Document MJV4Y6LOKZzwv465No9V57NDa

Mk TO: FROM: RE: DATE: Dennis Schwehr Roberta Straub^ Damages July 6, 1993 MEMORANDUM Here is another document which I came across and that may have implications for the damages calculations. I don't know if you've seen it or if you already have a copy. I realize this is not the best copy, but our original is no better. If you do already have a copy, please disregard this. I would be interested to know your thoughts on whether you believe this document is important or useful (or not) after you have read it. Thank you. ijs\nvpower\damagc3.mm9 i L -*JJ^ c-f-^ : .- 1-'' y' i SubmittsJ by: MAT IOMA L ELEC TE I*-*AL MANUFACTURERS ASS.V ATTO: 2101 L stroot, N.v;. Viashing ton, lj.c . 200 37 \ tm H^.i.ra m .l^ ^ y ranaB1Fy ffvtT.T NPC00036097 ? < i > 4 'i 1 * .-3<1 Ah -i- JINTRODUCTION Ths demise of the use of PCBs in transformers has resulted in the need for the industry to develop substitutes. Because r ijv? i that flammability is one of the primary criteria Cm 1he acceptance of these substitutesr studies of the flammability iavo been initiated by the industry. The studies have beer. ;.o_mci to be more complex than anticipated and extensive addiiional investigations will be required before fire safe trans formers can be approved using appropriate materials. The purpos ed this proposal is to provide short-range screening data to develop a basis for subsequent in depth studies that will lead io industry standards that v:ill protect the public and lrnit property damage and personal injury. 1 T- conduct ar. evaluation ci trar.sf armors co uibstitutes to os trolls/, fire ,,-aiecy s '.a.-ln'-fv -*.? >-rin and approved Les to. ;1- fbOGR.aM To conduct an ox"irr: .vuitai urov"-*'; t :* \ i ~ *" ubs Litu to u w Lt:h respec t to i:1amnab iJ.i1:y and c.;/ i y cor' L.:t '..*ith forthcoming standards proposed by Factory Mutual. Tn addi tion, catastrophic tranr-former failures will bo simulated and related to fluid proper tie a , gas content, of fluids, and trans former operating temperature. A few representative trans-formers will be "exploded" to confirm fire safety requirements. Obviously, additional stud.i es will bo required in order to achieve the NPC00036098 763360 dk -2overall purpose of the program as stated above. The DOEJ rprogram would provide the technical basis for the subsequent transformer studies. ` iwP.WK hACKGROUND For more than 40 years the transformer industry used Asha:-. Is ' . p o l y c h l o r i n a t e d biphenyls, PCBs) as an insulating and heat trans fer liquid in equipment that was required to be fire safe. Sine.: the early 1970's, environmental concerns mounted concerning the use of PCBs and ultimately the Toxic Substances Act mandated the cessation of manufacture and use of these-mater-ials. The search by industry for substitute materials ultimately led to the need 'or seme form of industry approval for fire safety, particularly when these substitute materials were used in transformers in inc-'O'- installr 11 or.s. In addition, further governmental regulations (" " eft M W 'JICGli, ) must be satis ricd. Since the industry is ro v : ith, ' <K 1i.jui:: :nculated trans fc-ra:ar t.'.at lias been shown to l:-e errs e*^pr*i.-.lEe, efforts and financial support, have been applied tt: an n on 1:a _ p roa ram c o nd u c t sd l. y r a: ` o l*y Mi; t u a L Ra sea r c h Co :;. o i:a r .;. :. irMRC) d iroc ted to the asca11is.::v.ent <:f shin darda for fire sai.e \ . he technical, difficultias ass .:ia ted vi th the qu?.nti t _rv.? -valuation of transformers insulated with substitute maicrials l.ava beer, much greater than anticipated and it is clear that- ex tensive additional experimental wort must be done before suit able standards may be established. The cost of full-scale transformer tests involving disruptive failure of the unit is very high.. The costs are prohibitive when NPC00036099 763361 ' .ne is employing the technique to qualify the fire safety of transformers incorporating PCD substitutes. We have found that `.*::plosive charges can be used to simulate transformer arc-faults v The purpose of this proposal is to relate the explosive disrup1:vu failure to arc-induced failure to he employed as a flamina- 5 t -ility criterion for transformers filled with PCB substitutes. PROGRAM ORGANIZATION It is proposed that the National Electrical Manufacturers Association (NEMA) wod_be the prime contractor. The program organization is shown on the following page. The Transformer Advisory Council, will provide transformer oriented inputs and guidance to NEMA on the overall program. Technical coordination of the subcontractors will be carried out Mr. H. E. Sheppard of Westinghouse and Dr. R. C. Osthcff of "encrai Electric. Mr. C. Miller of Factory Mutual Research Corporation and Mr, R. Gann of the National Bureau of Standard'ill act as technical consultants to 1JEMA on the overall rogrnr 763362 i w o\ o 763363 TECHNICAL PROPOSAL TASK I Analytical Studies Tho amount of energy involved in the arc-induced disruptive i til.ute transformers can be translated to equivalent explo: ives (in terms of total energy release and time duration). it. P. H. Barkan of Stanford University lias extensively studied ,;rc faults in distribution transformers.^ It is proposed to i-mploy Dr. Barkan to work with Mai La-Wright Corporation (Bailstc>r; Spa, ilev; York where the initial explosion studies were conducted; to relate electrical fault behavior in transformer tanks to that -f the u:;e of an explosive charge. Included in these studies will to the velocity of the fluid (and/or cover), the energy i.elease , the time duration, the pressure generated, and the tempratuiV' - -he fluids to evaluated. Although the initial -tudios will be concert traced on cylindrical distribution r.rr.nnurnicv -nn.k- v:ithout cores and coils, ir.athenstica1 cor ce1r.11.:.j ' i .0 rransfc -niers in re angular tank configurations '..ill 1* ili^elopea ;voa o -o.t in conjunction fh the efforts cf Dr. T. K. Alw.rwo si inghouse Transfer:.,er Medium Tower Division, Tlv^c studies will be mode to establish the relevancy cf these testing methods to transformer eredibility--e-f failure mod*. nd to support the use of explosives for the simulation of electric arc-induced catastrophic transformer failures. The (1) T. Ii. Barkan, B. L. Dansky, L. F. Ettlinger and E. 3% Kotsky, IEEE Transaction on Power Apparatus and Systems, Vol. PA-9':, No. 1, Jan.-Feb. 1976, p 17 et seq. NPC00036102 ^3364 similarity of. explosiva and electrical failures will be evaluated. Drs. Barkan and Alverson will recommend instrumentacin f :f; th" e-plo.sive tests, review the procedures, and aid in or rela Lion o f the data . TASK. IJ - Ianiticn Technique Investigation The large scale burn tests conducted at Factory Mutual / - ; ->-' ''` i-'.esearch Corporation (FMRC)i employed an aorular structure thar. contained n-heptane (approximately 1 inch of heptane on top of about 2 inches of water) for ignition of test liquids. "'he heptane was ignited and burned for about 15 minutes. The hepcane T-lamu heated the. fluid by both conduction and radiation until the fluid under test was ignited. .rtui-uitely, the heitane conti!Jscl to burn afte y t.vt *--Ft fiai*.i v.vij igr.ited, and this distorted ti:e heat re SUrtie-::ts. In addition, the physical act of adding s./, y easily have rose 11;d in contnxin.it.irr. of r. j lUid uni.,iS*--** t,v ^* Furthermore, once t'ne hepta ne was !qf it urnod l: cellular (also cal Ip:! polyhedr. disturbed /t kwv..-, -.... ** o T-r!1 his can. '-caul.t in `'distillation11 of the heytano into tho tes:: iquidf a.:ain causing possible errors in the hcat re1eusc mean- ;cements. It has\ been -observed by R. Gann 'oi: the Nation?.! , : jV" iiurcau oi S handsrdsKthat /*n improved iqnition technique should (2) G. 11. Markstein and L. M. Sommers, 4th International Symposium on Combustion, 1953, p 527 et seq. NPC00036103 763365 riifiniMWiair.'^tfi'irrfnnlW>IniatfimV if r11fI k -7- };*s developed,, e.g. an annular propane burner, annular electrical resistance heaters, etc. in order to eliminate the above- mentioned possible sources of error in the heat release measure- The improved ignition source will be designed and built i-/ NBS ; 1 / .v under , *, f the direction k1 / L r , / " i * ; * dg of, R. V ^ ** * A Gann. t ..<f ; / . i : y " -' : ,* * /** t 4'* -f ~* T7-.S?' Ill - Large-Scale Burns rf. - " * '*'*' '*.... It; is proposed that Factory Mutual Research Corporation tFMiiC) conduct a series of large-scale burn tests at the FMRC West Glncester, R.I. test site. These tests would employ 'the ignition technique developed by MBS and replicate burns would be conducted. The data developed in these tests would be utilized fo categorize the various fluids for use in indoor transfrrmers lor buildings with-various ceiling heights. In addition, an at- !.rmr-t v;1I ' be made to correlate the heat release data the *.arga-cvu\e burns with similar data obtained in the labor* Lory " -:tLa Z2 5t3 at FMRC. It is proposed that the following categories c iI>;id= v .aasure:; at the ViesL Glo-.:ester facility: 2 - silicones cr different flanunabil ities 3 - hydrocarbons of different flammabilities If is proposed that the heat--release-for each fluid bo-.*- IMilred-three-times, to determine the reproduoibiLity of the tost method. All tests will bo conducted with R. Gann of MbS provid ing consultation to FMRC personnel. ;'T7u Y- / c - ; Y ' V iv ..i< ' NPC00036104 763366 -8- TASK IV - Wright-Malta Tests j Tha actual tests involving explosives will be carried out; ( >t Wright-Malta Corporation (B;illston Spa, N .Y .} . Mr . Tod Butl'1', i `ho principal investigator joir the Malta site, has extensive xperinece with explosives, .In addition, the Wright-Malta torporation site is ideally situated and equipped to carry out ouch experiments. Extensive studies have been made of src-faul,t failures in transformers. The work of Barkan et al contaiTis experimenta 1 j :!ata which shows the basic energy release vs. tima characterise: cf arc-faults. Several concepts of arc fault simulation were considered; the most promising concept for simulation of the ncrg\ release was the propellant charge technique. The arc-fault energy release vs. time quickly de t---p-iinos i hac a lev; burning black powder propellant used conttcrcislv , y an ignition compound would be desirabl:;. Black powder i ..:c;u:..: \ tpy.v'r>1i-c inlly available in a largo nu;*.:our of -j;v.-r:-:y grades. .f:.: l si he \ .ates, and combustion tamper attires. Py proper blending :: i.:h j ;l*rules tIte aosired onerg j' re Lans? rat? can bo Jj preliminary studios, the charge was designed us ing the basic information contained in the Engineering Besign Handbook on Propellant Actuated Devices. Ciraph fr1 shows the appro;*: inate charge requirements for stco L p Ls ton actuators. UndtcT-t" <>;1 propellant actuatorSj'the heat loss to the transformerfluid is not clearly defined. The first powder charge tested was a 28 -j IF grade black powder. The combustion burning rate was slower NPC00036105 763367 . , /iu y -/ula '1(Zt u & t > * than desired in|the--water-tests. Therefore, a 50-50 mixture / of IF and 2F black powder was used. Barkan had taken fastax notion pictures of the top surface of the transformer oil duriuu *.i /i *yi'*1i`/ t */.-/ j..-/* T-ti, t t.\ /> "j-fauit failures*,'* The energy, release rate in the initial * xesrimont was slower than desired, so that the charge v/as t.iWL*L- *f 'Odifierl to -12.5 g. of 3F grade black powder. -Further measure ments of the pressure time curve in the initial phases of the proposed program may show that further charge modification may be necessary. Extensive data were available with an arc-fault failure npproxir.iitely 7" from the bottom of the 10 kVA transformer can. The propellant charge is located at the same location as the arc-fault. Fastax motion pictures of the transformer can -.over without the restraining band installed were used tc v-srif the L n e t z energy imparted tc the 1iquid. Graph -2 sh ; ;e:ie I area of Kinetic energy release that was des i*"*:;1 :*oti i [.;>etures will bo uaocl confirm that the proper nor .- i.s being imparted tn the transformer oil. nct i V.'.! hove shown that explosives may be used to ='..nul`tv : Induced faults with socompany.i ng cove r-blowing cn lu k A disti buti'jn transformer tank.?. In those cases where ignition of Mi oil occurred, the phonographic records showed a good analogy arc-induced cover-blowing experiments carried out by General Electric at its High Power Laboratory in Philadelphia. Th*> explosive charge simulated the arc energy and duration! in the Philndelph i.a tests. NPC00036106 763368  -10- Two other measurements of significance will be conducted during each transformer cover-blowing experiments. In each case, shallow (;a. 1" deep) discrete pans.-o-f combustible materials, -:-, /, r li l .g. hr m =founer oil, cardboard ,Vfete. will be placed in 5-fr-oh nerer.ienLs up to about 18 feet away from the transformer. Three rrays of such pans will be placed at 120 around the experimen1* 1 unit in order to .eliminate'the effects of low level winds. Such observations will provide an indication of capability of the. expelled dielectric to ignite adjacent combustibles. Further more, an array of temperature measurements above each tank will i=e made during the expulsion of the fluid. The array of temper ature measurements will be made in the region of 10-30 feet ?bove the tank in order to help correlate the temperature data needed *:; the FMRC heat release criteria. ::Iy very meager data exist concerning the fl.amr!.ility haractoristics of various dielectric iluids a fune (.heir t^nporatu res at the time of tank rapt;-,re*. The refc ;repose I:> utilivis ex ;lc;ivs t:> produce ccver-blowir. i: .ai_uree on 10 kVn tn containing the following fluids Trans former Oil - fot a corr.naris-':i standard j Hydrocarbon oils - e .q . FT Emp , a FAO, and a. bright stock 2 silicones oL' different viscosities, s.g. 50 vs and 100 cs. Each of these would be tested at initial liquid temper atures of 60UC , 100C, and 135C. In addition there are no data concerning the effects of the gases dissolved in the dielectric fluid (.that may be generated in the transformer) on NPC00036107 763369 & -11- LhG flammability. It is proposed two fluids be selected for testiny in which combustible gas is dissolved. We have chosen a gas composition consisting of 80S II2 , 10S C H ^ , and 101 ' hr- r>*n ihiy -representative oi gas cornoosiLicns that could be U .* -/ t /- * venerated from'jthetfluids in this test program. These- tests ill be run with fluid saturated with the gas mixture. An assessment of the change in flammability caused by the pressure f the gases will be made based upon these tests. Replication of the "blow-up" tests will be conducted on the fluid ? that yield significant information. At the present, the reproducibility of these tests has not been established. To this point in the proposed program only cylindrical tanks with the covers held cn by retaining rings will have been tested. [t is proposed to evaluate th:i effect of rsctano-ilnr i.an]; govv; ~ry with welded covers. The experiments wi 1\ 1rvolve fhe use of e-mail scale models in which the gcemetry will rs icprroro!:*:...rive of load conter transformer geometry. The sane -* onr.n as described above fur th cylindrical tanks 'will bo c i . - u d u c . Tour model tanks will be basted. ins'":! in the data and observations of the disruptive ;iilure of the cylindrical and rectangular tanks, as well as i i;eriictivo calculations, a selection of actual transformers (up Lo .cylindrical tanked units and up to 4 singledphase-padm.eunhed transformers) will bo failed-up explosive charges. -In- ^vv ti.Li* b f 1 T -Lbese^ases, the explosive chnrge/w&uld bo-^laced-in~an-artifici-*l r* / cavity in tlic hi-gtr-yol-fcaye winding*. Again, the same measuremomtp 'i>' "* described above would be included in the disruptive failure NPC00036108 763370 -12nxperiments. On--bhe-other* hand, 'should he analytical studies' provide the mathematical basis for including several 30 transformers, these would be substituted for the 10 units I1u<led o above. /'.gain, with the 30 transformers the same measurmeirients will be made as on the 10 fcVA can blow-up tests. NPC00036109 763371 J I, *,iJ- -SUb ITEMA TRi I -A-n--alyti_c- al Studies A. Single 0 50 fcVA DisL. i_____ B. Thre^0 5000 j kVA Load Center j II Ignition Technique , A. Method Dcvel. !____ E. Equip.Design | 1 C. Equip.Constr. J i III BH7;ns; A. FMRC Tests B. Data Analysis IV Wr ight-Ma 11a Tus tis A. Single-Phase j 10 kVA Tank ! Tests j B . Three-Phase Rectangular Load Center "odeIs Tests C. 10 c 20 Trans former' ! Lata Cor.::*'>lj- ' *:.".ion o 763372 rOHM.ER PCD REPLACEMENT STUDY JULY `.rjCUST I SEPTEMBER I Il IiI I i I i .1 I i 1 I I