel transformers then owned by the industry, some 22,469 were located in the utilities' distribution systems. R?C 1914 Report at 6. Other transformers in use on electric systems, numbering shout 20 million, customarily usad mineral oil as s dielectric fluid. As s result of manufseturi-g or servicing conditions, some of this equipment was found to contain concentrations of PCBs over 50 ppm.V V Approximately 10% of the mineral oil transformers were found to have FCB concentration levela above 50 ppm, while only II were found to have concentration levels above 500 ppm. 1992 Comments, vol. in *t 9. MOMS 213864 -5 In responding to the ANPR, L'SWAG focused ,,:s im gathering effort* only on higr. concentrat.cn, askarel trrnsfcr- = This approach was taken for several reasons. First, tne t.-;;t incidents that appeared to have triggered EPA's concern eac.n involved an askarel transformer. Second, there i* a general belief m the industry that the problems associated with Tineral oil fires are substantially different t.nan tftose raised oy EPA m the ANPR; .i.s., the significant concern :f a mineral oil fire is the fire itself and its potential to cause severe injury to people and property. In addition, as discussed m Section III, the potential production of chemical by-products free conversion of PCBs during a mineral oil fire is believed to be minimal. Finally, as shown in the 1392 RPC report see, e.c., 1992 Comments, Vol. Ill at 150-59), and as concluded oy EPA m the August 1992 rule (see, e.g.. 47 Fed. Reg. 37345-47 (1992)), the cost of the available risk reduction options regarding the use of mineral oil equipment, amounting to bill.on cf dollars, would be totally unreasonabla compared to the benefits acmeved. The 1914 RPC questionnaire, therefore, principally sought information in three areasi 1. The number and location of askarel transformers owned by the electric utility industry, including e.ns utilities' plans affecting the future use of this equipment; 2. The frequency, timing, location and circumstances of fir* incidents that occurred ovsr the last three years ard involved askarel transformers; the cost of clean up following such incidents, and the populations exposed to such incidents; and 3. The utilities' experience with retrofill options. MQNS 213865 -7- The complete results of the RPC survey are submitted herewith is Exhibit 2. A summary of the findings follow A. Inventory of Utility-Owned Askarel Trimfor-trt 1, Current Invntorv When first surveyed in 1981. ths utility industry was estimated to own 39,640 askarel transformers out of a total inverse of about 140,000 then in service. Of those 39,64C, some 57% or 22,469 were located on the utilities' distnbut.er. systems. Current data show that the electric utilities now own about 36,000 such transformers of which some 18,291 remain m service on the distribution systems. As concluded by RPC, the magnitude of the decline in the inventory is greater than normally would be expected, primarily due to phase-out programs initiated by some of the companies.V The general location of the askarel transformers on the distribution systems was found to be as follows;**/ (Cumber of Transformers Percent of Total Inside luilding Sidewalk Vaults Other 6,000 10,919 lr 372 32.9 59.7 7.5 Total 11,291 100.0 The overwhelming majority of indoor transformers (14.56) are located in vaults within the building, while 9.36 */ These programs are discussed in the next section. "/ RPC 1994 Report, Table 1 at 6. HONS 213866 -3- are located m separate transformer rooms; only g,n i;j installed in open areas. Natural ventilation to the outside predominates eve*, with regard to the indoor equipment, further, less t.na.n ha.i of the indoor transformers are installed in locations where there is direct access to air conditioning or heating system intake vents. In total, of the 6,000 indoor installations, less tr.an 60 transformers are found in vaults or rooms _n which such intake vents are located and less than 2,too are found where sjch intake vents are in the area immediately ad;acent to the transformer vault or room. v/ While other indirect routes to heating and air conditioning systems can he identified, it is consideraaly less likely that these routes would provide a pathway for significant smoke travel, a?: 1994 Report at 9. 2. Voluntary Phaae-Out Programs As indicated in the RPC survey, 76 of the surveyed utilities have askarel transformer phase-out programs m effect. Id. at 12. Based on these voluntary programs, it .s RPC 1914 Report at 9. Of the 12,291 transformers located in outdoor installations, approximately 1,350 are located immediately adjacent to building heati.-g and/or air conditioning systems that could serve as a Mins of building access for smoke/soot. Id- HONS 213867 9 estimated that tha number of distribution system askare. transformers will be reduced from 13.291 to 12,937 by me * - i of 1935 (a reduction of 29%), to 5,402 oy 1939 (a reduction :: 35%), and to 2, 927 by 1994. Id., Table 5 at 13. Thus, of ?.-.e distribution askarel transformers on line today, only 16% of them are expected to remain in use in ten years. Notwithstanding the adoption of these programs sy -a-, of the utilities, CSWAG be Waves tha: the eont.rued use of askarel transformers does net pose an unreasonable risk of injury to human health or the environment. These actions have been undertaken against a background of several concerns, first, there has been media and community expression of perceived risk from any exposure to PCBs. In addition, there have been two instances (Binghamton and San Francisco) ,,n which very high clean-up costs have been incurred, m par: because of the absence of any clean-up standards. The uncertainty of the financial consequences that could be impose: m regard to clean up, and from potential litigation, regard.ess of its merits, has resulted in business judgments to phase cut some of the askarel transformers. Me feel the actions the industry is collectively taking deserve some recognition and credit by the Agency ..n its analysis. At the same time, however, the initiation of these programs should not be viewed as industry agreement that any unreasonable risk exists or that further regulatory act.on is warranted. MONS 213868 -10- B, Frequency and Nature of Fire Incidents 1 ,, Frequency of the r.ieidtrti As shown by the RFC data (RPC 1984 Peport at 9-11'. the annual number of expected fire incidents involving i:.;.:/ ownad askarsl transformers is extremely low. Only eight such incidants have occurred ovar the last three years, indicating an annual estimated rata of 0.01%. Of these, nona occurred m industrial locations or in utility substation or generating facilities. In addition, despite the one dramatic event at Marxet Plaza in San Francisco m 1933, the actual amount of smoke travel associated with the incidents was negligible. I*, one incident no smoke escaped from the transformer, while smoke escaped from the transformer in the other seven mc.dent m five cases the smoke either did not enter the building (:-c outdoor installations) or entered only the area immediate.'/ adjacent to the equipment (three indoor installations). only two incidents involving utility-owned equipment d.d s-c<e travel into public areas of a building. These situations involved the Market FLaza/San Francisco and First Nations. Bank/Chicago incidents that are discussed more fully beicw. 2. exposure Levels The only data available regarding the chemical compos, in samples taken following fire incidents involving askarel transformers are the data obtained following the Binghamton, Market Plaza and First National Bank incidents. As acknowLedo by EPA in the ANPR, no PCDFs and no PCODs were found in the wipe samples from the bank building, 49 Fed. Reg. at 11073. While contaminant levels were found at the other two locat.cs USWAG believes that the significance of these findings HONS 213869 IT - n in 4, 4, i l i I 1 4* X tn 41 41 /) 41 II U 4 44 l| p 4 If o 1 I 4 It* c a. 41 r: X if 1 l. 13 TI u M M o 1; u Jl -- i 41 oo 41 r 41 *i X 4* *i l* t; 3 O M i: x 4 in ii t* 4 e 41 B *4 i r: 1* 1 ox If i) >4 41 ^4 o* Ii X X x u -- --4 H r4 CL --4 tu IU T X 41 Ii s, U tn V TJ o SJ o U U O c ip x ** X * *11 M o pii fc >1 -e m c: X -4 > > 44 ,--4 X ft u 1* 44 o c -4 4 i o tn C >4 if >. n u o 44 n ^i '-4 j: m c ti O *-4 M c> u in V if i .--4 O <s .-4 CL C -4 ii X n. %4 c o 44 CL If o 14 Ml X in C ** > N O X Ml CL aJ c -4 -4 H E c Ii X --4 c X 4-4 O TJ n M U > X UU u u> I) X J3 x cu c a 41 44 > m MX 44 Ml 41 o u ^4 O % u o If If oO 4 i x M u 44 M HI ll X M tr n. X X r-t *. i U TJ u z yE X 41 X m rx U oMM oX X6 m c x\ M HX O 0 I z ** o i 41 *4 V? * n U a MX u 41 n at o o X MX C X c U 1 o t: * I o Hi 41 o II *4 x II J3 uX -* *J ux >* u > 4i MX n e c i: x if o --4 c U a -C i 44 44 H X CL --4 If -4 x D Y C M XO 1 >* X) yu *4 i M 3 *2 -4 u >* O ^-4 0 U X 4 %4 x E `-4 41 ,--4 pi i X > uM o c c 4 c U 4 B *4 > M 41 II c x X If M y41 3 M X 44 MoO 1 44 MMO hX o O .V *-* U x > --4 m * cO a. M if M H < uo m n X 44 o j: if e m fL m --4 u --4 If u X * ii M t> u u o C -O i %i X 41 > < tj 41 II x <M if o -4 %i XI a o. y c H in * D r, o 44 4 If o--4 if c >- Ml oDhD 8. o TJ U MX -4 --4 M 4 X u if if c MX n 3 U X -4 H O U ocU u x *4 c; u -e & c >D O tj a o x 41 nm 6 tJ O m /I x *4 C X u tn <4 41 Xu c mx >* CL -4 If a Mm O. r| 3 X 3. Clean-Up C o a ti Survey data in d ic a te th a t tha average clean-up cost or an m a id e b u ild in g in c id e n t, excluding the M arket Plaza 4c*l TXJ -4 u X T:LJii n <-4: c X --4 tcxn XE dci X X tn c %4 o >M 4-oN --4 ulXl oo '-- u--4 MXifl L UX V* chM XiXf oo 3 < XuX & mQ mum CL d1 cX X-4 u * Vf M X%i hot>l m> --V4 3 e> M CL X X3 5 if M nOo-w m --Ne4 0. xh Xm c(**1,4 rm# u 9 XXO\ XX> --m>o4 e e *cd w4A If 9 AHQ M O TJ IXf -X4 u Xo XX M %4 ^>4 oc>-4 X X X --4 IXf u0X o --* <eu U mCO >P Oa 7> * >Mn3 K CH UM umK M) o4 1 M oo> M n a>+ uti E --C *--CmtF> E m -* (mJ o i*~4 m M M 4aoM*XMcMo Vc4C> e-Co>>-4* -a Pix Mi 4X9>--*<* \S*f>i--of MO -< O0 43 a. 3 a \* CO m <A X a X -12- 4. Factors Dacraasmq Fire Incidar.ts As indicated abova, phasa-out programs in plact w*.. substantially raduca tha numbar of askaral transformars tc.aeould bacoma involved in a fira ineidant. That daclina, eouplad with tha ovarall ineidant rata, shows conciusivalv that tha risk of a fira ineidant in tha futura on utility systams is vary low, as shown *n tha following tabla: Yaar 1984 1m 9m 35* 1987 1933 1939 1990 1991 1992 1993 1994 EXPECTED NUMBER OF ASKAREL TRANSFORMER FIRE INCIDENTS IN THE ELECTRIC UTILITY INDUSTRY Estimated # of Askarals** Estimated 1 of Fira Incidants 35,934 30,233 27,601 25,147 23,424 21,335 20,711 19,538 18,384 17,373 16,200 4.6 3.9 3.5 3.2 3.0 2.3 2.7 2.5 2.4 2.2 2.1 RPC 1984 Report, Tabla 5 at 13, and assumas 20-yaar average usaful remaining Ufa for subatation/ganaratirc facility transforaar. HONS 213871 13 The likelihood of *ueh incident* occurring also *s by the presence of protective device* (fuse*, switcnes or breakers) on many distribution systems. Such devices are installed to isolate the transformer m the event of a fault. Seventy-five percent of the distribution askarel transformers have protective devices on the low voltage side of the transformer, while 191 of the equipment have them installed on the high side. RPC 1984 Report at 12. C- Petroftiling Experience RPC sought data from the utilities concerning t.'.e.r experience with flush/refill and filtering processes to determine the frequency with which these methods have been used, the levels of PCBs achieved over time, and the costs cf the program. As indicated m the report, the results nave been extremely erratic and cannot be correlated with the methods used, the time elapsed or the expenditure made. RPI 1984 Report, Table 8 at 16. U5WAG believes, therefore, that there is no basis to alter the conclusion it reached two years ago that retrofilling askarel transformers is not, in the overwhelming majority of cases, cost effective. See usvcag IZ1 Reply Comments on Proposed Rules on the Use of Polychlorinated Biphenyls in Electric Equipment, June 24, 1982, at 22. Furthefcr there continues to be too little known regarding the electrical and fire resistant properties and toxicity of the substitute fluids to in any way compel their use. See Section IV. A., infra. MONS 213872 14 D. Impact of RPC lata on E?A Initial Sit.ratei The information provided by RPC demonstrates ;-a- Ja;a on which CPA based iti preliminary assessment of th* posed by the continued uie of askarel transformers are serious!, flawed. Fires involving these transformers are rare, isolated events. In addition, while abnormally high clean-up costs have been incurred in two cases (one of which did not mvo.-.-e a utility-owned transformer), the expected level of clean-.; costs is substantially less. 1. Frequency of fire Incidents The RPC data on ehe number of expected fire incidents contrast significantly with most of the estimates noted by -PA n the AN?R. If E?A continues to def.na a catastrophic mc.de-1 as one in which clean-up costs of $15-21 million can be exocotec. a catastrophic fire incident involving a utility-owned asxarel transformer can be expected approximately or.ee every 2.5 years. (This figure is baaed solely on the number of transformers adjacent to intake systems.)*/ RPC 1984 Report at 19. 20. I: the focus is placed on the number of incidents m which V This frequency rate was derived as follows: RPC ident.f.sd ~ 2700 indoor askarel transformers (6000 total indoor transformers (Table 1, p. 6) x 44.7| (the percentage of indoor transformers adjacent to intaxe vents}) and 1300 outdoor eskarel transformers (12,291 total outdoor transformers (Table 1, p. 6, and p. 8) x 10.91 4 the percentage of outdoor transformers adjacent to intake vents)) as being potentially involved in what EPA has defined as a catastrophic fire. 5ee general 1 / RPC 1984 Report at 19. These 4000 transformers were multiplied by the reported frequency rate of askarel transformer fires of 0.011 to arrive at 0.4 fires per year, the equivalent of 1 fire every 2.5 years. If measured against the total population of equipment, the frequency rate would, of course, be dramatically less. MONS 213873 15- smoke reached substantial portions of a build.ng, the expectancy rate based on the utility data is tnt same. Ags.-. it was only in the Market Plaza incident that smoke was carried in any significant way into the public building. 2. Clean-ua Casta As discussed above, the RPC data shew that the i*vel of clean-up costs averaged about $279,000 in indoor installations, and $27,700 in sidewalk vaults. These levels, rather than the aberrational $15-20 million figures, should ba used in analyzing the regulatory programs relating to the continued use of askarel transformers. If. REPRESENTATIVES CF THI 3INGHAMTQN, SAN FRANCISCO AND CHICAGO INCITSSTS The issuance of the ANPR was essentially prompted by three fire incidents that involved askarel transformers. These fires occurred in Binghamton, New York, San Franeiseo, California, and Chicago, Illinois. From its analyses of these fires, EPA appears to have concluded that PCS Transformer "fires pose relatively high risks" and "can result in relatively high clean-up costs." 49 Fed. Reg. at 11071. as Indicated in Section I, these conclusions cannot be supported based on the utility data eolleeted. In addition to seeking new data, EfA solicited comments on the representativeness of three fires in terms of the risks posed. This aspect of the Anpr is addressed below. USWAG believes that these incidents, as they art descr.sed by EFA, are anomalies for two reasons: (1) they are infrequent occurrences and (2) the potential risks are overstated, while MOMS 213874 16- :h Afney indicitti these fires art 't,1;* ms; w*;> characterized end well-researched incidents," 49 fad. sc at 11071' th knowledge, *s presented in the AN?R and r,,lt-a*.-s record, on the causes of the fires, the sources of contain:-.at.;-, the levels and extent of contamination, the exposure risks and the health consequences of that exposure is sparse, in add:-..;.-, the circumstances and results of the fires are significant; different from one another and do not, together, create a picture of significant risk from the continued use of askarsi transformers.*/ A. Fires associated with Askarel Transformers Are Rare CPA has presented data on the three fires. The data for two of the incidents shew t.-.e for-ation ard d,stnsj:.:r of PCIs and Pd combustion by-products. Based upon CSWAG's study of the frequency of significant events, only one util.ty incident was identified as catastrophic based on CPA's defi.t.t.;*, that being the San Franeiseo fire at One Market Plaza, May if, 1913. While the electric utility industry may ewr. only one- third of the askarel transformers in service, USWAG believes that its data indicate an infrequent occurrence of this type of event. Prom CHOUS* s survey, this indicates that a "catastrcp.-.r */ The three affected transformers on which CPA is focusing contained a high concentration of PCIs (654 or greater). C3WAG has previously argued that askarel transformers (containing PCIs) are different from mineral oil transformers that may be contaminated with PCIs. See 1912 Comments, Vol. I. In this rulemaking, USwAG believes that it is particularly important for the Agency to distinguish between the two types of equipment in order to make a proper evaluation of the risks. It is imperative that the Ageney formally acknowledge the difference m the preperation of its proposed rule. HONS 213875 -17- faiiure," *s ERA refers to it (49 Fed. Reg. it 11079), has the potential to occur approximately once in every 2.S years (3.4 events per year) for utility-owned askarel transformers, rpc 1914 Report at 19, 20. ERA must consider the frequency with which askarel transformer fires may occur in its risk evaluation. Equally important is the fact that the population of askarel transformers is declining in the utility industry. See Section I. Therefore, the chances that an askarel transformer will become involved in a fire are rapidly diminishing.^/ 3. Potential Risks Are Overstated Apart from the infrequency with which a Binghamton or San Francisco fire may potentially occur, the risks involved with such an event and continued use of askarel transformers are overstated as a result of the discovery of PCTFs and PCZOs at tne two buildings. It thus appears, m retrospect, that the initial finding of concentrations of PCDFs and PCSCs a: Binghamton led to an over-reaction and concern for public health protection from the use of askarel transformers. This over-reaction stems from the mere presence of two substances that are interpreted to be highly toxic. While the distribute" */ The Agency also oust not lose sight of the fact that the reason askarel fluids were originally used was because of their fire resistant properties. MOMS 213876 13- cf these by-products throughout the building m 3ir.gr.a.-tc,* and portion* of the building in San Francisco is unfortina;e, there is still such to be learned about the event* and tr.ere are many unanswered question* that need be addressed. Sc-e ci the facts surrounding the incidents indicate a far low*; hazard than EPA's analysis shows. 1. Causes of fires May Be Preventable The discussion of the cause of the linghamton fire n the ANPR and Versar's Report (Exposure Assessment; Fires Involving PCS Transformers) is very sparse. We merely k-ow, based on communications with the New York State Office of Government Services (which owns the equipment) and others that a fire occurred in the switch gear. G5WAG is aware that EPRI is planning to conduct detailed case studies cf Binghamton and San Francisco (RP 1263-20, Project Officer-. Ralph Komai) that should be completed by October 1984 .v ?.*s study's results should provide useful infcrmatier. on the actual causes of the events and thereby avoid any mis interpretation based only on current information. We thereftre suggest SPA await the results of this report before deter-.- *g whether any changes in the use rules are appropriate. V PCS Technical Report from EPRI, Number 8, attached -- hereto as Exhibit 3, prasants a summary of tha many relavant projects now in progress. MONS 213877 19- S.ailarly, for the San Francisco and Chicago f.raj there is little or no information presented on tha causes of tha events. No conclusion, therefore, can be reached tha: thasa avants and tha occurranca at Binghamton rasultad frem comparaola causas. 2. Lava Is of Contammar ts Fcr-ed Is Inters.sts-t Tha New York Stata Haalth Cepartrer.t was tha first group to datact and quantify FCBs, PCLFs, and PCTCs m soot samples takan from tha Binghamton fire. CSWAG notes the extreme variation in results from tha ANP* and tha Versa: report.V Thasa divargant rasults indicate a great uncertainty about tha levels of contaminants to which people potentially may have baan exposed. ' Sea AMP*, 49 Pad. *eg. at p. 11072 and Versar ~ Report at 4. HONS 213878 20 SPA not#* the variation m results between sa*:.Jr but fail* to foeu* on the real import of the problem. problem occurs because there is a huge disparity in res*, results fcr even a single sample. That sample, number 811711965, was tested by three analytical ehemists m four chemical analyses. Versar Report at (. The results are reported for 2,0,7,S-TCOF and 7333 as follows: 5mitn 1991a S~it 1991b Racce 19 91 Sta.l. 2,3,7,8-TC3F (ug/g) 273, 4 48 12 2, 3,7,8-TC3D (ug/g) 2-8, 2.9 1.2 0.6 There is more t.nan a 20-fold difference in result between Races ar.d Smith for 2,3,7,9-TC3~ and ~ore tna-, an S-fc'.d differe-re between Rappe and Stalling on 2,3,7,3-TC33. Notwithstanding these ra;or differences m resu.ts, ERA appears to rely on the results of soot samples first ta<e- from Binghamton by the New York State Health department. T'-es values of 273 ppa and 124 ppm for 2,3,7,8-TC3F are s*gnif.:a-: different from the values reported m a separate analysis cf the saatple by New York State, 48 ppm, and Or. Rappe, 12 ppm. 49 Fed. fief, at 11072. In light of these disparate results EFA cannot properly view the 273 ppm or 124 ppm levels of contamination as representative. HONS 213879 -21 Ev.dently. the analytical chemis:ry of wha: are toxic cc"pcunds of concern is arcirum. This .'a:; was confirmed m a reeent interview with Dr, Chr.stoff*r Rapp* cf the university of Trea, Sweden,*/ on* of the Binghamton investigators and a leading world export in the synthesis and a.-a.ys.s cf r-C"i and PCTFs, Dr. Rapp* indicated tret (1 :.',*:* are or.:., tn.re* laooratori*s capao.e of pe:f::-.-; t.-.e ana.ysis c: tn.ese compounds, (2) tn.e separation of the 2, 3, *,5 ,sc-er the 2. 3, 4, 8 isomer (the latter, m his assess**' t, a non-toxic isomer) was critical to proper quantification, a-i <2) the extraction of PCDOs and PCCFs frcm soot partieles ,s a rigorous operation requiring strerg solvents: because cf tms prcolem, n* suggested that the q.est.cn of whetn.er e by-preducts were actually biologically available to cause harm should be addressed. In light of Or, Rappe't remarks and the results of :* 811711955 sample, it appears that there is an mconaiste-.t, hot totally uncertain, knowledge scout the contarr-nation at Binghamton. Given the resulting levels and the signifioir.ee of the level of contamination in evaluating exposure risks, the actual risk posed say be significantly less than assured by EfA. 3. Confounding Sources of PCDfs and PCDOs A further complicating fact that has not been addressed 1/ Representatives of CJSWAG met with or. Rapp* in New York City on June B, 1914, to discuss his scisntific opinion on ll) sources of PCDQi and PCDFi, (2) linear by-product formation, (3) by-products of substitute fluids, and (4) human exposure. HONS 213880 -22 at all is she existence cf confound.ng sources of PCCFs a-d PCDCs, Recording to Of. Rapp#, i; n difficult to dis:.r.g.s.'. the or if:.', of PCOFi and Pno*, Conceivable sources arc polyvinyl chloride and flam# retardant chlorinated hydrocarbons, both commonly used substances in buildings. Althc-f the results of an inquiry into confounding sources could have a substantial bearing on the interpretation of potential r.s<s free. PCS fires, one questions why no investigation has been initiated on this matter. We also note tnat PCODs and PC0F9 ray be produced in other situations not involving PCSs, s.ch as barbecues, fireplaces and woodstoves in domestic dwellings C.ven the econcnc impact of any further regulation cf as fare 1 transformers, the Agency rust clarify whether ot^er sources of PCOFs and PCDDs are present, and, if so, hew that fact affected the cost and effectiveness of the clear-ip effort In addition, C5WAC believes that it would be relevant to fcr.sw what expected background levels of PCOFs and PCDDs m b*.ld.~gs might be for purposes of comparison. 4. Formation of Reaction Products USWAC also spoke with Or. Rapp* regarding the format.=of PCOFs and PCDDs following fires in which askarel transfer-era are involved. On the basis of this interview and other i-for-i". ,c available, USWAC believes that no conclusion can be drawn at this time regarding the linearity of by-produet production c: the quantitative conversion of PCia to other contaminants. On the subject of EPA's assumption that PCOFs and PCOOs are linearly formed from PCS mixtures. Or. Rapps felt that this could not be assessed because at present,no data are MQNS 213881 23- ava:labl. He stated that :.k.e exact -a:..-! :f t.-.e react.;.-, whether it i* monomoleeuiar or himolecuier, :> -.ct Jt.-cwr . :fact, more than one mechanism cay be operating. Therefore, model* used to simulate pyrolysis that show linear relatisrsn.es. according to Or. Rapp*, are not readily applicable to the f.re situations. These models have limitations and do not necessar-. mimic the effects of C i) a clean versus dirty ccrbust.c*. fla-e. (2) oxygen variation, (3) electrical versus enemies! energy, and (4) tempers cur*. Accordingly, witnout empirical data, the Agency may draw conclusions that misrepresent the severity or PCS combustion, and conversion. In addition, there is no adequate information upc.n whim to assess the quantitative conversion :i ?C3 asxare. f.uids to PCCFs and PC"s at a.ncnamtsn or San *rar.c.sco. Cnly sxetchy data have bean presented to date. The aralyc.ra. results we do have indicate the presence cf ore half to 43 t.res less 2,3,7,8-TCDF and of 13 c.mes to 53 tires less 2,3,7,8-TCSO in soot samples taken from One Market Plaza compared to Binghamton.^/ To C5WAG, this shows that B.nghamt;- and San Francisco are not comparable events, except m the qualitative identification of the compounds. The concer.trat.;_s are significantly different. EPA is aware that EPRI has contracted with the New York State Health Department to examine the pyrolysis and combustion of PCBs and retrofill fluids < RP 2028-4). "/ Th.s V Compare Table 2, p. 6 with Taole 5, p. 13, of the Versar Report. */ See Exhibit 3. HONS 213882 n 11 if} X Cl o t 1. M A ii ft 11 i ; t 11 4 *d * 1 # 4 4 4 *ri * M ft 4 ri Cl 11 > ut >. 1 i It O *4 M t tn l 4 *if 4 U Is V i t f: Hi r. n ;* i* o 4 4 > ft m m *4 O -1 U u i. U J 4 4 4 4 i: 1*. * & o u 1 ft ft Ii a4 4 ii 1* 1 :i i. M At 4 II n M *4 4 .* -4 C J c 4 m 4 ft* u 4 M 4 it Ik c r-4 (1 4 m 11 s u. Ii* 4 Is 4 a> Is Ik U 4 n F; u IP <1 4 4 i: O < bO IV V ii c c 4 Ii Ii 4 it ft u cv It * it 4 o *- a % * i Is hi Oi ' 4 -V il <4 4 i ft 4 4 ft VI %4 ii 4 4 4 i: > C ft 4 XI a *. m o a 4 r J3 o At ft p i m uo> 4n u 4 4 4 r1 ft 4 MO n a ft n X <4 U > a ft Hi il oO uo M n. c ft 4 ft a* o 4 u ft 4 4 T) 4 u 4a o. fj *tJ i; 4 > ft 4 4 At ft XM o i* -V Hi u 4 1 o. %4 <4 u o c j u O At ou < 4 N4 ft i 4 ft U -4 ft <4 ZJ u II At O j: m %4 c c 3 o JU u n -4 K M At D Z3 At 4 in >1 U A> Ou -* r -4 -4 o it o m 4 4 vi it c 4 <4 Z3 4 O --4 M 4H U 4 4 At 4 4 Im it 4 4 At 4 O At 4 it > ft O h it 4 4 4 ft a 44 OR a X ft M ft At 4 u tr J o 4c M ft -* TJ m i; 4 C ft* `4 4 4 m -^a u u ft C 4 C n 4 X r--1 4 m Hi n. O u At O % in 4 c Hi 4 u a. At ^4 o c* II M M M 4 c 1 At CUJ i; TJ 4 u (V . o > 4 HO 4 II U 4 A* c n 4 ft H At %4 ft ouc ft o 4u M -0 M 4 Hi c tr 1 o OH c KT* CO At J <4 U m it r. u A* -4 T) H *4 it *4 SB m M c: M 4 m C C-* -* U Di <4 U H t4 4 <4 4 A' H U u A* U C Cft ft* 4 -*4 % O %4 TJ EH it o u 4 4 -4 4 Vl oXI ft* it c 4 M 4 it m TJ 4 c Hi 4 ft O T) it XI . r M 44 P o Is 444 0 ri X aO ft ft 4 Ii it u 4 ft O vt a Xc 4 %4 Tf o ft m rc a a -- H OC a hi oc U u4 4 -c ft * ft C 4 M it u aS n4 oM M %4 o u 4 a w u 4 ft a o a 4 o Ec4 M ft At O o ft 4 3 o o U *4 ft it 4 pC M c c 4 % H ft 4 4 c -4 X o At o t d 4 it U o 4C XU A* -4 4 4 C it >* o*U U U %4 M a 4 a* o a O m l ft M it i--* 4 O >. TJ , 4 ft c D 4 o 4 O --4 M At 4 o '*4 o aM MJ m a ol 1 it !J 4 ft -4 A* c 4 4 ft X 4 *--4 oO P. m m M n CJ < X ul u 4 >. i: * D 4 o o ^4 CO ft E p O i) it P O 4 ft 4 4 4 fti 4 XI A C m CV a hi to*! i -* 0< M M At ^4 H %4 4 fi4 o H taj *4 C 4> H 4 oC it m z* At -q oO t n Cft* U c <4 4 r*> u 41 O 44 MM 4 Oi O r"i r> X TJ 1* 4 aM <-4 o 4 4 --4 u C4 4 N it ao tc M 4 <4 >. n 4 <4 4 i V U ft tu Di ft it C in *- 4 o4 u 4 4 4 ft u (7 hi N . I* Ii Ii n a ftc M it > 4 < %4 4 4 *4 4 O 4 O 4 E o 1 U u *4 n M * ^1 uu 4 H 4 c 4 i a a H o ftn m M V E ii a ftn it U 4 O IJ ft o o ou j: r; o Hi o.rv n t i it l n o ftn. * Q 4 M4 o >. 4 il 4 a4 XI p uX a i m <9 CD fO IK </> Z 5 -25- performed by EPA on the toxicity cf the Birghimtcn see: t.nat 'auliipl* exposures to scot from ?C3-Trar.sformer f.r-s have the potential to produce toxicity m :.he thymus, .-? hematopoietic system, the salivary gland duct epithelial, a-.d, possibly, the liver <Aef. 5)". 49 Fed. Reg. at 11073 . we..;* t.k,9 effects are not insignificant, the means of exposures to h_ma-s d.rmg an even: and following an event are total'.,, different. Because t.-.ere is so little infar.~at.cn availed.to simulate the exposure and effects on hura.ns, we suggest the Agency d.vert resources to this area before .t prepares fmd.-gs on exposure to PCB and PC3 mixture by-products m different redia by inhalation and dernal contact. C. Co-clusions Pecard.-r tke 3:'cka-tc-. 5a- ~rs-c.sc; and Chi race :.- = .de-ts EPA must weigh several factors i.n evaluating tne r.s<s actually posed from Binghamton, San Francisco and Chicago incidents and the representativeness of those circumstances These include: (1) The frequency of such events; (2) Future preventabi1ity and mitigation of the incidents; (31 The actual severity of exposure to fire fighters, ciean-up workers, and others; (4) The value of awaiting the completion of EPRX and (JSWAG studies prior to further aetion. We have demonstrated through the most recent CSWAG survey that the utility industry rarely experiences an askare. MGNS 213884 4 ,< o o *1 I u r 4 it <** c 4 W M o * o II 'I a <0 E At Hi o It 4 Hi M n M O <M VI c 41 h it it p a r* ,f O o Ht O ? -* J i: 4 u tr 4 i Hi m ii u M 4f * >a u 4 a 4 it C * u H o c -1 m 41 I' >a uM vi :J D n *-i /: i * it c E C-4L tr c it Hi O 1; H it 4 Hi <-4*i r~- 4 n 4: * i O M 4 it > <4 O R 4 L it 4 S1 m -i j; M T> C 4 VI It Hu o M VI c 4 It it L M4 VI V V# O Co 4 it (V *4 ri CL o CL VI it 4 o U ri T M VI * *1 4 U r. it tr c -4 c 4 -- u T> S 4 T> -4 CL V M VI H co H it 4 i--4 3 ou. CL u 4 3 *--I 4 > it /: u *1 Hi O it tt m m a 41 VI -i M i; It VIt) 4 t) c *! s, it II It* 41 li i; 41 4 in t . rt i; c n4 it ti VI 4 i j: C tr 4 <: *o i -t mu i: i 4 4 41 mM m :l u it c Ht 4 it u o - IL X> 4 4 *3 * Co 4 t> 4 --I ri o* 4 ii L 4 j: u M 4 CL Hi U --I 41 Si oii Hi s 4 4 Tl 4F4 l 4 . O i: t i tr 44 4 O i; VI H it J' r, tl Tl it *4 *o u Hi c: > O44 n <it VI , t; it o I* it Ml 4 4 41 * a it 3 4 IL Hi : 4 tr U o CL it 4a Ht -4 u o>H i> 41 T1 It 4 om3 it 4 - 4 4 ;l tr CL O 4 X O XI 4 4 VI Jj 444 j: n it 4 li M H 4 4 o it VI 4 >a 4 E oit > ii 44 M 4 Hi 4 4 VI > 4c 4 XI 4 VI It < it *--i IL 4 lil <4 U4 it 1 it U tl 4 4 UM 3m 4 a 4 J* O H M tr CL t: 1 S XI ,-4 o 4> j: <: m -t -27- (1) Mechanisms of PCCF and PCT3 far-**.:an from cudias of PCS- cn- and tetra-cnier.nated benzene pyrolysis and cemb-st.on; (2) Casa studies of Binghamton and San Francises, (3) Engineering Options for PCS Askarel Transferrer Risk Reduction; (4) Exposura assassmant for PCS Spills and Firas; (5) Risk Management of PCS Firas and Spills; (S) Utility Oaeision Framework for PCB Cacisior.s; (7) Studias on the Usa and Toxicology of ?C3 Substitutes among other important PC3 Research Projects. Many of tha projects will not be completed until Tacanoar 1984 or later into 1985. Yet they could have a sign.f.cant bearing cn the proper direction for E?A to ta<e determining tha need for further regular.cn of tha usa of askaral transformers. As we have shown, there are .rany uncertainties about tha risks associated with PCB fires. Similarly, there arc uncertainties about the risks associated with the use of new fluids.*/ Each of these considerat.cr3 muse be carefully assessed before any action is taken. What appears clear today, however, is that the three fires EPA evaluated as representative of risk do not adequately account for these uncertainties and do not present a sufficient bas.s on whieh to conclude that the continued use of askarel transformers pose an unreasonable risk of injury to human health or the environment. -----------------------'/ See Section IV.A, infra. HONS 213886 -is- ii: , mechanisms cf pcif avc pccc > major issue in cfiis rulemaxmg .1 the extent 12 which PCS* and PCS mixtures are converted 10 other compounds under pyrolysis and combustion condition*. As discussed mere fully m Sectien II.3.4., a definitive conclusion on this question cannot be drawn from the currently available data. Wh.ia tr.s focus of SPA's question was a i <i r t. transformers, tne Agency also ra.sed a more general quest.or. concerning fires involving mineral oil transferrers. :.n If5WAG' s interview with Cr. Rappe, he indicated that an aoundar.ee of seot would be formed from the combustion cf -inerai oil vhien would probably result m a very complicated -at.-.x. -era.se cf tms ccrplex -atr.x, .t -;rh: be vsr; difficult to extract PCCFs and PCCC* and fairly strong solvents would be required. He related experiences he had * at in spiking certain sediments with carbon 13-labelled PCCFs a-d finding it exceedingly difficult to recover the material, he anticipated that similar problems, along with separating Pills and PCOFs from other hydrocarbons, could result. In its kttwit, SPA examines laboratory studies and discusses them in terms of fire situations. EPA notes that chemical reactions which occur m a fire with Aroclors is far more complex than laboratory pyrolysis experiments. Yet, the Agency expects the same reactions to occur and chemical pred-cts to be formed. Further, EPA suggests that the lower PCS levels MONS 213887 -29- found m mineral oil "should have a minimal efface cr. ;-e reaction rate or product yield a: a given :rcera:-re'. 43 Fed. Reg. at 11074. Lower PCB levels rear lower PCOF lev.3 formed, according to the Agency's analysis. From our discussions with Dr. Rappe and our read.rg of the literature, L'SWAG believes that the Agency's first stats.---, that ?C3 ccToustion m fires is eerplex, is correct. ?/. react.cn mechanises and product yields are affected by -a-', variables other than just temperature. For example, it .ray re plausible to conclude that diluted PCBs or chlorobenzenes eanr.ot be converted to PCOFs or PCDDs, because mineral oil cr other insulating fluids would consume the oxygen available to nurture the fire. Accordingly, there is no information availar. today from which to conclude t.nat PCBs at low concentratic* levels can be transformed into PCDFs. I'SWAG believes that research work being performed ry EPRI, EPA, and possibly others is pivotal to any finding on the conversion of diluted PCBs and, for that ratter, even PCB askarel fluids encountering combustion. IV. OTHER RELEVANT FACTORS In Assessing the reasonableness of tre continued use of transformers containing or contaminated with PCBs when s.:r equipment may become involved with a fire, a number of ot.-.er factors must be considered. EPA has sought information cr twe of these factors, substitute fluids and exposure risk from contaminated water. A third factor, building fires general-/. HONS 213888 -30- also should be coniidrd. Each of these matters u addressed briefly belcw. A. Substitute Fluids for PC3a EPA has requested information on substitutes for ?C3 fluids. Specifically, it seeks data cn tne toxicity of substitute fluids m the event of a fire. Reviewing the available materials, L'SWAG concludes that not much is real!/ known aoout the properties of these fluids, their tox.city, or tne effect on them of pyrolysis or combustion. EPA identifies six major substitute dielectric flu.d types: silicones, high-temperature hydrocarbons, chlorinated hvdrocaroons, non-PCS askarels, fluorocarbons and mineral c.l, Several investigators have reviewed the scientific literature on aoth PCS capacitor and PCS askarel transformer fluids. EP1 itself has performed analyses through SRI International (Terr.a 1981) and more recently PEDCO Environmental Incorporated (jar.a 8, 1984). Similarly, EPRI is m the process of completing a study on PCS substitutes. See Exhibit 3. Although a reoe-t article published in Environmental Science and Technology (17:486A. 1913) concludes that some substitutes currently marketed have a low acute toxicity, appear to be more bicdegrad than PCSa and less likely to bioaceumulate appreciably, tr.ere is limited knowledge, especially in comparison with PCSs, or chronic and subchronic health effects. Thus, the normal day- to-day use of the substitutes generally has not been adequateLv evaluated from a health risk standpoint. HONS 213889 31- Sob* information is available, however, on tne effect of combustion. For example, during arcing cf a perchicrset.-.y.e-* transformer, the PEDCO study indicates the breakdown predicts include hydrogen chloride, carbon, carbon monoxide, and carte dioxide. In contrast, initial EPRI literature analysis shows that the predominant resulting product is phosgene. other hazardous decomposition products include hydrogen chloride ard carbon monoxide. This disparity indicates that a r.gcr:_s analysis of the potential health risks of substitutes and their by-products remains to be made. The current EPRI project should be completed by Decemser 1904. In addition, EPRI will be performing several studies the area of substitutes; 1. Pyrolysis and Combustion of PCS Contaminated Dielectric Fluids and Retrofill Fluids, such as tetrachloroethylene (RP 2020-4), scheduled for completion by Novemoer 1994. 2. Arc and Spark By-products {RP 1499-4-5) of Perchloro#thylene, scneduled for completion by August 1904.3 3. State-of-the-Art Toxicological Review of PCI Substitutes that will focus on pyrolysis and eoabustion products (RP 2020-12), scheduled for coBpletion by December 1904. Although these studies may not be definitive, they should help SPA in its deliberations. B. Risk from Contaminated water EPA solicits comments on the exposure risk from contaminated water resulting from PCB fires. From the limited information we have been able to obtain, it appears that botr. HONS 213890 -32- contammation of water and exposure lavala arc potential.,' nonexistent. Judging from a conversation with Pacific Gas i Electr.t Company, there was very little water used to dowse the transformer at One Market Plaza, and the little ultimately used was contained within the vault.*/ Further, it can be expected in future fires that firefighters will centre 1 j_=n types of electrical fires either by removing the source cf oxygen or by using carbon dioxide to control the fire. Therefore, USWAG believes there should be little contamination and no spread of contamination through water run-off because water is not required to control such a f.re. C. The Ccnsecuences cf 3uildi-c Fires The focus of the ANPR is on askarel transformer fires and the risks from exposure to the associated combustion by products. Given that askarel transformer fires do not bum as raging fires and in fact prevent the spread of fire, a convincing argument can be made that the installation of askarel equipment significantly prevents the direct loss of life from building fires and the damage costs that result frethose fires. In Binghamton, San Francisco, and Chicago, whether there was a real concern from PCI exposure or not. 1/ In the case of Binghamton, it is not apparent whether any water was used to control the fire. MQNS 213891 -33- Cher* w*s no less of life and people could he safely evacuated. USWAG believe* that in .taking any further risk assess**.".:, the measure of safety that askarel transformers provide must be weighed against the consequences of building fires, regardless of the type of transformers that serve them. Every year there are over on* million building fires.*/ In 1931, fires were attributed as th* cause of 300 deaths .n public and work environments at a cost of $3.4 billion. According to a recent article published in Civil Snqireen-a magarin* (May 1984, pp. 41-43), about 00% of the deaths that occur in building fires are attributed to th* inhalation of toxic gases emitted from certain burning materials. Thus we can esc.rate that 720 pecpl* died m 1901 from exposure to toxic fur.es. The toxic fumes result from the decomposition of plastics and other synthetic build.ng and furnishing materials. In a resolution brought before th* California Assembly, Assemblywomen Maxine Waters cites a test developed by a University of Pittsburgh toxicologist which "reveals tnat about 5 ounces of boning of polyvinyl chloride, one of th* commonly used synthetics in buildings, could emit enough hydrogen chloride gas to kill all the people in an averagesite bedroom in 10 minutes.***/ */ In 1901, for example, there were 1,003,775 building fires. See 1911 Database, Federal Emergency Management Authority, National Fire incident Report. H/ Civil Engineering/ASCE. Kay 1904, p. 42. HONS 213892 -3-.CONC1USION From the information provided in the RPC 1984 Report *nd elsewhere, USWAG believes that no unreasonable risk is posed to human he*1th or the environment from the continued use of transformers concerning or contaminated with PC3s. Many studies now being conducted will shed further light on many of the questions arising from the involvement of the transformers m fires. As of today, however, there is insufficient reliable information on which to place further restrictions on the use of the transformers under eonsideratio EPA should, therefore, issue a statement that further ruleiraki at this time is unnecessary. MONS 213893 :.m _/ ix .... .... 141 tea Uactxic lattltit* fetleeal taral tlactrlc Cooperative liMeiitisi leerlean FaU fwi UwelttiM leerlean tloetrle fewer terrlee Corporation Appalachia* rmr Cottar ColiMhue * Uaum Ohio tlactrle Colony Indiana a nlehlfoa electric Coepany teatteky Power Coapany Ohio rower Coapany Atlantic City electric Coapany ta Itinera Ca t tlactria Colony Caetrel tedaoe See 1 tlactrlc Carporatloe Control tlllnola U|bt Crayray caatral tlllaala fehlic Sarwtee Coepany Castral a Imu feat Corporation Control fewer Lifht Colony fehlic Sarvlco Coapany of StlahoH fouthweatarn tlactrle fewar caepany Moot Taiaa Otilltlaa Coapany Cincinnati See a tlactrle Colony CgaoAvoaltn tdlaon Colony Conrolldatod tdlion Coepany of Mow Tart, tna, Payton fewar t Llfht Coapany Duka ro*r Coepany hfiioni Llfht Colony Florida Power a Llfht Coepany Gulf ttataa Utllltlea Cappany ouitoa Lafntmf a fewar Coopany II Unoii fewar Coepany Indlanapella rawor a Llfht Coepany lowfillinoir Gar a (loctrie Coepany Iowa rowar a llfht Colony Iowa fehlic larvica Coepany Jaraoy Control rowar a Llfht COepany Laa utfalaa oayarteant of wator a fewar Middle leuth tarvlcn, tne, Arkema* rowar a Llfht Coepany Louiiien* rowar I Lifht Ceopany saw Orlaam fehlic larvic*. ine. Nlniooippl rowar a Lifht Coepany sononfanala rowar coapany now InfItno fewar Coepany Mow rork Itota llactrie a Gao Corporation hiofara son*** rowar eorperotlon Morthern Indiana fehlic fervice Coepany wartnarn Itataa rewar Caepany Ohio Id loan Coepany Ohio Vailay tlactrle Corporation Oklahoma Car a floetnc Coepany rarlfle 6aa t tlactria Coepany Fenneylvenie power i Llfht Coepany Philadelphia tlactria Coepany Pettea* tlactria Power Coepany Pvhlie larvica tlactrlc a Got Coapany PoMld larvica Caepany of Indiana, tne. louthara California tdlaon Coepany leathara Conpaay lervloea, tna. llabaea Power Caepany Gaerpia fewer Colony Gulf Power eaepany sltalarlppl Power Coapany taape tlactrle Coepany Toaaa Otilltlaa Oaoaratlnp Caepany The Cleveland tlactrlc llltaalnatlnf Coepany The Patrelt tdlaon Coepany Ttia Potaoae tdlaon caepany Toledo Ml tea Caepany union tlactrlc Coepany Vlrylalo tleetrle 1 Power Colony watt Pann Power Coepany wiaconain tlactria rowar Coepany wlaeaneln Power a Llfht Coepany wliceaein Puhlla larvica Corparatloo HONS 213894 (1 Ju ai?GRT Or t:-:z stud*. c? BY A5KABIL T*A.\'s;0?,> i,xr. RESOURCE PLANNING CORPORATION RPC 1225 Street NW Suite Woihtngron CC 2lX2s (202) 7C7.il 11 HONS 213695 o REPORT OF THE STUDY OF RISKS * BY ASSAREL TRa.N'SF OPJ'ER FIRIs Prepared For The Ed_son Elactnc Institute md The Utility Solid Waste Activities Group Prepared By Resource Planninff Corporation 1225 19th Street, N W Suite 50 Washington, D.C. 20036 June 19, 1984 HONS 213896 TABLE OF CONTENTS 1.0 INTRODUCTION...................................... _ 1.1 PCB Regulatory Background . . 1.2 Purpose of this Study ... ., 1.3 Study Methodology..................................... . 1.4 Organization of this Report ... ?\cz , 1 ; 2 4 2 0 SURVEY FINDINGS ... 5 2 I Survey Response .... 6 2.2 Askarel Transformer Population ... 5 2.3 Location and Environment of Distribution System Askarels 6 2.4 Frequency and Nature of Fire Incidents . . 3 2.5 Cost of Fire Cleanup ... . 11 2 S Utility Risk Reduction Measures . .12 2.7 Retrofit Experience . , u 3 0 COMPARISON OF SURVEY DATA WITH EPA MATERIALS 3.1 General ..... . 3.2 Frequency and Nature of Fire Incidents . 3.3 Cleanup Costs ..... 3.4 Comments Regarding EPA Cost-Effectiveness Analyses . . .. . 13 13 IS 19 20 APPENDIX A: LIST OF UTILITIES SURVEYED APPENDIX B: SURVEY QUESTIONNAIRE ii HONS 213897 LIST OF TABLES TABLE 1 TABLE 2. TABLE 3 table 4. TABLE 5. TABLE 6 T ABLE 7 TABLE 8 Location of Distribution System Askarels. Buildings in Which Transformers are Located . Location of Transformer Fire Incidents . . Average Cleanup Cost Estimated Reductions m Distribution Askarel Transformers Based on Current Phase-Out Programs........................................................ KVA Rating and Fluid Capacity of Retrofilled Transformers . ..................................... Flush Solvents used in Retrofits . Retrofit Results............................................... PAGE 6 9 13 14 15 16 iii HONS 213898 1 0 INTRODUCTION' 1.1 PCB Regulatory Background In August, 1982, EPA issued the PCB Electrical Use Rule authors.? the use of certain electrical transformers containing polychlorinated biphenyls (PCBs), Among other provisions, this rule allowed electric utilities to continue to use the approximately 40,000 askarel transformers then in service. In March, 1984, EPA issued an Advance Notice of Proposed Rulemalun? (ANPR) stating that the risks posed by fires involving PCB Transformer had not been considered during formulation of the 1982 rule, and that consideration of such nsks might result in the necessity for additional regulation. Accordingly, the ANPR solicited data specific to the risks of fires posed by PCB Transformers. 1.2 Purpose of this Study In the summer of 1981, Resource Planning Corporation (RPC), under contract to the Edison Electric Institute and the Utility Solid Waste Activities Group (EEI/USWAG), surveyed the 100 largest utilities m the United States. This survey collected data regarduig the use of PCBs by the electric utility industry, and was responsive to information requirements defined by EPA, EEI/USWAG, and the Environmental Defense Fund as part of an order issued by the U S. Court of Appeals for the District of Columbia Circuit. Although the 1981 survey collected substantial data regarding the use of PCBs and formed the basis for the August, 1962 rule, no data were collected regarding risks posed by PCB Transformer fires. This report describes the results of a 1984 survey of electric utilities specifically designed to obtain data relevant to the fire risk ANPR. The survey was again performed by Resource Planning Corporation under contract to EEI/USWAG. This recent survey provides substantial HONS 213899 empirical data regarding PCB Transformer fire risks, and should be particularly useful m describing the location and environment of askarel transformers and estimaung the frequency of PCB Transformer fires - two issues critical to EPA analysis of alternative regulations 1 3 Study Methodology In order to respond to the fire risk ANPR, three general types of information had to be obtained from utilities: detailed information regarding the location and environment of askarel transformers.1 information regarding transformer fires, and retrofill experience Four approaches to the collection of these data were considered, including: - Surveying those utilities that reported owning askarel trans formers m 1981 (90 of 100 surveyed) - Surveying the 10 or 20 utilities that reported the most askarels in 1981 Collecting some data (i e., fire data and retrofill experience) from the 1981 sample, and more detailed data on transformer location and environment from a small sample of utilities that are known to have such data available - Surveying those utilities that reported owning askarels In 1981.' and asking them to provide detailed data on a sample of their transformers. The advantages and disadvantages of these various approaches are relatively straightforward. Using the same utility sample as m 1981 would facilitate data comparisons, i.e., the 1981 transformer inventory could easily be compared with the 1984 inventory. However, it was deemed impractical to expect 90 utilities to promptly provide detailed information regarding the location and environment of each askarel transformer they own. Surveying the 10 or 20 utilities reporting the most askarels in 1981 would provide a somewhat more manageable sample, but the problem facing individual utilities of providing detailed data on hundreds of 1. Although PCBs were sold under a variety of trade names, they are commonly referred to by the generic name "askarel " Throughout this report, "askarel transformer" refers to transformers typically containing PCBs at levels of 80% or greater 2 HONS 213900 askarels would remain. The response rate would lately be low, par ticularly given the use constraints * Collecting some data from the 1981 sample and more detailed data from a small number of utilities (3-6) that are known to possess the data was considered a reasonable approach The surrey would be managtable, and data could be obtained on a substantial number of askarels within a relatively short time frame. The disadvantage with this approach ls its representativeness. It would be necessary to generalize to the industry based on a small group of utilities If the response of all sampled utilities were simihar, u could be reasoned that such comparability allowed reasonable population projections. However, if responses were dissimilar projections would be of questionable validity All of the disadvantages associated with the above approaches could be overcome if a random sample of utility industry transformers were selected and data obtained regarding their location and environment No single utility would be required to provide data on a large number of askarels, thereby alleviating their reporting burden, and reliable projections could be made to all industry equipment from a relatively small sample of transformers. Such an approach seemed the best methodology for obtaining data required to respond to the AN'PR. In refining this methodology, the 1981 survey data were reexamined in light of the ANPR data requirements Electric utilities have askarel transformers located in substations, generating facilities, and distri bution systems, However, the major concern stated by EPA in the proposed rule-making was the risk to the public of askarel trans formers, i.e., distribution system askarels. Given that expressed concern, and in the absence of any reports that askarel trans formers located in utilities1 restricted access areas have been involved in fire incidents similar to those that prompted EPA's ANPR,X 2. Substations and generating facilities also pose little or no risk, l e . there is no public exposure, equipment monitoring and inspection occurs frequently, and utility workers are both sensitive to the hazards posed by electrical equipment and trained in the use of PCB Nonetheless, the survey did seek information on fire incidents involving askarel transformers wherever the equipment was located 3 HONS 213901 data collection efforts were focused on the distribution system askarel transformers. In defining our study sample, it was determined that 52 utilities owned 99% of the distribution system askarel transformers reported m 1981 Therefore, it was decided to focus survey activities on these 52 utihues 1 Each of the 52 were asked to provide1 - General Data, including their current askarel transformer inventor;1 and risk reduction measures taken or planned ' - Fire Incident Data, including information regarding all fire incidents dunhjr~cne last three years which involved an askarel transformer, PCB Transformer, or transformers containing askarel substitutes. Askarel Transformer Data, including detailed information regarding the location and environment of a sample of their askarel trans formers. A sample of 450 randomly selected askarel transformers was deemed sufficient to make projections to the electric utility industry distri bution system askarel transformer population. Each of the 52 surveyed utilities was given specific instructions regarding the transformers for which data were to be reported. It should be noted that only trans former location and environment data were collected on a sample basis. All other survey data covered all askarel transformers owned by respondent utilities. For example, the data on fire incidents covers lU fire incidents in the last three years experienced by the sampled utilities. 1.4 Organisation of this Report The remainder of this report comprises two major sections. 2.0 SURVEY FINDINGS describing the data obtained via the survey of 52 electric utilities. 3. Appendix A provides a list of utilities surveyed, and Appendix B provides a copy of the survey questionnaire. 4 HONS 213902 3 0 COMPARISON OF SURVEY DATA WITH E?A MATERIALS analyzing the preliminary EPA regulatory analysis4 m light cf data obtained via the utility survey Putnam, Hayes & Bartlett, Inc., 1984. Preliminary Study and Cost-Effectiveness Analyses of Alternative Regulations for infloor PCS Transformers, EPA. Office of Pesticides ana Toxic Substances Fires), February 1984. MONS 213903 5 2.0 SURVEY FINDINGS 2 l Survey Response Of the 52 utilities, 51 provided data The single non-respondent indicated that insufficient time was available in which to respond The following results represent projection to the population of utilityowned transformers based on the results of the survey 2 2 Askarel Transformer Population Based on the 1981 utility survey, it was estimated that the electric utilities owned 39,640 askarel transformers, with 22,469 in distribution systems The 1984 survey indicated that the electric utilities nowown 35,984 askarel transformers, and the number of distribution system askarels has declined by approximately 18.5% since 1981 There are now an estimated 18,291 askarel transformers in service it. utility distribution systems. This overall decline as well as the decline in distr.bution system askarels is somewhat greater than would be expected via normal attrition, because a number of utilities have instituted accelerated phase-out programs. 2.3 Location and Environment of Distribution System Askarels Table 1 indicates the general location of the 18,291 electric utility distribution system askarel transformers TABLE 1: Location of Distribution System Askarels Inside Building Sidewalk Vault Other TOTAL Number of Transformers 5,000 10,919 1.372 18,271 Percent of Total J2 8 59 7 75 TuiTS5 5. The reported decline in the total number of utility-owned askarel transformers from 1981 to 1984 is somewhat less than the decline in the number of distribution system askarels, i.e., the 1984 survey reported more substation/generating facility askarels than the 1981 survey. It was determined that two utilities underreported their 1981 population of generating facility askarel transformers. The result of thiJ reporting error on 1981 equipment projections was less than 4%. 6 MONS 213904 Approxunately 10,919 transformers are located m outdoor sidewalk vaults, and an additional 1,372 are m other outdoor locations In the case of a fluid loss, 66.6% of all distribution system askare! transformers locations would contain the fluid and prevent release into the environment, 6 5% would contain the fluid m a catchment area. 7.7% of the locations would drain to sewer systems, and 19 2% would drain to surrounding soil, gravel or concrete For transformer cooling purposes, 94% of all distribution system askarel transformers are ventilated to an outside area, 4.5% vent to an indoor area, and less than 1% are in a sealed room or vault. Less than 20% of these ventilation mechanisms involve forced*air systems 2 3 1 Indoor Installations Of the 6.0CO transformers located in buildings, Table 2 provides aduonal information on the various types of building installations TABLE 2: Buildings in Which Transformers are Located Number of Transformers % of Total Residential (Condo, Apt. Hotel) Institution (School,Hospital,Church) Office Industnai/Mfg. Storage (Warehouse) ' Other Public (Store, Mail) Other* 727 318 3,232 180 180 727 636 12 1 5i 53 3 30 30 12 1 10 5 Total 6,000 100% * Primarily includes parking garages, and uninhabited industrial structures In those instances where transformers are located inside buildings, the average age of the buildings is 29 years, and although the average height of the building is 19 floors, over 50% of the transformers are in buildings of 11 floors or less. 7 MONS 213905 While 51 2% of the indoor transformers are installed in the basements of buildings, only 3 1% of the indoor installations are roof top instal lations, the Pemaiaing 45 7% are located on other intermediate floors within the building The overwhelming majority (84 5%) of the indoor transformers are located in vaults within the budding, 9 3% are located in separate transformer rooms, and the remaining 6 1% are installed m open areas Of the transformers that are located m a vault or transformer room, less than 1% have air conditioning or heating system intake vents in the transformer enclosure itself, and less than one-half (44.7%) have air conditioning or heating system intake vents in the area immediace.y adjacent to the transformer vault or room. Other than these direct routes, approximately 31.2% of the transformer locations are judged to provide some, albeit indirect, access to building heating and air conditioning systems The responses to the survey indicate these mthree: routes may include access to heatr.g and air conditioning through open windows in the building, holes in concrete walls and ceilings, spaces under doors, gaps around electrical conduit, etc. 2 3 2 Outdoor Installations There are approximately 12,291 askarel transformers m sidewalk vaults and other outdoor locations. In the case of a transformer fire. 10 9% of these transformer enclosures are located immediately adjacent to building heating and/or air conditioning systems which could serve as a means of building access for smoke/soot 2.4 Frequency and Nature of Fire Incidents The 1984 utility survey defined a PCB Transformer fire incident as a fire in which a transformer has been directly exposed to a fire, either from within the transformer or an immediately adjacent source. Using this definition, survey respondents reported that 8 PCB Trans former fire incidents have occurred in the last three years All of these incidents were reported to involve askarel transformers 8 MONS 213906 Using these data and the total number of askarel transformers owned by surveyed utilities, it can be estimated that o.01 percent of all askarel transfarmers owned by electric utilities can be expected to be involved in a fire incident each year Because so few incidents occur, it is difficult to develop reliable incident predictors, i.e . the small size of the sample does not allow the correlation of fire incidents with transformer age, KVA rating, type of building, ume of day, or any other measured variable. However, a review of the circumstances surrounding the 8 reported incidents may be useful in considering the likely results of future incidents Table 3 indicates the locations of utility industry askarel transformers involved in fire incidents dunng the past three years. TABLE 3: Location of Transformer Fire Incidents Location ___________ Outsice * sidewalk vaults Inside - office building Number of Incidents T _4 TOTAL 8 Because the risks of exposure posed by outdoor transformers may be significantly different than those of indoor transformers. each is discussed separately below. 2.4.1 Sidewalk Vault Incidents Four of the eight fire incidents involved transformers in outside sidewalk vaults. KVA ratings were reported for three of the four transformers (500, 750 and 1,000) and the ages of the four transformers were 8, 12, 15 and 20 years. Two of the fires occurred m the morning, one in the evening, and for one the time was not provided. All but one occurred on a weekday. . In two of the three incidents smoke escaped from the transformer enclosure via the sidewalk grate but did not enter any public buildings In one incident smoke did not escape the transformer vault. In one 9 HONS 213907 incident smoke escaped the vault and entered a nearby building In one of the three incidents, no fluid was released from the transformer, and in the otter three incidents, releases of 5, 25, and 60 ^aliens occurred Data regarding exposure of the public, utility personnel, and fire fighters were reported for two of the four incidents. Neither reported that more than five members of the general public were exposed to smoke/soot from the fire and the average number of firefighters and utility personnel exposed to smoke/soot was mne Limited data were reported regarding the lapse of time between fire start, de-energuation. fire department response, and extinguishment In one of the incidents, the transformer was de-energized approximately t' 10 minutes after the fire started, in one incident de-energuation required an hour, and in one incident as much as three hours may have elapsed. For one incident no data were reported regarding de-enerjization No data were available regarding the time required for fire department response. For one incident it was reported that two hours elapsed from fire start to extinguishment, and for another total time lapse was reported as three hours. PCB, PCDF, and PCDD test results were generally not available for the four sidewalk vault fires. For one incident PCB fluid levels of 13,000 ppm were reported before cleanup and 12 ppm after 2.4.2 Inside Building Incidents The four office buildings in which transformer fires occurred ranged from 1 to 48 floors. One building was 10 years old, one was 13 years old, and the ages of two were not reported. Transformer KVA ratings ranged from 500 to 2,000. The ages of two of the transformers were 10 and 13 years, with two ages unknown. In each of the four incidents occurring inside buildings, the transformer was enclosed in a vault, and in each incident smoke/soot escaped the transformer vault In three of the four incidents the escaping smoke/ 10 MONS 213908 soot only entered immediately adjacent areas of the building in one incident the smoke escaped the vault and was picked up by the air intake of an adjacent building Fluid was released from the transformer m each of the four incidents, with the quantity released ranging from 4 to 100 gallons and the average being 41 gallons. The extent of exposure of the public, utility personnel and firefighters to smoke/soot was generally not known However, for one incident no exposure to the public was reported, and for two incidents exposure of 12 and 18 utility person,-.*; and firefighters was reported. For one incident, PCS levels were reported in the soot before cleanup of 2-93 micrograms per 100 square centimeters and 1-21 after. Ancther reported 860-2,000 micrograms of PCS per 100 square centimeters of wipe area in soot before cleanup, with no after cleanup results available A third incident reported 18 micrograms per 100 square cenumeters of PCDF in the transformer vault, with no PCDDs or PCDFs found elsevher (after extensive tesunj). 2.5 Cost of Fire Cleanup When calculating the average cleanup cost associated with reported fire incidents, one incident stands substantially apart from all others Of the 8 incidents reported, cleanup of one is estimated to have cost $15 to 20 million. No other incident reported costs as high as S425,0`j To avoid the distortion of the one $15 to 20 million incident, Table 4 excludes that $15 to 20 million incident and indicates the average ost of cleanup associated with all other reported fire incidents. TABLE 4 Cost Type Analysis/Testing Cleanup/Decontamination Disposal Equipment Replacement TOTAL Average Cleanup Cost Sidewalk Vault $ 233 2,487 5,000 20.000 $27r700 Inside Building $ 16,625 144,375 8,750 109.175 $278,925 11 HONS 213909 2 6 Utility ?.isk Reduction Measures Survey data indicate that uoliaes have undertaken a variety of measures to reduce tk;likelihood of PCB fire incidents More specifically - 7S% of surveyed utilities have conducted nsk-assessmen: inspections of their askarel transformers *' 76% of surveyed utilities have askarel transformer phase-out proffTims completed or currently in effect - 35% of surveyed utilities have notified fire departments in their servile area of the location of askarel transformers - 20% of surveyed utilities have posted warning signs for firefighters m the approaches to askarel transformers - 75% of the distribution system askarels have a protective fuse, switch, or breaker on the low voltage side of the transformer, approximately two percent of which are oil-filled. These pro tective devices are designed to isolate (de-energue) the trans * former in the event of a fault in the low voltage line. - 19% of the distribution system askarels have a protective fuse, switch, or breaker on the high voltage side of the transformer, approximately 52% of which are oil-filled. These protective d-vices are designed to isolate (de-energi2e) the transformer m f_.e event of a fault in the high voltage line Askarel phase-out programs are currently in effect in 76% of the surveyed utilities These programs are primarily directed toward phasing out distribution system askarel transformers, although 13% of these utdities also have reported programs and plans for phasing out substation and generation facility askarels. Based on the phase-out programs voluntarily instituted among the utilities, it is estimated that the number of distribution system askarel transformers will be reduced from 18,291 to 12,987 by the end of 1985 This represents a reduction of 29%. By the end of 1989 the total will have been reduced to 6,102, a reduction of 65%. Table S shows the projected reduction m distribution askarels through 1994, given the phase-out programs currently in effect. 12 HONS 213910 TABLE S' Estimated Reductions in Distribution Askarel Transformers Based on Current Phase-out Proprams Year T55T 1985 1986 1987 1988 1989 199C 1991 1992 1993 1994 Estimated Number of Distribution Askarels ---------- IT23I----------- 12,987 10,792 8,780 7,499 6,402 5,670 4,939 4,207 3,658 2,927 Percent Remaminr 160 71 59 48 41 35 31 27 23 20 16 A variety of priority considerations are being taken into account by the utilities in their phase-out programs. The most common factors considered pertain to the location of askarel transformers within or m relation to the location of buildings. The type and location of the buildings in which these transformers are located are also being considered, with schools, hospitals, commercial and other buildings in congested areas receiving higher priorities. Several utilities are * prioritizing askarels for phase-out based on the assessed risk to food, feed, and water supplies. Other factors being considered are the age, condition, and voltage of the askarel equipment. Seventy-eight percent of surveyed utilities have conducted nsk assessment inspections of their askarel transformers to evaluate exposure risk. These inspections have often been performed in conjunction with phase-out planning. Ten percent of those utilities conducting risk assessments have increased the frequency of their askarel transformer inspections as a fire risk reduction measure. An additional 10% have sealed cable ducts and other holes in transformer areas in order to contain smoke in case of a fire. Other efforts reported to reduce fire risk include installing fire detection or fire suppressant systems within vaults, equipping vaults with a coolant sensor, sealing excess grating area or sidewalk vaults, and assuring that a C02 fire response vehicle Is available. 13 HONS 213911 2 7 Retrofill Experience Of the 52 utilities surveyed, 11 reported having retrofilled one or more askarel transformers, with a total of <18 aslcarel transformers retrofilled among these utilities The average cost of retrofuhnj reported by those providing cost data was 515,700 per transformer Approximately 45% of the costs reported were $5,000 or less, while the remaining 55% ranged oetween $10,000 and $<11,000 The range in KVA racing and fluid capacity of the askarel trans formers retrofilled is shown m Table 6. TABLE 6: KVA Rating and Fluid Capacity of Retrofilled Transformers Fluid Capacity (in srallons) $100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 TOTAL Transformer KVA (Nameplate Rati". z: 1017500 501-1000 lQGl-5uu0 1 'w . 700 1 1 00 1 14 9 2 25 012 3 023 5 00 l l 002 2 000 0 _0 0 _5 __5 22 . 12 15 49 None of the retrofilled transformers had nameplate ratings less than 100 KVA or higher than 5,000 KVA PCB levels reported for these transformers prior to retrofilling ranged from 400,000 ppa to 750,000 ppm. Over 60% of the transformers had starting PCB levels estimated at 700,000 ppm or above. In 27% of the retrofills, a flush/refill method was used, and in 35% of the cases a filtering process was used. The remaining 38% were retrofilled using both the flush/refill and filtering methods In several cases the filtering process was reported to be ineffective. 14 MONS 213912 whereas a other cases the flush/reftii process alone was considered insufficient. The fluid used is retrofilliny 4G% of the transformers was silicone RTeap was used in 29% and oil in 31% of the transformers. A variety of flush solvents were used m these retrofdls Table 7 identifies these solvents alony with the frequency of their use TA3LZ 7 Flush Solvents used in RetrofiUs Solvent N'umber of Retrofits Silicone Tnchlorethane Trichiorethane and Siiicone Chlorothane Tnchiorobenzene Oil Rtenp None No Response 3 4 I 10 7 11 1 3 4 TOTAL 49 Total does not equal 100% due to roundiny % of Total 6 8 2 20 14 22 2 H __ 3 98- In order to assess the effectiveness of the retrofill methods, respon dents were asked to report the PCB level after their 1st, 2nd. ar.d 3rd refill or filter chinye Data on PCB level were provided or. 40 transformers and are summarized in Table 8. 15 HONS 213913 TABLE 3 Retro fill Results Beginrung PCB Level** 1,000,000 1,000,000 750,000 750.000 700,000 700,000 TOO,000 "00,000 TOO 000 700.000 7CG,OCO 700,000 700,OCO 700.000 700,000 700,000 700 000 700,000 7C0.CCQ - , Li-'J 700,000' 700,CG0 600,000 600,000 600.000 600,000 600,000 600,000 600,000 500.000 500,000 500,000 400,000 400,000 400,000 400,000 400,000 400,000 400,000 400,000 Subsequent PC3 Levs (ppm) Type 1st Ond Ira Process* Flush/Filter Flush/Filter Flush/Filter B 703 3 1,251 F F -B 779 B 760 B 8,673 B 10,103 B 3,744 R 42,000 R 32,000 R S3.0G0 R 31,000 R 24,000 R 36,000 R 14.200 R 81,000 R 1,200 R 3,200 R 1.0C0 B 57,CC0 B 47,000 F 1,180 F 578 R 269,016 R 274,690 F 20,000 F 7,075 R ISO F 22 F 19 F 172 B 158,094 B 48,220 B 15,040 B 34,575 B 13,737 B 25.060 B 14,735 B 14,665 43 85 *788 130 2.8C6 2,753 6,443 3,000 2,200 4,600 5.000 1,800 3,200 32,200 3,149 11 25 5,000 550 10 7 95 - 105,840 30,942 10,157 23,087 9,171 16,817 9,849 9.822 237 316 20 27 472 6,913 6,175 3.416 4,510 400 2*0 850 2,100 560 300 4,500 37 10 91 Time Becn-E 3 3 13 3 5 5 5 5 5 30 30 6 6 6 33 29 11 5 33 3 ,, > <> 3 5 4 1 '2 2 2 2 2 0 2 2 * F * Filter; R * Flush/Refill; B * Combination ** For these transformers whose beginning PCB level was reported as askare a 70% PCB concentration was assumed. MQNS 213914 16 These survey data indicate that 13% of the transformers attained a PCB level of 5300 ppm after the first refill-filter change Of those transformers PefiHed/filtered a second tune, 23% attained a PCB level of 5500 ppm After the third refill/filter change, the PCB level m the 13 transformers were lowered to 510,000 ppm and 56% of these 13 attained a PCB level of 5500 ppm However, as is readily apparent the retrofill results are quite erratic. Results of the first refill/filter change range from 19 ppm to 274,690 ppm, the second refill/filter change from 7 to 105,840 and for the third refill/filter change from 10 to 6,913 There is no apparent correlation with ume, beginning level, cr process type In eight cases, the PCB levels increased between refills/filter changes rather than decreased. This was reportedly due to PCB ben? retained in the paper insulation and windings. Even where PCB levels are apparently decreased substan tially after three refills/filter changes, comments by respondents indicate the possibility of future increases due to continued PCB leaching Utilities have been spending an average of eleven months between first and third refill/filter change procedures. Thirty-seven percent of those reporting indicated spending two months or less Twenty-one percent, however, indicated spending over two years As stated, there appears to be no relationship between the length of time spent for retrofilling and the PCB levels reached. 17 MGNS 213915 3 0 COMPARISON OF STUDY DATA WITH EPA MATERIALS 3 I Genera] * With reyard to risks posed by askarel transformer fire incidents, E?A states in the ANPR. If EPA is not provided with adequate data, especially w the areas or the risks posed in the event of a firethe probability of these fires occurring, and the costs associated with clean-up following these incidents, EPA will make its regulatory judgments based upoo the data set forth in this document. These data indicate that PCB Transformer fires pose relatively high risks, occur with unknown frequency, and can result in relatively high clean-up costs This report provide* substantial empirical data regarding the frequency with which transformer fires occur and the costs of cleanup. A comparison of these data with the estimates developed in the preLmmary EPA analysis is presented below The A.N'PR also solicited general comments regirdutg the preliminary EPA cost-effectiveness analyses. As part of our study activities, the EPA analyses were reviewed, and the final section of this report provides our commentary. 3.2 Frequency and Nature of Fire Incidents In the ANPR, EPA presents a variety of estimates of the frequency of PCB Transformer fire incidents, ranging from a low of an expected 8 incidents per year to a high of 1,530 per year. In the preliminary study and cost-effectiveness analyses, the EPA consultant estimates that the number of expected annual catastrophic incidents will be between 0.6 and 11.*6 6. The preliminary EPA cost-effectiveness analyses estimates the lowprobability of a catastrophic incident as 0 0005 percent, the high probability as 0.01 percent, and the beginning (19841 transformer population to which these rates apply as 114,469. 18 HONS 213916 Developing data comparable to these esaaated frequencies is iiff.;.:: because do dear definition of fire incident or catastrophic incident has been developed by EPA In conducing the utility survey, a fu-e incident was defined as a fire in which a transformer has been directly exposed to a fire, either from within the transformer or an immediately adjacent source Based on this definition, survey data indicate that one fire incident per every 7,787 askarel transformers can be expectec each year This suggests approximately 5 fire incidents per year involving utility transformers If the definition of catastrophic incident is related to cleanup costs as was done in the preliminary EPA cost-effectiveness analyses, only one incident involving a utility-owned transformer m the Last three years would be classified as catastrophic Another way to estimate the likelihood of a catastrophic incident is to determine the number cf askarel transformers located in areas which might result in a Bingnam::n or San Francisco type incident, and estimate the likely frequency of j fire involving one of these transformers. Data obtained via the survey allowed reasonable estimates of these numbers Approximately 2,700 transformers are located inside buildings, in areas adjacent to rooms with heating or ventilation system intakes. Another 1,300 are in sidewalk vaults immediately adjacent to building heating and/or air conditioning systems. These 4,000 transformers might be considered as potential catastrophic incidents if the definition relates to the type of incident occurring in Binghamton or San Francisco. Given Lie reported frequency of askarel transformer fires of 0 01 percent, a catastrophic fire incident of the Binghamton/San Francisco type involving a utility-owned transformer can be expected approximately once every 2.5 years. 3.3 Cleanup Costs Insofar as fire incidents are concerned, .the preliminary EPA costeffectiveness analyses focuses solely on catastrophic incidents and assumes a cleanup cost of $20 million. Survey data indicate that the average cleanup cost for a non-catastrophic transformer fire incident outside a building Is $27,700 and inside a building is $278,925. The 19 MGNS 213917 single catastrophic incident reported in the survey was estimated to require S15 to 20 million in cleanup costs. 3 4 Comments Regard:nsr EPA Cost-Effectiveness Analyses A preliminary examination of the EPA cost*effecuveness analyses reveals four significant issues: - The estmaated frequency of catastrophic incidents may be overstated - The methodology used for calculating phase-out costs may be inappropriate - The assumptions underlying leakage rates used for calculating the quantities of ?C3s released frca askarel transformers ma\ be inappropriate. ' - When calculating PCB exposure, the benefits of spill cleanup are ignored, yet the costs or cleanup are considered The net effect is to overstate the cost-etfectiveness of additional reguiat.on 3 4.1 Freouencv of Catastrophic Incidents The EPA analysis uses a low probability of a catastrophic incident of 0 57 incidents per year and high probability of 11 incidents per year As indicated in section 3.2 of this report, survey data indicate a catastrophic fire incident involving a utility-owned transformer is likely to occur approximately once every 2.5 years (0.40 per year). The EPA low probability estimate is reasonable, however, the EPA high probability estimate substantially overstates the apparent nsks 3 4.2 Phase-out Costs The methodology used by EPA for calculating phase-out costs is not clear. The preliminary study indicates the methodology used for the 1982 rule is used with minor modification. We criticised the 1982 methodology in comments submitted to EPA on May 24, 1982. To the degree the 1984 EPA analysis has not incorporated changes resulting from our previous comments, the EPA estimated phase-out costs are understated. 20 HONS 213918 Estimating incremental phase-out costs shouJd contrast normal utility equipment replacement during the phase-out period with replacement which would be required dunng the same period as a direct result of the mandated phase-out. The costs associated with the incremental replacement units are the incremental phase-out costs inasmuch as they reflect only those costs associated with the purchase of units phased out pnor to the end of their useful life The 1982 EPA analysis estimated costs that would be incurred by utilities over a 30-year normal replacement period and compared those costs with costs incurred over a 6-year phase-out per.od. If the 1932 EPa methodology for calculating incremental phase-out costs was used m 1934, the EPA costs are understated 3 4 3 Askarel Transformer Leakage Rates Again, the 1984 EPA methodology for estimating PCB releases from askarel transformers is not clear, but is apparently the same as that used for the 1982 rule If the same methodology has been used, it can be summarized as follows. 1 The percentage of askarel transformers expected to fail m a year is estimated. 2 The percentage of failures expected tc result in PCB spills is estimated. 3 The percentage of askarel transformers expected to leak without failing is estimated. 4 The quantity of PCB fluid lost per spill is estimated 5. The quantity of PCB fluid lost per leak is estimated. 6. Using estimates from Steps 1-5 and an estimate of thenumber of askarel transformers in service, the total annual loss of PCBs is estimated. Conceptually, the approach is fine However, the methodology whereby estimates are developed by EPA for steps 1-3, i.e., the basis for determining the critical factors of failure, spill, and leakage rates, is not clearly defined in either the 1982 or 1984 EPA analysis. These 21 HONS 213919 data are cited as being based on data obtained from the 1981 EEI/L'SVAG sunrey and industry comments. No precise citations or methodology are provided.* The data collected by EEI via the 1981 utility survey do sot provide any information on failure rates inasmuch as reliable data on this were not available from the utilities. Similarly, no specific data were collected and reported by EEI regarding the per centage of failed transformers which spill or the percentage of trans formers which leak without failing. Although the EEI data may have been used in conjunction with other information to estimate these factors, the EPA analysis does not describe the methodology used 3ecause the procedures whereby EPA estimated spill quantities are substantially undocumented, it is difficult to assess whether the benefits calculated are realistic 3 4 4 Spill Cleanup A further problem with the EPA cost-effectiveness analyses is the treatment of PCB spill cleanup. Stated simply, although EPA recognises that utilities clean up PCB spills and attributes costs to these activities, they ignore the effect of cleanup when calculating benefits (defined by EPA to be gross pounds of PCBs spilled). In their 1982 analysis, EPA indicated that no empirical data were available which could be used to estimate the degree to which PCB spd] cleanup is effective. EPA also indicated that cleanup costs were estimated by the utility industry to range from $100 to $1,000,000. Despite the fact that no empirical cost data were available except this broad range of estimates, EPA made assumptions regarding the cost of cleaning up a PCB spill. These assumptions are apparently also used in the 19S4 analysis. Although the precise effects of these estimates on the EPA cost model cannot be determined, estimated phase-out costs are reduced by offsetting cleanup costs that are avoided as a result of phase-out, i.e., industry will avoid costs because equipment cannot spill if it is removed from service. Reducing estimated phase-out costs due to cleanup activities without considering the benefits (pounds spilled) due to cleanup, produces an artifically inflated estimate of the cost-effectiveness of phase-out. 22 HONS 213920 Existing EPA regulations governing the cleanup ar.d disposal of ?C3$ are sufficient^ to ensure that spill cleanup efforts by utilities remove m excess of 99% of all PCBs spilled from utility-owned equipment I;' EPA were to accept the effectiveness of existing regulations, the gross pounds of PCBs estimated to enter the environment without a phase-out would be 99% lower than current EPA estimates (EPA assumes cleanup has no effect on exposure). This would have the effect of substantially reducing the benefits attributable to phase-out, thereby substantially decreasing the cost-effectiveness of phase-out options7 7. Assuming a reasonably timely cleanup response, cleanup to any level below 1.000 pom should conservatively remove 99% of the volume of PCBs spilled. 23 HONS 213921 APPENDIX A LIST OF UTILITIES IN SURVEY HONS 213922 Alabama Power CwapaQy Appalachian Power Company Baltimore Gas & Electric Bos too Edison Company Central Power & Light Company Coloabus & Southern Ohio Electric Company Commonwealth Edison Consolidated Edison Co. of New York Delasrvs Power & Light Company Duke Power Company Duquesne Ligot Company Florida Power & Light Co Florida Power Corporation Georgia Power Ccrpony Idaho Power Company Illinois Power Company Indiana & Michigan Electric Company Kansas City Power & Light Company Kansas Gas & Electric Company Kentucky Utilities Company Los Angeles Department of Water & Power Louisiana Power & Light Company Louisvills Cos & Electric Co'Opauy Memphis Light, Gas & Water Division Metropolitan Edison Company Minnesota Power & Light Company New England Power/Massachusetts Electric Co. Niagara Mohawk Power Company Northern Indiana Public Service Company Northern State* Power Company Ohio Power Company Oklahoma Gas & Electric Company Pacific Gas & Electric Pacific Power & Light Company Pennsylvania Electric Company Pennsylvania Power & Light Company Public Service Company of Colorado Public Service Company of Indiana, Inc. Public Service Company of Oklahoma Public Service Electric & Gas Company Sacramento Municipal Utility District Southern California Edison Southwestern Public Service Company Tampa Electric Company Texas Electric Service Company ' Texas Power & Light Company The Toledo Edison Company Union Electric Company Utah Power & Light Company Virginia Electric & Power Company Vest Penn Power Company Wisconsin Electric Power Company Nonroepondent HONS 213923 APPENDIX 3 SURVEY QUESTIONNAIRE HONS 21392A * EDISON ELECTRIC INSTITUTE (EEI) and UTILITY SOLID WASTE ACTIVITIES GROUP (USWAG) ASKAREL TRANSFORMER SURVEY HONS 2U925 Introduction On March 23, 1984, the U S. Environmental Protection Agency (E?A) published an Advance Nouce"of Proposed Rulemaking (ANPR) regarding the use of PCBs in electrical transformers The ANPR indicates EPA concerns that inadequate consideration was given to the risks of transformer fires prior to issuance of the August 25, 1982, rule governing the use of electrical transformers containing PCBs. Is order to determine whether additional regulation is required, EPA has requested a vanety of information regarding transformers and fires USWAG has retained the services of Resource Planning Corporation (RPC) of Washington, D C , to accumulate the required data and assist m its presentation to EPA. As with the 1981 survey of PCB usage conducted by RPC for USWAG. the study is designed to cover a significant portion of the relevant equipment universe by scientifically sampling a limited number of the largest utilities Since a relatively small number of utilities are being asked to provide data, it is imperative that you fully participate to ensure the validity of the findings ALL RESPONSES PROVIDED IN THIS SURVEY DOCUMENT ARE CONFIDENTIAL THE RESULTS OF THIS STUDY WILL BE REPORTED IN THE AGGREGATE AND UNDER NO CIRCUMSTANCES WILL INDIVIDUAL RESPONSES BE ASSOCIATED WITH AN INDIVIDUAL UTILITY. To facilitate follow up for the purpose of clarifying responses, you are requester to indicate the name of your utility, and the name and phone number of a person to contact if there is some question m processing your responses Name of Utility: * Person to Contact: Phone Number: -2- HONS 213926 To meet the uae requirements imposed by EPA. ail responses must be received NOT LATER THAN MAY 15. 1984 Survey forms should bf mailed to EEI/USWAG Counsel c/o Tool Alien, Esq Wald, Harkrader A Ross 1300 Nineteenth Street, N W Washington, D C 20036 All questions or inquiries regarding this survey should be directed to' Rick Bell or Tom Florence Resource Planning Corporation (202) 797-1111 General Instructions This questionnaire is comprised of four major parts I GENERAL DATA II FIRE INCIDENT DATA III AS5ARSL TRANSFORMER DATA IV. RETROFILL DATA Each major questionnaire part has its own instructions Please read all instruc tions carefully as you complete the questionnaire. To ensure uniformity in the reporting of information, please use the following definitions Askarel Transformer: Any transformer designed and built with an insulation system using askarel fluid PCB Transformer: Any mineral oil transformer that contains 500 ppa of PCB or greater Distribution System Transformer: Any transformer not located in a substation or generating facility Transformer Fire Incident A fire in which, a transformer has been directly exposed to a fire, either from within the transformer or an immediately adjacent source. If the transformer was not fire damaged, then It should be considered a transformer fire incident only if there were visible signs of fire exposure to the inside of the transformer compartment or the extenor of the casing. -3HGNS 213927 I. general data This portion of .the questionnaire ls intended to obtain general mfonnaticn necessary to respond to the ANPR. Where information is not readily available please provide the most complete response possible within the tame hums of the study, 1. Indicate the number of askarel transformers owned by your utility as of March 31, 1984 Size--KVA (namelate rating) Distnbuuon System Substation Askarels______Askar els Generation Facility Askarels Total Askarels rtr-5oo ~ 501-1,000 vmi-s.m 2 If available, indicate the number of transformers containing1 askarel substitutes (i.e , silicon, RTemp, etc.) owned by your utility as of March 31, 1934 Size--KVA (aame- Distribution System Substitutes Substation Substitutes Generation Facility Substitutes Total Substitutes l'craro-- 50i-i', oer i:boi-5.bw >'i,ooo -- 3 Have you conducted a risk-assessment inspection of your askarel transformers' No Yes If yes, briefly describe actions taken or planned to reduce fire risk (attach additional sheet uf necessary)_ 4 Do you now have an askarel transformer phase-out program in effect0 No Yes ITTes, briefly "Hescribe the pheie out methodology, i e , pnonty of removal etc.; * 5 6 Whet year do you anticipate completion of the phase-out program0 5. Have fire departments in your service area been notified of the location of your askarel transformers? ____ No ____ Ye* 6. Are there any warning signs for firefighters posted in the approaches to askarel transformers? ____ No ____ Yes -4 HONS 213928 13 If known, give the approbate mxe Laps* from, a. the beginning of the fire to transformer de-energization b occurrence to fire deparmnent response c occurrence to extinguishment nrs hrs j.ls (indicate tenthsTf hour J H Approximately how old was the transformer involved yrs 15 Did the switch, breaker, or fuse on the low voltage side of the transformer open as designed'' __ no ___ yes __ no switch/breaksr/fuse on the lowvoltageside 15 Did the switch, breaker, or fuse on the high voltage side of the transformer open as designed? __ no __ yes __ no switch/breaker/fuse on the highvoltage side i < Was there any early warning device (alarm or sensor)' no yes If yes, what type? sra it function as designed? __ no ___ yes ia For incidents occurring inside a building, characterise the extent of smoke travel from me fire, __ smoke was restricted to the transformer enclosure (e g , vau,: or rcc- __ smoke escaped the transformer enclosure and entered immediately adjacent areas, but did not enter enure building ' __ smoke entered the entire building other; describe _______________________________ 19 For incidents occurring outside a building, characterize the extent of smoke travel from the fire: __ smoke was restricted to the transformerenclosure (eg, vault or room __ smoke escaped the transformer enclosure, but did notenter any public buildings __ smoke entered nearby public buildings __ other; describe 20 Excluding fire fighters and utility personnel, approximately how many people were exposed to smoke/soot from the fire' ___ people 21. Approximately how many firefighters and utility personnel were exposed to smoke/soot from the fire? ____firefighters/uulity personnel 22. Indicate the results of any testing' PCBs PCDDs Before After before After Cleanup Cleanup Cleanup Cleanup Trf. Fluid Soot PCDFs Before After Cleanup CleanuD 6 HONS 213929 II FIRE INCIDENT DATA "1 Tv-s Doraon of the questionnaire is designed to obtain data on ALL FIRE INCIDT\'~i INVOLVING ASKAREL TRANSFORMERS. PCS TRANSFORMERS OTTTiUNSrORM* = 1 " eOSTAI.VH.Vd AoiiA^kl^a-JSSTlTu-fSa' Sl'BKTfi TH-k'TAaT 3'VSaRS (aBRIL I 1 j3I->IAPCH 31. 15T7 A fire incident is defined as a fire m wtucn a transform? r has been direcuy exposed to a fire, either from within the transformer or an imaied.ately adjacent source If the transformer was not fire damaged, then it should only be reported if there were visible signs of fire exposure to the mside of the transformer compartment or the extenor of the transformer easing Complete ques tions 7-25 for EACH fire incident, Make additional copies of this fora as necessary 7 Indicate the si2e and type of transformer involved in the incident ___ KV* __ askarel ___ PCS (>500 ppm) ____ askarel suostitute (highTFs pen*' ITTskarel substitute, specify fluid type_____________________________ 8 Indicate the approximate time, day of the week, and date of the incident time day (Mon Tue etc ) month/day/year 9 Location of incident __ Inside building __ Residential (condo, apt., hotel) Institution (school, hospital, church) ~ Office __ Industrial/manufacturinf __ Storage (warehouse) __ Other public (store, mall) Ocner, specify _____________________ __ Sidewalk vault __ Other; specify __________________________ 10 If the incident occurred inside a building, indicate' age of building years unknown, can't determine location within building: basement roof other, specify____________ c. no. of floors in the building' ___ floors d environment of transformer; vault room open area other; specify 11. Cause of fire; ___ within transformer ___ outside the transformer; specify if known__________________________________ 12. Was transformer fluid released and exposed to the fire as result of damage to the transformer, a leak, or rupture? __ no ___ yes If yes, approximately how much fluid was exposed- __ gals (Indicate tenths of gallon if <1 gal.) -5MGNS 213930 23 3nefly describe the area cleaned and method of cleanup, i e , nail, fleer ceding' area, HVAC system, etc (attach additional sheets if necessary) Area Method of Cleanup 24. In the table below indicate approximate cleanup costs. When estimating cleanup costs be sure to consider manpower, equipment, transportation and any relevant overhead charges: Cost Type Cost Estimate Analysis/test costs: __ __ Site cleanup/decontamination: Disposal Vault/transformer/equip costs' --_______ _ ______________ Other (descr.be) TOTAL. 25 Have there been any lawsuits filed against your company as a result of this incident' ___ no ___ yes If yes, briefly describe nature and amount of damages claimed 7 HONS 213931 III. ASKAREL TRANSFORMER DATA In responding to te EPA data requirements, it is necessary to develop informa tion regarding the location and environment of askarel transformers UNLIKE twf FIBE INCIDENT PORTION OF THIS SURVEY, THIS PORTION RECTCTETT THAT YOU EXAMINE'AND KSFC ftTStnT SKaEL'.SCI ENTITYALLY SELECTED SAMPLE OV YOUff'BfHT'ftiaL'TIOirSYSTLM AsKaREL TRANSFORMERS, i e ASHA REL TRANSFORMER'S NOT LOCATED IK SUBSTATION5 OR GENERATING FACILITIES' Based on data provided in the 1981 PCS-survey, w have identifier the specific askarel transformers on which you are to report Please follow these instructions in identifying' those units- Prepare a list of all your askarel transfonners that are not located -n substations or generating facilities- This list can be m any sequence, i e., by serial number, locaaon"code, etc. Beginning with the transformer on the list, select every transformer until yoTTTiave identified a total sample of _____ transformers (When you reach the end of the list, go back to the beginning and continue your count until you identify the required number of transformers ) Proper selection of these transformer* is critical to the study If you have questions or problems in performing this task, call Rick Bell or Tom Florence (RPC) at (202) 797-1111 Once you have identified the specific transformers for which data is required, answer questions 25-40 for each transformer sencted Note that if your sample requires data on ter. transformers, you must complete ten sets of questions 26-40 Sufficient copies of this form have been included for your sample -8- MONS 213932 Answer the following questions for each transformer selected -n your sa.T-?i 26 Of the ___'transformers m your sample, where m the order was this transformer Selected (1st selected, 2nd selected, etc ) Selected 2? Indicate the nameplate rating of the transformer28 Indicate the askarel fluid capacity: KVA ________ 29. Specify the transformer location. __ Inside building __ Residential (condo, apt., hotel) Institution (school, hospital, church) __ Office __ Industnal/manufacrunjiff __ Storage (warehouse) __ Other public (store, mall) Other; specify ________________________ __ Sidewalk vault - ~ --' Other, specify ______________________________ _____________ 30 If transformer is inside building, indicate1 a. aye of building' years __ unknown, can't determine b. location within building' ___ basement __ roof __ other; specify___________ c no of floors in the building' ___ floors d environment m which located.____ vault ___ room __ open area ___ other; specify__________ 31 Check the response which applies to the transformer drainage path __ ' fluid would be contained, i.e., no drain, or dram has been sealed __ fluid would drain to contained catchmentarea __ fluid would drain to sewer system __ fluid would drain to surrounding soil other, specify: 32. Are vault drains required by local building codes? ____ no ___ yes 33. Check the response which applies to the transformer cooling ventilation system: __ no ventilation; transformer in sealed room or vault __ ventilation through grate in wall/ceiling to outside area __ ventilation through grate in wall/ceiling to indoor area other; describe:' __________________ _ 34 Is the transformer cooling ventilation system forced or natural? natural ___forced ___ none If forced, is it designed to stop in the event of a fire9 ___ no __ yes -9' HONS 213933 35, If tJus transformer is located inside a building, characterize the poten tial spread of^smoke and soot in the event of fire (answer a-d i a, Is there an air conditioning or heating system intake inside the transformer enclosure (e.g1 , vault or room uhere transformer is located;' ____no _____yes _____ transformer not enclosed b. Is there an air conditioning or heating system intake m the area immedi ately adjacent to the transformer enclosure (e g., vault or room)'' ____ no ___ yes ____ transformer not enclosed c Other than a yes answer to 35a or 35b above, would snoke/soot have other indirect access to the building air conditioning or heating system ^__ no __ yes TT"yes, descnbe access ________________________ ________ ______ ___ d Would smoke/soot have direct access to the air conditioning or heating system of any other building'' ___ no ___ yes If yes, describe access.__________________________ _________ 36 If this transformer is located outside a building, characterize the poten tial spread of smoke and soot in the event of are (answer a & b) a. Is there a building air conditioning or heating system intake immediate.;' adjacent to the transformer enclosure' ____ no ____ yes b Other than a yea answer to 36a, would smoke/soot have any other access to the air conditioning or heating system of nearby buildings'1 no ___ yes If yes, describe access: 3?, Is the transformer protected on the low voltage side by a fuse, $wi::h. or breaker? __ no yes , If yes, is the device oil-filled? ____ no __yes 38. Is the transformer protected on thehigh voltage side by afuse, breaker, or switch at the transformer location (exclude protection at the substation)'' ___ no yes If yes, is the device oil*filled? ____ no __ yes 39. Is there a smoke or fire alarm orother earlywarning deviceor sensor near the transformer? __ no __ yes If yes, describe the device_____________________________ 40. Is the transformer protected by s sprinkler system or other fire retardant system? ___ no system ___ sprinkler system ___ other system; descnbe ___ -10HONS 213934 IV RETROFILL DATA This portion of tq,e questionnaire is designed to obtain data on each askar-i transformer you have attempted to retrofill To the extent data are available, please complete questions 41 and 42 for each transformer Make additional copies of thus form is necessary 41 Indicate the number of askarel transformers your utility has retrofiiied __ _ transformers 42. The table below provides space for recording1 data on two transformers Please record the requested information for each transformer your utility has retrofiiied (or is retrofilimg) If space is required tor more than two transformers, please make additional copies of this fora When reporting costs please include all associated costs, inciui_-.g Labor, materials, disposal, and any additional fire protection costs result ing from use of the new fluid. If additional fire protection was required, please describe and provide specific cost estimate m the comments section a Transformer KVA (nameplate rating) Transformer 1 Transformer 2 b Fluid capacity c, Beginning PC3 level (pnor to retrofill) d Retrofit process (flush/refill or filtering) e Retrofill fluid f. Flush solvent PCS level after 1st refill or 1st cartridge change if filtering process) h. PCB level after 2nd refill (or 2nd cartridge change if filtering process) l. PCB level after 3rd refill (or 3rd cartridge change if filtering process) J- Time elapsed from initial draining to last refill (or cartridge change if filtering) (months) k. Approximate costs to date (consider labor, materials, equipment, transporta tion, overhead) Comments* -11- HONS 213935 Technical Newsletter Electrical Systems Division ?C3 TECHNICAL R?OAT ?SCi! E?3 NU2?, 8 i i , k n : > : s * s * *: n *ii' CH 5c-1 * ; * i *4 z * * I-5C3 sr iii HONS 213936 2028-3 1263-14 2028-4 2029-5 to 10 1 25 3- 1: e?^ ; ?DC? PROJECTS Re-oval of PCS* from Transformer Oil PCS Disposal Manual (2nd edition) Pyrolysis and combustion products of PCB, tri-, and tatra-chiorinated benzene CCMPlETIC'J _ cate ' June 83 Sapce-ber 1334 November .3:4 Analysis of PCDfs in insulating fluids State-of-the-Art Review: PCDDs and PCDFs m Utility PCS Fluid (CC 3308) Marcn 1935 Complete Decemoer 1933 mit:sat:c:( 1263-20 Documentation of building clean-up teenniques 31rgnamton and San Francisco; 1 26 3- 2 1 Eng.near mg options for PC3 transformer risk reduction r:sx assessment and management 1926-13 Exposure assessment for PCS spills and fires 1926-15 Riaic mangement of PCB fires and spills 2595 Decision framework for PCS decisions HEALTH EFFECTS 2374 Health effects assessment of PCBs SUBSTITUTE IHSULATIOS 1499-4,-5 Arc and spark byproducts of perchloroethylene 2028-11 Partial and complete combustion of transformer insulating materials 2028-12 (pending) 5tate-of-the-Art Toxicological Reviews of PCB Substitutes Octocer 1394 Octcoer 1544 Decemoer 1584 Decemoer 1334 July 1994 mid year 1985 August 1934 December 1994 December 1984 MONS 213937 Ti* Ie t Contracts: * Removal of ?C3s from Transferor Oil RP2J23 - 3 Cor tractors: Peter Way (-3 > Project Manager: Gil Addis Purpose t The objective is to develop and deronstrate cn a -i.;; scale a process capable of reducing ?CB contamination in transformer* below the tr.en EPA acceptaoil. ty level of *j -zr Scope of WorX: In an earlier project, RP2029-1, Ge-eral Electric, after testing m the laboratory cnose the solvent extraction process for scale-up. Other processes tested -ere: sodium reaction, not selected because significant cc-rrer;;a',, activitiy was under way; eiectron-beam irradiation; and cr.t.cal-f luid. The latter two were found to ce .-practical t*e laboratory. A relatively safe and economical solvent extract;:-, process was successfully scaled up to a 10 gallon per hour plant, and plans made to design a full scale plant (RP2025-3I. xoected Results; Successful operation of a large-scale solvent extraction plant Cor removal of PCSs from contaminated n.rera. oil would provide an alternative, economical means for decor.ta-.*.a Status A 500,000 gal/year plant has been designed and costed, The proposed host utility for this plant has completed site plans. The project is presently on hold to evaluate the E?A`s recent interpretation requiring a level of 2 ppm PCS in the treated mineral oil. Completion Date; 1985 / HONS Z13938 T.tle: PC3 DisposalManual, Second ;di:i0n Contract! * RP1293-14 Contractor: SCS Engineers Project Manager' Ralph Komai Purpose: This project willupdate the reference manual prepared m 1379 as a guide for utilities in evaluating disposal ootiors Scope of Wpr'c The Contractor will prepare discussions of tna available disposal requirements and disposal options, incl-d.r.g tne chertiosi treatment technologies which have beeen developed since the first edition. Chemical treatment will include bctn destruction and retrofill. There will be an update of the estimated PC3 inventory. Management of ?C3s also will ee addressed Sxoected Results: A revised manual will be available to the utility industry to assist in evaluating and planning management options. status: A draft report has been prepared and is in the process of being revised. The second draft will be distrtouted to selected industry people for technical review prior to finalization and publication. Completion date: September 1994 HONS 213939 l Tit'* m Pyrolysis ini Cc-b.st'tn of ?C2 f> )<<> i Corfact Contractor- P?4:2S-1 New for State Cepartrant of -*a'th Project "anager: Si 1 Adi's r*J Sccce c* *c-<: The cbsf/ation tnst PCBs -rder coroifons of oyro'ys's zr ecoustion tan be conve'-teo to twh more to*1,c polye-ltr.natsd dice':;'jrj^s (PC"s) n.j led :: -a'o,-:e"ed interest in the possibility of ref of ' 1'1 *,g c."5"t'y lo-.sa P'13 firi'tr-e-s with a fe'erfi; .n'lte'y to s.ch f >rf`:r-a:ions. hoe.er, a,an tna rest care'.,; rf:r'if ;*; prccesi w, 1 s; a 'esic.il exeunt of s33 in the new d'electri: 'Uid. A tie'-* is no reasor to enpec: that suen low concentrations of ?C5s *cda prcdjc* d's;r::;rv atj'y high concentrations Of PCDfs on pyrolysis or C0~CLSt':n. -Cweve*-, S'rca fe details of the mechanisms by which trrs conversion occurs are not we 1 estabiis-ed, it is dearly desirable to test this poss'tiluy e <;er'-e'--.r ly before launching a major retrofitting program. In a secono aspect o' f`s o' the level of formation of PCDFs from mineral oil ccntar'ratei with s!3s at a ' level i* of even more Importance because of the large number of :snta"irj:jj transforr,ers involved. Expected Pesults: We expect to determine the linearity of formation of PC" a a pyrolysis or combustion product with decreasing PC3 concentration In a he:; HONS 213940 liquid, eil^e* "i"eral oil cr m 1! fluid. St atus pyrolysis and careustion teciniq.es rav* teen ceveloced >i~Z i clean-uo -e fiois, particularly for pyrolyred :r coopusted cortatnirat*c oil, are nearing success. Analytics! results will fee a/sllaple in t'e ^.r future. '(cverper 192- HONS 2139^1 Ti:`, a: ` opciol f Eaqru.dipmPCeDnDt in Aslcjrel arc Conta-.iatad Contracts: Con,,tracto. s. RP2028-S,-10 rJaJd^iaann((l-7511(, GBear<teterallleEl(e-6ct)r,icNe(w-8*1o,rkIITSRtaIte(-r5fe*a. , .Research Triangle Institute (-101 Project Manager; Gil Addis Purees*; To improve t.ne state of the art oi ar.al/sis PCCfa Td determine under what conditions, if an/, PCDFs and reistet rtnpc-nds are found .n utility equipment. Sccse ct work; Better analytical separation between 2,3,",t,< PCD? a-d ctner less toxic congeners will be made by syr.c.-ea i:.ne a-i at.ting wit.n added 13G 2,3,',3,X corceners. ?CDr conte-t cf -t.lity equipment under various operating and con.camirar. r* srerarics will be determined. Laooratory arcing and corona tests will be made on askarel ar.d contaminated samples. Expected Results; Improved analytical protocol for separating 2,3,7,8,X PCDF and other congeners. Determination, of PCD? content of a rang* of utility equipment. Statust 13G spiked samples for analysis will be ready around July 1, 1984. A sampling plan for utility equipment has been prepared. Sample collection will start shortly. Completion Date; March 1985 HONS 213942 Con trac;: Contractor Project Manager: State-of-t,Ke-Ar: Review; ?U;rs and -CCPs in Utility PC3 fluid RP 1263-11 5CS Engineers Ralph KoT.ai Puroo3e: To examine and surmarire the state of the art of knowledge regarding PCCC3, PCCTs and other cyclic nydrocarsors r util-ty ?C3 fluids. The study also is to determine wnat conclusions could be made about formation conditions and to def.ne gaps in experimental data. Sccoe o: Work: See Purpose Expected Results: Compilation and critical analysis of information available in the literature. Status: Complete ICC 3308) Completion Date: December 1983 HONS 213943 6 Case Studies of ?C3 ?ransfor~er Fires. San Francisco, California, ana 3ingnarton, Mew York Contract; Ftp 1263 -- 20 Contractor.- IT Corporation/W. Corey Trench Project Manager: Ralph Kcmai Purccse: (l) to document re.Tedial clean-up techniques wmc.n nave been successfully and unsuccessfully applied to as<a:e. fire can tat i nation, (21 to generally document the technical ar,d tec.s.r*- related factors arising front an askarel fire, and (3) to define areas of future research relating to deconta.Tiration, Score of wor't: The project stay De summarized as collect.cn a-c re/iew of written documentation, development of an inter', .e- q,est .onnaire, interview of key personnel involved with clear issues at the two fires, data organization and review, and report preparation. Expected Results: Summary case studies will oe published wr.. provide utilities with information on what clean-up tecnmcue have been successfully used, the levels of allowable resid-al which were established for the respective buildings, hew tnev were arrived at, if understood, the basis of the incident, ar operational changes which have been instituted to reduce the impact of future incidents. Status: Contract has been sent for execution. Completion Date; October 1984, HONS 213944 1 Oes.gn Manual for P.ed.ctir- c: *.rcor-e PC3 Contamination Contract.: RP126 3-2 l Contractor West and 3ansen/TEI/Fenne Project Manager' Ralph Komai Purpose: This project will develop a manual to ass.st utility personnel m identifying contamination ca.ses ar.d gateways. select.cn of risx reduction tec.n.n .ques, a-d preparat.cn c lsclatic.n designs. Scope of Wcr'c The development of this manual will proceed through a sz-dy o historical airborne contamination, tn.e pat.nways ./irn nigh voltage operat.cn, des.g' :ecnr.*q-es f aircor-e isolation, cost factors and cost estimates, sensing systems, and the actual preparation and review of the manual. Expected Results: A manual will ce T3de available to the utility industry to aid them in planning risk red-ction engineer: measures for PCS transformers. Statusi contract has bean sent for execution. Completion Datet October 1984. MONS 213945 t Title! Exposure Assessment Methodologies for PC3 Spills and Tires Contract: AP1826-13 Contractor: Anderson Nichols West Project Manager; Abe Silvers Purpose! To develop methodologies to estimate exposure lev*i.s in soil, air ar.d water of PCB from PCB spills and to estimate exposure levels of firs products from utility transfor-ers containing PCB involved m fires. Scooe of Work; A fate model will be developed to estiTated exposure levels of PCBs from PCS transformer spills. A corpsrtmented model will be developed to estimate PCB fire products exposure levels. Expected Results: Code and documentation to estimate exposure levels. Status: In progress. Completion Patti December 1984 for draft of final report. HONS 213946 ? ' ;e A AlSK iSSeSS'en* vodal fQr "s :,rf irz ^r;;^::s m ilsot-icil ' " .............. ' *' Con: net a5:-:5-l; Cont ra;-.:r 5esourts Planning Corooraticn project `'mate". Ata Silve-s c .'-r:sa "o estrrate tne preSaoflity of a utility tra^'or-er :e-r in/oi/eO 11 a fme m c:r~er:i a 1, residential and Ous't; itr.ctj-ss. Store c* -P'v. In*or-e:icn -ill Se developed *rcn surveys to sst'-ita ;-a ;r:;>;r-:/ a t-aosfpr-er ;ir; involved m a fir*. I f. .5 rfS- ~ ts A final ^ecor: mti :rr':r-a:1:n on -.** -'s* :' j , S'itus In ;r:-;r*ss, Co-;'ey or rate- Cecerber 1984 for drift of 'inal reoort. HONS 213947 JO Title: ;;c:jicn : rirewcr'tj for i.ia-s;:': -e ti ?C3 Decisions Project no.- ."PZ555 Contractor. Project Manager: lirilie.i "ecus, Inc.: C-sn 3c.J and Ca/:d Denar are the principal investigators. t. Victor Hiemeyer Purpose: To develop ar.d test tools that utilities, ird-str/ gro-ps, s-t ot-er interested parties ran use to ti.r its.tots aooct ?C3 nst -anage-ent issues. Sc toe of wort: An analytical framework .s bei.-g developed :.-a*. car, trace t.te ,-eilth s.-d environmental effecta ste-rung fre- al:e:-at;ve ?C3 ris.< management policies. The fra.-ewert, i-p.a i.t tre fern of a computer model, li.t<s equipment managerne-t *.tn tne :.-rer of acc.certal releases, the t-~ter :: pecple i- different rategcries I'sucn as office verxers, ut___t. personnel, emergency workers) exposed to ?C3s and cents-.rates (such as TCTFs, TCCDs 1, to estimate tne expected "umber of healer, effects. It also estimates the costs cf the policy alt* so that tne extra health benefits can ce compared tc t.te extra costs. Parameters that are uncertain can be treated -t e.tner sensitivity analysis or with a full probabilistic treatment. xoeeted P.esults; A complete, documented and tested computer model that can be used in analysis of risk management aiterr.at. for electric utility PCB-containing equipment. Status-. The computer code implementing the framework has been implemented m prototype form. Completion Date: The model will be available in final ferm later in 1984. The prototype of the model is available .new wit the assistance of the developing contractor. // HONS 213948 Health Effects Assessment of PCis j-p C o n t am i r. a n t s Contract; RP 2374 Contractor: Clement Associates, Arlington, va Project Manacer: W. Weyze.n PurposeA comprehensive assessrent of tne potential nealt.n r.3k associated with cecupations 1 exposure to ?C3s and cc-ta~inancs, based cn available informat lc.n. Sccce cf Work; There are three parts to the study. Part ere deals with extracting from the literature data cn human and amrals on health effects associated with exposures. Part two involves characterization of the utility worit environment. Part tnree if. olves extrapolation of the health risks for persons exposed to PC5s and ccnta.Tir.ar.ts, for selected er.dgc.nts. Expected Results*. See objective. Status.- Part 1 has been completed. Part 2 is delayed b-t wi.l be completed in about one month. Part 3 depends on information collected m 1 and 2 has not started. Completion Data: Midyear 1985 HONS 213949 Title: . ire i Spar< 3y-prcd-cts c: CidA Contracts: 3_? L 4 9 9-4 , - 5 Contractors: Sprirgborn Laos. westmghouse Pro:ect Manager: Gil Add:3 Purpose; To determine arc by-products of psrcn ioroeth/i *r t f..led trar.sror-ers . = ccce u: Equipment, proced-res, a.-.d associated rat.cna.e nave bean ds-.eieped at Wes t irg.ncuse fer the stuay of arced prod,, of C2C1 ^ free tba point of view cf toxicity nazards. Spnngocrn will rake use of the equipment a-d techmquea to arc and to analyze the arc products from various cox.oi nations of C2CI;, electrodes and scl.d insulation -atena.. ''easurexent cf art *ne:c-; is 0: tv iced. Expected Pesults: Levels of potential 1; :cxic arc oroducts generated under varying operating conditions and using a r._-ser of materials combinations will oe determined. Amc.ng t.-.e arc products are CI2* HC1# ar.d CCC12Ot.ner products, such as carbc.n tetrachloride, hexachloroethar.e, carbon monoxide ar.d carbon dioxide, will a.ic be included in the analyses. statusi Arcing trials are oemg run on a regular basis a*d analyses proceeding accordingly. Most results will be ready by late June 1984. Completion Pate: August 1984 / j MQNS 213950 Ti;!e: Partial and Corslet# Cc.-buSti :n o' Trans'cr-e>- '.rs ,, Material* Con tract: . Contractor' RP2C23-11 hestlrighouse Project "snags'*. Gil Add'* s Puroese To t* amt re the products formed from either corolete or pa*-:'.!! co";.stion of liquid dialectics other than ?C3. Score t* Wt-y: The contractor wl11 Initially ccrtfjct a theoretita' e>nar-.c evaluation of coroustton raacticrs dtiii'v; a/ailap*e data "'-ese ca'cj'afons will tnen fce *xperimentally verified at selected <ey points ar: thus provioe a continuum of Information acpll cable to all slt-aicrs. The lnvesrt;att:n will include both theoretical and experimental effects at vanec combustion temperatures end varied oxygen levels. Tne ccmbust'on reac:i:*s cf insulating materials will be performed in the charter o' a tre--o-gra/i-ev": cell, which will allow monitoring of weight loss, rate of rise of tr'es-atj-e, reaction conditions, and Oxygen content. The combustion products will be tracre: for subsequent analyses or directed Into a detection device for on-str*an i's'.-sas The laboratory studies will use whatever sophisticated analytical techniques a--* appropriate (gas chromatography, mass spectroscopy, Infrared spectroscopy, etc.), to determine qualitatively and quantitatively each type of significant combustion product. /V HONS 213951 c^cectti L.-orcved knew1*.;;* :* :*e :*rva' ini c; products of Hcuid dielectrics. Status . Stirting Cc~pleti:n 2ett. February 1925 /i HONS 213952 It I i. Contract Contractorprcjec: 5titi-0f`the-Arc Tcx1;c';gi :ol s^vis* 3' 5C2 S -or. -..us s?::25-:2 (pe-di-g; SC; engineers y: Aadit Purpose To critically review the 11 texture on toxicology of PCS substitutes TVs m*or-ac on `5 -es:*i to avoid, sec* sure in trie *jture. a situation i'-' ir to re c^r-a": 3C3 fasetut. Sco:e 'i* -'fc '.1 peratorj search on toxicology sf PCS substitutes inc'um tyo'/s'.s 3": crt-stlon croducts. 3<:e:ted 'ii/'.s suastitjtes. s-mc*l review of :he toxicology 11 t-atjr 0' oc3 Sta:.s: Cort-stt being orepared. Cc-t'arc- Tate1 December 198* HONS 213953