Document DMM6pa7QJJL0wzavzq71QqyN

ABD00282039 CJ > l II A MEETING PEPCPT Part 1: Vinyl Chloride Peview S ,dy I. Purpose The purpose of the meeting was to discurs the lew it.dy oe \ r: coMf^cted by TRw with emphasis on input to tie fc' '<cwi~ cdflw*n: :f Existing and new air pollution control tec.-.r.p'c: es. Existing sources identified Curing tr.e c^gm*'. not regulated Oy the standard. rut Potential new sources of vinyl chloride e~' $ ` . Enforcement experience with the vir.yl cr'o"'-? '-CL-.-". Several other subjects were discussed as indicated :y t~e tf.ac'ec *<:>. 11. Place md Date U.S. Enviromenta1 Protection Agency Region V Air Enforcement Branch Federal SuiIding 230 S. Dearborn Street Chicago, 111. 60604 August 15, 1980 III. Attendees BruCe Varner - EPA Jim Bodamer - TRW Marge Cassidy - TRW ABD00282040 IV. Discussion There are seven vinyl chloride sources in Region v. Five of these sources are polyvinyl chloride (PVC) plants, one of which produces copolymer only and another produces a combinat ion o* cw^o y*''Cr 6 fi o homopolymer. The other three plants only produce r.cmopoiyme-. "he other two sources are PVC PSD Facilities with reactc* capacities greater than 50 gallons and less than 1,100 gallons, 'hey are cni_, subject to the sections named in 40 CFR Section 51.60(c). Following is a brief description of eacn source: Dow Chemical operates a ccpolvrer plant 'n : d 1 e" -, s"' car . Dow batch ids the coc: ^i approved a com: ~ec rei*r c^a i - 'c-a: to** ':ss sca'c: for the Cow plant. 7r? dpprc.?p T.etf,c;: :~i^re tat ,~e i i~ resin residual ar.a r?a-*tcr coering e". s t'; ** 3 zd '-c: e^cee sum of tne resin residual standard :cm re susce^-s 'C resin and 2.CC0 opm for C'spersion res`-^ and tne 20 z-zr -ea:: opening standard, anc tnat t"? -esin rec'dual zn-zz-z is exceeded. Cow -as demonstrated ~.~e carai'l't. to s:-'d r ccpo 'ye- s -: re^s `: r res ' n tp 1.1 sr~ an; r - e: ^ c t~ i ~e^ dispersion -es,n to 1,200 ppm :',2V o . F. Goocrith iterates to rVC plants in. re:*on -.-e r Henry, Illinois anc the ether in von La-e. '-e -en'-y and Avon l a e plants operate s^s oe^s i c j c^ d'-ctets lines for procucticn of home pc iy~e *. .or, Lar.e z^zi.za latex by :r,e dispersion process. Tr.e s-spens:;n s continodus 1 v str'ooec. 'his stripp-i^c `ecr'c1 h';_ de.e'ttet : 8. F. Coocricn operates at high tempera tores ,122-l'2-r. an.: short residence times. Seme suspension anc a'.', c i: :f'' ?r- resin grades ar.^zi oe continuously stripped reca-te : * *.~e * thermal instability. This nas teen a ccncerr p- -a-/ tc.rtfs the VC fltSHA? only requires an P.VCM level for tne t_,:e o* resin (i.e.. suspension, dispersion anc ot.ner) , net cne va'-'tu individua 1 graces. ABD00282041 v/:^ai!:^'.^rafi5-'e-,i'rea''tO',-a -'blowdown tank for steam sparging. Stripping.In the reactor only occurs at the Henry plant dispersion , ' Goodrich developed "steam-sweeo-ing" technology formatturning.the reactor-opening -rjndara when stripoing doesfriot occur in the reactor. The advantage to these two control techniques for reactor-opening ,vi' loss is that very little water is used, eliminating the neea for a large wastewater stripper. BorJen Chemical Company operates a-PVC plant in Illiopo-is, Illinois. The plant has an old and a new suspension rcmopclymer line with continuous'stripping technology, and an old dispersion r.H* ' and an old suspension copolymer line with batch strapping technology. Borden utilizes "steam piston11 technolocv for control of reactor opening emissions. This technique, as well as steam sweep., are considered to be as effective as water piston for reducing reactor-opening loss emissions. Borden has had some difficulty in attaining the 400 ppm residual VCM (R7CM) level for their suspension ccpclyner anc currently reaches this level'by taking a daily time-weighted average which includes suspension homopolymer production. General' Tire S Rubber {GTR Chemical) operates an older PVC suspension hcmopolymer plant in Ashtabula, Ohio. Batch 'stripping is,done external to the reactor and the plant just meets: the resin residual standard. This plant was used as a modeV during the original standard support study to develop -the reactor-opening standard. The plant uses the "water piston" method to purge the reactor prior to opening. The reactor is hydroulically filled with water, a vacuum is pulled '.'xyx- ABD00282042 h$ad. space-. are: exhausted^ .'water goes to a Was tewater method Is the -large amount '.generated'.: The.plant has experienced leakage of'VCM;-through.Valves.back into .the reactor prior to opening, but'this..Is not,inherent to water piston. This suspension ... .iplant.maynot.have '.been^..representative model on v/hich to 'oise-'we'standard, because several batches of suspension resin;can be produced between.openings, while dispersion resin plants must open each reactor after every batch. 6.; f.'.Goodrlch .operates a PVC R & 0 facility at the Avon Lake Technical Tenter that is adjacent to the Avon Lake plant. Diamond Shamrock operates a PVC R & D facility at the T. R. Evans Research Center In Concord, Ohio. These facilities are not subject to the resin residual and reactor opening standards. Control Technology, for Exhaust Gases Three of the five PVC plants in Region V have parallel back-up emission control systems that operate only when the*primary control is Inoperative. There are no secondary or in-series coi troi techniques being used. No carbon adsorption units are currently operating in Region V. At Henry and at Avon Lake, 8! F. Goodrich uses its own solvent absorption technology for primary control of exhaust gases, backed up by a Brule' incinerator. The incinerator and stack are lined with a ceramic material that allows the unit to reach 1600F within a couple of hours. A large surge tank is used to hold the exhaust g:es until the incinerator comes up to temperature. Since the incinerator operates less than five percent of the time, no scrubber is used for control of HC1. Exhaust gas from the Avon Lake Technical Center is controlled by the absorber and incinerator at the Avon Lake plant. Dow, Sorden, Diamond Shamrock, and General Tire all use incineration as their primary emission control. Dow experimented with a catalytic oxidation unit, but had problems caused by a lack of surge capacity. The batch polymerization process causes uneven flew of vinyl chorice to the incinerator, which results in temperature variations that sintereo, cr glazed, the catalyst. The current primary control is an oil and natural gas fired thermal incinerator. Dow has installed a surge tank to get uniform flows, and to avoid bypassing the incinerator wnen it is inoperative and thereby violating the exhaust gas standard. Dow plans to again attempt to operate .the catalytic oxidation unit. If they are successful, they plan to use.'it,'as .primary control, backed up by the thermal incinerator, since catalyticoxidation requires only 1/3 as much energy. Borden and Diamond Shamrock .have no back-up control system and General Tire uses a boiler for*,back-up control. Enforcement Experience ABD00282043 A"' *:? discussed: of v KiPSi':;'?Iji.^v , \. W. ./V - SKi; : .. ' sa - '*#m*fo* .i -._'V 5.>V'<'. v'?.' ;Y J&j/ % '.<: *1 jTvw' * .;' \* * . '' .* ' i'',) i'*** . .'. *>&' ?.r,, ?;*. : * ' * pressurized containment laining one reactor's at greater than i;the PVC reactor pressure, release but does not 'depends on mechanical :ia tin ry<\* j-.- ^ ...,v;... fry . f };, ABD00282044 known-,e* tjerna 1 .controls . ~ -Tia.tlorv reacts on ` and -Has' Charges; they could not v\Th'ese facts suggest .that something '- ;i^^^xi.s^;Tn7^the:'iprldces'^<h.efntstry which makes certain reactions 'ffio^^linerable -to the .nin-a.way reaction resulting in an efflefgency:\4'i s.charce. V : ;; .' t 'i* ' i. ' A.,common ..cause- o'f emergency releases is overcharging the reactor. The'r*e,4s.-apparently no 'reliable instrumentation that can be used.fqVj level indication.\ Dual metering of vinyl chloride and water ;can help prevent' overcharging. Additional suggestions for control of these releases include: Documentation pf the calculated discharge magnitude, not just a description of the method of calculation, should be required. Clarification that the emergency release standard applies to rupture disc discharges also is needed. Under current wording of the regulation, if a rupture disc, not in combination with a relief valve, ruptures and discharges, there is no reporting requirement. Rsactor opening standard. There needs to be a separate standard fcr reactors that are also being used for resin stripping, ft ere- 1s.no.provision in the standard for measuring and/cr calculating the emissions from.a reactor opening when the slurry is sitting In the reactor. This problem has been resolved .on a case-by-case basis in each Region. Reactor opening.records. "He records of all reactor openings Should be.maintained for a minimum of two years. The standard does hot.specify any time limit. This was an oversight during development of the standard. Od11y>astewater sampling. Oaiiy sampling of the inprocess wastewater following stripping should be required to ensure proper stripper operation and maintenance'. The standard only ABD00282045 -tt;ate<l * m further ^flUKSftV^Boirdfeh^tOr^ently gives a. weekly summary ^ .Bjt^afer-';ana1yil3;voluntar11y; It Is important 'to ^Xr^1^^n^'ny^i|mpTfn9^<lae' tirth*` large amounts of water being pUton method of reactor purge}. In the .development Of the vinyl chloride regulations, wastewater was not;considered to be a major source. However, compliance with the regulations has drastically increased the Quantity of Ihprocess wastewater discharged from many plants. .5 Semi-annual reporting. The semi-annual reporting requirement delays;any .enforcement action tnat may be taken. If there were more reporting of fewer data, enforcement would be easier and data more reliable. The minimum frequency should be quarterly. Also, there is no way to tell hew effectively exhaust gas controls are operating since only emission levels ?reater than 10 ppm are reported, but the levels may be much ess than this most of the time. 6, Testing methods. There is no performance specification for exhaust gas continuous emission monitors. Only an actual stack test shows*how reliably the continuous monitor is operating 7. Alternative continuous monitoring. There are no provisions in the standard for alternative continuous monitoring requirements. although alternatives (e.g. Incinerator operating parameters) seem justified for back-up exhaust gas controls. Other Suooestlons and Concerns The following other suggestions and concerns were discussed: For those PVC plants that have come into comcliance with the VC NESHAP, the current major sources of emissions from these plants are fugitive emissions and orocess units following the stripper (e.g., 20 to 40 tires as muen VCM allowed tc escape as compared to reactor opening losses and the 10 ppm level in exhaust gases). This calculation is from a PSC BAC7 analysis done by Region V, and is based on information in the standards support document. Borden Chemical is under the impression that RVCM losses from the drier' are equivalent to the 10 ppm standard for exhaust gases, and OAQPS had an ambient level in mind wnen they chose the 10 ppm standard. This point needs clarif: :ation. Transportation of VCM for PVC production is mainly ty railcar In Region V. Railroad strikes have occurred and sources may need emergency storage of VCM for stock-pilinc purposes at sites other than the PVC plant. Should tnere oe seme prov'sio-i for including such storage sites unoar the standard? ABD00282046 rtf*'- \ S'ftv-..' r'' ;; ,' g?V, :.'; fe.-, .;; IK>H|-^4. ^ "V ftjtf'S*.'*. famt- " ., V. ^ .../ . Msfe-T- .'Vv. - ; : -- |pV U- W;r' '' Blfii t -; ; - .V. ./ $&&?.,tK ' Hfe****^ ` R- & D Af . 'I Ci,^ m vl -ft v;