Document QgNQ9nzkQaJV1K5jpdy6rL0wE

To From INTEROFFICE MEMORANDUM Distribution M. R. Chmura Subject (LM*tin, Or|tniatln, or DMrtmnt| (Location, Organisation, or Oapartmaot) I note none o you were on the Distribution of the attached article. I have contacted Earl Primeau and Larry Allen to make certain they and their staff have the article and are aware of its implications. Calvert City has continuous O2 analyzers and monitors pH also. Escambia monitors O2 levels by periodic sampling. Larry Allen will investigate the desirability of continuous O2 measurement. Both plants use inhibitors. MRC/na Attachment M. R. Chmura RECEIVED MAY 08 1980 1320) AP00022480 (stscrfirfhoducti) INTEROFFICE MEMORANDUM MAR_ _3q C. NOVAK Subject Date 27 March 1930 Explosion in Huls PVC Plant ToDistribution________________________ _________ From____Nt FrilSPIastics.Techno!oov__ Distribution: {Uecatlon, Organization, or Oopartmtnt) {Location, Organization, or Department] p. p. Aguila t. w. A1 len J. J. Baliker P.L.T,, Brian A. F. Cantor R. E. Davis R. E. Gilbert A. E. Greene 6. D. He! ms R. E. Jones ' M. Langsam J. R. Lovett R. W. McGinnis J. H. McMakln^ J. C. Novak E. A. Primeau F. L. Riddle L. A. Schappell R. J. Siekkinen Please find attached a report on the vinyl peroxide explosion which occurred in Hills PVC plant December, 1978. This report is a pre print of a paper to be presented at the International PVC symposium In Cleveland. NF/nb Attachment ______________ N. Friis 1320) AP00022481 yinyl Chloride Peroxide Exp t, in A Vlny* *B. Tervlesch Chemische Kerke HUls AG, D-4370 Marl Introduction The formation of vinyl chloride peroxides and their effect on the polymerisation of vinyl chloride has variously been reported /1-47However, little has so far become known on the safety aspects assosiated with the formation of vinyl chloride peroxides in VC/PVC produc tion facilities /57. The paper reports on an explosion in the vinyl chloride recovery plant of Chemische Werke HOls AG which is thought to be due to a thermally induced decomposition of vinyl chloride peroxides. Description of the Recovery Plant and the Explosion Fig. 1 shows a simplified flow diagram of the plant erected in 1973 for. a capacity of 100 MT of recovered VCM/day. The unconverted VC obtained in the ga'seous state in various polymerization plants is accumulated in a gas holder, compressed to abt. 8 bar and 110C in a two-stage screw comp res sor with intermediate cooling and tEen condensed at temperatures of abt. 30C. aided during condensation for corrosion control. The condensed VC was scrubbed with water and then recycled to the polymerisation plant. The non-condensed gases were passed through a batch-type water-fed scrubbing tower and then passed to the VCM plant for further treatment. On the day before the explosion the safety valve of the gas scrubber was found to have a leak. Following this, the scrubber was blocked off and vented, the gas being delivered unscrubbed via a by-pass. The bottom of the tower was heated by steam at 3 bar to prevent freezing. The temperatures so obtainable inside the tower ranged from abt. 20 to '40C depending on conditions. On the day of the explosion it was originally intended to purge with nitrogen and then drain the water. However, before this work could be undertaken a severe explosion occurred in the plant on the morning of the 29th December, 1978. Within a few_seconds. the entire plant was on fire. The works fire brigade kept the fire from spreading to tee surrounding plants and, apart from this, allowed some 10 MT of VCM contained in the plant at the time of explosion to bum off in .a controlled manner. The facilities located upstream and downstream of the destroyed plant were blocked off immediately after the explosion, critical situations did not occur. Assisted by the fact that the plant was operated under remote control the explosion caused no injuries. On the other hand, the material damage was considerable. As shown in Fig. 2, the outdoor part of the installation was completely destroyed. The total material damage was in the order, Tnemittida of 2 million DM. Reconstruction of Cause' of Explosion .. * Investigations initiated immediately after the explosion showed without doubt that the explosion had occurred in the residual gas scrubbing tower. Fig. 3 conveys an impression of the brisance of the explosion. The bottom of the tower made of 10 mm steel plate was torn off, the cylindrical portion torn and folded up, and the feet compressed with enormous force. AP00022482 . in seeking the cause of the 'were discussed: & the following possibilities 1. Gas explosion of a vinyl chloride/air mixture 2. Runaway vinyl chloride polymeri2ation 3. 'Decomposition of vinyl chloride peroxide After taking account of all factors, possibilities 1 and 2 were ruled out. In particularthey could not explain the great violence of the explosion. The only possible cause left was the decomposition of VC peroxides. However, initially it was not understandable in what manner sufficient amounts of VC peroxide could have got into the scrubbing tower and why they decomposed just at that time. A comprehensive investigation program was carried out to clear up these questions. Sacked by the results of these investigations we now assume the following course of events: Phase 1 As a result of insufficient cooling in the condensers recondensation took place in the scrubbing tower with formation of a liquid VCM layer. Phase 2 The oxygen that had concentrated in the residual gases, possibly supported by further impurities present in the recovered VCM, gave rise to VC peroxides, which initially were A TM liquid Phase, 3 After venting of the scrubbing- towerthe peroxides insoluble in water precipitated and desposited as a tacky mass on the walls of the tower. Phase 4 The heat supplied to the tower to prevent freezing was sufficient to heat the contents to abt. 40C. At this temperature s'ow decomposition of the VC peroxide set in. Poor therm al conduction in the lumps of peroxide caused build- p of heat leading to another rise in temperature & d finally, at abt. 100c, the explosion* like decomposition of the peroxide Consequences for the Operation of VCM/PVC Plants In discussing the consequences to be drawn for the operation of VCM and PVC plants, a distinction should advisably be made between measures to be taken * o avoid peroxide formation and measures to be taken to safely remove peroxides already formed* From our investigations it has~meanwhile become-known that contrary to information from the ? terature no special impurities are required for the formation of peroxides from liquid VCM of usual specification with atmospheric oxygen in the presence of an aqueous phase. Therefore the fi.,.t-aa&ffl&SBfia&-g*cautlon to prevent peroxideeWffflRBSBSjffSfB^^^ 2rTc^^hec}^or^rfW^S^H!^HggB^SBB^ff VCM recovery plants it is recommended to rsly operating oxygen analyzers. Besides by-the addition of s3*ffiire`*inhiortors i" is possible tu cut"out peroxide formation to the greatest possible extent, if not entirely. Since oxygen cannot be comoU^eiv*exeiudec from VCM recovery plants, the use of The VC peroxides obtained by us were readily soluble and relatively stable, in chlorinated i Tiffin^**?1 chloride they were found Today the occasional AP00022483 formation of PVC especially in VCM recovery lanta is looked upon by us as a consequence of Intermediary formation of corresponding amounts of peroxides. Zn this manner VC peroxide can be removed safety. If, however, the peroxide has been precipitated from a solution, it is hard to handle. Water to some extent has_*. stabilizin effect, it_JAJfru* but even gram quantities of peroxi.ddeeTfrgggfftfetMSa ^sr. Against " >n, " `for flushing equipment, peroxides Decomposition finally brought about by hot eaustic soda solution all the ease can be taken to be of a thermal rather than hydrolytic nature. Accumulations of large amounts of peroxide involve the danger of explosion. Safe destruction of major amounts of peroxide is possible only in solution or in the presence of solubilizers, e.g., with alcoholic caustic soda solution. Literature Cited /17 07 07 6.A. Rasuwajew, K.S. Minsker, J. allgem. Chem. (Russ.) 28 (1958) 983 M. Lederer, Angew. Chem. 71 (1959) 162 J. Bauer, A. Sabel, Angew. Makromol.Chem. 47 (1976) 15 ai A-I* Kalinin et al, Chem. Industrie (1966) 1, 27 07 N.N., Case Histories af Accidents in the Chemical Industry, 3 (1970) 142 hills Modi(*d RVCM-Process and Situation Tt ?/79 at the Tima IF* Exofc&on occured Pig. 1 **. 2 i yig. 3 AP00022484 INTEROFFICE MEMORANDUM To A. *3. Santay Prom tiOEtrd l. Watson cc: H. R. Cheura R. E. GadomsM F. J. Ryan Date 12 July 1982 Subject Vlnvl Chloride Safety Meeting VCH Exposure Risk Quantification Corporate SafetyArexl art-own (LotttlM. OflMlutton, w DtMftnwnt) Chemicals Kanufactor1no/Attain V (LMitlw, OrpitlUlloii) ar 0MrtmMt) I have reviewed the Vinyl Chloride Exposure Risk Quantification study which was made In April, 1980 for the Trexlertown Hauls Ion Development Reactor Facility. I understand that you are planning to present an abstract of the methods employed by Air Products in this study to the September 1982 meet ing of the Vinyl Chloride Safety Committee. Ny only concern was the use of the 3.5 x lo'lrlterla outside of Air Products. You Indicated that all reference to this Internal criteria would be removed. The entire page would not be submitted--merely an abstract. I have no problea, then, with your plans to present the abstract In this manner. HLW:hes Howard L. Watson 020} AP00022485 / 7T iJ. PrtfrsJ'^ -rcEiveo I w** * 1981 AIR PRODUCTS 8 CHEMICALS. INC. VINYL CHLORIDE EXPOSURE RISK QUANTIFICATION STUDY TREXLERTOWN EMULSION DEVELOPMENT REACTOR FACILITY 6 .V / a Prepared by: L. C. Doelp R. E. Llnney R. W. Noel R. M, Ormsby Distribution: P. L. T. Brian A. E. Cummins P. Fong A. Greene J. R, Lovett C. McKinley D. J. Orr J. B. Pfeiffer F. J. Ryan M. J. Smith J. T. Wharton Issued: 1 April 1980 ------------------------1 AP00022486 VINYL CHLORIDE EXPOSURE RISK QUANTIFICATION STUDY Table of Contents 1.0 Executive Summary 2.0 Process Recommendations 2.1 VCM Release From Cylinder in Storage 2.2 VCM Release From Cylinder tn Process 2.3 Continuous VCM Release Frcm Pump 2.4 Continuous VCM Release From Reactor 2.5 Instantaneous VCM Release From Reactor 2.6 Continuous VCM Release From -Degasser 2.7 Instantaneous VCM Release From Degasser 2.8 Instantaneous VCM Release From Receiver 2.9 Continuous VCM Release From Incinerator 3.0 Risk Quantification 3.1 Fault Tree Generation 3.2 Frequency of Releases 3.3 Release Rates 3.4 Overall Frequency Rates 3.5 Criteria Appendix A Appendix & Appendix C Page 1 2 2 2 2 3 3 3 3-4 4 4 5 5 6 6 7 7-8 9-19 20-23 24-32 AP00022487 *1 <I >1 ' 1.0 Executive Summary A risk quantification study was conducted for potential employee exposure to vinyl chloride monomer (YCM) on the Trexlertown site, based on a fault tree analysis for the proposed emulsion development reactor system. Figures 1 and 2 show probability of exposure versus level of exposure curves for the three potential receptors chosen for this study. The risk of exceeding OSHA exposure criteria for each of the receptors is as follows: 0) Employees inside Allentown Labs Building (2) Employees walking between Allentown Labs and Trexlertown Lab #1 (3) Employees inside main administration building Frequency Rate, events/yr. 1.3 x 10" 3.7 x 10" 1 ;0 x i(T It should be noted that OSHA criteria for VCH exposure are not normally applied to workers outside of the actual process area. The probability of exceeding OSHA limits at any of the receptor locations is so low as to be acceptable, particularly since exposures at these levels have no known adverse effects on human health. AP00022488 >t 1 AP00022489 HA*A hRATE r ' a m w than X P P H - li te i, a v m t t / y u r IUZAM RATS fo r groater tkan X FMirlmitea. cventi/year n Ik AP00022490 * 2.0 Process Recommendations The following recommendations were Identified during the fault tree analysis and must be Implemented to achieve the calculated frequency rates for VCM exposure. 2.1 VCM Release from Cylinder in Storage 1. VCM cylinder size, type and connections should be standardized in order to facilitate a certification program'to ensure reliability. 2. PCTD should become responsible for the inspection and certification program of VCM cylinders before the Trexlertown development reactor becomes operational. Procedures and policies should be prepared v'.ja well in advance. The cylinders must be hydro tested every 5 years, per DOT regulations. 3. Sprinkler systems should be Installed In all VCM cylinder storage areas to control fires and secondary releases from other cylinders. 4. Use fusible plugs instead of relief valves on each cylinder. 5. Institute controls 'for-cylinder f+lling program as outlined in - orlolnal CE. Proceed with approval of cylinder filling CE. 6. Ensure from procedures (cylinder filling CE) and cylinder design that overfill Is not credible. 2.2 VCM Release from Cylinder In Process 7. All Ng requirements for the VCM system should be dedicated. 2.3 Continuous VCM Release from Pump 8. High pressure hydraulic relief should be set as low as practical on the VCM delay feed pump. 9. Install automatic valve on discharge of VCM delay pump and Interlock with the pump to ensure that VCM supply Is shut when pump Is off. This valve would then have to be opened before the pump would start. 10. Pulsations/vibrations caused by VCM delay pump should be checked at possible fracture points In the piping to determine If bracing Is adequate. (Measurements will be made by ElectroMechanical). 11. Chsnical sealed pressure Indicators should be used on all VCM feed lines. 12. Car seal priming valve on VCM pump closed. 13. Issue written permits for switching pump back to VCM service after being used for some other purpose. AP00022491 14. Color code VCM pipelines. 2.4 Continuous VCM Release From Reactor 15. Install check valve in line between reactor shot pot and reactor. 16. Process piping changes should be logged and drawings kept current to facilitate safety hazards reviews. Process changes should be reviewed and a formal review.held-evary^two .years. , 17. Reactor should be hydrostated prior to VCM series to ensure reliability of check valves as well as integrity of relief systems and vessel seals. Automatic vent on reactor must be tested during hydrostat. 2.5 Instantaneous VCM Release from Reactor 18. Install rupture disk under relief valve with a means to prevent pressure build-up between the two safety relief devices. 19. Relief valves and rupture disks should be Inspected annually. 20. Consider installing block valves Immediately down stream of all delay priming valves. 21. An approved back flow prevention system shall be installed to prevent emulsion from backing to potable/fire water supply system. 22. Develop fail-safe (open) operating mode for short stop addition valve, install pressure gauge on short stop pot. Install threeway valve and recirculation line to short stop pot. 23. Verify that automatic valve to receiver can maintain blowover rate. 2.6 Continuous VCM Release from Degasser 24. Degasser should be routinely hydrostated as often as the reactor to prove leakage Integrity and relief system reliability. 25. Consider orifice on Ng flow to degasser. 26. Install Interlock on bottom valve of degasser with reactor transfer valve. 27. Consider installing block valves Immediately downstream of all post-add priming valves. 2.7 Instantaneous VCM Release Degasser 28. Increase rating of rupture disk on degasser to pressure more near the vessel's certification of 300 psig. 2.8 Instantaneous VCN Release from Receiver 29. Eliminate or armor sight glass. 2.9 Continuous VCM Release from Incinerator 30. Reduce size of bleed valve between two block valves. 31. Detailed hazards review should be done on incinerator when new P&I's are available. 1 AP00022493 PSI drawings for the existing facility were used for the analysis. Although these were not in a final approved form, the team was able to identify the major causes for releases of VCM. A subsequent hazard review.addressing other potential hazards in addition to VCM releases in the emulsion development system has been recommended when final drawings are available. It was decided that the initial quantification for VCM releases would be done on the system as it now exists at Piscataway. However, recommendations which would affect release rates and the probability of releases were documented (see Section 2.0). The initial thrust was to determine where we are now and how this compared to a criterion so we could choose the most cost effective recommendations to meet that criterion. As a criterion evolved and the air dispersion calculations Improved in quality, a clearer picture evolved as to what recommendations had to be implemented. The final fault trees presented in the appendix assume the implementation of these recommendations. 3.2 Frequency of Releases Once the fault trees were generated it became necessary to assign frequencies for each of the primary events. Some of these could be obtained directly from our APCI Failure Rate Data bank (e.g. ' check -valve'failure;. - However for other primary- events -It.was necessary to rely on the experience of the people involved In emulsion processing to obtain frequency information (e.g., agitator seal failure). This understanding of the process was also critical for obtaining frequencies for operator errors (expressed as failures per demand) which required knowledge of the number of demands per year. The blending of two disciplines - emulsion polymer expertise and hazard quantification expertise was crucial for accurate representation of each primary event. 3.3 Release Rates The two major factors that influenced this parameter are the total quantity of VCM which can be released and the total time of the release. The quantity of VCM released for the major causes were determined from knowledge of the process. The time of these releases was also estimated based on experience and judgment. One other factor which was considered in the "release time" was whether the release was directly outside, such as a relief device operating, or whether the release occurred inside the process area, such as a spill. If the release did occur inside, the effect of room ventilation was taken into account when calculating the release rate. Smaller causes such as leakage from valves, pump seals and connections were also estimated and used in the analysis. Although these sources did not contribute to significant VCM exposure levels outside the emulsion development reactor building, they could contribute to worker exposure within the building and some of the recommendations from the preliminary hazards review deal with this concern. AP00022494 -/- 3.4 Overall Frequency Rates Appendix B describes the calculation of frequency rates for the individual releases from all sources In the development reactor systen. These calculations require the combination of dispersion models, probabilities of releases, magnitude of releases and wind direction data to generate probability of exposure versus level of exposure curves. This data was used to determine the total probability of exceeding OSHA exposure limits at the following three receptors: (1) Employees inside Allentown Labs Building (2) Employees walking between Allentown Labs and Trexlertown Labs #1 (3) Employees Inside main administration building Frequency Rate, event/vr 1.3 x io-5 3.7 x IO'* 1.0 x IO'6 3.5 Criteria We are concerned with the potential exposure of Air Products employees at this site to vinyl chloride. Trexlertown is the company headquarters, with a large number of office workers and administrative employees. These people, for the most part, are not aware that a small quantity of vinyl chloride is used at this site and they have not been asked to accept any risk of exposure. Employees working in the PVC laboratory and the emulsion development reactor building, however, have been educated about vinyl chloride and have accepted some risk of exposure. The only published cirteria for exposure to vinyl chloride are regulations generated by OSHA for workers in process areas. These regulations consist of the following: (1) A worker shall not be exposed to greater than 1 ppm vinyl chloride averaged over 8 hour workday (the Integrated exposure level would become 480 ppm minutes for a full workday). (2) A worker shall not be exposed to greater than 5 ppm averaged over any 15 minute period (the Integrated exposure level is 75 ppm minutes for this 15 minute period). Noncompliance with these regulations could result in a citation from OSHA. These criteria provide a conservative limit for exposure of employees outside the process areas. AP66622495 At present, there Is no accepted APCI hazard rate goal for exposure to carcinogens. However, the total hazard rates for exceeding OSHA limits are below the existing rate goal for fatal events. Since VCM exposures at or below OSHA limits have no known adverse health effects, the calculated hazard rates for exposures caused by the Emulsion Development Reactor System should be considered acceptable. AP00022496 APPENDIX AP00022497 ' JAppendix A Dispersion Calculations Continuous Roloase The continuous point source equation is used as follows: 60 10xA e 2uUoyozCo (z-h)2 _ (z+h^2 e 2oz + e 2az Qt total gas release rate, kg|sec o oo T ' Q + QA 0 Q-* active gas release-rate, kgjsac o associated air rate, kg|sec D * dilution factor o qmwt 0 c--I qtmw ' MW *= molecular weight of active gas MW- * MW of total gas O) (2) (3) CQ * density of release mixture 21.924 x MWt C " {460 TF) ^ k3Jm z * height of receptor, m h = height of source, m <4) APOO622498 x * distance from source, m y " lateral distance from centerline of plume path, m oy and 02 are calculated from Pasqul11-Gifford relationships using a virtual source, xy oy * cy(x + xv), m 02 oz(x + xv), m See Figure K An area is calculated such that at the source: 0 2y*2* x U x C0 * Qt where y* and z* are defined as follows: (5) (.6) 7) y* x 1 * Ti 2d*y 0 h 0 ^STT1 ozC0 u (8) 2* x SZT1lazC0Ji (9) Substituting 8 and 9 in 7 gives: o Qt oyoz "rar * **(*v><*v> 0 10) U * wind velocity, mjsec At x * 0, an *y is selected that satisfies Equation 10. This is equivalent to displacing the model source backward until, at the real source the cloud area times the wind velocity Is equal to the source rate. Equation 10 also satisfies the condition that the concentration at the real source is 100% of the emitted gas mixture. Equation 1, C}CQ gives the concentration, ppm at any point downstream of the real origin. For an outside observer the exposure E^ is given as: 1 AP00022499 11` AP00022500 E1 = " x T1 x q PPm min o where: (11) C|CQ steady state cone, ppm Tj duration of release, sec. r A similar exposure Eg is obtained when ^r is above a given threshold. C1 e2 * r~ x ti x to*ppm m1ri a 0 (12) 03) r* * threshold ppm o Exposure inside a building is obtained by treating the building as a well mixed tank. See Figure 2. The exposure above zero is given as: E1 60 X C O Ey * ppm min (14) ClC * concentration at ventilation intake of building, ppm *o Tj * duration of release, sec but not greater than Tg Tg * duration of residence in building, sec Concentration Inside the building for the two'time periods Is given as follows: o < t <Tj (14-A) AP00022501 Ei.L Co Co Ft Ft, V -u ppm (14-B) The calculation of Eg type exposure is illustrated in Figure 3. Eg is the area under the curve between T-j* and Tg*. Is the duration of the re lease. Eg *s siven as: E2 Ez' * ^ U5) Eg * exposure from to T-j U Eg * exposure from T-j to Tg* lc E,, * - v)+ -r e -FvT1 . -FTv ,* -j FV -0 f\a - V) e FTj* v -e Hi where 6V ` ? ,n - err (16) 07) 08) & P-- V - rln (19) -- * threshold concentration, ppm Lo concentration at vent intake, ppm When a continuous release is made through a vertical vent, additional credit is given to the height of the release as well as dilution achieved by the vertical jet. A top hat, neutrally bouyant jet model is used. See Figure AP00022502 AP00022503 The model Is based on conservation of mass and momentum. It Is assumed that the jet will be turned down stream when the jet velocity slows down to the wind velocity. Dilution Is achieved by air Inspired into the jet, a. Equations for the effective jet height, h., and dilution are: J * fe ft - ') !20) U molejsec gas at h. Jet Dilution * rp ----------------------------- ^ uw mole]sec gas at stack (21) b,, 8 radius of stack, m o U0 * velocity in stack, m|sec Uw 8 wind velocity, m[sec hj * effective jet height above stack height, m a 8 0.11 - 0.13, dimensionless Inspiration constant hj is added to the stack height, hs, to get the total release height. The moles of air inspired are calculated from the jet dilution and added to the o source rate of air, QA. AT A + Instantaneous Release The instantaneous point source equation is used as follows: _ (x-Ut)2 106DQt e 2oxI2 e (2*) ^2 axlayIaZlCQ 2 2oyl X [ (2 - *)Z . | e Zozl + e 2azl (22) (23) AP66622504 Qt = total gas released, kg CSy - Q + Qa Q active gas released, kg * air released, kg D dilution factor qmwt D * OjBT MW * molecular weight of active gas MWt * MW of total gas (24) oxI, oyl and ozl are calculated from Slade "Meteorology and Atomic Energy". Qt 3 The Initial cloud volume is yr- m . ^0 This cloud is assumed to have the dimensions of half a spheroid: x 2.15oxl y - 2.15ayl 2 - 2.15ozl of volume: (26) (27) (28) Qr 2it 3 x (2.15) axloylozl Lo J (29) Equation 29 is used to estimate the virtual source distance as follows: 1 3Q, C2b+d)' 2*C0(2.15)3a2c (30) where a, b, c and d are the coefficients given by Slade for the sigmas AP00022505 oxl oyl = a(x + xy)b oil - c(x + x/ (31) (32) Equation 23 gives the concentration, ppm, at a point down stream at x, y and z. As the cloud passes by the concentration varies with time. The exposure, , above zero is obtained by the integral: to T2 is the time period it takes for the cloud to pass by. I, - (34) T x - AoxI lZ U Combining Equations 23, 33, 34 and 35 gives: V 2ttpx1 J U x ^ , ppm min (35) (36) * yi2 10 DQt e (2it)d ^oxIoylozIC ( . Lidi|2 . 2ozI 2ozI e +e ppm (37) 0 is the maximum concentration, ppm, experienced as the cloud passes by. The Eg type exposure is given as: (38) Where Tj* to Tg* is the time period for when the concentration Is greater than the threshold, ^C--* . AP00022506 _ IJ SOU erf t, ppm min Where: t J (39) (40) The exposure In a building is obtained using an approximation. It is assumed that the cloud passes the Intake vent In such a short period that during this time no active gas is discharged from the building. The amount of active gas taken in is: T, *C6F H^+ 1 The initial concentration in the building is: (41) 5o %flz'c C,, ' V / t7 dt * 1 50 (42) Where J is given by Equation 37. The concentration will then decline as follows CB _ CBo Co C ft v (43) The exposure above 2ero Is given as: FT m \/2TroxIJ 1 ~ U 1 -e v x 1 , ppm min (44) T2 is the duration of residenceln the building. The amount of active gas taken in above the threshold concentration is: 'V - F dt *B Lo V (45) AP00022507 ''The Initial concentration is: FqXl J vU erf t (46) Exposure inside the building continues until T*t when the concentration r* falls below . T* * - -Sr In C* 7-To- Bo '7 The exposure Eg is given $s: Where \fl2yttioxI<3 U erf In C*|C. FT* 1 1 -e v (47) (48) (49) AP00022508 Appendix B Frequency Rate Calculations Probability of Occurrance, F Probability of occurance for each spill scenario was calculated using techniques described in the APCI Process Hazards Analysis Seminar notes. Failure rates were taken from industry sources and from operating experience*at Piscataway, The rates used represent the concensus of the fault tree team. Individual Probabilities, F The probability for each unique set of conditions (i.e., weather conditions, fan operation) was calculated by multiplying the probabiliti for each condition. ^wind^ ^weather occurance ) Frequency Rate,-events/year Because the probability of occurance was calculated based on a year's time and the F's are small, the frequency rate is approximately equal to F/year. Dispersion Angle of VCM Cloud. Cloud width at radius r is 4.30 OJ, r AP00022509 Probability of Wind being in Vent's Direction, Pw^nd The source is 32 NE at 38.1m From National Weather Service Data, the fractional time the wind is 32 NE - 45 is: 0,002336/degree of inclusion therefore 7wind * -002336 8 * 0.004672 arcsin ^ Probability of Weather Stabilities, ^ at^ The weather conditions have been modeled by these six conditions: Stability Class Wind Speed fr weal F 1.544 0.03 D 1.544 0.06 B 1.544 0.01 F 3.4 0.27 D 3.3 0.54 B 3.9 0.09 Probability of Fan Speeds, P^an The receptor has two air turnover rates. The lower rate occurs approximately 90S of the time while the high rate (100% turnover) occurs the remaining 10% of the time. These rates are independent of the emulsions development operations and are therefore considered to be random. Because different receptor fan rates had little effect on the longer releases, these probabilities were only considered for shorter releases. When the hypothesized releases occur inside the process room, the room ventilation system was also taken into account. The process room's dual capacity fans should switch to high turnover if there is a spill. The probability of the system operating successfully is 98%. AP00022510 SOURCE AND RECEPTOR LOCATIONS AP00022511 . .:&sn I'.INCY RATES FOR :ha VIOLATION - "" * 2.:s TM" * 8.::" 6.`~ Z.7- e.r- - " 2. :~r - 7.iZ= 3. r:f 7 .=S.zTL l.~ 0 Outside R&D 4.835 x 10'S 1.864 x 10 9 1.097 x 10"B 4.132 x 10"8 1.633 x 10"8 4.430 x 10'8 2.062 x 10'7 7.296 x 10`B 3.987 x 10"7 1.856 x 10`6 6.567 X 10"7 9.382 x 10`9 3.498 x 10'9 3.722 x 10-6 AP00022512 u> *XCcooOI. Q. . O' |a,< AP000225I3 Q uantity Released Probability s Frequency Rate AP00022514 CO RClCftSE FROM C FlINDER mCOoNcNEEsCtTED TD .G >E `'.; \ Q * 34.o ^ F- 7,1> *u> a ^ 0.002^ Q - 0.3. kj r` P a lot ns s -yri. r> yiyj f)n; !}/j <* A / d -92- J AP000225I5 -27- llo bOO Q ' 4,4fnw"* 0 * <1t O' ?.6^^|0oo O.XD7 * 6.0777C p $,577*/a .4 A P 0 0 0 2 2 5 16 I t I co CM AP00022517 *f AP00022518 *0 * onI AP00022519 -31*A I AP00022520 LZSZZOOOdV Memorandum cc: H. L. Watson The Safety and Health Work Practice Standard for Vinyl Chloride at Escambia has been fully approved. I will retain a copy In my files. HJStmcs 38641 Attachment roau i4I0i>/*ji RECEIVE APR - 3 (986 H. L WATSON AP00022522 SAFETY AND HEALTH WORK PRACTICE STANDARD - Vinyl Chloride ESCAMBIA PLANT REVISED 29 JANUARY 1986 'Mat* AP00022523 SAFETY AND HEALTH WORK PRACTICE STANDARD VINYL CHLORIDE ESCAMBIA PLANT REVISED 29 JANUARY 1986 Approvals: Chemicals Group Manager of Safety ,*V/'JjJ __ .__&*(?> Date `Manager of Manufacturing Services Date AP00022524 j. j. saQker MANUFACTURING MANAGER dSTe-- INDUSTRIAL X VICE MAN U FAdUA^ttG MANAGER "ES7T bste NOTE: The purpose of this Standard is to promote safe conditions through the avoidance of exposure and prevent accidents affecting the safety and health of employees, visitors, and contractors through proper instruction and imple mentation of safe operating procedures, so as to protect the well being of individuals and minimize product or property losses. It addresses the safety and industrial hygiene concerns Involved with the handling of vinyl chloride monomer. As the need for new procedures is recognized, it Is intended that they will be developed and added to this standard. T-rv. AP00022525