Document ypp6ax1YmBq39Ldb092625xOX

JOSEPH E. KELLER JEROME H HECKMAN CHARLES M,MEEHAN WILLIAM H. BOROHESANI, JR, ROBERT R TIERNAN WAYNE V. BLACK DAVID L HILL MARTIN W. BERCOVICI PETER M-NEMKOV JOSEPH E, HADLEY, JR. TAROLE C, HARRIS PETER THOMAS SMITH MICHAEL F MORRONE LARRY S. SOLOMON JOHN B. DUBECK CHRISTINE A. MEAGHER SHIRLEY S. FUJIMQTO LAW OFFICES Keller and Heckman 1130 lTTS STREET/ N.W. SUITE lOOO WASHINGTON, D.C, 20036 December 6, 1977 RECEIVED UtU 8 R.LGOITESMAN TELEPHONE 202 4571100 CABLE ADDRESS KELMAN writers DIRECT DLAL NUMBER (202) 457-1116 Re: SPI-PVC Safety Group SPl-PVC Manufacturing Technology Committee SPI-PVC Health Committee SPI-PVC Communications Committee SPI-PVC Lawyers Subcommittee Gentlemen Attached hereto is a letter from Harold Himmelman at Beveridge, Fairbanks & Diamond written to cover the Environmental Protection Agency's Draft Inspection Manual for vinyl Chloride. Due to the technical nature and bulk of this docu ment, copies are being supplied at this time only to the listed Voting Representatives and members of the Manufac turing Technology Committee. Having received this document only late Monday afternoon, we are sending it out without having an oppor tunity to review it. As always, if before we report to you further about the Manual you have any questions, comments or suggestions with regard to it, please do not hesitate to let tts know. Cordially yours Enclosures cc: (Without copies of Manual) Mr. William H. Bricker Mr. 0. P. Thomas Mr. Richard Fleming COLORITE 007280 HENRY L DIAMOND RICHARD M FAIRBANKS, HI ALBERT J, BEVERIDGE,XU Cary h SAISE A JAMES BARNES HAROLD HiMMELMAN Christopher h, buckley, jr, KEVIN M, FOLEY JONATHAN Z-CANNON ANDREW E MISHKIN CHARLES A. PATRIZIA SCOTT w, BOWEN law offices Beveridge, Fairbanks & Diamond One Farragut 5quare South Washington, D. C. 20006 TELEPHONE (202) 638-7800 December 5, 1977 CARL EARDLE OF COUNSEL CABLE ADDRESS " IN dlaw " BY HAND Joseph E. Hadley, Jr. Keller & Heckman Suite 1000 1150 Seventeenth Street, Washington, D.C. 20036 N.W. Dear Joe: 1 am enclosing herewith the vinyl chloride field in spection manual which EPA's Division on Stationary Source Enforcement has asked me to distribute to the industry. You will recall from my September 29, 1977 letter that the field inspection manual was originally prepared in draft by Research Triangle Institute. The manual, as you know, will be used by EPA inspectors during on-site visits to VCM, PVC and ethylene dichloride plants. In addition to regional EPA offices and Research Tri angle Park, SPI is the only other party to have received a copy of this draft as of this time. The Division of Sta tionary Source Enforcement wants all comments on the draft submitted by January 6, 1978. Currently, the Division plans no hearings on the manual and hopes none will be necessary. However, should there be major areas in need of further dis cussion, there is some possibility that an informal hearing or meeting might be held. EPA hopes to make the field manual final, and to issue it, within two or three months. Although I am sure that the entire manual will be of extreme interest, I should report that SPI has been especially requested to review Chapters 2 and 3. Sincerely yours HH: cl Enclosure cc: John R. Lawrence Robert Laundrie Harold Himmelman COLORITE 007281 NATIONAL EMISSION STANDARDS FOR HAZARDOUS AIR POLLUTANTS INSPECTION MANUAL for VINYL CHLORIDE Prepared by Michael F. Lamorte Contract No. 68-01-4141 Task No. 2 EPA Project Officer John R. Busik Prepared for U. S. ENVIRONMENTAL PROTECTION AGENCY Division of Stationary Source Enforcement Washington, D. C. November 1977 COLORITE 007282 ACKNOWLEDGEMENT This Inspection Manual was prepared by Mr. Michael F. Lamorte of the Research Triangle Institute. Project Officer for the Environmental Protection Agency - Enforcement Division was Mr. John R. Busik. The Task Manager was Mr. Richard Biondi assisted by Ms. Libby Scopino. The author appreciates the many contributions made by Mr. Biondi and Ms. Scopino during the preparation of this Manual. The author also appreciates the assistance of Surveilance and Analysis and of Enforcement Division personnel of District 6 in applying the Inspection Forms under field conditions. ii COLORITE 007283 TABLE OF CONTENTS ACKNOWLEDGEMENT LIST OFFIGURES LIST OFTABLES LIST OFINSPECTION FORMS LIST OFCHEMICALFORMULAS 1.0 INTRODUCTION 1.1 Background 1.2 Authority 1.3 Applicability ' 1.4 Definitions 2.0 EDO, VC AND PVC INDUSTRIES 2.1 Ethylene dichloride (EDC) 2.2 Vinyl chloride (VC) 2.3 Polyvinyl chloride (PVC) 3.0 PROCESS FLOW DESCRIPTION AND EMISSION POINT IDENTIFICATION 3.1 Ethylene dichloride--oxychlorination 3.2 Vinyl chloride 3.2.1 Hydrochlorination of acetylene 3.2.2 Dehydrochlorination ofethylene dichloride Page ii v ix x xi 1 1 2 z 3 6 5 7 8 g g 13 13 17 iii COLORITE 007284 TABLE OF CONTENTS (Continued) 3.3 Polyvinyl chloride 3.3.1 Suspension polymerization 3.3.2 Emulsion (i.e., dispersion) polymerization 3.3.3 Latex dispersion polymerization 3.3.4 Bulk polymerization 3.3.5 Solvent polymerization 3.4 Clarifying note on the balanced oxychlorinationdehydrochlorination process 3.5 Photographs of equipment 4.0 EMISSION MONITORING INSTRUMENTATION, RECORDS ANDREPORTS 4.1 Leak detection monitoring instrumentation 4.2 Leak detection monitoring recordkeeping 4.3 Routine leak detection and relief dischargerecordkeeping 5.0 INSPECTION PROCEDURES AND INSPECTION FORMS 5.1 Summary of compliance status 5.2 Checklist 5.3 On review of records 5.4 Pre-test equipment checklist for stack emissiontest 5.5 Equipment checklist for vinyl chloride concentration in inprocess wastewater,- resin, slurry, wet cake and latex samples APPENDIX A: Mean value calculation APPENDIX B: National Emission Stan iards for Hazardous Air Pollutants - Standard for Vinyl Chloride, October 21, 1976 REFERENCES 19 19 23 25 25 30 38 40 57 57 61 61 62 62 63 63 64 64 87 88 104 iv COLORITE 007285 LIST OF FIGURES 3.1(a) Ethylene dichloride process flow diagram using oxychlorination which involves the reaction of oxygen and hydrogen chloride with ethylene. 3.1(b) Ethylene dichloride process flow diagram using oxychlorination which involves the reaction of oxygen present in air and hydrogen chloride with ethylene. 3.2 Vinyl chloride monomer process flow diagram using hydrochlorination of acetylene. 3.3 Vinyl chloride monomer process flow diagram using dehydrochlorination of ethylene dichloride. 3.4 Polyvinyl chloride process flow diagram using suspension polymerization. 3.5 Polyvinyl chloride process flow diagram using emulsion (i.e., dispersion) polymerization. 3.6 Polyvinyl chloride latex process flow diagram using emulsion (i.e., dispersion) polymerization. 3.7 Polyvinyl chloride process flow diagram using bulk polymerization. 3.8 Polyvinyl chloride process flow diagram using solvent polymerization. 3.9 More recent polyvinyl chloride process flow diagram using solvent polymerization. 3 .10 Block diagram of a balanced oxychlorinationdehydrochlorination process. 3.11 Large capacity VC reactor using the oxychlorination orocess is located in the tall vessel. The oxychlori nation manual vent is the ooen ended pipe, rising above the top of the reactor on the left side. Page 10 11 14 18 22 26 20 31 33 34 39 42 v COLORITE 007286 LIST OF FIGURES (continued) 3.12 The tons of medium capacity, side-by-side PVC reactors using the emulsion process may be seen. Individual RD/SRV and manual vents may also be seen mounted on top of each reactor. 3.13 Small size PVC reactor with cover removed to clean reactor in preparation for next polymerization run. Prior to removing cover the PVC and water solution were removed and transferred to stripper. "Elephant trunk" is placed through the opening to vent vinyl chloride gas through "trunk" to the recovery system in order to be in compliance with emissions standard reactor opening loss. This type reactor is used in the susDension, emulsion and latex polymerization processes. 3.14 Medium capacity side-by-side PVC reactors using solvent process. Individual RD/SRV and manual vents connected to vinyl chloride recovery system may also be seen. Lower portion of reactors comprise the heating elements to raise the contents to the tem perature required for polymerization. 3.15 The top of medium capacity, side-by-side PVC reactors using solvent process. Foreground shows a motor valve for emergency venting through to the VC monomer recovery system. 3.16 The top of a small capacity PVC stripper vessel used in the suspension, emulsion and latex processes. The RD/SRV vent is connected to the recovery system. 3.17 A stripper column is shown that removes VC from a PVC-varnish solution resulting from the solvent polymerization process. 3.18 A typical wastewater stripper column is shown that may be found in EDC, VC and PVC plants. 3.19 The top of an EDC storage tank is shown with its vent. 3.20 Cylindrical, side-by-side, above ground VC storage tanks are shown. RD/SRV vents may be seen on pipe rack above tanks. vi 42 43 43 44 4^ 45 45 46 46 COLORITE 007287 LIST OF FIGURES (continued) 3.21 Spherical, above ground vinyl chloride storage tanks. RD/SRV vents may be seen perched on top of sphere. 3.22 Underground VC monomer storage tank area is shown. 3.23 Wash water stripper storage tank with its vent mounted at the top of the tank in an EDC-VC plant. 3.24 Upper portion of tar storage tank shown with its vent connected to recovery system in an EDC plant. 3.25 PVC suspension resin rotary dryer and dust collector. Drying is accomplished by the application of heat and rotary action. 3.26 PVC emulsion resin spray drying takes place in the cylindrical building. Large diameter feed pipe, seen at the left of the dryer, carries resin to the top of the dryer. Dust collector is perched at top of dryer. 3.27 An EDC light ends distillation column is shown with its RD/SRV vent. 3.28 An EDC heavy ends distillation column is shown. 3.29 The top of an EDC light ends column condenser showing the RD/SRV vent. 3.30 In the foreground the light ends dryer regeneration vessel is shown in balanced EDC-VC plant. 3.31 An EDC finishing column with the RD/SRV and its vent mounted on the top. (In some installations the finishing column performs the function of the light ends distillation column while in others it encompasses the functions of both the light and heavy ends columns.) 3.32 Foreground shows the insulated piping connected to the reactor refrigerator condenser vessel in a balanced EDC oxychlorination plant. 47 47 48 48 49 ag 50 50 51 51 52 52 vii COLORITE 007288 LIST OF FIGURES (continued) 3.33 Typical RD/SRV assembly with the vents from each connected to a manifold of the recovery system. 3.34 Dual RD/SRV vents mounted at top of spherical vinyl chloride storage tank. 3.35 Typical pump and double mechanical seal. 3.36 Top connections on railroad tank car shown with flexible hose attached for VC loading. Smaller diameter flexible hose in the foreground is con nected to recovery system. 3.37 Railroad tank car loading platform shown with pipe rack support for flexible hose VC feed and recovery system in VC plant. 3.38 Incinerator and scrubber stack showing the steam exhaust stream in an EDC-VC plant. 3.39 Incinerator and stacks of a PVC plant showing the platform (center stack) on which stack samples are taken to determine VC emission concentration. 3.40 Typical incinerator in a VC and PVC plant with bottom portion of stack shown at the right. 4.1 Schematic diagram of continuous monitoring system for vinyl chloride emissions. 53 53 54 54 55 55 56 56 60 vi i i COLOR!TE 007289 LIST OF TABLES 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 Most probable emission points identified by the keys in Figures 3.1(a) and 3.1(b) in the oxychlorination process for ethylene dichloride. Hydrochlorination of acetylene product gases and their boiling point temperatures. Most probable emission points identified by the keys in Figure 3.2 in the hydrochlorination of acetylene for vinyl chloride monomer. Most probable emission points identified by the keys in Figure 3.3 in the dehydrochlorination of ethylene dichloride for vinyl chloride monomer. Most probable emission points identified by the keys in Figure 3.4 in the suspension polymerization process for polyvinyl chloride. Most probable emission points identified by the keys in Figure 3.5 in the emulsion (i.e., dispersion) polymerization process for polyvinyl chloride. Most probable emission points identified by the keys in Figure 3.6 in the emulsion (i.e., dispersion) polymerization process for polyvinyl chloride latex. Most probable emission points identified by the keys in Figure 3.7 in the bulk polymerization process for polyvinyl chloride resin. Most probable emission points identified by the keys in Figure 3.8 in the solvent polymerization process for polyvinyl chloride and copolymers. Most probable emission points identified by the keys in" Figure 3.9 in the solvent polymerization process for polyvinyl chloride and copolymers. Page 12 15 16 20 24 27 29 32 36 37 ix COLORITE 007290 LIST OF INSPECTION FORMS Summary of Compliance Status Page 65 Checklist gg On Review of Records 75 Pre-Test Equipment Checklist for Stack Emission Test 81 EquiDment Checklist for Vinyl Chloride Concentration In Inprocess Wastewater, Resin, Slurry,Wet Cake and Latex Samples 34 x COLORITE 007291 LIST OF CHEMICAL FORMULAS Name Acetylene Chlorine Ethylene Ethylene Dichloride Hydrogen Chloride Oxygen Polyvinyl Chloride Vinyl Chloride Monomer Water Formula HC ~ CH ch2 = ch2 cich2 - ch2ci HC1 -( HA-C = CHC1 n H2C = CHC1 h2o xi COLORITE 007292 COLORITE 007293 1.0 INTRODUCTION 1.1 BACKGROUND Pursuant to Section 112 of the Clean Air Act (42-U.S.C. 1857), the Administrator of the U. S. Environmental Protection Agency (EPA) has added vinyl chloride to the list of hazardous air pollutants (40 FR 59477) and established a national emission standard (40 FR 59532) for facilities which manufacture ethylene dichloride, vinyl chloride, and/or oolyvinyl chloride. The NESHAPS regulations are applicable to plants Droducing the following: ethylene dichloride by the reaction of oxygen and hydrogen chloride with ethylene; vinyl chloride by any process; and one or more polymers containing any fraction of polymerized vinyl chloride. The regulations do not apply to equipment used in research and development if the reactor used to polymerize the vinyl chloride processed in the equipment has a capacity of no more than 0.19 m (50 gal). Research and development facilities containing a polymerization reactor capacity greater than 0.19 m (50 gal) but no more than 4.07 m (1100 gal) are exempt from all parts of the regulations except the 10 ppm limit. The proposed emission standards for existing and new plants were advanced in the Federal Register, December 24, 1975 for vinyl chloride in plants manu facturing ethylene dichloride, vinyl chloride, and/or polyvinyl chloride. Final standards (40 FRL 618-1) became effective October 21, 1976 and apply to existing and new plants [1]. EPA decided to regulate vinyl chloride because it has been implicated as the causal agent of angiosarcoma (a rare form of liver cancer) and other serious disorders, both carcinogenic and non-carcinogenic, in people subjected to occupational exposure and in laboratory animals exposed to controlled concentrations of vinyl chloride [2]. Reasonable extrapolations from these findings cause concern that vinyl chloride may cause or contribute to the same or similar disorders at present ambient concentration levels. The purpose of the standard is to set limits of vinyl chloride COLORITE 007294 emissions from all known process and fugitive emission sources in ethylene dichloride, vinyl chloride, and/or polyvinyl chloride plants to the level which insures a clean, healthy environment. This will have the effect of furthering the protection of public health by minimizing the health risks to peoole living in the vicinity of these plants and to any additional people who are exposed as a result of new construction [3]. This Inspection Manual contains the guidelines for the benefit of and the standardized procedures to be followed by EPA field inspectors or their designated representatives. In conjunction with the operator's testing and monitoring results and the required recording and record keeping of these results, the basic enforcement tools are readily available to properly trained field inspectors. The degree to which this portion of the NESHAPS program is successful depends critically on the effectiveness and efficiency with which inspectors conduct field inspections. 1.2 AUTHORITY Authority for promulgation of the NESHAPS standards and regulations of Air pollutants is contained in Section 112 of the Clean Air Act (42 U.S.C. 1857). It directs the Administrator of the U. S. Environmental Protection Agency to establish emission standards and regulations for hazardous air pollutants (40 FR 59477), to maintain a current listing of these pollutants, and to publish in this listing those pollutants that contribute to producing a hazardous condition to human life and to the environment. 1.3 APPLICABILITY The applicability of the NESHAPS standards and regulations is specified with respect to the function of the facility, product and process by which the product is produced. There are no exemptions to the NESHAPS vinyl chloride emission standards and regulations for production plants which employ reactors of any capaci ty to produce one or more of the following: ethylene dichloride by reaction of oxygen and hydrogen chloride with ethylene; vinyl chloride by any process; and one or more polymers containing any fraction of polymerized vinyl chloride [1]. Equipment employed in research and development of the polymerization 3 of vinyl chloride for which the reactor capacity is not greater than 0.19 m 2 COLORITE 007295 (50 gal) is not subject in any way to the vinyl chloride emission standards and regulations f1]. Equipment employed in research and development of the polymerization 3 of vinyl chloride for which the reactor capacity is greater than 0.19 m (50 gal) and less than 4.07 m (1100 gal) is subject only to the 10 ppm vinyl chloride emission limit into the atmosphere from each reactor, stripper, monomer recovery system, and mixing, weighing, and holding containers [1]. 1.4 DEFINITIONS The definitions of terms used in this Inspection Manual are precise and this precision is required to conduct inspections properly. The terms requiring this precision are defined below [1]. (a) "Ethylene dichloride plant" includes any plant which produces ethylene dichloride by reaction of oxygen and hydrogen chloride with ethylene. (b) "Vinyl chloride plant" includes any plant which produces vinyl chloride by any process. (c) "Polyvinyl chloride plant" includes any plant where vinyl chloride alone or in combination with other materials is polymerized. (d) "Slip gauge" means a gauge that has a probe that moves through the gas/liquid interface in a storage or transfer vessel and indicates the level of vinyl chloride in the vessel by the physical state of the material the gauge discharges. (e) "Type of resin" means the broad classification of resin referring to the basic manufacturing process for producing that resin, including, but not limited to, the suspension, dispersion, latex, bulk, and solution processes. (f) "Grade of resin" means the subdivision of resin classification that describes it as a unique resin, i.e., the most exact description of a resin with no further subdivision. (g) "Dispersion resin" means a resin manufactured in such a way as to form fluid dispersions when dispersed in a plasticizer or plasticizer/diluent mixtures. 3 COLOR!TE 007296 (h) "Latex resin" means a resin that is produced by a polymerization process that initiates from free radical catalyst sites and is sold undried. (i) "Bulk resin" means a resin which is produced by a polymerization process in which no water is used. (j) "Inprocess wastewater" means any water which, during manufacturing or processing, comes into direct contact with vinyl chloride or polyvinyl chloride or results from the production or use of any raw material, intermediate product, finished product, by-product, or waste product containing vinyl chloride or polyvinyl chloride but which has not been discharged to a wastewater treatment process or discharged untreated as wastewater. (k) "Wastewater treatment process" includes any process which modifies characteristics such as BOD, COD, TSS, and pH, usually for the purpose of meeting effluent guidelines and standards; it does not include any process the purpose of which is to remove vinyl chloride from water to meet requirements of this subpart. (l) "In vinyl chloride service" means that a piece of equipment contains or contacts either a liquid that is at least 10 percent by weight vinyl chloride or a gas that is at least 10 percent by volume vinyl chloride. (m) "Standard operating procedure" means a formal written procedure officially adopted by the plant owner and/or operator and available on a routine basis to those persons responsible for carrying out the procedure. (n) "Run" means the net period of time during which an emission sample is collected. (o) "Ethylene dichloride purification" includes any part of the process of ethylene dichloride production that follows ethylene dichloride formation and in which finished ethylene dichloride is produced. (p) "Vinyl chloride purification" includes any part of the process of vinyl chloride production that follows vi-nyl chloride formation and in which finished vinyl chloride is produced. 4 COLORITE 007297 (q) "Reactor" includes any vessel in which vinyl chloride is partially or totally polymerized into polyvinyl chloride. (r) "Reactor opening loss" means the emissions of vinyl chloride occurring when a reactor is vented to the atmosphere for any purpose other than an emergency relief discharge as defined in 61.65(a). (s) "Stripper" includes any vessel in which residual vinyl chloride is removed from polyvinyl chloride resin, except bulk resin, in slurry form by the use of heat and/or vacuum. In the case of bulk resin, "stripper" includes any vessel which is used to remove residual vinyl chloride from polyvinyl chloride resin immediately following the polymerization step in the plant process flow. 5 COLOR!TE 007298 2.0 EDC, VC AND PVC INDUSTRIES Polyvinyl chloride is a polymer employed in the fabrication of literally thousands of consumer and industrial products. The manu facturers of these products purchase the polyvinyl chloride resin from a relatively small number of producers. Polyvinyl chloride is polymerized from the vinyl chloride monomer. In turn vinyl chloride is produced from ethylene dichloride by cracking ethylene dichloride during dehydrochlorina tion or from the hydrochlorination of acetylene. In the following brief presentation, the attributes of ethylene dichloride, vinyl chloride, and polyvinyl chloride production facilities are discussed with respect to those considerations that are pertinent to plant inspections. 2.1 ETHYLENE DICHLORIDE (EDC) The principal process for ethylene dichloride production in the U.S. is oxychlorination which involves the reaction of oxygen and hydrogen chloride with ethylene [4]. This process accounts for 92% of the pro duction in the U.S. [4]. In 1974, there were nine plants using this process to produce 5.05 billion pounds of ethylene dichloride per year [6] This accounts for 74% of the total used for vinyl chloride production and approximately 77.5% of the 20.4 million pounds per year of vinyl chloride emission [6]. While refined ethylene dichloride is sold for other industrial uses, its major use. Is for vinyl chloride monomer production. Typically the economics of the industry dictates that an ethylene dichloride Diant be in close proximity to a vinyl chloride monomer plant in which case the shipping cost of large, continuous flows of pure ethylene dichloride to a vinyl chloride plant is minimal [4]. This has led to the concentration of these plants in Texas, Louisiana and the northern states [4], In most 6 COLOR!TE 007299 cases, the unconverted ethylene dichloride from the vinyl chloride plant is recycled back with the crude ethylene dichloride [4], Vinyl chloride emissions occur in an ethylene dichloride plant using the oxychlorination process from three sources: side reactions in the oxychlorination process and dissolved vinyl chloride in recycled hydrogen chloride and in recycled unconverted ethylene dichloride. Physically, the major vinyl chloride emission sources are the EDC reactor, EDC refining and the fugitive emissions that may be present. 2.2 VINYL CHLORIDE (VC) There are four processes to produce vinyl chloride monomer [6]. There are fifteen plants producing vinyl chloride having a production capacity of 6.8 billion pounds per year [6]. Two plants use the hydro chlorination of acetylene, nine use the chlorination-oxychlorination of ethylene (with oxygen from air) and dehydrochlorination, one uses the same process, except that pure oxygen is used in place of air, and three plants use the direct chlorination of ethylene and dehydrochlorination [6]. The main sources of vinyl chloride emissions in plants using the hydrochlorination of acetylene are reactor condenser vent (continuous), scrubber vent (continuous), heavy ends storage vent (intermittent), and assorted fugitive sources [6]. The sources of emissions from the chlorination-oxychlorination of ethylene and dehydrochlorination processes using oxygen from air or using pure oxygen are the same. The main sources of emissions are those dis cussed in Section 2.1 and not repeated here, purification system vents (continuous), scrubber vent (continuous), loading area (intermittent), and vinyl chloride emissions from the purification process; but the loss per unit product produced is greater from the purification portion of the plant. The process using direct chlorination and dehydrochlorination is used in vinyl chloride production when the manufacturing facility has other uses for the hydrogen chloride by-Droduct. This process also uses ethylene dichloride in a dehydrochlorination process to produce vinyl chloride from ethylene dichloride. 7 COLORXTE 007300 Z.3 POLYVINYL CHLORIDE (PVC) Polyvinyl chloride is produced by a polymerization process from the vinyl chloride monomer. There are five processes used to polymerize the monomer: suspension polymerization is the most widely used and accounts for 78% of the U.S. plant capacity; dispersion polymerization (i.e., emulsion) accounts for 13%; bulk polymerization for 6% [4]. The latex polyvinyl chloride is produced by dispersion polymerization and is sold and transported in a water suspension; solution polymeri zation is adaptable to a continuous process for copolymers and is used by one company in the U.S. [4]. Emissions may occur at any point of the processes, such as storage, reactors, strippers, mixers, weighers, blenders, recovery systems, inprocess wastewater, loading facilities, etc. [6]. 8 COLOR!TE 007301 3.0 PROCESS FLOW DESCRIPTION AND EMISSION POINT IDENTIFICATION The manufacturing processes covered in the vinyl chloride emissions Final Standard are discussed in this chapter with the aid of process flow diagrams. The emission points covered in the Standard are listed and keyed on the corresponding flow diagrams for each process. 3.1 ETHYLENE DICHL0RIDE--0XYCHL0RINATION* One of the major processes for ethylene dichloride (see Section 3.4) production in the U.S. is oxychlorination which involves the reaction of oxygen and hydrogen chloride with ethylene [4]. The oxygen may be introduced into the process in concentrated form or by an air stream as shown in Figures 3.1(a) and 3.1(b), respectively [6]. The two processes are similar and will be discussed together since the emission sources are identical. The generic reaction equation for the process is given by CH2 = CH2 + 1/2 02 + 2HCL---------------- CICHg - CHjCl + H20. (3-1) This reaction takes place in the reactor, shown in Figures 3.1(a) and 3.1(b), at elevated temperature. The oxychlorination process typically exhibits a 98" conversion to ethylene dichloride per pass from the reaction represented in Eq. (3-1) [6]. The raw materials (ethylene, hydrogen chloride and oxygen/air) are made to pass through a catalyst. In the presence of oxygen, the catalyst concentrates the- ethylene and chlorine allowing for the reaction, Eq. (3-1), to take place at a lower temperature. Because the reaction is highly exothermic, a water flow over the reaction tube surface is required to control the temperature in the reactor. The result is that steam evolves at the exit port of the water jacket. *While this process is usually identified as oxychlorination, it is more precisely an oxyhydrochlorination process. 9 COLORITE 007302 COLORITE 007303 Figure 3.1(a). Ethylene dichloride process flow diagram using oxychlorination which involves the reaction of oxygen and hydrogen chloride with ethylene. COLOR!TE 007304 Figure 3.1(b). Ethylene dichloride process flow diagram using oxychlorination which involves the reaction of oxygen present in air and hydrogen chloride with ethylene. Table 3.1: Most probable emission points identified by the keys in Figures 3.1(a) and 3.1(b) in the oxychlorination process for ethylene dichloride. Position Name | Key > Frequency ------------------------------------------- i--------------------- !---------------------------- Reactor Vent 1 j Continuous Reactor Refrigerator Condenser Vent 2 Continuous Wastewater from j Water Wash Column 3 Continuous Wastewater from Wash/Crude Storage i Wash Water Stripper Vent 4 5 Continuous Continuous , Wash Water Stripper Storage Vent Wash Crude Product Storage Vent Light-Ends Column Condenser Vent 6 7 ,, j Continuous Intermittent Continuous Finishing Column Vent Light-Ends Purification Column Vent Refined EDC Storage Tank Vent Heavy-Ends, Tar-Removed Column Vent 9 j ,n n 12 Continuous Intermittent Intermittant Continuous Heavy-Ends Storage Tank Vent 13 Intermittent Tar Storage Tank Vent Fugitive ' 14 Entire Plant Intermittent Intermittent/ Continuous ! 12 COLORITE 007305 The gas stream that exits the reactor contains ethylene dichloride in the form of a gas. In the process represented in Figure 3.1(a) this stream also contains ethylene, hydrogen chloride and oxygen; while that in Figure 3.1(b) is similar except that air rather than oxygen is present. The stream is passed through a hot water wash column to remove impurities from the ethylene dichloride, to maintain its gaseous state, and to introduce water vapor. The ethylene dichloride gas and water vapor stream are passed through a cold water condenser orior to entering the separation tank. The separated, purified ethylene dichloride is then transported to a storage tank [6]. The water from the hot water wash and from the separation tank is passed through a stripping column to separate ethylene dichloride and waste by-products. The separated ethylene dichloride is passed on to the EDC storage tank, while the waste by-products are disposed. The most probable emission points are listed in Table 3.1 and identified by the keys shown in Figures 3.1(a) and 3.1(b). The emission sources are listed to correspond with operational steps in the generalized hydrochlorination process. Fugitive emission sources (pumps, pump main tenance valves, pressure relief valves, samplers, etc.) are not identified because the locations of these sources are unique for each plant. The emission from any point, whether listed or not, depends on the operating condition (batch or continuous, reaction efficiency, etc.) of the plant at the time of inspection and in some cases on the immediate past history of operating conditions [6], 3.2 VINYL CHLORIDE 3.2.1 Hydrochlorination of acetylene In this process vinyl chloride monomer is produced by the hydro chlorination of acetylene. The reaction occurs between hydrogen chloride and acetylene at 85-141C in the reactor shown in Figure 3.2, in the presence of a catalyst, mercuric chloride on activated carbon [6]. The reaction is governed by the equation HC = CH + HCL ---------- H2C = CHCL. (3-2) 13 COLORITE 007306 VENT Figure 3.2. Vinyl chloride monomer process flow diagram using hydrochlorination of acetylene. COLOR!TE 007307 The reaction typically exhibits a 90% conversion to vinyl chloride per oass. As a result, acetylene and hydrogen chloride gases exit the reactor as well as vinyl chloride. These gases are compressed, cooled and pumped to a purification system consisting of a light-ends distillation column and a heavy-ends distillation column. In the light ends distillation column acetylene is separated and transported back into the reactor for another pass. Other volatile gases are passed to the vent gas reactor and vent condenser. Table 3.2 lists the main products and their boiling point temperatures. This shows that acetylene has the lowest boiling point temperature and is separated easily from the other compounds. Table 3.2: Hydrochlorination of acetylene product gases and their boiling point temperatures. Ga $ Symbol Boiling Point Temperature Acetylene Hydrogen Chloride Vinyl Chloride HC i CH HC1 H2C - CHC1 - 84C .. .................................... + noc L. - 13C Ethylene Dichloride ClCH2 - CH2C1 + 83C The reaction represented in Eg. (3-2) is exothermic. Therefore, the heavy-ends that are transported to the heavy-ends distillation column from the light-ends distillation column will contain some poly merized material as well as vinyl chloride monomer. The vinyl chloride is distilled, subsequently liquefied and placed in a storage tank. The heavy-ends are transported to a temporary storage tank and ultimately incinerated. 15 COLORITE 007308 The most probable emission points are listed in Table 3.3 and identified by the keys shown in Figure 3.2. The emission sources are listed to correspond with operational steps in the generalized hydro chlorination process. Fugitive emission sources (pumps, pump maintenance valves, pressure relief valves, samplers, etc.) are not identified because the locations of these sources are unique for each plant. The emission from any point, whether listed or not, depends on the operating conditions (batch or continuous, reaction efficiency, etc.) of the plant at the time of inspection and in some cases on the immediate past history of operating conditions. Table 3.3: Most probable emission points identified by the keys in Figure 3.2 in the hydrochlorination of acetylene for viny' chloride monomer. Position Name Key Frequency Reactor Condenser Vent 1 Continuous Scrubber Vent ! VCM Condenser Vent 2 Continuous 3 Continuous VCM Storage Vent Heavy-Ends Storage Vent 4 Continuous * 5 Intermittent VCM Loading Vent Fugitive 6 11 - 1 , Entire ! 1 Plant Intermittent Intermittent/ Continuous fr Intermittent is used in the sense that if the rate of incineration is greater than the rate of storage, emission is probably non-existent. When the reverse is true, emission will be continuous. 16 COLORITE 007309 3.2.2 Dehydrochlorination of ethylene dichloride The production of vinyl chloride monomer by dehydrochlorination (removal of hydrogen chloride) involves the thermal dehydrochlorination of dry ethylene dichloride. Thermal dehydrochlorination is sometimes referred to as a cracking process. Vinyl chloride results when ethylene dichloride is placed in a cracking furnace at approximately 510C [4], For most efficient operation, the furnace is packed with a catalyst such as pumice or charcoal. Typically, the conversion efficiency per pass is 94 to 97%. The cracking of ethylene dichloride to obtain vinyl chloride monomer is used exclusively in those processes in which ethylene dichloride is the intermediary crude. This includes the chlorination-oxychlorination of ethylene process using concentrated oxygen or oxygen in air (described in Section 3.1.1) and the direct chlorination of ethylene process (not covered in the Standard). The cracking process is the same regardless of the process used to obtain ethylene dichloride. The reaction is represented by the equation ClCH2 - CH2C1 -------- H2C = CHC1 + HC1. (3-3) In an integrated ethylene dichloride-vinyl chloride plant, where the oxychlorination process is employed to produce ethylene dichloride, the hydrogen chloride by-product in Eq. (3-3) is returned to the ethylene dichloride reactor as an input crude shown in Eq. (3-1) [4,6]. Figure 3.3 shows the process flow diagram [6], The ethylene dichloride is transported as a liquid (boiling point +83C, see Table 3.2) and is vaporized completely prior to entering the cracking furnace. The hydrogen chloride is removed in the cracking furnace. The gas flow exiting the cracking furnace consists of mainly vinyl chloride, but also present are ethylene dichloride, hydrogen chloride and other hydrocarbons. The 17 COLORITE 0073X0 RECYCLED TO oo Figure 3.3. Vinyl chloride monomer process flow diagram using dehydrochlorination of ethylene dichloride. COLORITE 007311 quenching column uses liquid ethylene dichloride to liquify the ethylene dichloride and the hydrogen chloride while the gaseous vinyl chloride passes on to the partial condenser and condenser. Hydrogen chloride gas also flows along with the vinyl chloride since hydrogen chloride is not very soluble in ethylene dichloride at the quenching temperature. The liquid ethylene dichloride is transported to the washed crude storage in the ethylene dichloride section of the plant and ultimately transported back to the cracking furnace for another pass. The vinyl chloride and hydrogen chloride gases are cooled, compressed and pumped to a purification system consisting of a light-ends distillation column and a heavy-ends distillation column. In the light-ends column the vinyl chloride is separated from the hydrogen chloride. The hydrogen chloride is recycled to the ethylene dichloride reactor if the oxychlorination process is used. Other volatile by-products are passed to the vent gas reactor. The reaction represented in Eq. (3-3) is exothermic [6]. Therefore, the heavy-ends and the light-ends distillation columns will contain some polymerized material as well as vinyl chloride monomer. The vinyl chloride is distilled, subsequently liquefied and placed in a storage tank. The heavy ends are transported to a temporary storage tank and ultimately incinerated. The most probable emission points are listed in Table 3.4 and identified by the keys shown in Figure 3.3. The emission sources are listed to correspond with operational steps in the generalized dehydro chlorination process. Fugitive emission sources (pumps, pump maintenance valves, pressure relief valves, samplers, etc.) are not identified because the location of sources is unique for each plant. The emission from any point, whether listed or not, depends on the operating conditions (batch or continuous, reaction efficiency, etc.) of the plant at the time of inspection and in some cases on the immediate past history of operating conditions. 3.3 POLYVINYL CH-.ORIDE 3.3,1 Suspension polymerization The suspension polymerization process is the most common process for polyvinyl chloride production [4], The resin produced is sometimes referred 19 COLORITE 007312 Table 3.4: ____ Position Name Most probable emission points identified by the keys in Figure 3.3 in the dehydrochlorination of ethylene dichloride for vinyl chloride monomer. . Key Frequency 1 Recycling of EDC to Washed Crude Storage Light-ends Distillation Column Vent Heavy-ends Distillation Column Vent Heavy-ends Storage Vent VCM Storage Vent Light-ends Column Waste Water Quench Column Vent Fugi tive 1 2 3 4 5 6 7 Entire PI ant Intermittent Intermittent Intermittent . Intermittent Intermittent Continuous Continuous Intermittent/ Continuous 20 COLORITE 007313 to as suspension resin. Polymerization of vinyl chloride requires the mixing of weighed amounts of vinyl chloride (weighed amounts of a comonomer where desired to change the polyvinyl chloride properties), catalyst, water and suspending agents (soap) [4]. The raw materials are mixed in a clean glass or stainless steel lined reactor. Air is removed from the reactor by a steam jet or vacuum pump. Reaction temperature is controlled by either cooling or heating, depending on the details of the process used. The reaction is initiated by the catalyst. As the reaction proceeds, polyvinyl chloride is produced in particle form. Agitation is employed to prevent a slurry from agglomerating in the reactor and the suspending agent disperses the vinyl chloride droplets. The polymerization process is allowed to continue until &5 to 90% of the vinyl chloride has polymerized; this requires approximately 6 hours [4], At this time, the reaction is stopped because the process is not as controllable. If allowed to continue beyond this point, the product becomes increasingly less uniform with respect to molecular weight and, therefore, the physical properties are less uniform. The vinyl chloride residue is in vapor form in the reactor, dissolved in water, and/or trapped in the polyvinyl chloride granules. Figure 3.4 shows the process flow diagram for suspension polymerization [4] Vinyl chloride is supplied to the weighing station from the vinyl chloride source (plant or tank car) and from the recovery system. The comonomer (if required) for a batch is weighed in its own weighing station. In addition to the weighed quantities of vinyl chloride monomer and comonomer, the initiator catalyst and suspending agent are transported to the reactor. The polymerization take^place at elevated temperatures and at a pressure 1,1 in the range of 5.1 to 6.8 atmospheres. The-sesidue from th^j^eactor is * transferred to the^stSpper tarU<^jdeT^thej^ s removed""and ' tr>ir$ported t&^flie vinyl chloride recovery system. The contents of the reactor (polyvinyl chloride granules, gases, water and initiator catalyst) are subjected to a stripping process to remove unreacted vinyl chloride. The stripping process may take place in the reactor or the contents may be transferred to a stripper vessel where stripping takes place. Vinyl chloride is stripped by the application of heat alone or by the application of the combination of heat and vacuum. The vinyl chloride and 21 COLORITE 007314 COLOR!TE 007315 Figure 3.4. Polyvinyl chloride process flow diagram using suspension polymerization. ,,i . * other volitized chemicals driven off are transferred to the recovery system. The vinyl chloride is obtained from the distillation column and passed on to the recovery system storage tank. The other gases are vented, and included will be some vinyl chloride. The remaining slurry is transported to a slurry blend tank where several batches are blended together to obtain a more uniform product [4]. The blended slurry is then pumped to a centrifuge where most of the water is removed. The wet polyvinyl chloride is then dried to remove remaining water and vinyl chloride. The resin is then transferred to storage or bagging stations. Table 3.5 lists the most probable emission sources for the suspension process with the key referred to Figure 3.4. The Keyt^ldenotes the recovery system vent through which the noncondensable gases are vented. The venting may take place manually or automatically and may be intermittent or con tinuous depending on the system design and the pressures employed. Keys'2 ^nd(3 are the weighing stations for the vinyl chloride monomer and comonomer. High pressure may build up at these stations, in which case the emission will contain vinyl chloride. The reactor emissions, Key;4j, may arise when a run-away condition develops and due to residual vinyl chloride vapor will always arise when the reactor manhole is opened for cleaning. A run-away condition typically develops in the event of a power failure, equipment failure and/or operator failure. When this occurs, the reaction for the batch may require a temporary or a complete termination either by automatic pressure valve release or by manual venting4]. Keys; 5Jtf6^n^T)are all associated with the stripping, blending and drying operations of the polyvinyl chloride slurry. In each of these operations vinyl chloride emissions may occur [4]. Keys^fT)andy_9yare points where emissions may occur in the storaqe and bagging of the resins, while Key 10)is the emission Doint from the waste water treatment process. 3.3.2 Emulsion (i.e., dispersion) polymerization The emulsion polymerization process uses equipment basically similar to that of the suspension process described in Section 3.3.1 [4]. The resin produced is sometimes called an emulsion resin. Emulsion resins can be polymerized at lower temperatures and at a higher rate than suspension 23 COLORITE 007316 Table 3.5: Most probable emission points identified by the keys in Figure 3.4 in the suspension polymerization process for polyvinyl chloride. POSITION NAME Vinyl Chloride Recovery Vent Vinyl Chloride Weighing Tank Vent Comonomer Weighing Tank Vent Reactor Vent and Opening Loss Stripper Vent Slurry Blend Tank Vent Centrifuge Vent Product Col lection Vent Silo Vent Waste Water Dryer Exhaust Bulk Loading Fugitive KEY FREQUENCY 1 Intermittent/Continuous 2 Intermittent 3 Intermittent 4 Intermittent 5 Intermittent 6 Intermittent 7 Intermittent 8 Intermittent 9 Continuous 10 Continuous n Intermittent 12 Entire Plant Intermittent Intermittent/Continuous 24 COLORITE 007317 resins. However, emulsion resins are also more sensitive to heat and shear stresses. When subjected to either or a combination of heat and shear stresses, the resultant changes in the resin's physical characteristics may make it unsuitable for use. The resin obtained from the emulsion process is of smaller particle size than obtained from the suspension process [4]. The emulsion process flow diagram is shown in Figure 3.5 and the cor responding keyed emission sources are presented in Table 3.6. A study of Figures 3.4 and 3.5 show that the processes are identical in the following ways: batch reactor process; water is used as the suspending medium to suspend liquid vinyl chloride; all process equipment is identical except for the dryer. In addition, the suspension process uses a centrifuge to aid in drying while the emulsion process does not. In the emulsion process soap and water are used as the emulsifier. The emulsion process differs from the suspension in the following ways: more soap is added to the slurry in the reactor which stablizes the monomer droplets and results in the absence of agglomerates; a spray dryer ; is used because it does not produce excessive temperature or shear stresses during the drying, while the rotary, flash, or fluidized bed dryer used in the suspension process may produce these stresses [4]. 3.3.3 Latex dispersion polymerization Latex resins are produced by the emulsion process [4]. The latex resin is polymerized vinyl chloride monomer suspended in water. It is sold and transported in this solution form. The process is identical in almost all ways to the suspension and emulsion processes. It differs in that there is no drying process step and more soap is added in the reactor than for the emulsion or suspension processes [4]. The result is a latex resin which is a colloidal suspension of polyvinyl chloride. The process flow diagram and most probable emission sources are shown in Figure 3.6 and Table 3.7, respectively. 3.3.4 Buik polymerization The bulk polymerization process is a batch process and consists of two polymerization steps [4]. In the pre-polymerization reactor there is liquid vinyl chloride in the presence of a polymerization initiator. The reactor is of a similar design to that used in the suspension process. The conversion to polyvinyl chloride from vinyl chloride is in the range 7 to 25 COLORITE 007318 COLOR!TE 007319 Figure .1.5. Polyvinyl chloride process flow diagram using emulsion (i.e., dispersion) polymerization. Table 3.6: Most probable emission points identified by the keys in Figure 3.5 in the emulsion (i.e., dispersion) polymerization process for polyvinyl chloride. POSITION NAME Vinyl Chloride Recovery Vent Vinyl Chloride Weighing Tank Vent Comonomer Weighing Tank Vent Reactor Vent and Opening Loss Stripper Vent Slurry Blend Tank Vent Spray Dryer Vent Product Collection Vent Silo Vent Process Water Bulk Loading Fugitive KEY FREQUENCY 1 Intermittent/Continuous 2 Intermittent 3 Intermittent 4 Intermittent 5 Intermittent 6 Intermittent 7 Intermittent 8 Intermittent 9 Continuous 10 Intermittent ..................................................... .......... .... ' " 11 Intermittent ........`"""i Entire Plant Intermittent/Continuous 27 COLORITE 007320 COLOR!TE 007321 Figure 3.6. Polyvinyl chloride latex process flow diagram using emulsion (i.e., dispersion) polymerization. Table 3.7: Most probable emission points identified by the keys in Figure 3.6 in the emulsion (i.e. dispersion) polymerization process for polyvinyl chloride latex. POSITION NAME Vinyl Chloride Recovery Vent Vinyl Chloride Weighing Tank Vent Comonomer Weighing Tank Vent Reactor Vent and Opening Loss Stripper Vent Slurry Blend Tank Vent Product Collection Vent Loading Station Fugitive KEY 1 2 3 4 5 6 7 8 - Entire Plant FREQUENCY Intermittent/Continuous Intermittent Intermittent Intermittent Intermittent Intermittent Continuous Continuous 1 Intermittent/Continuous 29 COLORITE 007322 12% [4]. This suspended polyvinyl chloride in liquid vinyl chloride is then transferred to a larger, high pressure, horizontal-type reactor. To this is added more liquid vinyl chloride and initiator. This reactor, sometimes called an autoclave, serves as the post polymerization reactor resulting in a reaction efficiency of approximately 85 to 90% [4]. The post polymerization reactor is more rugged and the agitation more vigorous than the pre-polymerization reactor [4]. The post-polymerization reactor must be cleaned after each batch; less frequent cleaning is required for the pre-polymerization reactor. The bulk process is similar in most aspects to the suspension process. Since there is no water or water vapor in the suspension, low temperature (-35C or -31F as opposed to 7C or 44.6F in the suspension and dispersion) condensers may be employed in the recovery system. Moreover, the drying operation is not needed, and no in-process waste water system is present [a]. The remaining monomer in the post-polymerization vessel may be removed by a number of processes. The monomer may be removed by vacuum alone. Another method is to introduce steam into the autoclave and the steam and released vinyl chloride are removed by vacuum. The steam-vacuum procedure may be repeated as many times as required to meet the emission standard. Those plants using the steam-vacuum process require wastewater stripping to bring the wast water in compliance with the emission standard. Whichever method is used, the recovered monomer is placed in a temporary holding tank and recycled back to the pre-polymerization reactor through a filter [4]. The process flow diagram and the most probable emission points are given in Figure 3.7 and the keyed emission points listed in Table 3.8 [4]. 3.3.5 Solvent polymerization The solvent polymerization product is considered a speciality product and is a small segment of the total PVC industry [5]. However, it serves a large number of important needs that usually involve thin PVC coatinas, such as for the food and beverage industries [5]. The early developed process is shown in Figure 3.8 and the more recently developed process is shown in Figure 3.9 [5], The processes are similar in nrost aspects, differing 30 COLORITE 007323 NITROGEN Figure 3.7. Polyvinyl chloride process flow diagram using bulk polymerization. COLORITE 007324 Table 3.8: Most probable emission points identified by the keys in Figure 3.7 in the bulk Polymerization process for polyvinyl chloride resin. POSITION NAME Pre-Polymerization Reactor Vent Post-Polymerization Reactor Vent Monomer Holding Tank Vent Recycle Condenser Vent Pressure Reducer Vent First Bag House Vent Second Bag House Vent Reject Vent KEY FREQUENCY 1 Intermittent 2 !1 Continuous i 3 | Continuous 4 ! Continuous 1 5 1 Continuous j 6 ; Continuous j 7 ] Continuous 8 Continuous | Fugitive Entire Plant Intermittent/ Continuous i 1 1 32 COLORITE 007325 COLORITE 007326 Figure 3.8. Early polyvinyl chloride process flow diagram using solvent polymerization. VCM ~w VAC Reactors Solvent V. N* VCM Recovery Liquid Products 4U^> Monomers and Storage Area ^ VCM ^---------------------------------------- - ^ Vinvl Acetate and Solvent Solvent and Vinyl Acetate o > _> ton 9 Silo V Packaging COLORITE 007327 Figure 3.9: More recent polyvinyl chloride process flow diagram using solvent polymerization. more in the technological developments of recent years than in the basic process operations. The early process is described below followed by the points of difference. The early process flow diagram is shown in Figure 3.8 and the cor responding source emission points listed in Table 3.9. The emission points in Table 3.10 correspond to the process flow diagram shown in Figure 3.9. The comonomers, initiators and solvents are continuously introduced into the reactor. The process flow shows that the VCM, comonomer and initiator are introduced into the reactor along with the solvent, usually n-butane [4]. The comonomer is almost always vinyl acetate. The continuous process provides a degree of turbulence among the constituents in the reactor that results in a copolymer conversion efficiency approaching 100%. A continuous copolymer stream is drawn off from the reactor and filtered. The filter cake is passed on to a flash evaporator where it is dried and the monomers recycled. The solvent is drained from the filter and recycled back into the reactor along with the recovered monomers. During any one pass of the solvent stream, some solvent is lost to the process; therefore a solvent make-up stream is also required [4], Vinyl chloride emissions from the reactor area in a continuous process are relatively lower than from batch processes [4]. There is some evidence to suggest that the vinyl chloride is more easily removed from the resin than in other polymerization processes. The most important difference between the processes is that the earlier process produced a slurry as a result of polymerization while the later process produces a solution. The stripping operation from solution gives a lower VCM concentration and typically produces less emissions. Moreover, inprocess water comes into contact with the polymerized material after the stripping operation. This avoids the requirement of an inprocess water stripper, and the inprocess water storage and treatment equipment should not give any emissions. Finally, drying is accomplished by means of hot air rather than flash evaporator. 35 COLORITE 007328 Table 3.9: Most probable emission points identified by the keys in Figure 3.8 in the solventpolymerization process for polyvinyl chloride and copolymers. POSITION NAME KEY FREQUENCY Receiving Tank Vent 1 VCM Storage Vent Reactor Vent Flash Evaporator (Stripper) 2 3 4 Solvent and Monomer Recovery 5 Bag Filter 6 Bag Filter-Screen Grinder Filter Grinder Filter-Screen 7 8 9 Storage Silo Bulk Loading 10, 11 12, 13 Fugitive Entire Plant Continuous Continuous Continuous Continuous Continuous Continuous Continuous Continuous Continuous Continuous Intermittent Intermittent/ Continuous 36 COLORITE 007329 Table 3.10: Most probable emission points identified by the keys in Figure 3.9 in the solvent polymerization process for polyvinyl chloride and copolymers. POSITION NAME ............... Reactor Monomer Condenser Resin Drying Silo Solvent and Vinyl Acetate Condensor KEY 1 2,3,4 5 6 7 FREQUENCY Continuous Continuous Continuous Continuous Continuous Fugitive Entire Plant Intermittent/ Continuous 37 COLORITE 007330 3.4 CLARIFYING NOTE ON THE BALANCED OXYCHLORINATION - DEHYDROCHLORINATION PROCESS Oxychlorination and direct chlorination are the two major processes used for ethylene dichloride production [4]. EDC plants operate a balanced process which consists of a vinyl chloride plant and a direct chlorination plant operated adjacent to and in coordination with the oxychlorination plant [4]. The Standard covers the oxychlorination, but not the direct chlorination process. However, this strict interpretation is not the intent of the Standard [4], The Standard promulgated on October 21, 1976 covers any and all portions of any process that receives, transports or processes recovered ethylene dichloride from the oxychlorination plant, even though these processes may also be used for materials from the direct chlorination process. A block diagram of the balanced process is shown in Figure 3.10. Typically ethylene dichloride refining is common to both the direct chlori nation and to the oxychlorination plants. Therefore, the ethylene dichloride crudes are refined through the same equipments. In Figure 3.10 the EDC refining equipment is shown to be part of the EDC oxychlorination plant. The crude produced from the oxychlorination plant may contain vinyl chloride monomer. Therefore, the common receiving point of the monomer and all down stream parts of the direct chlorination plant are subject to the Standard. That the crude from the oxychlorination plant may contain vinyl chloride arises because of recycled ethylene dichloride and because the recycled hydrogen chloride used is a by-product of the monomer cracking. In Figure 3.10 this is shown by the flow of HC1 and EDC from the dehydrochlorination to the oxychlorination reactors and EDC refining, respectively. While the vinyl chloride monomer is present in small concentrations in HC1 after having been subjected to a stripping process and to even lower concentrations if the ethylene dichloride crude is also stripped of the monomer, equipment malfunctions may give rise to significant monomer concen trations in the crude produced in the oxychlorination plant. In the absence of equipment malfunctions, the monomer is easily stripped during the HC1 and EDC stripping operations. This is true because the monomer boiling point is -13C while it is +110C and +83C for HC1 and EDC, respectively (Table 3.2). 38 COLORITE 007331 VCM Figure 3.10: Block diagram of a balanced oxychlorinationdehydrochlorination process. 39 COLORITE 007332 The economics of the PVC industry demands that large producers of EDC employ the balanced oxychlorination - dehydrochlorination type plants [4]. From Eq. (3-3) it is seen that for each vinyl chloride molecule produced in the cracking of EDC, one HC1 molecule evolves as a by-product. This forms the HC1 stream in Figure 3.10. However, in the oxychlorination reaction, governed by Eq. (3-1), two HC1 molecules are required for each EDC molecule produced. The HC1 by-product provides at a maximum one-half the EDC required by the cracking furnace. Therefore, the direct chlorination process must supply at least one-half of the EDC required in a balance type plant. Presently, 95% of the EDC annual production rate is produced in balance type plants. 3.5 PHOTOGRAPHS OF EQUIPMENT The technology employed in the EDC, VC and PVC industries is of a relatively higher technical level than Inspectors normally encounter in other plant inspections subject to the Clean Air Act. In addition, specific processes differ for the same general product class from plant to plant. In many cases, if not in most, the equipments used were constructed according to unique specifications. The operation of the equipments may be different, in which case the materials, size, shape and their placement in the plant may also be different. Typically, the Inspector is not able to draw from experiences gained of previous inspections to the degree that is commonly done in other industries. To assist the Inspector in conducting complete and efficient plant inspections, photographs of equipments are shown in Figures 3.11 to 3.40 from which VC emissions are more likely to occur. From photographs of general types of equipment such as reactors, strippers, storage vessels, etc., it will be less difficult to determine the function of more specialized equipment types. No attempt has been made to present photographs of dif ferent types and sizes of equipments because this would require in excess of 200 photographs. Table 3.11 gives the figure numbers and the equipment category of those selected for reproduction in this Manual. 40 COLORITE 007333 Table 3.11: Figure number of photographs and corresponding equipment category. Figure Number 3.11 to 3.15 3.16 to 3.18 3.19 to 3.24 3.25 and 3.26 3.27 to 3.32 3.33 and 3.34 3.35 3.36 and 3.37 3.38 to 3.40 Equipment Category Reactors Strippers Storage Vessels Dryers Distillation Vessels RD/SRV Double Mechanical Seal Railroad tank car loading and unloading Incinerators 41 COLORITE 007334 Figure 3.11 Urge capacity VC reactor using the oxychlorination process is located in the tall vessel. The oxychlorination manual vent is the open ended pipe, rising above the top of the reactor on the left side. Figure 3.12 The tops of medium capacity, side-by-side PVC reactors using the emulsion process may be seen. Individual RD/SRV and manual vents may also be seen mounted on top of each reactor. 42 COLORITE 007335 COLORITE 007336 Figure 3.13 Small size PVC reactor with cover removed to clean reactor in preparation for next polymerization run. Prior to removing cover the PVC and water solu tion were removed and transferred to stripper. "Elephant trunk" is placed through the opening to vent vinyl chloride gas through "trunk" to the recovery system in order to be in compliance with emissions standard reactor opening loss. This type reactor is used in the suspension, emulsion and latex polymeri zation processes. Figure 3.14 Medium capacity side-by-side PVC reactors using solvent process. Individual RD/SRV and manual vents connected to vinyl chloride recovery system may also be seen. Lower portion of reactors comprise the heating elements to raise the contents to the tem perature required for polymerization. Figure 3.15 The top of medium capacity, side-by-side PVC reactors using solvent process. Foreground shows a motor valve for emergency venting through to the VC monomer recovery system. Figure 3.16 The top of a small capacity PVC stripper vessel used in the suspension, emulsion and latex processes. The RD/SRV vent is connected to the recovery system. 44 COLORITE 007337 COLOR!TE 007338 Figure 3.17 A stripper column is shown that removes VC from a PVC-varnish solution resulting from the solvent polymerization process. Figure 3.18 A typical wastewater stripper column is shown that may be found in EDC, VC and PVC plants. 46 COLORITE 007339 Figure 3.21 Spherical, above ground vinyl chloride storage tanks. RD/SRV vents may be seen perched on top of sphere. COLORITE 007340 Figure 3.23 Wash water stripper storage tank with its vent mounted at the top of the tank in an EDC-VC plant. COLORITE 007341 p A Figure 3.26 PVC emulsion resin spray drying takes place in the cylindrical building. Large diameter feed pipe, seen at the left of the dryer, carries resin to the top of the dryer. Dust collector is perched at top of dryer. 49 COLORITE 007342 Figure 3.27 An EDC light ends distillation column is shown with its RD/SRV vent. so COLORITE 007343 COLOR!TE 0 0 7 3 4 4 Figure 3.29 The top of an EDC light ends column condenser showing the RD/SRV vent. Figure 3.30 In the foreground the liqht ends dryer regeneration vessel is shown in balanced EDC-VC plant. COLORITE 0 0 7 3 4 5 Figure 3.31 An EDC finishing column with the RD/SRV and its vent mounted on the top. (In some installations the finishing column performs the function of the light ends distillation column while in others it encompasses the functions of both the light and heavy ends columns.) Figure 3.32 Foreground shows the insulated piping connected to the reactor refrigerator condenser vessel in a balanced EDC oxychlorination plant. COLOR!TE 0 0 7 3 4 6 Figure 3.33 Typical RD/SRV assembly with the vents from each connected to a manifold of the recovery system. Figure 3.34 Dual RD/SRV vents mounted at top of spherical vinyl chloride storage tank. COLORITE 0 0 7 3 4 7 Figure 3.35 Typical pump and double mechanical seal Figure 3.36 Top connections on railroad tank car shown with flexible hose attached for VC loading. Smaller diameter flexible hose in the foreground is connected to recovery system. Figure 3.37 Railroad tank car loading platform shown with pipe rack support for flexible hose VC feed and recovery system in VC plant. Figure 3.38 Incinerator and scrubber stack showing the steam exhaust stream in an EDC-VC plant. 55 COLOR!TE 007348 I Figure 3.39 Incinerator and stacks of a PVC plant showing the platform (center stack) on which stack samples are taken to determine VC emission con centration. Figure 3.40 Typical incinerator in a VC and PVC plant with bottom portion of stack shown at the right. 56 COLORITE 007349 4.0 LEAK DETECTION MONITORING INSTRUMENTATION, RECORDS, AND REPORTS A requirement of the Standard is that an EPA approved leak detection and elimination program be operational. This requirement includes an installed continuous leak detection monitoring system, routine leak detection monitoring with a portable hydrocarbon detector and a leak elimination plan. The Standard also requires recordkeeping (recording and retention for at least two years) of data relating to leaks detected by one of several ways. Conducting a meaningful inspection for the determination of com pliance of vinyl chloride emissions in typical EDC-VCM-PVC plants presents a number of unique challenges. The first is that the technology used requires a relatively high degree of expertise. The second is that the plant area, from raw materials to finished product shipping, is measured in acres rather than in square feet and may extend from below ground level to several hundred feet above ground. Added to these is the difficulty in making a definitive determination of some equipments with respect to the specific process(es) or function(s) they serve. Therefore, the NESHAPS Inspector needs to resort to complementary methods, in addition to normal inspection procedures, to make a determination of compliance. To circumvent these unique challanges, a NESHAPS inspection, of necessity, will need to rely on the in-plant continuous monitoring instrumentation and the records obtained therefrom. In this chapter, the typical continuous monitoring system, the resultant records and reports are described and discussed. All are required to be in compliance. 4.1 LEAK DETECTION MONITORING INSTRUMENTATION The standard promulgated on October 21, 1976 requires that continuous monitoring detection and measurement instrumentation be permanently installed in a plant in which vinyl chloride may be emitted to the atmosphere. 57 COLOR!TE 007350 Typically, the instrumentation consists of a vinyl chloride or hydrocarbon measurement instrument, mini-computer, computer program for data acquisition and data reduction, data terminal and assorted 1/4" stainless steel tubing, solenoid valves, vacuum pump?, etc. The measurement instrument may be a gas chromatograph or, if the owner/operator assumes that all hydrocarbons measured are vinyl chloride, an infrared spectrometer or flame ion detector or an equivalent or alternative method. The monitoring instrumentation is based on area (i.e., volume) sampling. However, some plants will also monitor fugitive emission sources such as pump seals, agitator seals, couplings, etc. The typical instrumentation system uses one measurement instrument with a system of tubing that serves to draw air samples from an area of the plant into the detector air sample chamber where a measurement of vinyl chloride concentration is made. The concentration value is then transmitted to the computer memory to be printed out on the computer terminal at a later time. The time interval between measurements is 1 to 3 minutes. Usually, each monitoring point is measured in sequence -tfnd the sequence K unchanging. When all points associated with a measurement instrument have been measured for vinyl chloride emissions and transmitted to the computer memory, the computer program provides for each measurement to be printed in tabular form. Each measurement is identified with respect to the time of the measurement and location within the plant. Depending on the computer program, average vinyl chloride emissions for each point on the basis of shift, day, week and month may also be printed out and become a part of the record. In some instances, the data terminal does not print any measurement made unless the vinyl chloride concentration is greater than some defined level, such as 5 ppm. Plants that employ this system will record each measurement for each point on the measurement instrument's printer. It too becomes a part of the required recordkeeping. Usually the concentration is measured at each point at least every 25 minutes. Each plant sets its own threshold level for the purpose of defining a leak. In most cases, two consecutive measurements from the 58 COLORITE 007351 same monitoring point equal to or greater than the concentration threshold value are used in the definition of a leak. The concentration threshold level is the definition of a leak for the leak detection monitoring system. This definition requires the approval of the EPA Administrator and it is set at a level compared with the vinyl chloride background concentration. When two consecutive measurements ata point indicate a leak, plant personnel assigned to the "Leak Detection Patrol" investigate with a portable instrument the region of the plant in which the monitoring point is located. The monitoring system is required to be calibrated daily by one of two methods, described in 61.68(c). Someplants reserve one of the points in the sequence of point measurements for calibration. Thus, the instrument is calibrated in each sequence of measurement. Some plants may use more than one measurement instrument when a large number of points are being monitored. The number of points under observation by an instrument ranges from 9 to 19, while the total number in a plant ranges from 9 to 76.' However^ EPA approval is required with If respect to the position and minimum number of points. A schematic of a monitoring system is shown in Figure 4.1 for which there are n-air sample inlets. When a solenoid valve is activated, it allows an air sample to be drawn by a vacuum pump from a point in the plant into the detector air sample chamber. The pump operates continuously, evacuating the manifold of the previous air sample so as not to influence the vinyl chloride concentration measurement of the next air sample. In addition, the volume of the air sample drawn prior to actual measurement is sufficient to effectively purge the detector air sample chamber of any residue from previous air samples. The switching of the solenoid valves may be accomplished by one of two methods; a mechanical or electronic timer where the sequence is unchanging; a computer-controlled system where the sequence may be changed according to a program. Those installations using timers usually activate one solenoid at any one time. Computer-controlled systems may have sophisticated programs where one or more solenoids may be operated to more quickly assess the emission(s) in one or more plant areas. 59 COLORITE 007352 Sample Inlet Probe Location 1 Valve Figure 4.1: Schematic diagram of continuous monitoring system for vinyl chloride emissions. 60 COLORITE 007353 4.2 LEAK DETECTION MONITORING RECORDKEEPING The owner/operator is required to record data, retain the records on location for a minimum of two years and to make available, upon request of an EPA representative, those records obtained from the continuous Joafr* detection monitoring system. Specific data of the detected leaks will be recorded and retained which pertain to the location within the plant, vinyl chloride concen tration and the date and approximate time of measurement. 4.3 ROUTINE LEAK DETECTION AND RELIEF DISCHARGE RECORDKEEPING The owner/operator is also required to record data, retain the records on location for a minimum of two years and to make available, upon request of an EPA representative, records obtained of leaks detected during routine monitoring with a portable hydrocarbon detector and for relief discharges from reactors. Specific data of each leak detected during routine monitoring will be recorded and retained which relate to location within the plant, vinyl chloride concentration, date and time of measurement, cause of each leak, and the action taken to repair or eliminate each leak. Specific data of each relief discharge from reactors will be recorded and retained which pertain to the source, nature and cause of the discharge, the date and time of the discharge, the approximate total vinyl chloride loss during the discharge, the method used for determining the vinyl chloride loss, the action that was taken to prevent the discharge, and the measures adopted to prevent future discharges. In addition, a daily operating record for each reactor, including pressures and temperatures, will be recorded and retained. 61 COLOR!TE 007354 5.0 INSPECTION PROCEDURES AND INSPECTION FORMS The plant's compliance with the Standard will be determined through periodic inspections by an EPA representative, in addition to data required in the semi-annual reports on stripping and reactor opening loss and on emission tests. The inspection entails a physical inspection as well as a review and assessment of the plant's records. In the following presentation, a number of forms have been developed for the purpose of aiding the inspector to make a definitive inspection in the most efficient and expeditious manner. If the forms are used properly, all major facets in the standard will be covered. In fact, the questions and statements that make up the forms mirror the major facets of the Standard. Moreover, the questions and statements have been couched so that answers and responses may elucidate or reveal noncompliance conditions. Ideally, all questions and statements on all forms will be answered completely during the regular or periodic inspections. However, some plants may excel in complying with one or more sections of the Standard. An inspector having become aware of this fact may elect not to pursue questions and statements relating to that portion of the Standard. The frequency of the inspections is not recommended in this document, because this is a function of personnel being available on a regular basis. Besides, experienced inspectors are the best judge of the required inspection frequency. The forms described below are self explanatory and provide a sequence that an inspector may choose to follow during an inspection. 5.1 SUMMARY OF COMPLIANCE STATUS This form serves to summarize on one page the status of the plant with respect to the major facets of the Standard. The questions and statements serve to determine where emissions occur and at what levels, the emission control devices used for each emission source and the waivers that have 62 COLORITE 007355 been issued. This form will be most useful to personnel of the Enforcement Division having responsibility to make a determination of compliance. 5.2 CHECKLIST This form contains 29 question and statements requiring responses. They are grouped under the following categories: General, Leak Detection Monitoring System, Stack Emission Monitoring System, Portable Instrument, Leak Detection and Elimination, Discharges to the Atmosphere, Fugitive Emissions, Reactors and Furnaces, Control Devices, Stack Emissions, Inprocess Wastewater and Reactor Opening Loss. The responses require the inspector, in most cases, to actually view and verify. In the case of instrumentation, there is provision for the inspector to determine the accuracy of the instrumentation and/or system calibration. It is strongly recommended that a calibration be performed on one or more monitoring points of the leak detection monitoring system. In doing so, the inspector automatically checks the integrity of those monitoring points. It may be more practical for the plant's personnel to provide the responses to No. 23 rather than the inspector. However, the inspector should be present during the time the sample is being prepared and data obtained. It is strongly recommended that the inspector observe in its entirety the plant's procedure to determine the reactor opening loss. 5.3 ON REVIEW OF RECORDS Due to the maturity of the PVC industry and the economic environment that results, EDC, VCM and PVC are typically large capacity plants. One plant may cover many acres and extend up to 100 to 200 feet above ground level as well as below ground level. It may be physically impossible for an Inspector to Inspect all parts of a plant in person. It is strongly recorrmended that the central point of the Inspector's focus be placed on the records that the plant is required to maintain, and particularly the leak detection monitoring system. Therefore, the review and assessment of the plant's records is an important aspect of every inspection. The items in the form provided for the review of records is designed to elucidate compliance at the major or most probable emission points. 63 COLOR!TE 007356 5.4 PRE-TEST EQUIPMENT CHECKLIST FOR STACK EMISSION TEST The stack emission test required In the Standard, Test Method 106, is very clear and precise on the equipment and apparatus to perform the test. This form is designed to ensure that both inspector and plant personnel are reminded of the entire equipment and apparatus list required. It also serves to document alternative or equivalent equipment or procedures. 5.5 EQUIPMENT CHECKLIST FOR VINYL CHLORIDE CONCENTRATION IN INPROCESS WASTEWATER, RESIN, SLURRY, WET CAKE AND LATEX SAMPLES Test Method 107 is also clear and precise on the equipment and apparatus required to analyze samples in a head space analyzer. This form is designed to aid the inspector and plant personnel in conducting analytical tests that conform with the Standard. 64 COLORITE 007357 COLOR!TE 0 0 7 3 5 8 JUmiR. 01 CG,..1 L. nNl.' S ,ATo'J- (SCS) NESHAPS INSPECTION FOR VINYL CHLORIDE EMISSIONS COMPLIANCE - PARAMETER OR ITEM Emission Source 1 Where ducted? Atmosphere or Control Device 2 Emission Control Device Description In Existence Date to be Installed 3 Frequency of Emissions Cont./Intermit./Emergency 4 Applicable Regulations 5 Emission Standard ppm by vol. or ppm by wt. or gm. per Kg. 6 Estimated Emissions ppm by vol. or ppm by wt. or gm. per Kg. 7 Monitoring Requirements 8 Waiver Applications 9 Date Waiver Issued 10 Remarks on Compliance Emissions & Emission Tests Waivers Recordkeeping -- Ir-- ,-Jctii.,. Inspector's Name Finn's Name_____ ~ Finn's Address EMISSION SOURCE STATUS CHECKLIST NESHAPS INSPECTION FOR VINYL CHLORIDE EMISSIONS COMPLIANCE Inspection Date_________________ Inspector's Name________________ GENERAL 1. Firm's name______________________________________________________ 2. Firm's address___________________________________________________ 3. Process designation/product identification (check one or more): EDC: oxychlorination Q balanced N.A. VCM: Q hydrochlorination dehydrochlorination N.A. PVC: suspension Q dispersion latex N.A. bulk solution 0 copolymer 4. Rated and average annual reactor/cracking capacity EDC: Design; Normal Max.; Actual VCM: Design; Normal Max.; Actual PVC: Design; Normal Max.; Actual LEAK DETECTION MONITORING SYSTEM 5. Permanent leak detection monitoring system [61.65(8)(1)J a) Total number of points monitored: b) Number of measuring instruments:j Type(s) c) Time Interval to cycle all points: d) Definition of a Leak: e) Lower detection limit (LDL) of instrument________________________ f) Measurement sequence: Q unchanging Q program controlled g) Data reduction (check one or more) O none hourly shift daily Q weekly monthly Q other (specify) 66 COLOR!TE 007359 6. Calibration of leak detection monitoring system [61.65(8)(iii)]: a) List instruments and data from monitor points which were tested for calibration. (Attach sheets if more space is required.) Date & Time Location in Plant Instrum. Cal. VCM Instrum. VCM Percent Action Iden. Concen. Concen. Deviation Required b) Calibration method used (check one or more) None Paragraph 61.65(8)(1ii)(A) - Test Method 106 - 5.2.1, 5.2.3 Paragraph 61.65(8)(1ii)(B) Other STACK EMISSION MONITORING SYSTEM 7. Emission (i.e., stack) monitoring system / a) List emission (i.e., stack) sources Location in Plant Ducted Processes Continuous or Sampling, Other Number of Monitoring Points b) Lower Detection Limit (LDL) of instrument 67 COLORITE 007360 8. Cal bration of emission monitoring system: a) List instruments and data from monitor points which were tested for calibration. Date & Time Location in Plant Instrum. Cal. VCM Instrum. VCM Percent Action Iden. Concen. Concen. Deviation Required PORTABLE INSTRUMENT 9. Calibration of portable hydrocarbon detector(s): a) List instrument(s) which were tested for calibration Date & Time Location in Plant Instrum. Cal. VCM Instrum. VCM Percent Action Iden. Concen. Concen. Deviation Required b) Calibration method used (check one or more) None Paragraph 61.65(7)(iii)(A) - Test Method 106-5.2.1, 5.2.3 Paragraph 61.65(7)(iii)(B) Other 68 COLORITE 007361 LEAK DETECTION AND ELIMINATION 10. Leak detection and elimination process [61.65(8)] a) What is the frequency of the "Leak Detection Patrol"?. b) Check typical points and equipment which are required to be routinely checked for leaks by the "Leak Detection Patrol" in the following: reactor cracking furnace stripper EDC purification recovery system mixing weighing holding tank separation tank blending tank storage - raw storage - finished storage - heavy storage - water light ends col. heavy ends col. wastewater col.. water wash col. water quench col. wash water stripper dryer condensers RD/SRV Others: control devices pumps compressors agitator(s) loading unloading flanges valves filter strainers centrifuges holding bins silos c) What is the average elasped time between the determination of a leak by personnel of the "Leak Detection Patrol" and the plant's personnel taking corrective action for the purpose of eliminating the leak? Small leaks:______ Large leaks: ~ d) What is the average elapsed time between the monitoring system alarm becoming activated and the plant's personnel taking corrective action for the purpose of eliminating the leak? Small leaks: Large leaks: DISCHARGES TO THE ATMOSPHERE 11. Safety relief valves (SRV) and other emergency type (i.e., manual, vents, are: (check on a more) D connected to the recovery system through a common header (i.e., manifold) vented to the atmosphere D is manual vent ducted to recovery system 69 COLORITE 007362 12. If a pressure gage Is located between rupture disc (RD) and safety relief valve (SRV), list points in the plant where the gage indicates a higher than normal pressure, typically 0 to 5 psig. (Note: While in common use in plants, a pressure gage between RD and SRV is not a requirement of the Standard.) Identify those RD/SRV points which are in PVC plants by check (/) in column 3. RD/SRV Identification Location In Plant PVC Service Inspector's Comments 13. List vents, other than emergency types, which are vented to the atmosphere and not connected to a recovery system, and which are suspected of having had short and/or long periods of emissions exceeding the limits specified in Sections 61.62, 61.63, 61.64 and 61.65. Vent Identification Location In Plant Location of Nearest Monitoring Point(s) FUGITIVE EMISSIONS 14. Investigate and witness the plant's standard operating procedure [61.65(c)] for fugitive emission sources [61.65(b) (1), (b)(2), (b)(5), (b)(6) and (b)(7)] and list any possible deficiencies. Pertinent Fugitive Emission Source Location In Plant Possible Deficiency 70 COLORXTE 007363 15. Are plant personnel following the established standard operating procedures? Yes No 16. If the response to the above is "NO", list pertinent fugitive emission source, location in plant and action required where the established standard operating procedure is not being followed. Pertinent Fugitive Emission Source Location In Plant Action Required 17. List equipment, location in plant, identification in the process of any pump, compressor and agitator which is not equipped as seal-less or with a double mechanical seal or which does not duct emissions through a control system. Equipment Location In Plant Identification In Process Inspector's Comments 18. Incoming raw material received by: 0 pipeline Q Truck Q Rail 19. Finished product material shipped by: Pipeline 0 Truck 0 Rail 71 Q Barge 0 Barge COLORITE 007364 reactors and furnaces 20. Give the number of reactors for each capacity. EDC: Capacity Number PVC: Capacity Number VCM (Hydrochlorination): Capacity Number 21. Give the number of cracking furnaces. VCM{dehydrochlorination): CapacityNumber CONTROL DEVICES 22. List control devices, location in plant, major entering streams, major exiting streams, vent location. Control Oevice Location In Plant Major Entering Streams Major Exiting Streams Vent Location 72 COLOR!TE 007365 STACK EMISSIONS 23. List plant's performance specifications of stack emission continuous monitoring instrumentation: a) Stack identification ; b) Mean value, , calculated from a series of absolute measurements made by using the equipment specifications and procedure of reference Test Method 106 (see APPENDIX A); c) Number of measurements used to calculate the mean value; d) Accuracy, percent of the mean value; e) Calibration error, percent of each calibration gas mixture value; f) Zero drift (2 hr.), percent of calibration span; g) Zero drift (24 hr.), percent of calibration span; h) Calibration drift (2 hr.), percent of calibration span; i) Calibration drift (24 hr.), percent of calibration span; j) Response time, ________(Time required from the insertion of a known vinyl chloride concentration gas sample into the stack and the stack instrument indicating a value 95% of the known vinyl chloride concentration). 24. If more than one stack was tested, and if the other test data were significantly different from the above data, use attach sheets to provide the information obtained from other stack tests. 25. Were the stack emission tests made under conditions of maximum production rate? Yes No If the response is No, give the percent of maximum production rate under which the stack emission tests were made;percent. 26. Were all stack samples analyzed within 24 hours? Q Yes Q No If the response is No, give elapsed time, to the nearest hour, from taking sample to its being analyzed:hours. 73 COLORITE 007366 27. List deviations or substitutes from reference Test Method 106 equipments materials and procedures which are required in conducting the stack emission tests. Equipments/Materials/Procedure Identification Deviation or Substitutes From Test Method 106 Inspector's Comments INPROCESS WASTEWATER 28. Identify any inprocess wastewater stream and location in plant which is mixed with another water stream prior to the reduction of vinyl chloride concentration to 10 ppm or less. Inprocess Wastewater Stream Identification Location In Plant Identification of Process Step Inspector's Comments REACTOR OPENING LOSS 29. Briefly describe the plant's procedure to determine the emission due to opening the reactor. 74 COLORITE 007367 ON REVIEW OF RECORDS NESHAPS INSPECTION FOR VINYL CHLORIDE EMISSIONS COMPLIANCE Review Date______________________ Reviewer's Name Firm's Name_________________________ _______________________________________ _________ Firm's Address EMISSION (STACK) 1. Review of hourly Dyes Dno average from continuous stack emission. 2. Number of stack emissions (continuous emissions for 1 hour or more) which exceeded limits specified in Sections 61.62, 61.63 and 61.64: 3. Were each of these emissions properly and accurately reported in the appropriate Semi-annual Report? Yes DNo 4. If the response to No. 3 is No, list date, time, emission point location, duration, estimated integrated emission, and the cause or causes of emission. (Attach sheets if more space is required.) Date & Time Location In Plant Duration Est. Integrated Emission Cause or Causes Action Taken LEAK DETECTION SYSTEM 5. Review of (check one or more) dhourly Q shift Qdaily monthly average from leak detection system. ~ 75 Dweekly COLOR!TE 007368 6. Number of times one or more monitoring points indicated emission levels exceeded the value for the leak definition (See CHECKLIST 5 (d): 7. List date, time, monitoring point location, duration, estimated integrated emission and the cause or causes of emission of those leaks in No. 6. (Attach sheets if more space is required.) Date & Time Location In Plant Duration Est. Integrated Emission Cause or Causes Action Taken ATMOSPHERIC DISCHARGES 8. Number of emergency discharges to the atmosphere: _______________________ . 9. Were each of those emergency atmospheric discharges properly and accurately reported? Yes Q No 10. Were adequate preventative and/or maintenance procedures taken to prevent and minimize emergency atmospheric discharges? Yes No 11. If the response to either or both No. 9 and No. 10 is fto, list date, time, emission point location, duration, estimated integrated emission, and the cause or causes of the improperly and/or inaccurate reporting and the inadequate preventative and/or maintenance procedures relating to the emergency atmospheric discharges. (Attach sheets if more space is required.) Date & Location Time in Plant Duration Est. Integrated Cause or Inspector's Emission Causes Comments 76 COLORITE 007369 12. Determine from plant records whether temperature, pressure, flow rate(s) and/or other process variables and/or if equipment failures (i.e., reduced flow rate of cooling water, defective temperature controller, etc.) gave rise to the necessity of the emergency discharge(s) from PVC reactors. List the date, time, location in plant and the condition(s) which appear to have produced the need for an emergency discharge(s). Date & Time Location In Plant Conditions 13. List the dates and time when similar or equivalent conditions existed as in No. 12 and for which an emergency discharge was not reported. Date & Time Location In Plant Conditions 77 COLOR!TE 007370 14. Review the leak detection records of those monitoring points nearest the location(s) and down stream in No. 13 on those dates and times where conditions existed which appear to give rise to the necessity of emergency discharge. List date, time, location in plant and the emission levels of the nearest monitoring points. Date & Time Location In Plant Emi<>sion Level of Monitorina Points 15. List the date, time, location in plant from the tabulation in No. 14 where the owner/operator appears not to be in compliance with the NESHAPS vinyl chloride standard and its amendments. Date & Time Location In Plant Inspector's Consents 78 COLORITE 007371 REACTOR OPENING LOSS 16. Review analytical records of vinyl chloride concentration in polyvinyl chloride to determine "reactor opening loss" of reactors (and stripper where applicable)* prepolymerization and post polymerization vessels. List date, vessel identification, and batch identification where emissions exceeded standard. Date Vessel Batch Identification Identification Inspector's Comnents xI ! RESIN, SLURRY, WET CAKE AND LATEX SAMPLING 17. Review analytical records of vinyl chloride concentration in polyvinyl chloride resin, slurry, wet cake and latex to determine the weighed average residual vinyl chloride concentration In all grades of poly vinyl chloride resin processed through the stripping operation on each calendar day. List date, vessel Identification, and batch identifi cation where emissions exceeded standard. Date Vessel Batch Identification Identification Inspector's Comments 79 COLORITE 007372 RECORDKEEPING 18. Review and assess the recordkeeping as required in the standard in the listing below, and comment on each with respect to completeness, form and ease of reference. Record Stack emission monitoring Leak detection monitoring Leaks detected by "Leak Detection Patrol" Stack emission tests Reactor opening emissions Inprocess waste water emissions Resin, slurry, wet cake and latex emissions Inspector's Cormients - 80 COLORITE 007373 PRE-TEST EQUIPMENT CHECKLIST FOR STACK EMISSION TEST (TEST METHOD 106) NESHAPS INSPECTION FOR VINYL CHLORIDE EMISSIONS COMPLIANCE Pre-Test Meeting Date Inspector's Name Firm's Name Finn's Address.____________ 1. Probe a) Is probe made of stainless steel, pyrex glass, or teflon tubing? b) What is temperature of stack?___________ c) Does probe have glass wool plug? Yes No ?.. Sample line a) Is sample line made of teflon? Yes No b) Is a new unused piece used for each series of bag samples? Yes No 3. Quick connects a) Are 2 male and 2 female connects used? Yes No b) Are they made of stainless steel? Q Yes No c) Does the pair for the bag have ball checks? Yes Q No d) Are they assembled as required? Q Yes No 4. Rigid container a) Is container leak proof? Q Yes No Unknown b) Does it have a cover to protect contents from sunlight? Q Yes No 5. Sampling bags a) What material are bags made of?____________ b) Are bags of 100 liter capacity? Yes No Unknown 81 COLORITE 007374 riHk 6. Needle valve Will needle valve allow proper adjustment of sample flow? Yes No 7. Vacuum pump a) Is pump of the leak-free type? Yes Q No b) Does pump have a minimum capacity of 2 liters per minute? Yes No 8. Charcoal tube Does a charcoal tube follow pump to prevent admission of vinyl chloride to atmosphere? Q Yes Q No 9. Flow meter Does the flow meter have a capability of measuring flow range from 0.10 to 1.00 liter per minute? Q Yes No 10. Pitot tube and manometer a) What type of pitot tube is used? b) Is pitot tube attached to probe? 0 Yes Q No c) Will an Inclined manometer be used? Q Yes No 11. List substitutes for equipments and materials required in Test Method 106 in conducting stack emission tests. Equipment/Materials Identification Substitute Inspector's Comments 82 COLORXTE 007375 12. Is plant in compliance with Test Method 106? Q Yes No 13. If the response is No, list action iterns which must be completed prior to Test date in order to conduct stack emission tests: Action Item 1:__________ _____________________________ 2:____________ _____________________ 3:_______________________________ ____________ ___ 14. Test Date: 83 COLORITE 007376 EQUIPMENT CHECKLIST FOR VINYL CHLORIDE IN INPROCESS WASTEWATER, RESIN, SLURRY, WET CAKE AND LATEX SAMPLES (TEST METHOD 107) NESHAPS INSPECTION FOR VINYL CHLORIDE EMISSIONS COMPLIANCE Inspection Date____________________ Inspector's Name___________________ Firm's Name_________________________________________________________________ Firm's Address_______________________________________________________________ 1. Sample bottles a) Are the sample bottles of 60 ml (2oz) capacity? Yes No b) Do the bottles have waxed lined screw on top? Yes No c) Do the bottles have electrical tape or equivalent to prevent loosening of bottle tops? Yes No 2. Vials a) Are vials of 50 ml capacity? Yes 0 No b) Are they equipped with sealed Teflon faced Tuf-Bond discs for water samples? Yes O No c) Are they equipped with seals and caps, Perkin-Elmer Corporation No. 105-0118 or equivalent? Yes No Unknown 3. Analytical balance a) Is it capable of weighing reproducibility to + 0.001 gram? Yes No Unknown 84 COLOR!TE 007377 b) What is the weighing span in the region of weight that is used in Test Method 107? 4. Syringe a) Is its capacity 100 pi? Q Yes No Q Unknown b) Is the model Precision Series "A" No. 010025 or equivalent? Yes No Unknown 5. Vial Sealer a) Is the Model, Perkin-Elmer No. 105-0106 or equivalent? Q Yes No Unknown 6. Gas Chromatograph a) Is the Model, Perkin-Elmer Model F-40 head space analyzer No. 104-0001 or equivalent? 0 Yes No Unknown b) List substitutes used for the following: 2 m x 3.2 mm stainless steel column; contains 0.4% carbowax on carbopak A (or Carbopak B) Perkin-Elmer No. 105-0133 or equivalent: 7. Thermometer a) Range 0 to 100C, with accuracy + 0.1C, Perkin-Elmer No. 105-0109 or equivalent. Yes No 8. Sample Tray Thermostat System a) Perkin-Elmer No. 105-0103 or equivalent. Q Yes 0 No 9. Septa a) Sandwich type, for aitomatic dosing, 13 mm, Perkin-Elmer No. 105-1008 or equivalent. 0 Yes 0 No 85 COLORITE 007378 10. Integrator - Recorder a) Hewlett-Packard Model 3380A or equivalent. Q Yes No 11. Filter dryer assembly a) Perkin-Elmer No. 2230117 or equivalent. 12. Soap Film Flowmeter Yes No a) Hewlett-Packard No. 0101-0113 or equivalent. Yes No I I 86 COLORITE 007379 APPENDIX A: MEAN VALUE CALCULATION The mean value calculated from the reference method (Test Method 106) test data measurements is used as a norm to assess the stack emission continuous monitoring instrumentation. The mean value of the data set is calculated according to the following expression I I (A-l) where 4U Xj = The i--absolute measurement obtained from reference Test Method 106, n * Sum of the n- absolute measurements, i=l n 9 Number of absolute measurements, X = Mean value. I 87 COLORITE 007380 APPENDIX B NATIONAL EMISSION STANDARDS FOR HAZARDOUS AIR POLLUTANTS Standard For Vinyl Chloride 88 COLORITE 007381 Z+r J --; ?>f , ..... J * *^ J*` ^ ` * - ..v r*'-:imn+TrJ1 ,Ti -Vr^-s;:-^? ; -..v..-;r^~;3r - ' ,,*--v::. ' - .-j- -si , .* *$?**. Sir* '*T~ THURSDAY, OCTOBER 21,1978 ' -*v~A * .7. ... ^~. V;j iLl ''l ; ' _ . .*-7* .i,iJ-;_ r-#> ^, - t. *:*!*** >" _ '*>:. -V.'J.-1': :?- ^ - - i~s- ' 4 <4- * * m,T_ '*r*. v'^r* 'i.'JC: ~;<!F:*~-`1 -* ,s> : ^ TSfekrtF^ES':** ENVIRONMENTAL ^ssap^jr:.\ -'*"'* - :^;:-t# , V --*Wr.'H----.'./ -> - :i'i5'?7r' -' PROTECTION -:t*5 L v,--> ^ ij^gcassKi*jv^i*ii^s.-r -*-. > * 2S2MV: AGENCY ; STg--' -. -1*-*i^*>.' \ V*..***><" . -V^' ' .Kr-" . "L" , * ' '.^k' . " . '-7--"." ` 5 : - *^,.i.W.~W..S.'*-' -- ''*;V V NATIONAL EMISSION STANDARDS FOR' \ HAZARDOUS XirV' f POLLUTANTS -~ ; Standard For Vinyl Chloride 89 COLORITE 007382 46360 RULES AND REGULATIONS Tit!* 40--Protection of Environment CHAPTER I--ENVIRONMENTAL axe required ta.be captured trolled. '.r con greatest water usage would be employed^ and that there would be no recycling PROTECTION AGENCY SUBCHAFTX* C--AJ* fWOttiXMO- [KU. 6XS-11 Exalts do EmomiRU, Ixmsb of water. There is no regulation which, would require water recycling. Accord-- ' EPA prepared a document entitled th* tog to SPi, the control system utilizing Quantitative Risk Assessment for Com the most water will not be used gener PART 61--NATIONAL EMISSION STAND* munity Exposure to Vinyl Chloride which ally by the industry and economic fac ARDS FOR HAZARDOUS AIR POLLUTANTS estimates the risk from vinyl eWorld* tors will causa plants to recycle much Standard for Vinyl Chloride : exposure to populations living in the vi cinity of vinyl chloride-emitting plants On December 24. 1975. under section. - before and'after Implementation of com* of the water. Therefor* according to SPlth* impact of the standard on water consumption will be negligible 112 of the Clean Air Act, as. amended (42 tools to meet the standard. There arena The environmental impacts of the O.S.C. 1857), the Environmental Protec tion Agency (EPAJ added vinyl chloride to the list of hazardous air pollutants<40 FR 50477) and proposed a national dose-response data for the concentra tions of vinyl chloride found In the am bient sir. Therefor* assessments of risk, at ambient levels or exposure were ex promulgated standard may be summer- .`SS toed a* follows; The primary environ-""^-*? mental Impacts of the standard are ben--. (35*3 rdclal and will consist of vinyl chloride a-s emiAinn standard for It (40 Pit 59532). The standard covers plants which manu trapolated from doae-response data from higher levels of exposure using both a mating reductions of approximately 94 percent .at ethylene dlchlorlde-vtnyl facture ethylene dlchlorid*. vinyl linear model azuf; a log-problt model. chloride plants and 98 percent at poly- chloride, and/or polyvinyl chloride - - Extrapolations made with each of three . ting! chloride plants. Percentage num EPA decided to regulate vinyl chloride models entailed using different seta of bers &w both source categories are based because it has been Implicated as the assumption* Because different aasunro-- otr a* estimated 90 percent reduction in. causal agent of angiosarcoma and other tinns can be made in extrapolating to . fughhe emissions., and-1974 emission serious disorders, both cardnoganlo and noncardnogenle, in people with occupa tional exposure and in animal# with ex perimental exposure to vinyl chloride. Reasonable- extrapolations from these finrUny cause concern that vinyl chlo ride may cause or contribute to the same or similar disorders at present ambient. low doses, the health risks ara reported.. to range* - - It wee estimated that 4.8 mflBbn peo ple live within 5 miles of ethylenedicho-- ride-vinyl chloride and polyvinyl eWo rld* plants and, that the avenge ex posure around thee* plants before Instal- --Isring of etmtrote to meet the standard r.V - -" . .The potential secondary environmen tal Impacts of the standard are either lnslgtdfleant or will be minimized with outadditional action, except for one ad verse impact Hydrogen chloride Is al ready emitted by process equipment at ethytee dichloride*vinyl chloride plants - air levels. The purpose of the standard la to minimim vinyl chloride emlsiion*- la 17 parts per billion. The expoaur* lamia for uncontrolled plants were an-- and by other petrochemical plants In the cnmptmw when ethylene dlchlortde- from ah known, proctea and fugitive ciliated baaed on mtimated 1974 snl* vtnyl chloride plants are typically lo -mivdnn sourcss In ethylene dichlorid*> atom levels.. Using the linear dose-re cated. An Incinerator used to attain the vinyl chloride and polyvinyl chloride - sponse model SPA found that tiro standard at an ethylene dlcbloride-vtnyl plants to the level attainable with, best rate of initiation of liver angiosarcoma chloride plant could Increase hydrogen available control technology. This will among people living around uncontrolled., chloride emissions by several fold. Typl- - have the effect of furthering the protec plants is expected to range from lees than, roily,however, due to the corrosion prob tion of public health by minimizing the. one to ten esses of liver angiosarcomer lem* which would otherwise occur both health risks to the people living In the per year, of exposure to vinyl eWorld* on plant property and in the community, vicinity of these plants and to any addi The log-problt model gave predictions,, plants use .scrubbers to control already tional people who are exposed as a result that are 0.1 to 0.01 times this rot* This existing hydrogen chloride emission* of new construction. wide range is an Indication of tbs na- Hydrogen chloride emissions, resulting interested partiee participated hi the certalnties in extrapolation to low doaest Hero control of vinyl chloride emissions rulemaking by sending comments to EPA. Dus to the long latency time observed lz* are expected to be controlled for the The comments have been carefully cone cancer cases resulting from vinyl eWorld* same reason, tt even a moderately effi sldered, and where determined by the -exposure; increases Initiated by exposure cient scrubber (98 percent control) were Administrator to be appropriate, changes this year will not be diagnosed until the- used to control the hydrogen chloride have been, made to the regulation se pro 1990*s oc later. Vinyl chloride is also es emiMfcw resulting from Incineration of mulgated. timated to produce an equal number of vinyl chloride emissions, the Increase in StracMAsr or: i Sxuisaib primary eancero at other sites, for a total hydrogen chloride emissions from a typ of somewhere between lea than one and ical ethylene dlcbloride-vtnyl chloride In ethylene dlchloride-rinyt chloride twenty case* of cancer per year of ex- plant doe to the standard would be re plants, the standard limits vinyl chloride poeur* among residents around plant* - duced to 35 percent. However, EPA plans emissions from the ethylene dlchlorida The number of these effects- is expected to further evaluate the need to control and vinyl chloride formation and purl- to bereduced at least in proportion to th hydrogen chloride emissions, since dif flcaHon processes to 10 ppm. Voe the ox- reduction In the ambient annual avenga fusion model results indicate that under ychlorination process, vinyl eWorld* vinyl chloride concentration, which 1* "Womb-rose" meteorological condition* --m i^inin are limited to 0-2 g/kg ot ethyl expected to be 5 percent of the uncon the hydrogen chloride emissions from ene dlchlorlde product trolled levels after the standard Is toa- the process equipment and the incinera In polyvinyl chloride plant* thestand-- plementad. tor combined would cause maximum am aid limits vinyl chloride hlesions from Changes m the standard since pro bient concentrations of hydrogen chlo equipment preceding and Including the posal do not affect the level of control ride hr the vicinity of ethylene dicblo- stripper in the plant proceos flow to 10 required. Thu* the environmental Im rids-vinyl chloride plants to be In the ppm. Emissions from equipment follow pact of the promulgated standard I* same range or somewhat higher than ing the stripper are to he controUedby with on* exception, the same, as that existing foreign standards, and National stripping dispersion resins to 2000 ppm described In Chapter 8 of Volume I of Academy of Sciences (NAS) guidelines and other resins to 400 ppm. or by using the Standard Suitport and Environmen forpublic exposure. . equivalent controls. Vinyl chloride emis tal Impact Statement. According to data sions from reactor opening are to be re submitted by the Society of Plastics In Economic Impact duced to 0.02 g/kg polyvinyl chloride dustry, Inc; (SPD, the Impact on water tn accordance with Executive Order product. consumption In the draft environmental 11821 and OMB circular A-107. EPA In both ethylene dlchlorlde-vinyl chloride and polyvinyl chloride plants, relief valve dischargee and manual vent ing of gases are prohibited except under emergency conditions, fugitive emissions Impact statement was overstated. In es timating the Impact on water consump tion, EPA based its estimates on worst case condition* That Is, EPA assumed that those oontrol systems with the carefully evaluated the economic and Inflationary Impact of the proposed standard and alternative control levels and certified this in the preamble to the proposed standard. These Impacts are FfOHAl M61STM, VOL 41. NO. 103--THUWDAY, OC70SI* II, 1*70 90 COLORITE 007383 RUUS AND REGULATIONS 46561 discussed. In Chapter 7 of Volume X of . Act,, a public hearing wag bald on the the possible exposure patterns, these the Standard Support and Snvironmen- proposed standard on February 3. 1978, cases cannot be ignored in the evaluation tal Impact Statement, comments on tbs In Washington. D.C. Presentations wen of the potential public health problems. proposed standard bars resulted In only made by the Environmental Defense With regard to the alleged overstated one major change in the economic Im Fund the Society of the Plastics indus emission levels, the unconmolled emis pact analysis. EPA estimated that there try. Inn. Dow Chemical Company. Dia sion levels reported by EPA were based would be four plant, closures as a result mond Shamrock Corporation, and Air on 1974 data. This qualification was ol the promulgated standard. Ot the four Products and Chemicals. Inc. Copies of stated wherever emission data were pre plants identified ss possible closure can the comment letters received, the public sented. EPA recognizes that emissions didates. one has given notice that It no hearing record, end a summary of the have been reduced since that time, and longer produce* polyvinyl chloride and comments with EPA's responses are stated this in the preamble to the pro the other three have Indicated that they available for public inspection and copy posed standard. EPA decided not to do not intend to close as a result of the ing at the EPA Public Information Ref- gather more recent, data on emission standard. erenoa Unit, Room.2922 (JEPA Library); lavels, because these emission levels are The economic Impacts of the promul 401 M. Street. SW,, Washington. D.C. m expected to change, and gathering the gated standard, may be summarized aa addition, copies of . the etmmums sum data would take considerable time both follows: The total capital cost for exist- ,, mary and Agency- responses may be ob on the part of EPA and on the part of tag plants to meet the standard la esti tained. upon written.request from the Industry. Since the purpose of the stand mated to be $198 million, of which. 115 Public Information Center. (P3&-21S). ard is to minimize emissions, these more' t"1"1"" is for ethylene dlchloride-vinyl Environmental Protection.. Agency. 40L', current data would not affect the stand chloride plants and $183 million la for M Street, SW,, Washington. D.C. 20480 ard Itself. The 1974 emission levels were- polyvinyl chloride plant*. SPA estimates. (specify Standard.Support and Enc(ran- . also used is diffusion modeling to prelect that thee*. plants will have to spend $70 , mental Impact-L Statement, Emission maximum ambient air concentrations mUlinn per year to maintain the required Standard lor Vinyl Chloride, Volume 11} _ around uncontaolled plants. These maxi- emission levels, .in addition, tbs total. capital cost for gristing plants to meat' ths EPA's 1983 water edtuant guidelins Srmmmcawr Counsg sna-C8swesa.ro iV.'s >. nmm-air concentrations would probably be lower If 1978 emission levels were used. This imuld reduce the relative Impact - limitations is expected to bs $83 million.- ' <D DecUkm to Bsf vhnlchloride at a-' of the standard below that described in - and the total annualized operation cost haoardou*. atr poOdant. m general, the the Standard Support and Environmen is $17 million The costa to the industry ; commenten did not contest EPA's deci tal Impact Statement, but would not of meeting the OSSA standard cannot bo. sion to list vinyl chloride sa a hazardous affect the basis of the standard itself. Quantised at this time, but theyare eg,-" air- p-uii.T.s However, three oommant- , (2) Approach for Reyvlattny Vinyl pected to overlap to some degree with the: ora- (two companies and. one Federal Chloride Under Section 112. Two ap costs to meet EPA's fugitive emission-, agency) argued that XPA placed undue proaches other than using best avail regulations. The costa of meeting thA- fmphMi- on factoca'suggesting that vinyl able control technology were suggested fugitive emission regulations are Included: -hUHH- presented a health risk and by the eommenters for regulating vinyl In the total costs cltsd above for meeting Ignored `factors suggesting, that no sig chloride under section 112. The first was the promulgated regulation. Broken cot nificant risk was'involved, under section to ban polyvinyl chloride products for. separately, the capital cost of meeting. 112, however. EPA could remove vinyl which substitutes are currently available the fugitive emission regulations la $3T chloride from the list ot hazardous air and to gradually phase out other poly mMlon and the annualized cast Is $3A pollutants only If information ware pre vinyl chloride products ae substitutes minion. .. ' . sented to EPA tbat-ahows that vinyl are developed. The standard is sot expected to detv > chloride la dearly not a hssarrinns air In the preamble to the proposed stand construction, of new ethylene dlchloride- . r"n,lt*"* * m the ard EPA specified its reasons for not set vinyl chloride plants or most types af conunmfcmmmirTi tbs iwumr^^w ^14. ting a zero emission limit for vinyl new polyvinyl chloride plants. For ons notprovide conclusive evidence thaivtnyl chloride as follows: (l) There are bene type of polyvinyl chloride plant (diaper-. rfn-ySSe hi was heiarrlena elr nollulant. ficial usee of vinyl chloride products for sion process) that represents 13 percent of the tadustxy production, the standard wh(4*gm)>mouaubeueMe oirmcpohnwtrlbiuhtwasftfotridrTetairtihr^orr which desirable substitutes are not read ily available; (2) there are potentially would significantly, deter the construe- prows, that the health risk factors em adverse health and environmental im tion of smaller plants. phasised by EPA wars insignificant. It Is estimated that the price ot poly-, . Several othar eommenters agreed with pacts from substitute* which have not been thoroughly studied; (3) there are a vinyl chloride resins wUhrise by apprat- XPA'e decision to list vinyl chloride as a number of employees, particularly in the mately 7.3 percent in order to maintain hasardoue air pollutant, but argued that fabrication industries, who would be precontrol profitability and also to re-- SPA had overstated the health problem, come at least temporarily unemployed; cover the total annualtasd control costs the taels, and the projected wad (4) control technology is available necessitated by the standard at ethylene ambient air concentrations around un dlchloride-vinyl chloride plants sad poly- - controlled plaate. With regard to the al which is capable of substantially reduc ing emissions of vtnyi chloride into the vinyl chloride plants. This increase is leged overstated health problem, the atmosphere. , estimated to translate Into a maximum --Miwnto* tatL tar -vairmta. that the EPA agrees that substitutes do exist or consumer price taersaee fit goods fabri UiL worker SPA discussed as having could be manufactured for most poly cated from polyvinyl shkrVie resins of been expoeed-ftTviiiyl chloride levels low vinyl chloride uses. However, in general, approximately 3d iswt Recovery of er. t-h-n those usually encountered In these substitutes do not have some of the efluent annualized coats plus mainte polyvinyl chloride production has been more desirable characteristics of poly nance of precontrol profitability la esti dropped from the Rational Institute ot vinyl chloride, such as nonflammability. mated to add approximately 2 percent to Occupational Safety and Health's listing If vinyl chloride and polyvinyl chloride polyvinyl chloride resin prices and result of workers with snglosarcomA EPA were banned, other substitutes with in an additional maximum consumer agrees that there are queetlons concern these more desirable characteristics price Increase of 1 percent. ing the level of expoeure and in some would likely be developed. There is a risk Poetic Psaxmesiioir cases the pathology of these cases not that these substitutes would also bavs Involved directly in polyvinyl chloride adverse health or environmental effects. During the public comment period. SO -mv vinyl-chloride production. These un Since control measures are available comment letters on the proposed stand certainties are stated in the appropriate which can reduce vinyl chloride emis ard were received. There were 24 from footnotes of the Scientific and Technical sions by 90 percent or more. It does not industry: 3 from environmental groups; Aseeament Report on Vinyl Chloride and seem prudent to reduce emissions by the 13 from Federal. State, and local agen Polyvinyl Chloride (STAR) where the remaining percentage and take the risk cies; and 8 from individual citizens. As angiosarcoma cases are listed. However, of introducing new untested chemicals required by section 112(b) (1) (B> of the in spite of these uncertainties, in view of into the environment. ) FEDltAl MQISTM, VOL 41, NO. JOJ--rHUSJOAY, CTOUft 31, 1474 91 COLORITE 007384 r562 RU1ES AND' REGULATIONS - 'i Another approach suggested hr the capacity of no more than 0.19 m* (SO venting will he permitted as a last resort eommenters -was to bsse the standard for gall. Reactors In this range can gen* before the relief valve opens. The same f* individual emission point on cost eraHy be famd in a laboratory, whereas notification procedures are required for reus benefit, Several of the fugitive the larger reacton are typically pilot manual vmting to the atmosphere as are l-iission sources were named specifically scale facilities. Emissions from laboratory required for relief discharges. as ones for which the costs of control scale equipment are relatively small, and There are several changes In the nu were substantially higher than the bene* application of the controls required by merical emission limits in the promul j 3. Although EPA did determine a costoeflt ratio for the controls required the standard would be expensive and im practical EPA also decided to exempt re* gated standard. Except for the standard for reactor opening loss, these changes tor a number of emission points, EPA search and development facilities con simply Involve conversion to the Interna does not believe such a ratio is an appro* taining reactors.greater than 0.19 m* (SO tional System of Units (SI). There was ;--late basis on which to set a standard. gal) and no more than 4.07 m* < 1100 gal) an error Involved In the original calcula ctlon 111 of the Clean Air Act provides In capacity from all parts of the standard tion used Wderive tile standard for reac the development of standards based- except the 10 ppm limit for reactors. tor opening. Correcting this error dou un best control technology (considering *trippers, monomer recovery syatons, and bles the allowable it is em costs). Even under section 111, however, mixing, weighing and holding containers. phasised that-the change In this stand r rndards are not based on a fine bal*. EPA decided not to require these facul ard Is a correction, and not a change In j clng of coats versus benefits. Instead,, ties to meet other parts of the standard the Intent for the degree of: control re 1 its are considered in terms of the af because of the technical problems in quired. ' fordability qf the control technology re- - volved in doing-' so, 'Tor example, the The proposed 'standard required the1 ired to achieve a given ____leve_l standard tar reactor evening la based In Installation of a rupture disc beneath r4 the economic impact of payshi* . part on reducing the frequency of open- each relief valve to prevent leakage from, rndards on the mdustxy in qu< " lng the reactor.' Research and develop the relief valve. A provision-has been -~on. Unlike section 111, section 113 does ment reactors have'to-be opened after added to the promulgated-; standardise not explicitly provide for considaratioa every batch far thorough cleaning. Also, that a rupture disc is not: required tf pf costs, so it would clearly-be inappro- stripping technology-is? developed Indi the relief valve is tied- Into a process Use ! { late to consider costs to a greater exit under section-UZ than would be vidually for each resht-fn research and development equipment-Tberefore, at or recovery system. In this- ease, any leakage from the relief valve would be cone under section in. As discussed in tainment of the stripping limitations in contained. . ' the preamble to the proposed standard the research and development equipment The regulation; -for obtaining .-vinyl ~r vinyl chloride, EPA believes costs would not always be possible. Tbs 4.07 chloride samples has been changed to an ly be considered under section 113, but m*. (1100 gal) flgurd~tias selected as an operating procedure. The - proposed Jy to a very limited extent; Le-. to assure that the costs of control technol ogy are not grossly disproportionate to r ie amornri of emission reduction ! ihleved. In comparison with other nitTM pomtA the costs of controlling the fugitive emission sources mentioned >y the commontem are relatively small r imposed with the amount of a*"1--1"" ' duction achieved. Several commentars recommended adding to the regulation a provision for ,-sceas -miMirm, during startup, shut- - upper cut-off point because there are no commercial reactors smaller than this. - (4> Smiaion UmUs. The only majorchange m the emisslorr limits between .proposal and promulgation Is the addi tion of a provlstoa for emergmey manual venting of vtnyl ohlortde from reactors to the atmosphere;-The proposed stand ard prohibited an manual venting to the atmosphere. In the preamble to the pro posed, standard. EPA invited Interested persons to comment on whether permit ting venting to-the atmosphere standard stated that tbere were to be no emissions from taking the samples. Several commentary pointed out that the use of the word "no" -would make this regulation impractical to enfovow. There fore, the promulgated standard specifies the operating procedure which EPA orig inally intended to be need to control this source. This revision is only a change in wording and does not represent a change tn the level of the standard. . The regulation for taking samples baa also been revised to apply only to sam >wn, anrf epa considered could result In overall lower emissions. ples containing at least 10 percent by ds comment, and decided that this There are several methods available for weight vinyl chloride. This Is consistent addition is not necessary for the vinyl chloride standard. Startup and shutdown rM the process has wisenflally so effect i emissions to the atmosphere for poly- preventing relief dischargee from reaotore, one at which is manual venting of part of the reactor contents for purposes of cooling and reduction in pressure with the other parts of the standard which apply to equipment "In vinyl* chloride service;" "In vinyl chloride serv ice" - distinguishes between, situations nyl chloride production, and technology exists to avoid excess emissions during startup and shutdown at ethylene di- lioridevlnyl chloride plants. We do not within the reactor. The higher the tem perature and pressure-within the reac tor, the greater the amount of vinyl chloride which has to- be removed to where vinyl chloride Is clearly Involved and situations where vinyl chloride Is a minor component or contaminant, and as defined in promulgated 161.01(1) neve plants should be allowed to emit -brta* the reactor under control. Manual means that a piece of equipment con iceas emissions dicing malfunctions, and therefore are requiring them to shut down immediately. (3) Selection of source categories. In ic preamble to the proposed standard -PA recognised that some nail research and development facilities may exist where the emissions of vinyl chloride are significant and covering these facilities nder the standard would be unnecessary and Inappropriate. However, EPA did not venting can be dene at a lower pressure than the pressure required to open the relief valve. Por this reason manual vent ing can result nr lower emissions than would occur by allowing the reactor to discharge through the-relief valve. Pur- thamore, a manual vent valve Is under the control of an operator and can be closed. A relief valve may become clogged with resin and not close. The result would be loss of all the reactor contents. tains or contacts-cither a liquid that Is at least 10 percent by weight vinyl chlo ride or a gas that is at least 10 percent by volume vinyl chloride.The proposed standard required a vinyl chloride monitoring system for continu ously measuring vinyl chloride levels both within ths plant (for leak detection) and within stacks. The proposed standard did not outline required specifications for the monitoring system, except that It was to have sufflclmt information available to The contents of a reactor can be manNearly define which facilities should he -ually vented to a gasholder or other hold- analyze the samples with gas chromatog raphy, or if all hydrocarbons were as xeluded from the standard, and ncouraged interested parties to submit such information during the comment period. Based on the Information sub mitted, EPA decided to exempt poly- lnyl chloride reactors and associated lng vessel. However, in some cases, such sumed to be vinyl chloride, with infrared as during severe weather conditions, sev _3pectrophotometry, flame ion detection, eral reactors may be out of control at or equivalent It required that each plant one tins There, would- be Insufficient submit a description of its monitoring holding capacity under these conditions to manually vent the contents of all ths system to EPA. so that EPA could deter mine whether It was acceptable or not. Comments were received indicating a equipment from applicability of all parts reactors to a gasholder. Therefore, when need for EPA to specify some criteria for of the standard if the reactors are used all other measures to prevent relief valve judging the acceptability of monitoring a research and development and have a discharges have been exhausted, manual systems. The accuracy of the monitor- RPMAl BKIJT, VOL 41, NO, 305--'mUSSDAY, OCTOBH 21, 1974 92 COLORITE 007385 RUUS AND REGULATIONS 46563 ins system would bw related to too fro* (g) (3) (11), and 61.70(0 (2H1> so that made the same as for other rains and ' quency of calibration. Therefore, EPA measurement of the-vinyl chloride levels others requested that they be made less has in tho promulgated stand- in the resins is to be mads immediately stringent EPA decided not to make the aid requirements lot tha frequency of after stripping la completed rather than standard for stripping dispersion resins calibration and procedures to be canted as the rain Is being transferred out of the same m for other resins because there out in the calibration of the monitoring the stripper. This allows a plant to carry Is sufficient evidence to Indicate that instrumoau. out operations in a stripper after strip these resins are more difficult to strip The portable hydrocarbon. detector re ping lias been completed but before It Is than other resins. With regard to mak quired by the proposed standard was re transferred out of. the stripper. This is ing the stripping levels for dispersion quired to have a sensitivity of s ppm. consistent with the original intent of the rains leas stringent, only one of the eight Comments were >* that standard. manufacturers of dispersion resins spe instruments In this sensitivity range are The regulation for loading iwi unload cifically commented that the dispersion delicate and require continuing mainte ing lines in 161.88(b) (1) has been re rosin standard should be made less nance. The portable hydrocarbon detec vised to clarify that.tt applies only to stringent. Only two of several grades of tor is required for leek detection and for lines that are disconnected after each dispersion rosins made by this company measuring vinyl chloride concentrations loading, or unloading operation. Perma cannot meet the 3.080 ppm limit. The inside tho equipment before opening it nently installed pipelines that are opened proposed standard takes into considera A 5 ppm sensitivity is opt needed in. infrequently for'inspection or mainte tion that some resins are more difficult either case, and the required aensittvitj nance. for example, ax* covered by the to strip than others by providing, for has been changed to 10 ppm in the pro opening.'of equipment regulation rather averaging among different resins. mulgated standard. *' The proposed standard Contained / thA& tbt loiidlb(r> isA .mloidlfif. Uoi rful*tioiL - y . .,(8). TttUng.. reporting, and reccrrdkseplny. There are. several relatively single regulation for compreeeocs. The '' The regulation'for tniiwctee minor changes In the testing, reporting, promulgated standard has separate regu- water In the proposed standard could end. recordkeeping requirements. A pro- teOcna for rotating and. reciprocating have been misinterpreted to require In elston has been added to 181.87 which compmseeia This le consistentwithhav- dividual. treatment - oCT^wastewater requires that stack gas.sampLea taken in* separate regulations for rotating and ' streams- SecUnnSlAS(b)(9) ay of tho with. Teat Mathod 108 arete be analyzed reciprocating pumps In- both thepro- -promulgated standard, eternise that within 34 hours. This Is consistent with posed and promulgated standards. ' - wastewater steeama thakare.required to the requirement^, to the proposed Test Section 81.M of the proposed standard* be treated <ie. those containing greater Method 188. The promulgated standard provided for the me of equivalent meth than 10 ppm vinyl chloride) can be com also specifies that to averaging the re ods of control which have been approved bined to be treated. However,.-waste sults of- the three runs required by Test by SPA., The promidgated standard re- water streams that contain greater than Method 108, a time-weighted avenge 1s qutra that the plant owner or operator 10 ppm vinyl chloride cannot be com to be used. - submit a request for determination of bined with wastewater streams that con One conunenter requested that the equivalency within 30 days of the pro tain less than 10 ppm vinyl chloride be* oxygen content and moisture content be mulgation date if the alternative control lore treahnent: La. dilution cannot be specified for the 10 ppm concentration method Is Intended ae the Initial means yuaad to most the standard.*. standards. The proposed standard speci of control. The parpoee of this is to pro >, The commontars recommended several fied thatthe vtnyl chloride concentration vide time for SPA to evaluate the method changes In the emission. Emits whleh is to be corrected to 10 percent oxygen before the plant has to be In compliance have not been incorporated into the (wet basis) If combustion is used as the (for existing sourese, 90 days after the promulgated standard-. These are dis control measure. In the promulgated promulgation date). SPA also suggmte cussed la the following paragraphs. standard, thia requirement has been ex that this request for determination of , It was reeommsnded that the requlre- pended to all control measure*. equlvalaney be eceompaaled by a re*~~nient. for double mechanical seels on A provision has been added to the quest for waiver at oompilsaco pursuant pumps. compressors, and agitators be re promulgated standard which states that to section 113(e) (1>(B>(11> of the Ask moved because the single seals currently If a reactor la also used ae a stripper, the The request tor a waiver for compliance used on this equipment have small aaals- reactor opening --may be deter should provide tor tho ease where VA aiens and are more rellahl* than double mined immediately following the strip determines that a method la not equiv mechanical seals. SPA la aware that each ping operation, IT a reactor is also used alent and the plant neads to pnrcheee fugitive emkslon- source, suchas one as a stripper, the rain Is In the reactor other equipment. In no case will the pump, taken by itself cantro relatively when It Is opened. This means that vinyl waiver of compliance be extended beyond small emissions. Fugitive emissions con- chloride to the rain which has already two yean from tho date of promulga sUsrod as a whole are a significant been stripped to acceptable levels can tion. source of missions, however.and the in escape from the rain and become pert There are several wording clarifica tions which have been made in the pro mulgated standard. The definition for "In vinyl chloride serrieW 00.81(1)) has been clarified by ^**g that it equipment that contacts vinyl chloride as well as equipmmte that con tains vtnyl chloride This would Include tent of the standard Is to reduce these. Double mechanical seal pumps are com* manly used to the Industry for emission reduction. Sealless pumps or equivalent systems are available as options to double The cwmmenters recommended to* i rnaslni the avenging time for the 10 of the reactor opening lose. It Is EFA's intent that once a rain has been stripped to the required levels, that additional controls are not required. Odder the new provision, vtnyl chloride escaping from the rain after It has been stripped to acceptable levels le not counted as part of rrirtnr oDeafau Ion.. such equipment as agitators. Words have been added In II0US2, 81.83. and 81.64 to clarify that the 10 ppm emission limits do not have to be met when equipment has already been opened in compliance with the regula t-io.n.tfotrh.oTpesnTinTg of equipment. Equlp- ppm limits * the limits for reactor opening and stripping to 39 days. Some at the commonters apparently thought that the 10 ppm limits bad to be met on an Instantaneous basis. However, stooe the performance test for determin ing compliance consists of three runs for * minimum of an hour eacnh., tnhe aver- A section requiring continuous moni toring of stack emissions has been added to the promulgated standard The con tinuous monitoring of stack emissions was required to the proposed standard. The addition of e specific paragraph for emission monitoring serves only to clarify the requirement. 2eqSuiLpmSe5nt*JreSgSuleatiiEon? canJcKonStaiZn lmLoJrei' than 10 ppm vinyl chloride and would be in violation of the standard If this statement were not Included. Mia* time for the 10 PP Umit is atlernt three hours. Increasing the averaging tfam, to 30 days for any of the emission limits would permit higher peak emis sion levels. EPA has determined that thia The standard has been revised so that the initial report requires a "description" rather than a "detailed description" of the equipment used to control fugitive emissions.. Several commenters pointed The requirements for stripping poly Is neither desirable nor necessary. out that a detailed description would vinyl chloride resins to specified levels Some commentary requested that the contain proprietary information. EPA have been revised to IS 81.84(e), 51.87 stripping levels for dispersion resins be agrees that a detailed description In the RDtXAI. HOISTfl, VOL. 41, NO. 20S--THURSDAY, OCTOBER 21. 1974 93 COLORITE 007386 r:Initial report la unnecessary. It addi ttional information la needed. EPA can r'obtain it under section 114 at the Act and I "atmheenptlatnntacccaonrdreaqnuceeswt citohnf4id0eCnPtiUal treat-' Part 2 for information It believes to be proprietary. I The proposed standard required that a semiannual report be submitted every . 130 days. The promulgated standard specifies dates for the submittal of the reports. It also specifies that the'lint semiannual report does not have to be i submitted until at least six months after -RULE5 AND REGULATIONS' - ' m section 8,4 of Test Method 104 the requirement for an automatic integrator nae been replaced with a requirement for a due Integrator or planlmeter for meas uring peak. area. Thu change U in re sponse to a comment which atatee that automatic Integrators are unnecessarily elaborate and expensive. A new section 8.8 baa been added to Test Method 104 which requires deter mination. of the water vapor content of the sampling bag by measuring the-zfnblent temperature and preseur* near the bag. The vinyl chloride concentration of (2) Vinyl chloride-by any process, and/or -v <3) One or more polymers containing any fraction of polymerized vinyl chlo ride. <b> ThU -subpart does not apply to equipmmt used In research and develop ment if the reactor used to polymerize the vinyl chloride processed tn the equip ment hae a capacity of no more than 0.19m* <50 gal). (c> Sections of thU subpart other than 161.84(a)(1), (b), (o). and (d) do not apply to equipment used tn research and the initial report la submitted. tire bag can then be: reported on a dry development if the reactor used to po The standard has been revised to *Um- basis. A provision for etwtnr'y the rigid lymerize the vinyl chloride processed in i--tnate the requlmneit to record the cause container for leaks hae b6an added to the equipment has a capacity cf greater i of any leak detected by the vinyl chlo- section 7.4 of Test Mstm-y los. than 0.19 m* (80 gal) and no more man 1 ride detector, the action taken to repair the leak, and the amount of time re quired to repair the leak. ETA la com* r~ cerned only that leaks are detected and. repaired. That this has been done can be ' established by locking at the atrip chart record of measurements made by the vinyl chloride detector. These records are . . The only change te-Teet Method 107 U the provision to. Section 5.3-2 for us* of Carbopak C m*,wen as Carbopak A. - Aoiwoezrrz 6eeuen:iu or the Cleea Air Set ea added by seau4(e) of Pub. L. Si-404, 44 sat less (43 tJAti.' 1SS70-7; Section 114 eg the mean Air Act. as: Added by aea 4(e) at Puh. u S1-S04. to Stem 1S47. and amended 4.07 m* <1100 gal). 9 61.61 Definidoi - Terms Used to this subpart are defined to tbe Act. to suhpart A of-this part, or tn this lection u follows?- ^ 7*: ' j- (*) - "Ettijlen* dJchioridi*. plant" t tx^ dudes any- plant which produces ethyl rtffl required for the portable hydrocar by Puh. L. to-SlS, mmS(a) (4). SS Stai. 384 ene dlcblorlde by reaction of oxygen end bon detector however. - --- : (43 USA ltoTo-O)* fwtluu 301(a) eg toe hydrogen chloride with ethylene: Several asnmaitatera reeomaaided that the companies be allowed an extra --two weeks to submit to EPA data from the Initial performance test They also Cleon Air Act. a> ammded by eee. 18(e)(3) g run. i*. ai-eov m etaa mi (ta use. isojg (a)) .. -TDstediOctober ti'i975. --~ <b) "Vtnyl. chloride plant" Includes any plant which produces vinyl chloride by any process. , -- - . . (c) "FolyvlnyL chloride plant" Includes recommended that they submit the data by regular-mail rather than registered man. EPA baa not adopted either of these - recommendations A source la supposed to be in compliance with tbe standard within 90 days of the promulgation at the standard. The standard requires that the emission tests be done wlthm the r 90 day period, and permit* an extra 34 ) days for determination of results. The purpose of using registered mall Is to document the fact that emission data have been mt and received. This way if the results are lost in the mall, there win be no question that they were sent. <S) rest method. Test Method 104 has been changed to recognize that on a gas _chromatograph equipped with a Chrom060rb 103 column, acetaldehyde may interfere with the vinyl chloride peak. When a Ample Is expected to contain acetaldehyde, a secondary column as do- scribed in section 4.3.2 must be employed. \Smm spectroscopy or another absolute analytical technique is required to con firm the vinyl chloride peek obtained with the gas chromatograph, only If peek - resolution with the secondary column Is not successful. In section 4.1.4. alumtntaed Mylarbegs can be substituted far Tedler begs. EPA now has data to allow this substitution, Qttisu** ;: any plant where vtnyl chloride alone or to combination .with other material* is ig A&mMatratat. polymerised. -,1-.^..-,^-.. 'Part 41 eg Chapter Z, Title 40 at ths- (d) "Slip gauger" " a gauge which Ood*-ef Federal Regulations U amended hae a probe that moves through the gas/ as follows: The table of sections for Part liquid Interface to a storage or transfer 41 Is amended by adding a list of sections vessel and Indicates the level of vtnyl for new- Subpart- V: and Part 41 Is amended by-adding'a new Subpart P reading asfollows: -sr- Onleileii Standard far Vhed chloride in the vessel by -the physical tote of the mq.*Hqi the gauge dis charges. .... . (e> "Type of resln.7 means the broad classtilcation of resin referring to the See. '-.tg.,* --- ' basic manufacturing process for produc 41AO Applicability.. - ing that resin. Including, but not limited 41A1. Deflnltlona.-.1 - - 1A3 lalsuea steadied tor ethylene dl-- nhlwetAw 41.4T- tnUemn rtasdant tor vinyl chloride *--**' 4144 ggileahm etandanS for polyvinyl chlo- w . tide plants *<- ... to, the suspension; dispersion, latex, bulk, and solution processes. - --' - <f) "Grade at resin" means the sub division of resin classification which de scribes it as a unique resin, Le., the most exact description of a resin, with no fur 41AS - Kndmkm ataadsid tar ethylene, to* ther subdivision. __. chloride, vlnyi chloride and poly- <gl "Dispersion resin* means a resin ' vinyl ehiottde plenta 41AS Squivalent equipment and procedurea. 61.47 SSUatiun taels. 1 Sl.ee *--!-- monltcrun. S1A* Initial report. 51.70 q-nl-"--*-1 report 41.71 - gem illeeping.. manufactured to such away as to form fluid dispersions when dispersed In a plasticizer or plaetlcizer/diluent. mixturn. - - - (h) "Latex resin" means a resin which is produced by a polymerization process Auinuauii gentian 113 of the Clean Atr Act ee added by me. 4(e) of Pub. L. l-Ot. to StaS. 14S8 (43 USA 18870-7); aeatlon 114 egof the CUan Air Act. as added by aeo. 4(a) Pub. L 91-404. M Stet. 1SS7. and amanded which initiate* from free radical catalyst sites and Is sold undided. (1) "Bulk resin* means a resin which la produced by a polymerization process to which no water Is used. provided that the samples are analyzed by Pub. L. 93-319. me. 8(a) (4), 84 StL 344 _ (J) "Inprocess wastewater" means any within 24 hours of collection. (43 USA 18870-9): eoction 901(a) Of the water which, during manufacturing or in section 8.1.3 of Test Method 104 the requirement to use "oxygen gas** has been replaced with "oxygen gee or air. as Clean Air Act, aa amended by tee. 18(c) (3) Of Pub. L. 91-404, 84 Stat. 1713 (43 UA.O- 18478(a)). - processing, comes Into direct contact with vinyl chloride or polyvinyl chloride or results from the production or use of required by the detector." Several corn- Subpart F--National Emission Stsndsid any raw material. Intermediate product, mentors stated that most gas chromato for Vinyl Chloride finished product, by-product, or waste graphs are designed to us* hydrogen and air for their flame detectors. When used, g 61.60 Applicability. in tills way. they are capable of detect (a) This subpart applies to plants product containing vinyl chloride or polyvinyl chloride but which hae not been discharged to a wastewater treat ing 0A ppm vinyl chloride In air. This la which produce: ment process or discharged untreated as sensitive enough for monitoring the 10 (1) Ethylene dlcblorlde by reaction of wastewater. ppm emission limits stipulated In the oxygen and hydrogen chloride with <k) "Wastewater treatment process" standard. .. ethylene. includes any process which modifies FEDERAL ttCIJTE*. VOL 41, NO. 205--THUI5DAY, OCTOIIS 21, 1474 94 3 3 COLORITE 007387 RULES AND REGULATIONS 4tjobo characteristics such as BOD. COD, TS3, (>1,(53 Emir*ion . ttandurd for vinyl stripper (or the reactor If the plant has and pH. usually 1or the purpose of meet chloride plants. no stripper) In the plant process flow is ing effluent guidelines and standards: It An owner or operator of a vinyl chlo not to exceed 10 ppm. except as provided does not include any process the purpose ride plant shall comply with the require In 3 61.65(a). This requirement does not of which is to remove vinyl chloride from ments of this section and 3 61.65. apply to equipment that has been water to meet requirements of this (a) Vinyl chloride formation and puri opened, is out of operation, and met the subpart. fication: The concentration of vinyl requirement in f 61.63(b) (8) (1) before U) `-In vinyl chloride service" means chloride in all exhaust gases discharged being opened. that a piece of equipment contains or to the atmosphere from any equipment (d) Monomer recovery system. The contacts either a liquid that Is at least used In vinyl chloride formation and/or concentration of vinyl chloride in all ex 10 percent by weight vinyl chloride or a purification is not to exceed 10 ppm, ex haust gases discharged to the atmos gas that is at least 10 percent by volume cept as provided In 3 61.65(a). This re phere from each monomer recovery sys vinyl chloride. quirement does not apply to equipment tem Is not to exceed 10 ppm, except as (m) "Standard operating procedure" that has been opened, is out of operation, provided in 3 61.65(a). This requirement means a formal written procedure offi and met the requirement In ] 61.65(b) cially adopted by the plant owner or (6) (1) before being opened. - does not apply to equipment that has been opened. Is out of operation, and met operator and available on a routine basis to those persons responsible for carrying out the procedure.. (n) "Run" means the net period of time during which an emission sample Is collected. to) "Ethylene dlchlorlde purification" Includes any part of the process of ethyl ene dlchlorlde production which follows ethylene dlchlorlde formation, and tn which finished ethylene dlchlorlde Is produced. ' y'vL- (p) "Vinyl chloride purification" In cludes any part of the process of vinyl chloride production which, follows vinyl chloride formation and in which finished vinyl chloride is produced. 61.64 Emission Mandanl for polyvinyl chloride pisnu. An owner or opermtor.-of a polyvinyl chloride plant shall comply with the re quirements of this section and 1 61.65. .. (a) Reactor: The following require ments apply to reactorsr - - (1* The concentration of vinyl chlo ride in all exhaust gases, discharged to the atmosphere from each reactor is not to exceed 10 ppm except as provided in paragraph (a)(2K-of this.section and 3 61.65(a). ..-Jr-'- (2) The reactor opening lose from each reactor Is not to exceed 0.02 g vinyl chlorlde/Kg (0.00002 lb vinyl chloride/ lb) of polyvinyl chloride product, with the requirement In | 61.66(b) (8) (1) be fore being opened. (e) Sources following the stripperCs): The following requirements apply to ' emissions of vinyl chloride to the at mosphere from the combination of-all sources following the strtpper(s) Cor the reactor(s) If the plant has no strip- peris)] in tha plant process flow In cluding bub nob limited to, centrifuges, concentrators, blend tanks, filters, dry ers. conveyor air discharges, baggers, storage containers, and lnprocess waste-' water: -. (1) In polyvinyl chloride plants using stripping technology to control vinyl chloride emissions, the weighted average residual vinyl chloride concentration in (q) "Reactor" Includes any vessel in. the product determined on a. dry solids all grades of polyvinyl chloride resin which vinyl chloride Is partially or totally basis. This requirement applies to any processed through the stripping opera polymerized Into polyvinyl chloride. vessel which is used as a reactor or as tion on each calendar day, measured (r) "Reactor opening loss" means the both-a reactor and a stripper, m the immediately after the stripping opera emissions of vinyl chloride occurring bulk process, the product means tile tion is completed, may not exceed: when a reactor Is vented to the atmos gross product of prepolymerization and (1) 2000 ppm for polyvinyl chloride phere for any purpose other than an postpolymerization. . -i - dispersion resins, excluding latex resins; emergency relief discharge as defined In , (3) Manual vent valve discharge: Ex > <U> -400 ppm for all other polyvinyl 161.65(a).'' cept for an emergency manual vent valve chloride resins, including latex resins, (s> "Stripper" includes any vessel m which residual vinyl chloride Is removed from polyvinyl chloride resin, except bulk resin, in the slurry form by the use of heat and/or vacuum. In the case of bulk resin, stripper includes any vessel which is used to remove residual vinyl chloride from polyvinyl chloride resin immediately following the polymeriza tion step in the plant process flow. discharge, there is to be no discharge to the atmosphere from any manual vent valve on a polyvinyl chloride reactor in vinyl chloride service. An emergency menual vent valve discharge means a discharge to the atmosphere which could not have been avoided by taking meas ures to prevent the discharge. Within 10 days of-any discharge to the atmosphere from any manual vent valve, the owner or operator of the source from which the averaged separately for each type of res in; or .~ (2) In polyvinyl chloride plants con trolling vinyl chloride emissions with technology other than stripping or in addition to stripping, emissions of vinyl cchiclotrcidde: to the atmosphere may not (I) 2 g/kg (0.002 lb/lb> product from the stripperis) (or resctor(s) if the plant has no stripper(s) 1 for dispersion 61.(52 Emission standard for ethylene discharge occurs shall submit to the Ad polyvinyl chloride resins, excluding latex rfii'hloride plants. ministrator a report in writing contain resins, with the product determined on a An owner or operator of on ethylene ing information on the source, nature dry solids basis; dichloride plant shall comply with the requirements of this section and I 61.65. ta> Ethylene dichloride purification: The concentration of vinyl chloride la all exhaust gases discharged to the at mosphere from any equipment used In ethylene dlchlorlde purification Is not to exceed 10 ppm, except as provided In 5 61.65(a). This requirement doee not end cause of the discharge: the date and time of the discharge, the approximate total vinyl chloride loss during the dis charge. the method used for determining the vinyl chloride lose, the action that was taken to prevent the discharge, and measures adopted to prevent future dis charges. (b) Stripper: The concentration of vinyl chloride in all exhaust gases dis charged to the atmosphere from each (II) 0.4 g/kg (0.0004 lb/lb) product . from the strippers [or reactor(s) if the plant has no sbrlpper(s) ] for all other polyvinyl chloride resins, including latex resins, with the product determined on a dry solids basis. 3 61.65 Emiiuivn >lunddnl fur oil,,' 1,-ne dii-liloriile, vln.>t chloride and poly vinyl chloride plant*. An owner or operator of an ethylene apply to equipment that has been stripper Is not to exceed 10 ppm, except dlchlorlde. vinyl chloride, and/or poly opened, is out of operation, and met the as provided in 3 61.65(a). This require vinyl chloride plant shall comply with requirement in 5 61.65(b) before being ment does not apply to equipment that the requirements of this section. opened. (b) Oxychlorination reactor: Except as provided in 1 61.65(a), emissions of vinyl chloride to the atmosphere from each oxychlorination reactor are not to has been opened, is out of operation, and met the requirement in I 61.65(b) (6) (1) before being opened. (c> Mixing, weighing, and holding containers: The concentration of vinyl chloride In all exhaust gases discharged (a) Relief valve discharge: Except for an emergency relief discharge, there is to be no discharge to the atmosphere from any relief valve on any equipment in vinyl chloride service. An emergency relief discharge means a discharge which exceed 0.2 g/kg the 100 percent ethylene to the atmosphere from each mixing, could not have been avoided by taking bichloride product from the oxychlorl- weighing, or holding container In vinyl measures to prevent the discharge. With r.ation process. chloride service which precedes the in 10 days of any relief valve discharge. HDWAl IE6I5TC*, VOL 41, NO. 305--THLH5DAV, OCTOSH 31, 1*Z 95 U Ii '> 1 fl>' COLORITE 007388 I""4606$ RULES AND REGULATIONS - the owner or operator of the source from system from which the concentration of ride Is to be reduced so that the equip 'm _which the relief waive discharge occurs vinyl chloride in the exhaust gases does ment contains no more than 3.0 percent ] shall submit to the Administrator a re not exceed 10 ppm; or equivalent as by volume vinyl chloride or 0.0850 m` (25 port m writing containing information provided In { 61.68. gal) of vinyl chloride, whichever is on the source, nature and cause of the (ill) Rotating compressor: Vinyl larger, at standard temperature and discharge, the date and time of the dls- chloride emissions.from seals on ell ro pressure: and r charge, ihe approximate total vinyl ehlo- tating compressors in vinyl chloride (U> Any vinyl chloride removed from -^'*2 I ride loss during the discharge, the meth- service are to be minimized by installing the equipment in accordance with para I od used for determining the vinyl chlo compressors with double mechanical graph (b) (8) (1) of. this section is to be ride loss, the action that was taken to seals, or equivalent a* provided in 161.86. ducted through a control system from prevent the discharge, and measures if double mechanical teals are used, vinyl which the concentration of vinyl chlo - adopted to prevent future dischargee. chloride emissions from the seals are to ride In the exhaust gases does not exceed (b) Fugitive emission sources: ba minimized by mtuntjininy the pres 10 ppm, or equivalent as provided in < 1) Loading and unloading linear Vinyl sure between the two seals so that any <81.68. chloride emissions from loading and un- leak that occurs la Into the compressor; (7) Samples: Unused portions of sam leading line* which are opened to the by ducting any vinyl chloride between ples containing at least 10 percent by r~atmosphere after each loading or un the two seals through, a control system weight vinyl chloride are to be returned loading operation are to be minimized from which the concentration of vinyl to the process, and. sampling techniques as follows: chloride in the exhaust gasea does not are to be such that sample containers in (D After each loading or unloading exceed 10 ppm; or equivalent as provided vinyl chloride service are purged into a .operation and before opening a loading in 9 61.66. *' doeed process system. - ... . or unloading Una to the atinoephere, the <lv) Reciprocating compressors: Vinyl (8) Leak detection and elimination: quantity of vinyl chloride in all parts of chloride emissions from seals an all re Vinyl chloride emissions due. to leaks- each loading or unloading lina that are ciprocating compressors in vinyl chloride from equipment in. vinyl chloride service to be opened to the atmosphere is to be service are to be by installing are to- be minimized by instituting and --reduced so that the parts combined con double outboard seals, or equivalent as implementing a. formal.'leak detection tain no greater than 0.0038 m* <0-13 ft*> provided, in 1 81.88. If double outboard and elimination program. The owner or of vinyl chloride, at standard tempera seals are used, vinyl- chloride emissions operator shall submit a description of ture and pressure; and ... from the seals are to be minimized by the program to the Administrator for (11) Any vinyl chloride removed from maintaining the pressure between the approval The program is to be sub --a loading or unloading line In accord two seals so that any leak that occurs is mitted within 43 days of the effective ance with paragraph (b) (1HI) of this lnto ttoe compressor; by ducting any date of these regulations, unless a waiver section is to be ducted through a control vinyl chloride between the two seals of compliance is granted under { 61.11. system from which the concentration of through a control system from which the If a waiver of compliance is granted, the vinyl chloride in the exhaust gases does concentration of vinyl chloride In the program is to be-submitted on a date not exceed 10 ppm. or equivalent sa pro exhaust gases does not exceed 10 ppm; scheduled by the Administrator. Ap vided in 16LM. or equivalent as provided in 1 81.68., proval of a progianr will be granted by <2> sup gauges: During loading or un <v> Agitator; Vinyl chloride emissions the Administrator provided he finds: loading operations, the vinyl chloride from seals on all agitators in vinyl chlo - (i> It Includes a reliable and accurate emissions from each slip gauge In vinyl ride service are to be minimized by In vinyl chloride monitoring system for de ! chloride service are to be minimised by stalling agitators with double mechani tection of major leaks and identification j ducting any vinyl chloride discharged cal seals, or equivalent as provided la of the general area of the plant where a from the slip gauge through a control i 61.68. if double mechanical seals are leak is located. A vinyl chloride monitor system from which the concentration of used, vinyl chloride emissions from the ing system means a device which obtains -vinyl chloride In the exhaust gasaa does seals are to be minimised by maintaining air samples from one or more points on not exceed 10 ppm. or equivalent aa pro the- pressure between the two seals so a continuous sequential basis and ana- vided in f 81.08. that any leak that occurs is Into the agi lyres the samples with gas chromatog (3) Leakage from pump, compressor, tated vessel; by ducting any vinyl chlo raphy or. if the owner or operator as and agitator seals: - ride between the two seals through a sumes that all hydrocarbons measured -- (1) Rotating pumpe: Vinyl chloride control systen from which the concen are vinyl chloride; with Infrared spectro emissions from seals on all rotating tration of vinyl chloride In the exhaust photometry flame ion detection, or- an pumps In vinyl chloride service are to be minimized by installing seailes pumps, pumps with double mechanical seals, or ' equivalent as provided fin 181.88. H double mechanical seals are used, vinyl chloride emission from the seals are to be minimized by maintuning the pres sure between the two saala so that any leak that occurs Is Into,the pump: by ducting any vinyl chloride between the two seals through a control system from which the concentration of vinyl chlo ride in the exhaust gases does not ex ceed 10 ppm; or equivalent as provided in {61.68. (11) Reciprocating pumps: Vinyl chlo ride emissions from seals on all recipro cating pumps tn vinyl chloride service are to be minimized by Installing double outboard seals, or equivalent as provided in f 61.88. If double outboard seals are used, vinyl chloride emissions from the seala are to be minimized by maintaining the pressure between the two seal3 so that any leak that occurs Is Into the pump; by ducting any vinyl chloride be gases doe* not exceed 10 ppm; or equiva lent as provided In I 61.66. (4> Leakage from relief valves:- Vinyl chloride emissions due to leakage from each relief valve on equipment in vinyl chloride service are to be minimized by installing a rupture disk between the equipment and the relief valve, by con necting the relief valve discharge to a process line or recovery system, or equiv alent a provided In 181.68. (5) Manuel venting of gases: Except ms provided in 1 81.64(a) (3), all gases which are manually vented from equip ment in vinyl chloride service are to- be ducted through a control system from which the concentration of vinyl chloride in the exhaust gases does not exceed 10 ppm; or equivalent as provided in ! 61-68. (6) Opening of equipment: Vinyl chloride emissions from opening of equipment (including loading or unload ing lines that are not opened to the at mosphere after each .loading or unload ing operation) are to be minimized as follows: (1) Before opening any equipment tor equivalent or alternativemethod. (11) It includes a reliable and accurate portable hydrocarbon detector to be used routinely to find small leaks and to pin point the major leaks indicated by the vinyl chloride monitoring system. A portable hydrocarbon detector means a device which measures hydrocarbons with a sensitivity of at least 10 ppm and Is of such design and size that it can be used to measure emissions from local ized points. (ill) It provides for an acceptable cali bration and maintenance schedule for the vinyl chloride monitoring system and portable hydrocarbon detector. For the vinyl chloride monitoring system, a daily span check is to be conducted with a concentration of vinyl chloride equal to the concentration defined as a leak ac cording to paragraph (b) (3) (vi) of this section. The calibration is to be done with either: (A) A calibration gas mixture pre pared from the gases specified in sections tween the two seals through a control any reason, the quantity of vinyl chlo 5.2.1 and 5.2.3 of Test Method 106, or MDMAL ESCIITM, VOi. 41, NO. 205--TMUMDAr, OCTOSEi 21, 1070 96 COLORITE 007389 RULES ANO REGULATI NS 46567 < B > A calibration gas cylinder contain- vinyl chloride in equipment a*4.75 m* of emission test results and other data Ins the appropriate concentration of vinyl chloride. If a calibration gas cylin der is used, the analysis must be trace able to the National Bureau of Stand ards or to a gravlmetrically calibrated 01250 gal In volume for which an mis sion limit is prescribed In 9 61.65(b) (6) (1> prior to opening the equipment end using Test Method 106, a portable hydro needed to determine emissions. (g> Unless otherwise specified, the owner or operator shall use test Test Methods In Appendix B to this part for each test as required by paragraphs vinyl chloride permeation tube. carbon detector, or an equivalent or al .(g)(1), (g)(2). (g)(3), (g)(4), and i iv i The iocatlon and number of points ternative method. The method of meas (g) (9) of this section, unless an equiva to be monitored and the frequency of urement Is to meet the requirements in lent method or an alternative method monitoring provided for in the program are acceptable when they are compared 9 61.67(g) (SHI) (A) or (g)((5)(i)(B). has been approved by the Administrator. If the Administrator finds reasonable with the number of pieces of equipment 61.66 Equivalent equipment and pro grounds to dispute the results obtained in Tinyl chloride service and the sice and cedures. ' - by an equivalent or alternative method, physical layout of the plant. `T> It contains an acceptable plan of action to be taken when a leak Is de tected. <vi> It contains a definition of leak Upon written application from an own er or operator, the Administrator may approve use of equipment or procedures which have been demonstrated to his he may require the use of a reference method. If the results of the reference and equivalent or alternative methods do not agree, the results obtained by the reference method prevail, and the Ad which is acceptable when compared with - satisfaction to be equivalent In terms of ministrator may notify the owner or the background concentrations, of vinyl reducing vinyl chloride emissions to the operator that approval of the method chloride In the areas of the plant to be monitored by the vinyl chloride monitor ing system. Measurements of background concentrations of vinyl chloride in the. areas of the plant to be monitored by the vinyl chloride monitoring system are to be included with the description of the program. The definition of leak for a given plant may vary among the differ ent areas within the plant and is also to change over time as background con centrations In the plant are reduced. _ f 9) Inprocess wastewater: Vinyl chlo-^ atmosphere to that* prescribed for com pliance with a specific paragraph of this subpart. For an existing source, any re quest for using an equivalent method as the initial measure of. control is to be submitted to the Administrator within .30 days of the effective date. For a new source, any request for using an equlvalent method is to be submitted to the Administrator with the application for approval of construction or modification required by 9 01.07. " - - - J previously considered to-be equivalent or alternative Is withdrawn. U) Test Method 106 is to be used to determine the vinyl chloride emissions from any source for which an emission limit Is prescribed In ft 61.63(a). or (b> 9 61.63(a), or ff 61.64(a) (1). (b). (c). or <d>. or from any control system to which reactor emissions are required to be ducted In f 81.64(a) (2) or to which fugi tive emissions are required to be ducted In 9161.66(b) (lHtt), (b)(3), (b)(5), (b)(8) (11),or (b)(9)(H).' ride emissions to the atmosphere from 61.67 Emission testa. (I) For each run, one sample Is to be inprocess wastewater are to be reduced as follows: \- U) The concentration of vinyl chlo ride in each inprocess wastewater stream containing greater than 10 ppm vinyl chloride measured Immediately as it leaves a piece of equipment and before' being mixed with any other Inprocese wastewater stream Is to be reduced to no more than 10 ppm by weight before being mixed-with any other inprocess wastewa ter stream which Contains less than 10- (a> Unless a waiver of emission testing is obtained under 9 61.13, the owner or operator of a source' to which this sub part applies shall test emissions from the source, -- - - - (1) Within 90 days of the effective date In the case of an existing source or a new source which has an initial startup, date preceding the effective data, or (2) Within 90 days of startup in the case of a. new source. Initial startup of which occurs after the effective date. collected. The sampling site Is to be at least two stack or duct diameters down stream and one half diameter upstream from any flow disturbance such ae a bend, expansion, contraction, or visible flame. For a rectangular cross section an equivalent diameter Is to be determined from the following equation:-- equivalent dinwter.2 width) 1 lengtl+width The sampling point in. the duct ts to ppm vinyl chloride: before being exposed (b) The owner or operator shall pro be at the centroid of the cross section. to the atmosphere, before being dis charged to a wastewatsr treatment proc vide the Administrator at least 30 days prior notice of an emission test to afford The sample Is to be extracted at a rate proportional to the gas velocity at the ess: or before being discharged untreated the Administrator the opportunity to sampling point. The sample- 1s to be as a wustewatee. The paragraph does' have an observer present during the test. taken over a of one hour, and apply to water which Is used to displace . (c> Any emission test Is to be con is to contain a minimum volume of 50 vinyl chloride from equipment before It ducted while the equipment being tested litem corrected to standerd conditions. Is opened to the atmosphere la accord ance with 9 81.64(a) (2) or paragraph (t (fli of this section, but does not apply to water which is used to wash out equip ment after' the equipment has `already been opened to the atmosphere In ac cordance with 9 01.64(a) (3) or para is operating at the maximum production rate at which the equipment will be op erated end under other relevant condi tions as may be specified by the Adminis trator based on representative perform ance of the source. (d) Each emission test Is to consist (II) For gas streams containing more than 10 percent oxygen, the concentra tion of vinyle chloride as determined by Test Method 106 Is to be corrected to to percent oxygen for determination of emissions by using the following equa tion: graph tb> (6> of this sactlon. 01> Any vinyl chloride removed from the inprocess wastewater St accordance with paragraph (b> (9) (1) o< this section is to be ducted through a control system from which the concentration of vinyl chloride in the exhaust gases does not exceed 10 ppm, or equivalent as provided in 5 61.56. <c> The requirements in paragraphs <b>(l>. 1 b) (2), (b)(6). (b)(6), (b)(7) and (b> (3) of this section are to be In of three runs. For the purpose of deter mining emissions, the average of results of all runs Is to apply. The average Is to be computed on a time weighted basis. (e> All samples are to be analyzed within 24 hours, and vinyl chloride emis sions are to be determined within 30 days after the emission test. The owner or operator shall report the determinations to the Administrator by a registered letter dispatched before the close of the next business day following the deter C*iwbera- C:20.9--p10e.r0cent Oj Ctfarreciw*) concentration ol rinyl chloride In tbi Mtan* pm, eocmtod to 10 per* DOTMfltOVTftlfc. Ci-Tne concentration ot vtnyl chloride ns measured by Test Method iotf, a0A*PeCeaf oiyftti in tho ambient air ot standard conditions. tOA*Percent otyfea m the ambient air at standard condition*, minus tbt 10 pccoMtC omm to which tho correc tion is betas Bute, Percent OiPercent osyfea ia tbo exhaust cos ns corporated into a standard operating mination. procedure, anr made available upon re (f) The owner or operator shall retain measured by Reference Method 3 in Appendix Aot Part 80< this chapter- quest for inspection by the Administra at the plant and make available, upon (111) For those emission sources where tor. The standard operating procedure Is request, for Inspection by the Adminis the emission limit Is prescribed In terms to Include provisions for measuring the trator. lor a minimum of 2 years records of mass rather than concentration, mass FfOCRAL RfpISTEf, VOL *1, NO. 203--THURSDAY, OCTOBIR II, IW* 97 COLORITE 007390 J~ 4636S Rin.ES AND REGULATIONS' emissions in kg/100 kg product arc to r be determined by using the following equation: ,, [C* (2.60) 0 10-] tlOOl that Is also used as a stripper, the deter mination may be made immediately fol lowing the stripping operation. (1) Except aa provided in paragraph (g)(9) <11) of this section, the reactor detection, or an equivalent or alterna tive method. The vinyl chloride monitor ing system used to meet the requirements in 1 61.69(b) (8) (I) may be used to meet the requirements of this section. wr.rf*: r cx-ki Tinjl cUartlWlOO kc pmdnrt. C-Tt* >DntitiUw at najl oDiofld* w nUMune i by Tot MvtSod los. VC I at rinyl eRlorM* it an* itDMpban and d (J-Volonnolo flow nil Id affix u d*Urmlnl by opening loss la to be determined using (c) A dally span check, is to be con the following equation: ducted for each vinyle chloride monitor ing system used. For all of the emission ,, W (2.60) (10-) (Cb) sources listed in paragraph (a) of this where . YZ. section, except the one for which an emis- slon limit is prescribed in l 61.62(b), the r-- fttltitm MttBod 3 of AppondU A t* Pan w of ttfl* dapur. 10-' - CouthMoo ItcUMM ppm. Z-Pndnctleant* Cnfit). C-tt Ytay1 cbtaridftamiartaai/fcf pfoduct. dally span check is to be conducted with jyCftp*ity of tharapetorla m*. "5.G0Dot^ofTlgl eUoddi N oat Atmosphere Rad a concentration of vinyl chloride equal to 10 ppm. For the emission source for (2) Test Method 107 Is to be used to determine the concentration of vinyl CoaTirsloo footer far opoL Cbwtppqt by^eotomft etayl chloridt m datmnlsad by which an emission limit is prescribed In - . Tot liftthod lOA'or a portable hydrocarboo i 61.62(b), the daily span check is to be P~ chloride in each inprocess wastewater, stream for which an emission limit is detector whleb motsam hydrocarbons with a itaattMty ef at least 10 pptt. --y--Namber of ano* the reeator was last ' conducted with a concentration of vinyl chloride which is determined to be prescribed in J 61.63(b) (9) 0). (3) Where a stripping operation la opened to thestmoephfe. _ z^Awfart ks of potyetoyk eUoride produced par .. bftteh In UMtrmnbvrfbtchejnic the n*ctor determined to be equivalent to the emis sion limit for that source based on the ,, used to attain the emission limit in 18L- 64(e), emusinnp are to be determined i using Test Method 107 as follows: waslRffftp--dtothofttmospher^ emission teat required by f 67.67. The (A) If Methodl l'06^is used to deter calibration is to be done with eitherr (1) A calibration gas- mixture pre (i) The number of strippers and sam mine the concentration of vinyl chloride pared from the gases specified in sections ples and the types and grades of resin to (Cb) ,. the sample isr-tn be withdrawn at 5.2.1 and 5-2.3 of Test Method 106, or . ^be sampled are to be determined by the a constant rate with a probe of sufficient (2) A calibration gas cylinder con Administrator for each individual plant length to reach the vessel bottom from taining the appropriate concentration of i the Urns of. the test based on the the manhole. Slamplea are to be taken - vinyl - chloride. If a. calibration gas plant's operation. for. 3 miwiitM within. 6 inches of the ves cylinder is used, the- analysis must. be 01) Each sample is to be taken lmme- sel bottom. 9 mtwwt-- pear the vessel traceable to the- National. Bureau of - diately following the stripping operation. center, and 5 minutes near the vessel top. Standards or to a gravtmetrically cali (ill) The corresponding quantity of . (B) If a portable hydrocarbon detec brated vinyl chloride permeation tube- . material processed by each shipper is to be determined on a dry solid# basis and by a method submitted to and approved : by the Administrator. (Iv) At the prior request of the Ad ministrator, the owner or operator shall provide duplicates of the samples re quired in paragraph (g) (3) (1) of -this section. (4) Where control technology other than or in addition to a stripping opera tion 1$ used to attain the emission limit r~in 161.64(e).-emissions are to be deter| mined aa follow#: d) Test Method 106 is to be used to determine atmospheric emissions from, ah of the process equipment simultane ously. The requirements of paragraph, (g) (1) of this section are to be met. (11) Test Method 107 is to be used to determine the concentration of vinyl - chloride in each lnprocess wastewater stream subject to the emission limit pre scribed in i 61.64(e). The mass of vinyl tor Is used to determine the concentra tion of vinyl chloride (Cb), a probe of sufficient length tn reach the vessel bot tom from the manhole is to be used to make the measurements. One measure ment will be made within 9 inches of the vessel bottom, one near the vessel center and one near the vessel top. Measure ments are to be made at each location until the reading is stabilized. All hydro carbons measured, ere . to be assumed to be vinyl chloride. -(C) The production rate of polyvinyl chloride (Z) ie to be determined by a method submitted to end approved by the Administrator. (U) A calculation based on the number of evacuations, the vacuum involved, and the volume of gas In the reactor Is hereby approved by theAdministrator as an al ternative method far- determining reac tor opening loss for poetpolymerization reactors in the manufacture of bulk resins. 9 61.69 Initial report.. - (a) An - owner or operator of any source to which this subpart applies ha.ii submit a statement in writing notifying the- Administrator that, the equipment and procedural specifications in 51 61.65 (b)(1), (b)<2h (b)(3), (b)(4), (b)(5), (b)(6). (b)(7),. and (b)(8) are being implemented. (b) (1) In the case- of an existing source or a new- source- which has an Initial startup date preceding the effec tive date, the statement is to be submit ted within 90 days of the effective date, unless a waiver of compliance is granted under f 81.11. along- with, the informa tion required under f 81.10. If a waiver of compliance is granted, the statement Is to be submitted on a date scheduled by the Administrator: (2) In the ease of a new source which did not have an Initial stamp date pre ceding the effective date, the statement Is to be submitted within 90 days of the chloride in kg/100 kg.product hi each 3 61.68 Emi*i*on monitoring. Initial startup date. in process wastewater stream is to be de termined by using the following equa tion: Cg t tiny! kf pcotaet. coocanaatfon of Ttnjl chloiido > mssinrsd i .EmwbayteTr&itrtrMauiMnVMlir.1d0o7.toRaJ&odiQACMtilaiM* . witb DMUxod which hi* bora rabmittcd to ad appro**} by th* Administrator. ' li)"1 - Coaversioniftetov for ppm. Z-ProdaetloG rats Ocy/br). detorainod in accord- one* with ft method which baa baan submitted and tpproTod by th* Admlnifttrfttof. (a) A vinyle chloride monitoring sys tem is to be used to monitor on a con tinuous basis the. emissions from the sources for which emission limits are pro scribed hr f 61.62(a) and (b>, 1 61.63(a), and i 61.64(a)(1), (b>. (c),and (d).and for any control system to which reactor emission are required to be ducted In i 61.69(b) (1) (il), and (b)(2), (b)(9). (b> (6) <ll), and (b) (9) (ii). (b) The vinyl chloride monitoring sys tem(s) used to meet the requirement in paragraph `a) of this section is to be a (c> The statement is to contain the following information: (1) A list of the equipment installed for compliance, (2) A description of the physical and functional characteristics of each piece of equipment. . (3) A description of the methods which have been incorporated into the standard operating procedures for meas uring or calculating the emissions for which emission limits are prescribed in 19 61.65 (b) (1) (1) and (b) (6) (1). (4) A statement that each piece of (5) The reactor opening loss for which device which obtains air sampels from equipment is Installed and that each an emission limit is prescribed in i 61.64 one or more points on a continuous piece of equipment and each procedure <&) (2) is to be determined. The number sequential basis and analyzes the samples is being used. of reactors for which the determination is to be made is to be specified by the Administrator for each individual plant at the time of the determination based on the plant's operation. For a reactor with gas chromatography or, if the owner or operator assumes that all hydrocar bons measured are vinyl chloride, with infrared spectrophotometry, flame ion 61.70 Si niif.nnu.il report. (a) (2) is to be determined. The number source to which this subpart applies shall submit to the Administrator on Septem- FEDERAL REGISTER, VOt. 41. NO. 205--THURSDAY, OCTOBER 21, 1976 98 COLORITE 007391 KU1U ANU JlBOULAIIONi 40309 ber 15 and March 15 of each year a report In writing containing the Information required by this section. The Ant semi annual report is to be submitted follow of 8 hours for each grade of resin which is being processed, whichever is more fre quent The sample la to be taken as the resin flows out of the stripper and iden may be made immediately following the stripping operation. 361.71 Recordkeeping. ing the first full a month reporting period after the initial report is submitted. (b)(1) In the case of an existing source or a new source which baa an initial startup data preceding the effective date, the first report is to be submitted within 180 days of the effective date, unless a waiver of compliance Is granted under tified by resin type and pads and the (a) The owner or operator of any date and time the sample was taken. source to which this subpart applies shall The corresponding quantity of material retain the following Information at the processed by each stripper over the time source and make it available for inspec period represented by the sample during tion by the Administrator for a mini the eight hour period, is to be recorded mum of two years; and identified by resin type and grade (1) A record of the leaks detected by and the date and time it represents. - the vinyl chloride monitoring system, as 3 61.11. If a waiver of compliance la (ill) The quantity of material proc required by 3 61.65(b) (8), including the granted, the first report is to be sub essed by the stripper la to be determined concentrations of vinyl chloride as mitted on a date scheduled by the Ad on a dry solids basis and by a method measured, analyzed, and recorded by the ministrator. . submitted to and approved by the Ad vinyl chloride detector, the location of (2> in the case of a new source which ministrator. each measurement and the date and ap did not have an Initial startup date pre (lv> At the prior requeai of the Ad proximate time of each measurement. ceding the effective data, the first report ministrator, the owner or operator shall (2) A record of the leaks detected is to be submitted within 180 days of the provide duplicates of the samples re during routine monitoring with the initial startup date. -. quired in paragraphs (c) (2) (1) aim (c> portable hydrocarbon detector and the (c) Unless otherwise specified, the (2) (ID of. this section. action taken to repair the leaks, as. re owner or operator shall use the Test (v) The report to the Administrator by quired by 3 61.65(b)(8). including a Methods to Appendix B to this part to the owner or operator is to Include the brief statement explaining the location conduct emission testa as required by vinyl chloride content found to all the and cause of each leak' detected with paragraphs (e)(2) and (c)(3) of this sampler required: in paragraphs (O (2) the portable hydrocarbon detector, the section, unless an equivalent or an alter (i) and (c) (2) (ill of this section, aver date and time of the leak and any action native method has been approved by the aged separately for each type of resin, taken to eliminate that leak measured in Administrator. If the Administrator over each calendar day and weighted ac accordance with 3 81.68. finds reasonable grounds to dispute the results obtained by an equivalent or al ternative method, ha may require the use of a reference method. If the results of the reference and equlvalmt or alterna cording to the quantity of each grade of resin processed by the stolpper(s) that calendar day. according to the following equation: ~` (3) Pbr the relief discharges ' from reactors subject, to the provisions of 3 61.65(a), a daily operating record lor each reactor, including, pressures and temperatures............... tive methods do not agree, the results 2. Appendix B is amended by adding obtained by the reference method pre Teet Methods 106 and 107 as follows: vail. and the Administrator may notify Itenme lOS--Dtminxinas or Vnm the owner or operator that approval of CnsiDt ROM STATtOHMT SO UlCkS the method previously considered to be equivalent or alternative is withdrawn. - (1) The owner or operator shall in clude in the report a record of any emis sions which averaged over any hour period (commencing on the hour) are +r"_ "" * /* Qrr "' A-Mwur seaesgi eonseotradw ti type T, rede is ppm. _Q-T_oPttstol pde,eldaoketti.oa eftype r. Ribs ever tbejabeiw nmoDtrcnoM ' Performance of this method should not bo attempted by psrsons unfsmliur witn th* operation of a gas chromatograph, nor by those who an unfamiliar with source sam pling. as there are many dstaUs that an in excess of the emission, limits pre scribed in I) 81.62(a) or (b), 181.63(a). or 3} 61.84(a) (1), (b), (e), or (d), or for1 any control system to which reactor emissions are required to be ducted to I 61.64(a) (2) or to which fugitive emis sions are required to bo ducted to 181.88 Ti-Trpe of RMs: Jmt& ... m where a Is t*e*l number cf redo typee pnrdaesd doing tbs 2A beurpedoA ' Af-Cuseatrstlan o( vtsvt OM1> Is cos sample of rrade 0, RMS, is ppw. F--PimlasMse o< mk <h reMn wptmntsd br lbs tapis, is kg. 0i-Orads at rsIs. eg, ft, 0%. ssd (h, o-Tatak number aTpUm < nMs pnSso the 244ttar period. beyond the scope of this prssentetlon. Can must be sxerclsed to prevent exposure of sampling personnel to vinyl chloride, a car cinogen. - 1. Principle and Applicability. ' 1.1 An intagrated bag sample of stack gns- contalnlng vinyl chlarlda (chlorosthylsns) Is subjected to chromatographic analysis, (bMIXU), (b)(2), (b)(8), (b)(6)(ID.or (b)(9) (11). Tho emissions are to be meas ured in accordance with 181.68. (vt) The owner or operator shall re tain at the source and make available for Inspection by the Administrator for (2> In polyvinyl chloride plants far ` a minimum of 2 yean records of all data' which a stripping operation is used to needed to furnish the Information re attain the emlsison level prescribed to quired by .paragraph (c) (2) (v) of this 1 61.64(e), the owner or operator shall section: The records are to contain the include in the report a reoord of tho following information: vinyl chloride content la the polyvinyl (A) The vinyl chloride content found chloride resin. Test Method MT Is to be used to determine vinyl chloride content to all the samples required to paragraphs (e) (2) (1) and (c) (2) 01) of this section, as follows: identified by the resin type and grade using a flame Ionisation detector. m The method Is applicable to the meas urement of vinyl chloride In stack gases from ethylene dichlortde, vinyl chloride and poly vinyl chloride manufacturing .processes, ex- cept where the vinyl chloride is contained in psrtlcuiate matter. 2. Range and Sensitivity. The lower limit of detection will vary ac cording to the chromatograph used, values reported include 1 X 10-' mg and 4 X 10-' mg. 3. Interferences. --, Acetaldehyde, which can occur in some (1) IT batch stripping is csed. one rep and the time and date of the sample, and vinyl chloride sources, will interfere with the resentative sample of polyvinyl chloride' 1 (B) The corresponding quantity of elnyl chloride peek from the Chromosorb 1Q2 resin is to be taken from each batch of each grade of resin immediately follow ing the completion of the stripping, and grade and the date and time the batch is completed. The corresponding quantity of material processed to each stripper batch is to be recorded and Iden tified by resin type and grade and the date and time the batch is completed. <ii> If continuous stripping is used, one representative sample of polyvinyl polyvinyl chloride resin processed by the strlpper(s), identified by the resin type and grade and the time and date it represents. (3) The owner or operator shall in clude to the report a record of the emis sions from each reactor opening for which an emission limit Is prescribed in 3 61.64(a) (2). Emissions are to be deter mined in accordance with 3 61.61(g) (3). except that emissions for each reactor column. See sections 4.3.2 and 6.4. If resolu tion of the elnyl chloride peak is still not satisfactory for a particular sample, then chromatograph parameters can be further altered with prior approeal of the Admin istrator. If sltermtlon of the chromatograph parameters fails to resolve the vinyl chloride peak, then supplemental confirmation of the vinyl chloride peak through an absolute analytical technique, such as mass spectro scopy, must be performed. 4. Apparatus. 4.1 Sampling < Figure l). chloride resin is to be taken for each are to be determined. For a reactor that Is 4.1.1 Probe--Stainless stesl. Pyrex glass, grade of resin processed or at intervals also used as a stripper, the determination or Teflon tubing according to stock temper- FEDMAl 2COISTG2, VOL *1, NO. 203--THUX5DAV, OCTOIIR 21, 1V76 99 COLORITE 007392 46570 RULES AND REGULATIONS sturv. each equipped with a glees wool pluf K remote partleuiet* matter. 4.14 Sample line--Teflon. 0.4 mm outside diameter, of sufficient length to connect prop# to bag. A new unused piece la employed for each series of bag samples that constitutes an emission test. '' 4.1.3 Male (3) and female (3) stainless steel quick-connects, with ball checks (one pair without) located aa shown In Figure 1. 4.1.4 Tidlar bags, 100 liter capacity--To contain sample. Teflon begs are not accept able. Aluminized Mylar baga may be used, provided that the samples are analyzed within 24 hours of collection. 4 14 Rigid leakproof containers for 4.1.4, with covering to protect contents from sun light. 4.1.4 lleedle valve--To adjust sample flow rate. 4.1.7 Pump--Leak-free, "i"1".'" capac ity 2 liters per minute. - 4.14 Charcoal tube--To prevent admit- sion of vinyl chloride to atmosphere in vicin ity of samplers. . 4J4 Plow meter--For observing sample flow rste; capable of measuring a flow range from 0.10 to 140 liter per minute. 4.1 JO Connecting tubing--Teflon, 6.4 mm outside diameter, to sample train (Figure 1). 4.1.U Pitot tube--Type S (or equivalent), kttached to the probe so that the sampling flow rate can he regulated proportional to the stack gas velocity. . 44 Semple recovery. 44-1 Tubing--Teflont 9.4 mm outside diameter, to connect bag to gaa chromato graph sample loop. A new unneed piece is employed for each Miles of bag samples that constitute* an emission teat, and la to he dis carded upon conclusion of analysis of those bags. 44 Analysis. 4.3.1 Gas chromatograph--With flame Ionization detector, potanttasnstrlc strip chart recorder and It to ID ml heated sam pling loop In automatic sample valve. 4.34 Chromatographlo ooluina Otalnles* steel. 3-0 'A 34 mm. containing 80/100 meets Chromoeorb 103. A secondary eolum of OB SP-M, 30% on 90/90 mesh AW Chromoeoeh P, stainless start, 34 m X 34 usm. will bo required If acetaldehyde la present. If used, the SF-96 column la placed aftar the Chrosno- sorb 103 column. The combined oolumni should than be operated a* 110*C. 444 Plow meters (3)--Rotameter type, 0 to 100 ml/ mm capacity, with flow control valve*. 44.4 Gas regulators rue required gaa cylinders. 4.3 3 Thermometer--Accurate to one de gree centigrade, to measure temperataro of heated sample loop at time of Semple injec tion. 4.3.8 Barometer--Accurate to S nun Hg, to measure atmospheric pressure around gee chromatograph during iample analysis. , 44.7 pump--Leak-free. Minimum capac ity loo mi/mln. 4.4 Calibration. 4.4J Tubing--Teflon. 8.4 am outalde diameter, separate pieces for sach calibration concentration. 4.44 Tedlar bags fllxtsen'-lnch square size, separate bag marked far arch calibra tion concentration. 4.44 Syringe--0.5 ml, gaa tight. 4.4.4 Syringe--50*1, gas tight. Mention of trade on specific prod ucts does not constitute endorsement by the Environmental Protection Agency. 4.4.5 now meter--Rotameter type, 0 to and the attenuator setting. Record the lab, 1000 ml/mln range accurate to 1%. to oratory pressure. From the chart, select the meter nitrogen In preparation of standard peak having tha retention time correspond gas mixtures. ing to vinyl chloride, aa determined in flec 4.4.9 Stop watch---Of known accuracy, to tion 74. Measure the peak area. Aw. by use dm* gaa flow In preparation of standard gaa of Ha, and a dlso Integrator or a planlmetar. mixtures. Measure the peak height. Ha. Record Aa and 5. Reagents. It la necessary that all rea the retention time. Repeat the injection at gents he of chromatographic grade . least two times or until two consecutive vinyl 8.1 Analysts. chloride peaks do not vary m area more than 5.1J. Helium gaa or nitrogen gaa--Zero 5%. The avenge value far thaoe two areas grade, for chromatographic carrier gaa. will be- used to compute the beg concentra 5J4 Hydrogen gaa--Zero grade. tion. 5.1.3 Oxygen gas, or Air, as required by the detector Zero grade, 54 Calibration. 34J Vinyl chloride.- W.9 + %--For prep aration, of standard gaa mixtures. ' 344 Calibration cylinder* (3), optional-- One each of 60. 10 and 5 ppm vinyl chloride In nltrogvn with certified analyst*. Analysis must be traceable to MBS (National Bureau of standards) or to a gravlmstrically cali brated vinyl chloride permeation rube. 34-3 Nitrogen gaa Zero grade, , for prep aration of standard gas- mixtures. 8. Procedure. . ,w. ` 4.1 Sampling. Assemble the sample train at in Figure 108-1. Perform a bag ltah check according to Section 7.4t Observe that all connection* between the bag and the probe are tight. Piece the end of the probe at the centroid or the- stack and start tha pump with-the needle valve adjusted to yield a flow of 04 lpm. After a period ot time suffi cient to purge the line several time*, ha* elapsed, connect the vacuum line to. , the bag and evacuate tha bag until tha rotam eter Indicates no flow. Than reposition tha sample and vacuum llnae and begin tha ac tual sampling, keeping the rate proportional to tha stack velocity. Direct the gaa exiting the rotameter away from sampling peraonnaL At the and of the sample period, shut off the pump, disconnect the sample line from tha hag. and disconnect the vacuum line from the bag container, protect the bag container front sunlight. -' - 64 g---pt- storage. Sample bags must- be kept out of direct sunlight. When at all pos sible. analysis la to he performed within 34 hours of sample collection. Compare tha ratio ot Ha to Aw tor the vinyl chloride sample with the same ratio for the standard peak which la closest In height. As a guideline, if these ratios differ by mors than 10%. the vinyl chloride peak may' not be purw (possibly ecakaldehyda la preMnt) and the secondary column should be em ployed (sew flwotlon. 444).. 84 Measure the ambient temperature and barometric pressure nea* the bag. (Assume the'relative humidity to be loo percent.)- From-a water saturation vaporpreeeure table, determine the-recced and water vapor con tent of the bag. .. / i.' '.ti- .a 7. Calibration, and Standards: ; 7.1 Preparation at vinyl chloride standard gas mixtures. Evacuate a alxteen-lnch square Tedlar beg that has petted a leak check (described fit Section 7.4) and meter In 34 liters of nitrogen. While the bag is ailing, use the 0.5 ml syringe to inject 350*1 of 39.3+ % vinyl ehMnlde through the wall of the beg. Upon withdrawing the syringe needle, im mediately cover the resulting hole with a piece at adhesive tap*. This gives a concen tration of 50 ppm of vinyl chloride. In a Ilk* manner use the other syringe to prepare dilu tions having 10 and 5 ppm vinyl chloride concentrations. Place each bag on a smooth surface and alternately depress opposite sides of th* bag 50 timer to further mix the gases.. 74. Determination of vinyl chloride re tention time. This section can. be performed lnmitanfoualy- with Section 74. Establish chromatograph conditions identical with those in Section 84, above. Set attenuator to l 1 position. Flush the sampling loop with zexo helium or nitrogen and activate 84 Sample recovery. With a piece of Tef 'the sample valve. Record the injection time, lon tubing Identified for that beg. connect a the sample loop temperature, the column bug Inlet valve to the gaa chromatograph temperature, the. curler gaa flow rate, the ample valve. Switch the valve to withdraw chert speed and the attenuator setting. gaa from the bag through tha sample loop.. Record psalm sad detector responses that Flumb the equipment so the temple gaa occur ID th* absence of vinyl chloride. Main paaeaa from the sample valve to the laak-frae tain conditions. With the equipment plumb pump, me then to a charcoal tub*, followed ing arranged identically to Section- 84, flush by a o-ioo mi/ram rotameter with How oon- the sample loop tor 30 seconds at the rate of trol valve. 100 ml/mln with one of the vinyl chloride 4.4 Analysis. Sat the column temperature calibration mixture* and activate the sample to 10O* c the detector temperature to 150* C, and the sample loop temperature to 70* O. When optimum hydrogen and oxygen flow ratse have been determined verify and main tain these flow rates during all chromato graph operations. Using zero helium or nitrogen aa the carrier gaa, establish a flow ram.in the range consistent with the manu facturer'* requirements for satisfactory de tector operation. A flow rate of approxi mately 40 ml/mln should produce adequate separations. Observe the base line periodi cally "* determine that the note* level baa stabilized and that base line drift has ceased. Purge the sample loop for thirty second* at the rate of 100 ml/mln, then activate the sample valve. Record the injection time (the position of the pen on the chart at the time of sample Injection), the sample number, the valve. Record the- Injection time. Select the peak that corresponds to vinyl chloride. Meaaur* the distance on the chart from the Injection time to the time at which the peek maximum occurs. This quantity, divided by the chart speed, la defined as the retention time. Record. 7.3 Preparation of chromatograph cali bration curve. Make a gas chromatographic measurement of each standard gaa mixture (described in Section 7J) using conditions Identical with those listed in Section 6.3 above. Flush the sampling loop for 30 seconds at the rate of 100 ml/mln with each standard gas mixture and activate the sample valve. Record C,. the concentrations of vinyl chlo ride Injected, the attenuator setting, chert speed, peak area, sample loop temperature, column temperature, carrier gas flow rate, sample loop temperature, the column tem and retention time. Record the laboratory perature, carrier ga* flow rate, chart speed pressure. Calculate A,, the peak area multl- FIOERAl MGISTM, VOL 41, NO. 205--THIMS0AY, 0CT08M 21. 1*78 TOO COLORITE 007393 tlillS AND M0UIA71 NS 46571 i by the iitiQuihi aatting Hifwtmii two injection ana* era within sg, then pi** 1 ' -*,,J > . ._ . tliaw point* n O.. ma Dm other oonoantatlou ] bwu piotm dn* * --n ourr* through tbo point*. Man calibra tion dally, or before ui attar not aaa at bag aamplaa, tblOmr 1* mar* fnquent, UT'-V,'.:--c/ -' . t 7.4 Bag look check*. Whll* pecfocmaoo*. of ttu* notion la required mbaaquanl to bag uta. It la alao adnaad that 14 bo> performed prior to bag tiaa. Aftar each tag maka tot a bag did not dereiap leak# aa fouoae. To laak check. connect * weta* manomatte and rnaa eurtae the bag to 8-K> cm H, (3-fcta BLO>. Allow to etand for- IQ mlniit.r Aity inTi^- man* In tbn water mam ' i leek. Alao obaok tba rigid m thk muuMTa tar s . , . t %.c" (Notat An-alternaHto laak ebaek t to pramurlaa tba bag to 3-10 cm K.O mi `Wsw""br5r-V'. v.3~4 lg Mfi and-aUov to rtand orerolght. deflated bag Indicate# a. leak.) JVr ` y.-.'. SsjMnotMsiwtelf " - '"t- ' - --r -'f. 1. - (1 . .'-0 - mmpto bag In Mb Hd.eonaHnr. plaea a ' . -' ` rTW"'iS'VVe;?.2 rotameter ts-Uao between tba bag and U> -;----'!L^toxV ` o-t, -iky pomp buat. Irammae ebag PetRna of tba fSSMS .7" a - -m rotamotarterrglatoraaru flow whan tbo bog appaaio to bo empty mmoateo a. laak. - is ^.baa.l*4o *" mmorn.,.' g r*-!--.*--! i- &1 .DatarmlitO' tbo- anmpto paafc arm- ao *>- fOUaw-. ' o*>"."s ------ tS-t,.-w .* -- ..IT.'lt * iggvibtlno' I04-L: atg and 4X10-* mg With paopaa --twwwa., the upper Limit may ba astaodad ae. beaded. A.aTba aaoel* pHk aa. knmawiindpataia Ar^Tba irtinninaa ttatar- . A Praenion and Saproductbilltf. . la lniarlabontoty aonpamon. baewaan mean, labermtodaa of three tmln ----p1-- 8J Vlnjl eblorido eeooanMatloiH, rtom. tba calibration oui a dmotbad In 7J, abora, aaiaet tba vatoa 0< O, tba* ta^onda to d,, tba aaapio put in*. Cbi- oulata C, aa tbUowa: _ _ ^V. each aplH into tbfa* parm yielded a --* darlatlon of 3.83% for a eample with a mean , vttb tbi*>of tt>awttoA aboulrt of 2A9 ppm, 4,18% for a eample with a main of 1.88 ppm. and 533%- tor V aample with a . at a cm- nbwnningrapb,- nor by- mean of 83.88 ppm. ..... r wbo ara nnlbmiliar vita, mmpllng a*> -- A Safety. C.P.Tr Ct- preran* aapoamo oc So not rMaaaa Ttnyt ehlortdo to the labora tory atmorphera during preparation of ttasdVeotntg or purging with vctf/alr mi*- BqunUon.lO*-3 ... ttnyl 3-PrHMtplaandAppheabUty. mint bo bold to t.mtnlmm. When thoy ara mqalrad. the tapor moat be rooted ` -Tbr valarTapatcaaladtottba tnaaUan malytag - C^Tba aoaoaonMaoat abgl dMlt la tba bag- ** -'"^tTnnnr-TTTTtn rf Tlarl tlikalila InitiuM br a-WlBl IVWVM PfMnM) CM UDMMliy-SM___B___| ncnrdil itawiia aMwtn.anft Ti-Tba aaaala kap nawaiat ua tar abokm vala aa tba baa < aaaum, *t g.-TbHj*.labemiety yam at tarn at aaabdi, am 7%-Th*. ntmn fata pit t Boa, tba wopli. laao - IX Ibo baaM for tbla matbod ralatta to tba vapor iqtinMm vblcb tt aatabUahad batvtnt RVCM. pvo, tpatn. wmtar, and atrin a Bleat d ayitam. It bat bean damonatostad that tba nvcM is a PVO main wm aqulUbrmto in a doaad va*aal qolta rapidly, proridad tbnttha tatapmntnraief tbo PVO malm la ualntalnad abotv tba glaan tmnaUtsn.. tauiparatgra at tbat appcttte rmbt. . t A fhia pfoodihim U, wiltiMt fOf dali^ tbo nnyl cblorlda monomer (VCU) ta outers* air. Vlayt ehtorida. eras at low ppm lerela. meat norar boraotad tbalde tbo laboratory. After rial* bare- bean analyzed, tbo pimuit wttbln tba trial muaa be Tented prior to mmorai hem the instrument turn table. Vlala mat ba Tented into an aettramd. charnaA toba tmag a hypodermle needle to prorant rtleaw of Ttnyt nblorlrta into tba laboratory atmoephar*. The enamel must ba roplaoad prior to- Tlnyl chloride break 8through. l . Apperatua. .. ..-t-1'*- iiaipmtnia ittmd dmtag MUlaailiin. TC. eontmt of inpnwaaa vaatawmtaa tamplea, U * -v*^ and tba taaldual Ttnyl eblorldr nyioomar JA Bottlaa--80 ml (3 oa), with wand #, Batarancaa. 1. Brown, D. W. toy, x, W; and Stapbanton, M. R. "Vlayt Cbtortda Ubnttodag Maartba b. 7. Goodrich Cbmtloal Cmnpany in LonurlU*, Kantncky." BagMg IV, Vh. Ban* ronstantal Protaction Igag, Bormnnw and Analyala Dtrlalon, Athaag flanrili Jtma 24. 1874. X "Sralnatlon. of A CoUaatton and Analy tical ptocadnra for vinyl CbloMdo tn Atr," by O. D. Clayton and Aaaoctataa. narambar 13. 1074. SPA Contract MO. 88-03-1400. Taak Otdar No. 3.27A Xaport OK. 73-VCL-l. (SV0U) oontant of polyrlnyl cblorlda (PVC) raalna, vat cake. Hurry, and lata* aamplaa. It cannot be uaad for polymer in fuatd form, fucb aa abaat or cube*, if a raaolutloa of tbo Uotd temw on tope, for PVC aamplaa. 9-1-3 Vlala 80 ml Bypo-Ttala,1 taalad with TMm faced Tuf-Bood dlaea far water an Ttnyl rfiinrida pee* u not aadafaccory for a 833 Bectrlcal taper or equlraleoA to particular eample, than rbmmakurapb prarant looaenlnf of bootla top*. pannutam may bo altarad with prior tp- 83 Sample raooTary. proral of tba Admlniaaater. If that* la mac A33 Vlala--With aaala and oapa. Perkm- ron to baUara tbat aom* otbar bydrocarbon Bmar Corporation Mg 108^)11% or equlravltb an Identical retantlon thno la prtatnf lant. In tba aamplg than aupplaaiaatal con firma- 833 Analytical btltnco ^'ptM* of tlon of tba Ttnyl paak tbrougb an a*tgbln( to -0A01 grain. , abaoluta analytical tacbnlqwa. aucb aa maaa 833. Synnga, 100 spaetrcacopy, tbmiid b* parfotmad, "A" Mg 010038, or equiTalanA 3. *** of Btatvmary Some* 3. Santa and BanaltlTtty. Emlaiioa Matbod f Vinyl Chloride," by MldTMt PABaarcb Inadtnta, 1070. IPA Contract Mo. 88-03-1098. Taak Order No. 7. Uta lower limn of detection of Tlnyl chlo ride will Tory aeeordlmf to tbo ehtomab> graph uaad. Valuer reported lnelnde 1X10-* iMontlon of trad# natnee on apaolflo prod uct* doaa not eooatttoto aadonamattt by tbo BtyUonmontat piotaetton Agaooy. 1 FSOItAL UCISTH, VOA. 41, NO. 208--THU850A7, 3C7CMI 21, 1978 101 COLORITE 007394 146573 RULES AND REGULATIONS' 3-3-4 Vial SNlr, PerUn-Erne/ No. 103- 0103 or equivalent Si1 Analysis. 3-3-1 Ou chromatograpte--Ferkln-Zlmar Corporation Model F-40 "H-t**! ljzer, Ho. 104-0001. or equivalent. 5-3.3 Chromatograpblo column--stain- less iM, 3 mx3-3 sun, 0.4% Cuboni 1500 an Csrbopak A. Ferkln-Slmer Corporation Ho. 100-0133, or equivalent. Curbopax a can be used In place of Carbopek A. 622 Thermometer--0 to 100* C, aoeurato to 202* a, Perkin-Emer No. 105-0100 or equivalent 1 5-3.4. Sompl* tray thermostat eywtsni Perkln-KIm-r No. 105-0103, or equivalent. 5 3.3 Septs--sandwich type, taw Men waling and weighing, using a 100 2 syringe. In tbo otae of dispersion resins, the cup cannot bo used. Tbo sampls u instead weighed approximately in an aluminum dish, teenarmed to tbo tond rial and weighed accurately In tbo risk Tbo sample is then placed lb tbo terkkuBmer bead tpaoo rwi- lyaer (or equivalent) and conditioned for one hour at M)*C, Nora: 3oom aluminum rial capo bare a center section which must-be removed prior to placing into temple tray. If not removed, tenons damage to tbo injection needle will occur. . - - 722 Suspension realn slurry and wet cake tempise Slurry must bo Altered s-*-g. a small Bncbncr funnel, with vacuum to yield wet cake The Altering process must be con tor--Ignite the detector according to tbe manufacturer's lnxtrucoons. 72.1.4 Amplifier halemw Balaaoe toe amplifier according to tbo p*rtTl**- Lllr-,>t instruettona. -- 722 Programming the chromatograph-- Program the chromatograph as follows: a. I--Doting time--Tbe normal setting is 3 eeconda. b. A--Analyvls time Tbe normal setting la a minutes. Certain types of samples con tain high boiling matrelaia which can eauto- Interference wtih too vinyl chloride peak on subsequent eaalyaee. la these omss tbo analytes tone must be adjusted to eliminate to* intnferene*. An Automated beckflush yxtem ran alas be used to solve this prob lem. - -. betto rinsing. 13 mm, Pvrkln-Slmer No, 106- tinued only so long an a steady ttruam of o. B Washing--Tbe normal totting la 02 1003, ar equivalent water la eritlng-trom-tba funnel, xxceasive mlnutm. . .--, , 53,9 Integrator - recorder -- Bewlstt - nitration time oobld result In some loss of d. W Wtebiuieuon tlino Tbo TMi set- Packard Model 33S0A. or equivalent. VCM- The wet cake remple (0.10 to 42 grams) tlng Is 02 minutes.. - - . xk ,u..< 53.7 Tutor drier assembly (3)--Perkin- Is addad to a teredorial '(Including septum a. X-7-Number of analyses per tample--The Bair Ha. 2330117. <* equivalent 5.3.3 Samp 91m flowmeter--Hewlett Pack and aluminum rap)- and Immediately nt1--* normal setting la ju-- * - --a. . Sample weight la then determined tb 3 deci . 722 Trepantton of eampla turntable--Be ard No. 0101-0113, ar equlvelsnt. j - - mal piece* Tba sample h than placed In tbo fore placing any sample into, turntable, bo-. . 6.4 CUimtUB. ------ .- .W3"-'-'v Pvrtin-glmar hood spaceanalyear (or equiva certoln that tbo center section-of the alu- 5.4.1 Eegulatore--for required gee eyin- lent) and conditioned lew-one hour at so*o. mtoum oap ha* been removed. Tb* numbered Irm. A sampl* at wet- cake lo used to determine ample bottle* should be pieced In too oar- 3. Beegento - ' - iANe;`*4i*J*ih3 - * " % 7*T''rt-- * -T3 (total solids)-This la required for ealeu-- responding numbered positions In too turn- . i 8A Analysis. lattng tbe BTC3t-: - -yrqjSA tablet Insert aamnlee In the following order*. 1 5-1J Hydrogen gas urn grad* -t---- - --722 Dlsparteon lealn^aluiry samples.-- Poeltloue-1 2 3--Old 3000 ppm standards 111 mntta (u-<ao tMa' . *'i . This material should notba filtered. 8ampis for enqdttlontng Thera ere necessary only 8.12 4It rsm grille w-'-*-,, mutt bo thoroughly mixed. using a tarsd efter'th* analyzer hae-not been used for 34 S-3 Calibration. -- * teal (tnotudlng --pr" aluminum cap) hour* or longer. ., ... _____ I 6-2_l Standard. cylinders (4)---on* Nik add. approximately S--drops (025 to 029 Position 3--60 ppm standard^freshly prr- . r>f 50, 500. aooo. and 4000 ppm vtnyl chloride- grams) of slurry or latex: using a n-i-- pared. - -i . . .-- , tn nitrogen, trim oertlfled analysis. dropper. Tbit should be. done immediately PcelUon_4--too ppm standard, freshly pre-'' 7. Procedure. " ' '.f-' after mixing. Seal the teal as soon as possible. pored. - -711 Sampling. _ t __ > Determine sample weight, accurate to 0201 Position 5--2000 'ppm standard,' freshly*. grams- Total sampla-watght must not exceed prepared.- - - - -- ^ 71.1 ptc sampling IHaw the raaln or 020 grams. OraCMat tb* teal fatjana boar lorry to How tram a tap on toe tank or Uo- at fio-c m tbs anotynn. Datstnnna tbo TS Fotttum 01--4900 ppm standard, frealdy pre-. P4MdL _ * , nil taa top Una baa b*an wall perged. Xa-. on tha slurry Maple (SeoMea 722). PoetttonJT--Sampl* Be. 7 (This u too first toed a 40 ml aampla bottle tmd the tap. On. 722 ~ Znproaass .waotawatsr tampls^- sample of tosdoy, hot la given aa7 to be con- end Immediately tightly cap tba bottle, wrap Uktng a tarsd teal (InoMMlng septum and tttest with toe turntable end the Integrater 1dacoteal tapa around tba cap and bottle to aluminum cap) quickly add approximately printout.) prevent the top tram loosening. -Place an oo of water using a toedUrine dropper. Seal After all mmplea hare been poeitloned. 1s- Identifying labal on mob bottle, and record tbo teal as soon - as possible. Determine tert toe second set of 50. 500, 3000, and 4000 tb* data, Oma and aample looatloa botb an sample- weight accurate, to 0201 gram. Con--- ppm standards. Samples. Including stand cb* bottlea and la a log book. dittos tbo teal fur two boun at to'O m the- ards mot bo conditioned in toe bsto of 712 Water sampling--Prior to non tbe - antlyatr. , .ic>rf. 50 ml vials (without tbe dleoa) mart be . 72 Analysis. - ----S - jr..-'- M* O for l. hour (not to exceed 5 hours). 72.4 start ebromatogrepb pngnia-- - capped with aluminum fall and moused ** T22 Reparation at gas chromatograph-- teben all eamples. including standards, have (00`C for at least ooa hour to destroy or Tnttall tba cbromatognpblo column and con^ bean conditioned at 00* C for 1 hour, start remora any organic matter that aould In dttion ororalght at UflfOi Do not connect tbe tha analytes program according to toe manu terfere with analyte*. At tbe mmpllsg loca exit cod of tbo oohrnm to the detector while facturers' Instructions. Thera Instructions * tion flu tba vlala bubble-lire, to orerflowtng conditioning. , must ba carefully followed when starting ao that a convex mcnlaena forma at tbe topi - ~ 7222 now-- rate adjustments .Adjust Tba r treat eater la dsaplaoad aa tba seaHag flow rates as follow*t -`. -i" due is carefully placed, Teflon aide down, an s. Hlnogcn earrtsr gad get regulator on tbe opening at tbe rlaL Place tba aluminum cylinder to raod 50 pat* Bet regulator on teal over tbo dleo and tba neck at tba rial chromatograph to 12 kg/cm*. Normal flow* and crimp Into pleor. Afllx an Identifying at this pressure shouid be 25 to 40 cc/mlnute. label on tbe bottle, and record tba date, time, Check with bubble flow meter. and sample location botb on tba male and b. Burner sir supply--Set regulator on cyl In a log book. All tamplss mat bo kept re- inder to raod 50 pelg. Set regulator on Crljerated until analysed. chromatograph to supply ate to burner at a 72 Sample reeerrery, Semptoamuat bo not rate between 350 and 500 ce/mlnutc. Cheek awltbin 24 boon. _ 72.1 Ream sample*--1Tbe weight at the. resin used must be between 02 end 42 gram* An exact weight moat bo obtained (--0201 gram! for eecb sample In tbo oaoo of sus pension resins a rolumetrlo cup can be pre pared which will bold tbs required amount of sample. Tbe sample bottle is opened, and the cup volume of ream Is added to tbo tand Sample rial (Including septum and alumi num cap). Tbe rial is Immediately sealed wttb bubble flowmeter.--v -- - 3. Bydregen supply act regulator on cyl inder to read 30 pci* Bet regulator oa chromatograph to supply approximately 35*5 oc/mlnuta. Optimise hydrogen flow to yield tbo moot sanilttr* detector response without extinguishing the flams, check flow with bubble meter and record this flow 72.12 Temperature- - adjustments--Set temperatures as follows: a. Oran (chromatographic column), M* and the exact sample weight Is then obtained. Beport this rains on the dsta ahset as It U required for calculation of RVCM. In tha case of relaarety dry main samples (water b. Dosing Une, 140* c. a. Injection block, 140* C. d.-Sample chamber, water temperature. content <02 weight %). 100 1 of distilled 0*C12*C. and stopping program to prevent damage to toe doalng assembly. .--* 722 Determination of total solid* (TS). For wot oaks, slurry, ream solution, and PTC lawn samples, determine TS for each ample by accurately weighing approxim ately 3 to 4 grams of sample In an aluminum pan before and after pi-ct-g in a draft oven (109 to no* C). Samples must be dried to nooctant weight. After flret weighing re turn too pan to too oven for tTthat% pe riod at time and then reweigh to verify com plete drynsm. TS 1* then calculated a* too final ---p1- weight divided by Initial tam- pie weight. -_ 8. Calibration. Calibration h to be performed each eight- hour period when the Instrument la used. Bach day, prior to running eamples, tb* col umn should be conditioned by running two of too previous days 3000 ppm standards. 82 Preparation of Standard*. Calibration standard* are prepared by fill ing to* teal* with tba vinyl chloride/nitro gen standards, rapidly lasting too septum and sealing with to* aluminum cap, Uie a ate Inl ms steel bn* from tb* cylinder to tbs Vial. Do not use rubber or tygon tubing. Tbe . water must bo injected into tbo rial, after 72.12 Ignition of flams Ionization detec sample Una from toe cylinder must be I MDMAl. KIOIlTIfi, VOL 4), NO. 305--THURSDAY, OCTOSEK 31. 1976 102 COLORITE 007395 RULIS AND REGULATIONS purgad (Into.hood) for oarers! mJnutaa prior to Mllni Mala, Altar purging, reduo* tha Sow rmto to approxlmataly 500-1000 eo/mln. Placa and of tuning into MaT'tnaar bottom) and attar ona minuta alowly ramora tubing, placa aaptum In rial aa aoon aa poaalbla to mini* I 1.1 lm t -VVU wtaiM, oa (2Xf). . I iWttpli ooUlBf Um tbaa (tf% wiitr* nlaa mining air with aampla. Altar tha aland- ard Mala are aaaltd. lhjac* 100-1 or dtntUad wmtar. c. ""% (* 197X10-*' , 5.988X i( 10-*) U Preparation of chromatograph calibra tion surra. Equation 107-3 Prepare two SO ppm. two 500 ppm, two 3000 ppm. and two 1000 ppm atandard aamplaa. TIm taDawtsg iMnl acusiiaa au bt bm! Car tar oapl* which malalcs Veil, PVC ttdtenua. ' Run tha calibration aamplaa In axaotly tha . aama mannar aa' regular aamplaa. Plot A*, tha latagrator ana counta for taeh atandard A.J_ .-.X*7 aampla re c, tha coucantratlaai of Mnji chlorldo In aaoh atandard tampla. Draw a. Una of baaa da through tha point*. - 0. Calculation*. - . Pd Baaponaa factor.. - . Pt tta oaUbratlon curra daacrlhad In ' _ f' * -L r - -r-r.-Ti j_.;. Equation 107-4 r -1 , JL . .SaoatonjJ, ahora, aaiaot tha ralna of C. . that reareaponda to A. for aaoh aampla. oom- v ' it" ` puta tha raapooM factor, Ht, for aach aampla, j^r-.'Raaulm ealculatad using Equation 10T-* aa follows- . .. _ -ff*repreaant eoacantretlon baaan on tha total ` i^-- " i,, , 'A *'" ' - Ssjoampl*. To obtain reaulta baaad on dry FVO ' " B/a-rfi Equatloit 1--0--7-----1^>oontauAdlrld* byia.. 9-3 Haaldital rtnyl chlortda motto oanWaRon. or rlnyt ehlorkt* mnnnmtr oonoaatratlon. . Caloulat* Om aa followa: w b KmpfciM t* ito feOwiat: 10~* f (2.06AX lfr-)j Equation 107-S ; C^***3*') ig'Sii; Baaldual Vinyl Chlortda Ronomar Oon- v - Eouatlon 107-2 * Pl7MnyI Chlorlda Rrelna and Wat nquawn iu< a ^.CiKa Sample*. B. P. Ooodrtch Chamical Co. standard-Twt precadure No. 10O4-T. B. P. tnaaatntlia of rflqrl In pom, P,--labantary_______ -. Ooodrtch Technical Cantar, Aron Laha, Ohio. ^ January 30, 197A * filam impailua *t - ... JT* 2 arena. A. h. "Tha SolubUity of Vinyl J/.-Mataaaiar waMh* d< VCR (U). V'.-Vojmc ofaapaapOaaa (Mai team* boa mwph rr'ChloftdP in Polyrlnyl Chlorlda' ACS-DlrlZ.tLoa at Polymar Chamntry, Polymar Pre> prlnM tS (3): IPT. 1P7A ff-Oatcaattanl -I,-Hmwoy.i fLf-*awmxaaioM ate VCR la PVO at A tor VCR bio - (aporedmata) ws JT-*JXl<re-ff- a arc. TV-XquUlbradoatam 3. Barenn A. B-."rha Diffusion of Vinyl Chlorldo In Polyrlnyl Chlorlda.- ACS-Dtrtaton at Polymar chaaahtry, Polymar Pro* ; print* I* (2): 303,1PTA - A Banna, A. R. U B. Crldar, C. J. TOma- h*K and J. M. Whuaay. Ansiyds for vinyl -fac nr,a-V*OoOWD. AP.-7Mna.Rg. Chlorldo In FVO Powdan by Haad-8poo# Oaa niininaliigioiilij ~ in lia imlillaliol I D00.7S-30PM PUad 10-30-79:8:43 lm] - _ - . *i. nociAi mum, vol <i, no. ioj--ihuksoay, oaotn 31, i7 103 COLORITE 007396 REFERENCES 1. National Emission Standards for Hazardous Air Pollutants: Standard for Vinyl Chloride, Environmental Protection Agency, Federal Register, Vol. 41, No. 205, Thursday, October 21, 1976. 2. Scientific and Technical Assessment Report on Vinyl Chloride and Polyvinyl Chloride, Environmental Protection Agency, EPA-600/6-75-004, June 1975. 3. Standard Support and Environmental Impact Statement, Volume 2: Promulgated Emission Standard for Vinyl Chloride, Environmental Protection Agency, EPA-450/2-75-0096, September 1976. 4. Standard Support and Environmental Impact Statement: Emission Standard for Vinyl Chloride, Environmental Protection Agency, EPA-450/2-75-009, October 1975. 5. Speciality Vinyl Chloride Resin Processes Effects of Governmental Regulations, R, N. Wheeler, Jr., Union Carbide Corporation, South Charleston, West Virginia 25303, September 1, 1976. 6. Vinyl Chloride - An Assessment of Emissions Control Techniques and Costs, 8. H. Carpenter, Environmental Protection Agency, EPA-650/2-74-097, September 1974. 104 COLORITE 007397 CODEN CEPRA SEPTEMBER, 1977 Vol. 73, No. 9 SCOPE m s?rrMcE -rj? oci 17 Hero Petrochemicals Outlook What's Happening with Cyclohexane........................................................................ Demand for U.S. cyclohexane should grow at an overall rate of 7%/yr., cyclo hexane for nylon libers should increase 6%/yr., and cyclo demand for engineer ing, plastics, and exports should increase 10%/yr. Aromatics in Europe............................................................................................................ The 1980s are expected to be an era of benzene over-capacity in Europe, a sit uation that could be exacerbated should petrochemical complexes be devel oped in the Middle East. 70th Annual Meeting to be Held in That `Wonderful Town'.............................. Included in the program of 117 sessions and more than 500 papers is an all-day symposium on legal issues facing the chemical industry. 25 29 99 Plant Loss Prevention . Environmental Factors vs. Flare Application......................................................\ 41 People living near a plant often view a flare as a tremendous waste of their irreplaceable energy resources rather than as the safety device it actually is. Explosion Pressure Relief................................................................................................ 45 The use of pressure relief devices in fluid bed dryers, granulators, filters, dust collectors, and containers can improve operating safety. , A Survey of Vapor Cloud Incidents........................................................................... 1 54 Trends toward plants with larger capacities, and higher pressures and oper- ^ - - ating temperatures makes understanding how to prevent vapor cloud explo sions imperative. Olefin Plant Safety During the Last 15 Years...................................................... 64 Good design alone cannot keep a plant safe. The quality of its methods of op eration, training, communication, instruction, etc. are of equal importance. Explosion in a Naphtha Cracking Unit..................................................................... 69 Most of the 106 people injured in this explosion (14 died) were hurt by flying glass from broken windows. Amoco's New Generation of Pilot Plants..................................................................... The plants are basically data factories operating at steady-state and, in some cases, dynamically, to provide essential process information. 73 Hazard Evaluation of Self-Accelerating Reactions................................................ Frequently, the design problem is not the lack of test data, but the estimation of conditions in the system. 80 Improving Material Flow from Bins.............................................................................. The proper equipment can insure steady, measurable flows. 82 Hydrogen via Steam Reforming of Naphtha............................................................... Growing demand for ammonia fertilizer has meant an increasing demand for hy drogen as a major raw material. Removing H2S from Geothermal Steam.......................................... ............................. The economics of this process appear to be comparable to the relative cost of flue gas desulfurization system for fossil-fuel fired boilers. 87 93 DEPARTMENTS Books............................... 6 Speak Out...................... 37 Washington Newsletter. 39 Institute News............ 110 People.............................. 112 What's New..................... 117 Data Service..................124 Future Meetings .... 136 Employment............... 141 Professional Services . 160 News & Notes of AIChE.................. 162 CEP September 1977 1 COLOR!TE 007398 ee 00 fcl--: Figure 1. An Alaskan flare that provides smokeless burning without using smoke suppressants. PLANT LOSS PREVENTION Environmental Factors Vs. Flare Application People living near a plant often view a flare as a tremendous waste of their irreplaceable energy resources rather than as the safety device it actually is. R. Schwartz and M. Keller, John Zink Co., Tulsa, Okla. Prior to the development of the first successful smokeless flare in the 1950s, (/) the people who designed this equip ment were mostly concerned with obtaining positive igni tion, and gave little or no thought to the results of the ensuing combustion. That first smokeless design, as re markable a technical breakthrough as it was for its time, now pales in comparison to the equipment used to meet current environmental regulations. Today's flare system designer must not only cope with ^ Circle Data Service Card No. 68 CEP Septenther 1977 smokeless requirements, but must also investigate the total environmental impact including thermal radiation, noise production, ground level concentration of combus tion products, visible light, and in some instances the aesthetics of the design. Meeting the increasingly stringent environmental requirements has necessitated large re search and development expenditures by the John Zink Co. The flare system's most dramatic impact on the en vironment is its potential for production of very large 41 COLORITE 007399 flames and enormous clouds of smoke. Such emissions can be seen from many miles and can quickly attract the attention of neighbors, public authorities, and environ mentalists. Current environmental requirements force the plant designer to route more of the vented gases into the flare system, resulting in larger flare sizes even though the capacity of the attached plant may be the same as previous designs. In addition, the use of larger components in plant design has increased the amount of gas that must be handled smokelessly by the flare. Normally, the flare designer's first environmental con sideration is whether the waste gas will produce smoke. Considerable research into this question indicates the weight ratio of hydrogen to carbon of the waste gas is one of the key factors. (2) However, it has been found that identical H/C ratios will in one case produce smoke, and in another burn without smoke production. Careful in vestigation of the data has shown that instantaneous H/C ratios in the combustion zone may greatly exceed the average, due to the lack of mixing of waste gases. This ob servation is supported by flow studies which show tur - "-if- '* t V. bulent flow, with Reynolds numbers as high as 50,000, fai led to promote mixing of gases. (3, 4) Presence of the liquid phase of hydrocarbons in the re lief vapor can also render H/C ratio predictions invalid. Liquid droplets as small as 15 microns can negate smoke less equipment operation, with larger particles producing even greater consequences in the form c f fallout of burning liquid particles in addition to unwarranted smoke pro duction. Careful consideration must be given to knock-out drum design and location. The traditional approach to smoke suppression is the in jection of steam into the combustion zone. However, there is not total agreement as to the chemistry or physical phenomenon associated with this injection. Since black smoke is clearly an indication of unburned carbon escap ing from the combustion zone, the carbon must be com pletely burned or chemically combined to prevent its escape. Steam injection leads to two endothermic reac tions that offset the desired end; the water gas shift (C + H20 - CO + H2), and steam reforming chemistry (C, Hr + H,0 = XCO + ZH2). Regardless of which reac tion dominates, the end result is the elimination of smoke. Other factors that also lead to the success of steam in jection allow the use of smoke suppressants other than steam. These factors include the eduction of air into the combustion zone, and turbulent mixing, which increases reaction rates and elevates reaction temperatures. These two factors make possible the use of high pressure gas, air, or other smoke suppressants, but with.overall lower ef ficiencies in terms of pounds of suppressant to pounds of hydrocarbons as a result of the loss of the aforementioned chemical reactions. Another apparent factor is the total kinetic energy in the combustion zone, exclusive of chemical energy release from the combustion process. Such energy levels can be created from high velocity discharge and/or from adding additional energy from a secondary source. The total energy required is a function of gas composition, burner design, quantity of gas being burned and other factors. Ex tensive experience is required with gases of similar com position to accurately predict the required energy level. While most smokeless designs use some suppressant to obtain smokeless burning it is possible to achieve smoke less burning without a suppressant as long as a satisfactory energy level can be produced. Pioneer work in the Zink test facilities has produced a proprietary method for main taining energy levels under virtually infinite turndown ratios. Designs obtaining smokeless burning without sup pressants are operationally successful in capacities up to several million lb./hr., Figure 1. Where a secondary energy source is required, the most common choice is low pressure air. When properly utilized, the low-pressure air can boost kinetic energy tr the smokeless burning level, as well as provide primary r to the combustion process. Additional advantages of type of flare, Figure 2, dozens of which are in succ operation, are their lower overall installation and operating costs. Smoke suppressant control Response time lag in application of err a problem that plagues flares old and application of a smoke suppressant was a strictly manual operation. Tb operation leads to at least thr 1) smoking prior to operator re over steaming of the flare w noise that occurs before op reduction; and 3) waste of stea*. to gas flow cessation. COLORITE 007400 , Numerous attempts have been made to automate the ap plication of smoke suppressant. Some of the problems in automatic control design have been: 1. Sensitivity to low velocities. 2. Ability to withstand high velocities and impact loads. ; 3. High maintenance due to fouling. 4. Reaction to varying combustion characteristics. f 5. Lag time. 6. Ability to install and maintain the device during flare operation. A remarkable breakthrough in the design of smoke sup pressant control has been made by the Shell Development Co. (5) This entirely new concept utilizes a grade-level optical sensor to detect, in a sense, a combined measure of flow, composition and suppressant effectiveness. Burning characteristic changes such as variations in H/C ratio that cannot be effectively detected by velocity and/or density measurement are readily recognized by this new system. Figure 3. The new system eliminates all of the design problems listed above. Accurate prediction of thermal radiation is becoming increasingly important with larger plants and correspond ingly larger flare relief loads. Ever increasing "emergency" vent rates, and economically unfeasible flare heights, have made it more important that cooperative studies between the user and manufacturer be employed to determine flare height. Several methods have been proposed for calculating flare heights. The overriding consensus of the authors of these methods is that flares, like any combustion process, cannot simply be scaled up from correspondingly small tests. Numerous large-scale tests have been used to develop a proprietary computer program for radiation prediction. Among the factors that are considered are quantity of gas to be burned, composition, smoke suppression method, flare burner design, and ambient conditions. The question of what constitutes an acceptable radia- 1 S tion level is equally important in determining flare heights. The John Zink research facilities have been used to de- Figure 4. A combination low level and i termine the physiological limits of a person wearing nor- elevated flare system. COLORITE 007401 mal plant clothing, i.e., hard hati long-sleeve shirts, and gloves. This "human testing" has disclosed that radiant impact levels of 1,500 B.t,u./hr,/sq. It. can be tolerated for an in definite period of time by an active worker. An impact level of 1,650 B.t.u /hr./sq. ft. requires limiting of ex posure time to five m nutes or less or the use of additional clothing. Impact levels as high as 3,000 B.t.u./hr./sq. ft. were tested. Limited over-exposure resulted in skin reac tion similar to a mild sunburn. Many neighbor complaints are caused by flaring noise. Noise control is a social as well as a legal responsibility. Flaring noise can be attributed mainly to two sources-- smoke suppressant injection and combustion. Careful orifice design (6) can greatly reduce the suppressant in jection noise on steam flares. However, in practice the minimum size of suppressant orifices is limited due to plugging by line scale, etc. Additional gains in suppressant noise reduction have been made by improving the ef ficiency of suppressant usage and by the use of the highly responsive optical smoke suppressant control previously discussed. Combustion noise is related to energy release. Smokeless burning results in an ohvious increase in the energy re leased and the potential for greater combustion noise generation. Therefore, it is inevitable that open smokeless flaring will result in some combustion noise. In chemical plant applications, where gases other than hydrocarbons are being flared, careful considerations must be given to design. In this case there is no substitute for experience with a similar application. Open burning is dependent on the combustion process being sufficiently exothermic to maintain ignition. Flares burning in the open air transfer a tremendous amount of heat away from the flame and as the temperature levels depress the com bustion process can cease. The amount of energy--Low Heating Value (LHV)-- necessary to maintain ignition varies. There is no fixed value for the minimum level required, however, experience allows prediction of the minimum LHV. In cases where the waste gas does not contain sufficient energy to support combustion, it is necessary to make use of special endo thermic flare designs. Flaring of gases such as chlorinated hydrocarbons, phos gene, and sulfur compounds require additional considera tion. Burning of chlorine containing compounds will result in production of HC1, and sulfur compounds produce S02 and S03, all of which have legal threshold concentration limits. Flares, unlike oxidizers and furnaces, do not adapt themselves to post combustion removal of these con taminants; therefore, the approach has been to design suf ficient stack heights to provide an allowable ground level concentration. responsibility in many areas of the U.S. at this time, it is in several foreign countries. The light of an elevated flare focuses attention on the process plant, leading to many complaints that might otherwise go unregistered. In addi tion, with today's energy conscious environment, neighbors often view the flare as a tremendous waste of their ir replaceable energy resources rather than as the safety de vice it actually is. Plants have the responsibility of limiting flaring activity by tightening operating procedures and practices. However, this tightening of procedures cannot completely eliminate some day to day flaring. The alterna tive to visible flaring is a low-level enclosed burning system. In addition to providing a hidden flame, there is the equally important benefit of noise reduction. Full load noise levels of less than 70 dBA adjacent to the flare have been obtained. For most plants it is desirable to use a flare system that combines a low level flare with an elevated flare, Figure 4. Day-to-day and startup flaring loads are burned in the low level flare while infrequent high volume emergency loads are handled by both low level and elevated flares. When properly designed such combined systems can prevent visible flaring for virtually all of the plant's operating time. Elimination of light and noise as a focal point is a proven public relations asset to the plant operator. (11) Today's designer of flaring systems is faced with increased public awareness of the flare's impact on the environment. Equipment selection and system design require careful consideration of safety and environmental aspects. Im proved communication and early cooperation between flare manufacturer and plant designer is essential to suc cessful flare system design. # Literature cited 1. Zink,J. S., and R. D. Reed, U.S, Patent 2,779,399, "Flare Stack Gas Burner," 2. Reed, R. D., "Furnace Operations," Gulf Publishing Co , Houston (1973). 3. Narayan, B, C., "Experimental Study of the Rates of Turbulent Mixing in Pipe Flow," M.S. thesis, Univ, of Tulsa, Tulsa, Okla, 4. Mukhtar, A. A., "Transient Analysis of Turbulent Gas Mixing in a Y-Branch Pipe," Ph.D thesis, Univ. of Tulsa, Tulsa, Okla, 5 Schmidt, T. R., "A Smokeless Flare System," Ecolibrtum, Shell Oil Co., 6, No. 1 (1977). 6. Seebold, J. G., Hydrocarbon Proc,, 51* 143 (1972). 7. Gifford, F. A.t Nuclear Safety, 2(4)47 (1961). 8. Pasquill, F,, Meterological Magazine, 90,1063 33^49 (1961). 9. Briggs, Q, A., "Plume Rise," AEC Critical Review Series. TID 25075 (1969). 10. Peters, J. M., Pollution Engrg, p. 25 (October, 1973). 11. Powell, D, T , and R. . Schwartz, "Operating Experience with a Low- Level-Type Flare." API Preprint 59-72. Presented at 37th Midyear Meeting of the Div, of Refining, American Petroleum Institute. Ground level concentration The prediction of ground level concentration is a de veloping area with several approaches available. Dis persion models such as the Pasquill-Gifford (7, 8) method and plume riser models such as Briggs method (9) seem to lend themselves to flare applications, providing one recognizes the exothermic characteristics of the flare. However, also to be noted is the fact that Briggs method is based on measurements from stacks, not flares. A modi fied approach has been offered by Peters. (10) If Peters method is used, one must remember to reduce the heat re lease by an appropriate level due to radiation transfer from the flame, and if appropriate, due to smoke. Visible light and noise from an elevated flare can be an extremely sensitive problem with the neighbors of a process plant. Although elimination of light is not a legal 44 R. Schwartz, who holds B.S.M.E, and M.S.M.E. jph P/ degrees from the Univ. of Missouri, is general W manager. Flare Div,, .John Zink Co. He has been , - r ^ employed in various engineering positions m the jR company's Heat Transfer and Steam Power Plant Design Depts. WJk M. Keller, an engineer in the Flare Div. of John Zink Co-, is a graduate of the Univ of Tulsa. Prior to his employment with Zink, he was a sales engi neer of polyethylene fusion equipment with a Tulsa-based manufacturing firm. CEP September 1977 COLOF.ITE 007402 f PLANT LOSS PREVENTION A Survey of Vapor Cloud Incidents Trends toward plants with larger capacities, and higher pressures and operating temperatures make understanding how to prevent vapor cloud explosions im perative. J. A. Davenport Industrial Risk Insurers Hartford, Conn. r- ... Table 1. Valuation factors for petroleum equipment. Year Factor Year Factor 1976........ .......... 1.00 1975........ .......... 1.06 1974....... .......... 1.18 1973....... .......... 1.38 1972....... .......... 1.44 1971........ .......... 1.49 1970....... ........... 1.60 1969........ ..........1.71 1968........ ..........1.79 1967........ ..........1.87 1966........ .......... 1.95 1965........ ..........2.01 1964........ ..........2.04 1963........ .......... 2.06 1962........ .........2.07 1961........ ..........2.10 1960........ ..........2.10 1959........ ..........2.13 1958....... ..........2.16 1957....... ..........2.22 1956....... .......... 2.40 1955....... ..........2.62 1954....... .......... 2.70 1953....... ..........2.74 1952........ ..........2.78 1951....... ..........2.78 1950........ ..........2.97 1949....... ..........3.06 1948....... ..........3.06 1947....... ..........3.35 1946........ ..........4.05 1945........ ..........4.89 1944....... ..........4.93 1943....... ..........5.10 1942....... 1941....... .......... 5.59 1940....... 54 Vapor cloud explosions have in recent years been the pre dominant cause of the largest losses in the chemical and petrochemical industry. Because of trends toward plants of larger capacity, higher pressures, higher temperatures and greater inventory holdup, these losses have been increas ing both in frequency and severity. Obviously, as a major insurer of chemical and petro chemical properties, our company has a great interest in vapor cloud losses. We have been gathering data on vapor cloud incidents for about two years in order that methods may be developed for: 1. Reducing the number and severity of these losses. 2. Estimation of loss potential from incidents of this type. 3. Analysis of the cause of these losses to aid in educat ing our own engineers and our insureds. In an earlier work 0) concerned with vapor cloud in cidents, no differentiation was attempted between the fol lowing types of incidents involving large releases of flam mable materials: 1. Confined explosions. 2. Boiling liquid expanding vapor explosions (BLEVE's). Table 2. Type of facility for incidents where overpressures were created. Number % Industrial................ Transportation...... Other....................... Total.................... ........ 75 ........ 18 ........ 7 ........ 100 CEP Septemoer 1977 COLORITE 007403 m , V*"* V-'-' $' Type of Plant Table 3. Type of industrial plant for incidents where overpressures were created. Total Number of Incidents: No. % Number Incidents with $ Data Available No. % Corrected Property Damage Loss 8 million % Average Loss, $ million Petrochemical... ......24..... ..... 75.0.... ......19............ 70.4...... .......285.3.............. 60.9... ........ 15.02 Refinery............. ...... 8.... ...... 25.0.... ...... 8........... 29.6...... .......183.4............... 39.1... ........ 22.92 Totals............. ......32 .... ..... 100.0.... ......27 ........... 100.0...... ....... 468.7.............. 100.0 Table 4. Source of spill material for industrial incidents where f>ppr-\' overpress1 ures were created. ' Number &; Total Number Incidents with $ Data Corrected Property Damage Average Kz " Source of f. Spill Material of Incidents No. % Available No. % Loss $ million Loss, % S million *- From Process Equipment. ...24........... . 75...... ....21............ 78.... ........ 329.4............. .. 70. ......15.7 > From Storage Tanks.......... ... 4........... . 13...... .... 4............ 15.... ........ 130.0............. .. 28. ......32.5 t c From Transportation Vehicles within Plants .. ... 3........... . Unknown............................ ... l........... . 9...... .... 2............. 3...... .... 0............ 7.... 0.. . i........ 9.3........... .. 2, ...... 4.6 LL Total................................. ...32........... .100...... ....27...... ......100.... ........ 468.7............ ..100. ......17.4 3. Open-air, unconfined vapor cloud explosions. The survey by Industrial Risk Insurers was concerned with only the third category. We were careful to exclude a number of notable losses (for example, the several poly vinyl chloride plant explosions) which were in the other two categories. Also included in the survey were incidents involving large clouds that upon ignition did not create noticeable overpressures and incidents where large clouds did not ig nite. (See Appendices A, B, and C.) Since large-cloud fires that do not create overpressures and large clouds that do not ignite do not receive the same degree of publicity that the vapor cloud explosions receive, these lists are not con sidered to be as complete as the listing of vapor cloud explosions. For a number of years Industrial Risk Insurers has been using a calculation method to estimate the potential for probable loss in plants where a vapor cloud hazard exists that uses the following criteria: 1. The maximum credible spill is equal to the contents of the largest process vessel or train of vessels not readily isolated. Storage vessels and major supply or fuel pipe lines are not considered. 2. The amount of material vaporized for hot flashing liquids is equal to a ratio of the superheat of the material (CpAT) to the heat of vaporization (AHv). Materials with a boiling point below 70F are assumed to vaporize 100%. 3. The explosive yield is equal to 2% of the theoretical heat of combustion potential in the cloud and is expressed as an equivalent quantity of TNT using a heat of combus tion of 2,000 B.t.u./ lb. for TNT. 4. A maximum peak overpressure from the explosion is 5lb./sq.in. CEF September 1977 5. Damage to the plant is estimated from published data sources. (2,3) While we realize that this is a simplistic approach, we feel it is necessary because factors such as wind velocity and direction, weather conditions, degree of confinement, spill rate, and operator response are unknown prior to the incident. Our calculation method, therefore, assumes con ditions approaching those encountered with explosively dispersed clouds. Data sources Information on the incidents surveyed were gathered from published sources and from private communications. The latter include internal Industrial Risk Insurers loss reports and information supplied by numerous individuals in industry and insurance organizations. Published sources many times do not contain sufficient data to analyze adequately the cause and effect of a vapor cloud incident. Many of these published sources appeared many years ago, before the possibility of an open air fireball explosion was considered. In addition, they were often written to explain the cause of the incident and do not include information on explosion effects and size ot spill. Private communication sources at times gave addi tional information to supplement published sources. The information in this survey is the best available to Industrial Risk Insurers at this time. Some of this iniormation is only a best estimate. If other organizations or individuals have additional or more complete or correct in formation, we would be happy to receive it. \\ e intend to update this survey on a regular basis. 55 COLORITE 007404 Table 5. Mode of release for incidents where overpressures were created. Mode of Release All Incidents No. % Industrial Incidents No. % Transportation Incidents No. % Other Incidents No. % Vessel Failures.............. 13... ......... 30.3,............ 7......... ... 21.9 .............. 5............ .. 62.5...... ....... 1............. .. 33.3 Piping, Valves, or Fittings Failures......... ............ 26... ......... 60.5............. 23.......... ... 71.9 ... . . 3. . 37 5 n 0 Release from Venting Facilities...... ............ 3... ......... 7.0............. 1......... ... 3.1 ............. 0............ ... 0 ...............2............. .. 67.7 Unknown ......... ............ 1 ... ......... 2.3... ... 3.1 .............. 0............ ... 0 ...... ....... 0............. .. 0 Total.............. ............ 43... ......... 100.0... .......... 32......... ...100.0 ............. 8............ ...100.0...............3............. ..100.0 The incident tabulations contained in the vapor cloud survey in the appendices are in three parts: 1) overpressures created; 2) ignition--no overpressures; and 3) no ignition. Each tabulation is arranged chronologically. As the in cidents were gathered, a random number was assigned which is shown as the code number. This number is re tained for ease in future updatingof the survey. This num ber is also used as a key to the data sources listing that fol lows the tabulations. The loss figures are for property damage only. The actual amount of the loss at the time of incident and the amount of the loss if it had occurred in 1976 are given. Valuation factors used are given in Table 1. These are standard factors used by Industrial Risk Insurers for petroleum and petrochemical plant equipment. In the "available data" column of the appendices are given, where possible, information on the following: I) cause of the incident, 2) spill source, 3) amount of material spilled, 4) size of cloud, 5) location and nature of ignition source, 6) wind conditions, 7) equivalent quantity of TNT, and 8) explosive yield or efficiency factor. The amount of the material spilled was in some cases given by the data source and in other cases was calculated from spill source data (size of opening, pressure, and time) or cloud size using previously published calculation methods. (4,5) The TNT equivalent was, in some cases, given by the data source. In others it was calculated from information on the explosive effects by previously pub lished calculation methods and data sources. (2,3) Where both the amount of material spilled and the TNT equivalent could be determined, a yield or efficiency factor was calculated utilizing a ratio of the heat of combustion of the material to the heat of combustion of TNT. Glass breakage data usually is a poor indicator of over pressures from vapor cloud explosions since this damage occurs in a range of 0.5 to 1 Ib./sq. in. and the correlation Code No. Year Table 6. Selected industrial incidents. Material AHC Weight of Material B.t.u./lb. in Cloud, lb. Total B.t.u.'s Actual Loss, $ million Corrected Loss to 1976, $ million 61......1954.. ..Acrolein ......11,830... ... 3.. ...1961.. ..Cyclohexane ......18,676 ... 76......1962.. ..Ethylene Oxide ......11,482... 40,000 8,500 38,000 . 4.73 x 109 .. ... 2 ...... 5.4 1.59 x 10 .. ... 0.7 .... ...... 1.47 .. 4.36 x 108 .. ... 8 ......16.56 18......1964.. ..Ethylene ......20,276... 200 to 500 (4 x 106') .. \ to L ... 3.2 .... ...... 6.5 ll x 107J 4......1965.. ..Ethyl Chloride 62 ......1966.. . .Methane 17......1966.. ..Butadiene 6......1967.. ..Isobutylene ...... 8,246... ......21,502 ... ......19,200 ... ......19,367 ... 38,000 600 850 20,000 .. 3.13 x 10s .. ... 0.9 .... ...... 1.8 .. 1.2 x 107 .. ... 4.8 .... ...... 9.4 .. 1.6 x 107 .. ... 0.016.... ...... 0.03 .. 3.87 x 10a .. ...17.5 .... ......32.7 13......1968., -<C9 fl.9 x lO9') ......19,000 ... ...100,000 to 200,000..... j to k. ...28 13.8 x 109J ......50 40.. ...1969.....Naphtha + H2 77......1970.. Cm + h2 19.. ...1971.. , .Ethylene 32.. ...1971.....Ethylene 22.. ...1971.. ..Butadiene 9.. ...1974.. ..Cyclohexane ......19,000.. ......19,000... ......20,276 ... ......20,276... ......19,200 .. ......18,676 ... 50,000 250,000 8,000 1,000 27,000 120,000 .. 9.5 x 10* .. ... 3.5 .... ..... 6 .. 4.75 x 109 .. ...30 ......48 .. 1.62 x 109 .. ... 2.6 .... ...... 3.9 .. 2.0 x 107 .. ... 6.1 .... ...... 9.1 .. 5.18 x 10s .. ... 0.245.... ...... 0.365 .. 2.2 x 109 .. ...70 ......82.6 8.. ...1974.,->cs (Unsaturated) ......19,000 ,. 60.. ...1975.....Ethylene ......20,276 .. 74.. ...1975.,...h2 ......51,571 .. 16,800 12,000 665 -.3.192 x 108 .. ...13.2 ...........15.6 .. 2.4 x 108 .. ...40 .......42.8 .. 3.4 x 107 .. ... 2.75 .... ...... 2.91 56 CET r,ti'tember i9?7 COLORITE 007405 Figure 1. Amount of loss as a function of cloud energy content explosively dis persed vs. drifting clouds. to TNT explosive effects is less accurate at low overpres sures. (6') Where glass breakage was used to calculate the TNT equivalent, this fact is indicated. Overpressure incidents Since the incidents where overpressures were created are of more immediate concern and most likely are more complete, the conclusions that follow are only from that tabulation. The other two tabulations are presented for in formation only. Type of facility. Table 2 is a breakdown of the type of facility involved in the incidents where overpressures were created (industrial, transportation and others). Where the incident involved transportation vehicles but occurred within an industrial plant, the incident is included in the CEP September 1977 industrial category. Type of industrial plant. Table 3 indicates the type of industrial plant where the 32 industrial losses occurred (petrochemical or refinery) and an average corrected loss amount where loss amount data is available. Note that while only 25% of the losses occurred in refineries, the average property damage loss was $22.92 million, com pared to $15.02 million in petrochemical plants. Source of spill. Table 4 shows the source ot the spilled material in industrial plants. While 75% of the incidents and 70% of the corrected loss amount were due to spills from process vessels, the average corrected loss amount was only $15.7 million compared to S32.5 million tor in cidents caused by spills from storage tanks. Because ot the greater quantities of materials usually found in storage tanks compared to process equipment, a greater average COLORITE 007406 Table 7. Estimated losses using Industrial Risk Insurers calculation method. Material Spilled AH, B.t.u./lb. Weight of Material Spilled, lb. Total B.t.u.'s Property Damage Loss Estimate, $ million Propylene & Propane......... ............. 342,000............. ...........6.8 X 109.......... ................ 29.6 Benzene................................... ............. 322,000............. ...........5.6 X 109........... ............... 11.2 Ethylene Oxide..................... ............. 56,000............ .......... 6.7 X 10s.......... ............... 8.0 Propylene................................ ............. 46,000............ ...........9.1 X 10s.......... ............... 17.0 Ethylene.................................. ............. 66,000............ ...........1.3 X 109.......... ................ 33.2 Propylene................................ .............. 66,000............ ...........1.3 X 109.......... ................ 47.8 c,.............................................. ............. 39,000............ ...........7.6 X 10s........... ............... 15.2 Butadiene............................... ............. 42,000............ ...........8.5 X 108.......... ............... 13.5 Butadiene............................... ..............20,200............ ............. 10,400............ ...........2.1 X 108........... ............... 5.0 Vinyl Chloride....................... .............. 8,239............ ............. 9,000............ ...........7.4 X 107.......... ................ 8.5 Ethyl Chloride....................... .............. 8,822............ ............. 180,000............. ...........1.6 X 109.......... ................ 13.8 Ethylene Oxide..................... ............. 12,024............. ............. 70,000............ ..........8.4 X 10s.......... .................16.4 Cyclo-hexane........................ ..............18,846............. ............. 26,000............ ...........4.9 X 1G9.......... ................ 11.8 Vinyl Chloride....................... .............. 8,239............ ............. 24,000............ ...........2.0 X 108.......... ................ 12.3 Vinyl Chloride....................... .............. 8,239............ ............. 61,000............ ...........5.0 X 108.......... ................ 12.0 Vinyl Chloride....................... ........... 8,239............ ............. 26,000............ ...........2.1 X 108.......... ................ 9.8 Monochlorobenzene ........... ..............11,754............. .............. 4,000............ ...........4.7 X 107.......... ................ 9.0 loss amount is expected. Mode of release. The mode of the release leading to in cidents where overpressures were created is summarized in Table 5. For industrial plants, the largest number of incidents were caused by failure of piping, valves, or fittings. As would be expected, the largest number of transportation incidents were due to vessel failure, reflect ing the number of tank car derailments with consequent puncture of the container. Loss amount and heat content of cloud. Table 6 is a list ing of 18 selected industrial incidents where overpres sures were developed. Included in this list are incidents where both loss amount and spill quantity information were available. The heat of combustion for the material spilled is used to obtain a total heat content of the cloud. The data from Table 6 is presented in Figure 1, in which clouds produced by catastrophic vessel failure (ex plosive dispersal) are compared to drifting clouds. Note that there is reasonable correlation for the explosively dis persed clouds, but a wide variance for the drifting clouds. This is to be expected since explosively dispersed clouds would be less affected by partial confinement, winds, ambient temperature, humidity, and topographic features. Note that nine of the drifting cloud incidents fall above and five fall below the line drawn through the explosively dispersed cloud points. This would suggest that explosively dispersed clouds do not represent the most severe con dition. The various factors affecting drifting clouds would tend to cause them to be more severe or less severe than the ex plosively dispersed clouds. For example, incident code numbers 18 and 32 were small leaks in polyethylene plants where surrounding buildings gave a degree of confinement to the explosion. Therefore, a small leak was able to pro duce a loss of significant magnitude. Incident code number 9 was a large spill from a large pipe with almost im mediate ignition. Dispersion of this cloud would approach explosive dispersion, but because of the congestion in the plant there was a degree of confinement Loss potential. Table 7 is a listing of 17 probable maxi mum loss estimates for plants where a vapor cloud hazard exists insured by Industrial Risk Insurers. These represent 58 a random sampling of our chemical and petrochemical risks. The estimates were made using the calculation method previously summarized. The points from Table 7 are plotted in Figure 1. Note that these points show less deviation from the "explosive dispersal line" than do the drifting cloud points. This cor relation would indicate that the calculation method de scribed predicts the explosively dispersed vapor cloud which, in turn, is not the "worst case," but approximates an average condition. While it is conceded that the spread of points in Figure 1 indicates a data deviation that makes questionable any conclusions reached, the data is the best available, at least to the author. The conclusions are at best generalities and should be viewed as such. It is only through thorough analysis of each incident, both large and small, that an understanding of the factors affecting the probability that a vapor cloud explosion will occur and that it will be severe if it does occur can be determined. I urge those who have information on additional in cidents and additional information on incidents included in this survey to forward it to the author. Confidentiality of the source will be strictly observed. Data submissions should include, if possible, the fol lowing: 1. Date and location of incident, 2. The type of process. 3. Material spilled. 4. Size, height, and shape of cloud, 5. Amount of material spilled. 6. Source of spill. 7. Size of spill opening. 8. Pressure and temperature of materials in process. 9. Weather conditions (wind, temperature, humidity). 10. Overpressure effects (description of damage, prefer ably with photographs). 11. Amount of property damage loss. In conclusion There are many factors that affect the development of a vapor cloud and the explosive forces, if an explosion oc- CEP September 1977 COLORITE 007407 cuts. The data presented in this article are insufficient to reach any firm conclusions concerning these factors. Gen erally, it may be shown that explosively dispersed vapor cloud explosions produce property damage that is a func tion of the amount of material dispersed. Drifting-cloud vapor-cloud explosions do not show any consistency. The vapor cloud loss estimation material used by Industrial Risk Insurers predicts an explosively dispersed cloud and is reasonably consistent with the actual incidents in the study. The explosively dispersed cloud is not the most severe nor, for that matter, the least severe incident, but is rather an average incident. Acknowledgment I would like to thank the many people in the chemical, petrochemical, and oil industries who contributed much of the information presented in this survey. Without their valuable assistance, this survey would be far less defin itive. Literature cited 1. Strehlow, Roger A.. "Unconfined Vapor Cloud Explosions--An Over* view," paper presented at the Nth Symposium (International) on Combustion at Pennsylvania State Univ (August 1972), 2. Br&ste, W. C., and D. W. Simpson, "Guidelines for Estimating Ex plosion Damage," in "Loss Prevention/' Vo). 2, A CEP Technical Manual AIChE, New York (1968). 3. Stephens, M. M., "Minimizing Damage to Refineries from Nuclear Attack, Natural and Other Disasters," Office of Oil and Gas, U.S. Dept, of the Interior (1970). 4. Brasie, W. C., "Guidelines for Estimating the Hazard Potential of Chemicals," paper presented at the 81st National Meeting, Loss Pre vention Symposium, AIChE, Kansas City, Mo (1976). 5. Decker, P. E.. "An Analytical Method for Estimating Overpressure from Theoretical Atmospheric Explosions, ' paper presented at Annual Meeting of the National Fire Protection Assn. (1974). 6. High Pressure Polyethylene Committee, The Assn, of Petrochemical Industries, Japan, "Experiment on Ethylene Explosion," paper pre sented at the 74th National Meeting, AIChE, Tulsa, Okla. (1974). Code Year Appendix A. Vapor cloud survey, overpressures created. Location Type of Process Material Released Actual Property Loss, S million Property Lose Corrected to 1976, $ million Killed/^ Injured Available Data 65 .. .1921 - .England 25 .., .1948 . .Germany 61 .. .1954 . .West Virginia 37 ,.1954 . Portland, Ore. 42 . .1955 . .California 33 .,. .1956 . .New York 44 . . .1958 - -Oklahoma 47 . ..1959 ..Georgia 3 ....1961 . .Tex** 76 . . .1962 . .Kentucky 48 . . .1962 , .New York 1 . .1963 . .Louisiana 63 . . .1964 . .Nevada 18 . . .1964 . .Texas 7 .. 1965 . Louisiana .. Hydrogen ? . .Dimethylether ? Acrolein 2 ..LPG .. 0.33 . .Gasoline Plant . .Butane ..04 . .Polyethylene . .Ethylene 7 . .Refinery -- . .Cyclohexane Oxidation . .Propane ..LPG . .Cyclohexane .. 0,1 7 .. 0.7 . .Ethyl Amines . .Ethylene Oxide 8 . .Propane . .Ethylene * -- . .Methane or Ethylene . .Hydrogen . .Polyethylene . .Ethylene .. 0.2 3.9 7 ,, 3.2 . .Ethylene ,.Propane .. 32 ? ? 5.4 0.89 1.0 ? 0.2 ? .. 1.5 .. 16.6 .. 0.4 .. 8.0 ? .. 6.5 6.4 . i/? . .Break-up of Dirigible ZR-2 over Hull, England re leased hydrogen which ignited. Windows broken within 2 mi. Blast heard 5 mi. Tremors felt 50 mi. . Many . .Rail car filled "bubble-full." Car overheated in sun. No relief valves. Car rupture explosively dis persed car contents forming vapor cloud which exploded. .. 0 .6,000 gal. of acrolein explosively dispersed into a vapor cloud when tank car ruptured. Acrolein was contaminated causing polymerization which overtaxed car relief vent. 7 ,400 ft. diameter cloud formed from broken safety relief valve stem on tank car Two storage tanks 60 ft. away distorted indicating 3.5 to 7 Ib.,sq. in, overpressure. 7 . Rupture of 6 in. line at 500 lb./sq. in. in a gasoline plant led to explosion and fire that destro>ed all outside processing equipment including towers, pressure vessels and piping. ? ,40,000 cu. ft. at STP released from polyethylene re actor vent at 20,000 lb./sq. in. Aerial explosion created 1/2 Ib./aq. in. overpressures. ..0/1 .. .Flexible coupling failure. Ignition at boiler Blew employee into ditch. ..23/? .. 1/0 . .Train wreck punctured tank car. Rural area. NFPA report states there was an explosion. . .Valve failed at reactor bottom releasing 6,000 gal. spill. Cloud contained 8,500 lb of vapor. Control room failed. Estimate overpressure of 2 lb sq in. .. 1/9 at 100 ft. from point of spill. . Day tank containing 6,500 gal. of ethylene oxide be came contaminated with ammonia. Tank rup tured, explosively dispersing ethylene oxide into air. Cloud immediate!}, ignited creating explosive forces equivalent to tons of TNT as evidenced by damage. ..10/17 . .Truck accident ruptured 6,876 gal. tank. Cloud 400 to 600 ft. in diameter, 80 ft high exploded, Low .. 0 overpressures created. . .Sight glass or metering station failure, Explosion of cloud resulted in 1 ton TNT equivalent. Ensuing .. 0 fire of long duration. . .Venting experiment accidently resulted m ignition of 200 lb. of hydrogen. Explosive force estimated to be 60 lb. of TNT. .. 2/17 , .Failure of 3/8 in. compression fitting on 1.000 to 2,500 lb./sq in. ethtlene line in pipe trench re sulted m spill of 200 to 500 lb. Explosion equaled 0.12 to 0.30 tons of TNT. Explosion in courtyard between buildings giving partial confinement. .. 0/12 . .Vented cold liquid into carbon steel flare header. Continued on pose 60 CEP September 1977 59 COLORITE 007408 Appendix A. Vapor cloud survey, overpressures created, (continued) Code Year Location Type of Process Material Released Actual Property $ million Property Loss Corrected to 1976, $ million Killed/ Injured Available Data 4 , ..1965 ..Louisiana 62 , ,. 1966 . .West Germany , .Ethylene Hydrochlorination -Ethylene . .Ethyl Chloride . .Methane .0.9 . 48 1.8 9.4 17 . . .1966 . .Louisiana . .Synthetic Rubber . .Butadiene . 0.016 0.03 6 . . .1967 , .Louisiana . .RefineryAlkylation Unit Isobutylene .35 13 . ..1968 . .The Netherlands . .Refinery . -C^or lighter .,28 Hydrocarbons 40 . . .1969 . .England , .Refinery-Cyclic Naptha + H2 Hydroformer ,3.5 24 , , .1970 . .Missouri -- -.Propane ? 65 50 6.0 7 77 . ,. 1970 . .New Jersey .Refinery, H-oil , .>Cl0 Hydro- Unit carbons + H2 .30 48 19 . ..1971 ..Louisiana . .Ethylene .Ethylene .2.6 3,9 32 . , ,1971 . .Texas 22 . . 1971 ..Tern* .Polyethylene . .Ethylene ..Butadiene . .Butadiene .8.8 13.1 . 0.16 .. 0.24 23 . ..1972 ..Illinois . J.PG 94% Propylene .7.5 10.8 15 . . .1972 . Brssil 49 . . .1973 . .Japan 52 . . .1973 . .Japan 68 . ..1974 ..Florid* 9 . . .1974 . .England 59 . . .1974 . -Illinois . .Refinery . -Ethylene ..VCM - . .Butane . .Ethylene .Vinyl Chloride ..Propane .8.4 .16 *> .1.0 12.1 22 7 1.2 . .Cyclohexane Oxidation . .Cyclohexane .70 83 .. . .Isobutane . 18.4 21.7 38 . . .1974 ,,Mississippi . .Butane 77 .0 Rupture of recycle line on reactor led to spill of . 3/83 . 3/2 6.000 gat prior to ignition. Explosive force of 0.15 to 0 20 tons of TNT was estimated. . -Ignition occurred within seconds after pipe failure. Logical source of tgmtion 200 ft. from leak. 600 lb. of methane released. Explosion equivalent to two to three tons of TNT. . .Valve on reactor bottom opened accidentally on air . 7/? . 2/75 7 ? failure. Ignition source 50 ft. from point of spill. Spill about 200 gal. prior to ignition. OverpreasuresofO.5 to 1 Ib./sq. in, were estimated. ...Failure of 10 in, gate valve bonnet on line led to re lease of about 20,000 lb. of isobutylene. Explosive force of 12 tons of TNT estimated from glass breakage which would give a 10% yield. . .Breaking of water-oil emulsion in slop-oil tank caused tank froth-over and failure, 50 to 100 tons of hydrocarbons formed huge cloud. .Failure of 16 in. vapor line at 250 Ib./sq. in. led to formation of large cloud with ignition within 10 sec. Estimate 50,000 lb. spilled. .122,000 lbs. at 942 Ib./sq. in. spill from a pipeline rupture. Cloud 1500 ft. long, and 10 to 20 ft. high detonated (according to U.S. Bureau of Mines) with a force equivalent to 50 to J5 tons of TNT. 0/4 .. Failure of reactor operating at 2500 Ib./sq. in. due to localized overheating caused large cloud of about 250,000 lb. which, upon ignition, created widespread explosion damage and fires. Over pressures were highly directional. Peripheral damage indicates an explosive yield of up to 50 tons of TNT. A yield factor of about 4% is indi cated. 0/8 ...Failure of truck tanker with frozen relief valve at about 400 Ib./sq. in. released about 6,000 lb. of ethylene. The equivalent of about 1/2 ton of TNT did widespread damage to light construction components. 3/7 Release of about 1,000 lb. of ethylene from 1/2 in. pipe connection to larger pipe. Three explosions occurred. Some confinement from building walls around an alleyway, 1/6 ...Failure of pump. 12 mi./hr, wind caused cloud to dnftdOOft. before ignition from unknown source S-l0 mm. after release. Estimate a spill of 27,000 lb. Explosion damage consisted of broken win dows, displaced cooling tower internals and ce ment asbestos siding. . 0/230.,.118,000 lb. of propylene released when rail cars collided in humping operation in switchyard. Rolling car created an elongated vapor cloud cov ering about 5 acres. Two explosion centers were identified, U.S. Dept of Transportation states that the explosion may have been a detonation, .37/53 ...Frozen open water draw-off valve on bottom of storage sphere resulted in cloud which exploded. . 1/7 .Excessive pressure and temperature from process upset caused valve flange to fail, . 1/23 ...Broken valve yoke caused 1,100 ton spill creating 53.000 cu. ft. cloud. .0/0 .Hose from 9,000 gal. tanker failed causing 40 ft high vapor cloud. Explosion destroyed two ware houses, crushed cars and broke windows in four block area, .28/100 ,,.Poorly constructed and installed 20 in. temporary pipe failed releasing 60 tons of materials at311*F. and 136 Ib./sq. in. Ignition in 25 to 35 Bee, created 20 to 30 ib./sq. in. overpressures. Explosive force estimated to be equivalent to 15 tons of TNT. . 7/356 .Accident created 22 in. x 26 in. hole in end of 30.000 gal. rail car releasing contents. Ignition of the 1/2 x 3/4 mi. cloud was 8 to 10 min. after the collision. Based on damage, the apparent yield of the explosion was between 200 and 400 tons of TNT. An unusually high conversion factor (25 to 50%) is indicated, . 0/24 . .Overfilling of salt dome storage well created a cloud 1.25 mi. in diameter. Two explosions occurred. The second was 800 to 1,000 ft above grade and 60 CEP September 1977 COLORITE 007409 Ta Appendix A. Vapor cloud survey, overpressures created, (continued) Code Year Location Type of Process Material Released Actual Property Loss, $ million Property Loss Corrected to 1976, S million Killed/ Injured - Availab'e Data 57 ,. ,1974 . .England Polyethylene ..Ethylene 54 ...1974 ..Texas 5 .. .1974 . .Texas . .Butadiene Isoprene Unit . ,>C5Hydrocarbons 11 ...1975 ..Belgium 60 ,..1975 ..Holland 70 .. .1975 , .Germany 74 ...1975 . .Calif, .Polyethylene Plant . .Ethylene .Ethylene , .Propylene .Refinery-Cyclic . Naptha Hydroformer .Hydrogen Hydrogen 79 .. .1975 . .Czechoslovakia or 1976 51 .. .1976 . .Texas .Ethylene .Ethyl Alcohol Light Hydro* carbons Ethylene 13.3 13.2 26.5 40 2.75 15.7 .. 15.6 .. 29,1 42.8 1.1 2,91 was the most severe- Glass breakage to 7 mu, houses damaged to 900 feet. 0/2 . .Leak from flange joint or broken thermowell in high pressure piping caused spill of 2,000 to 6,000 lb. in about 30 min. Windows broken in 200 yd. ra- dius. Roofs and ceilings in four buildings dam* aged. Electrical breakers tripped in process areas and boiler house. Part of gas confined in open top barricaded area. 1/235 . .Humping accident in railyard resulted in puncture of 34,000 gal. tank car. Amount of spill in 2 to 3 min, before ignition not known. Ignition by lo comotive 600 ft. away. Estimate 2 to 3 Ib./sq. in. overpressure at 1,000 ft. from point of rupture. 2/9 .Piping failure led to spill of 16,600 lb. Explosion knocked out deluge systems and electrical feed to fire pumps. Most damage from ensuing fire. Maximum overpressure estimated from damage was 3.5 Ib./sq. in. at 120 ft. Explosive equivalent of 1 ton of TNT was estimated giving a yield fac tor of 1%. 6/13 Fatigue failure of a vent connection on a compressor suction line let ethylene escape which formed a gas cloud in an open area that ignited from an un known source. Highly directional pressure waves caused widespread structural damage. ..14/104 ..Failureof level controller on column caused cold liquid to pass out of relief valve and into carbon steel flare header which cracked. Cloud of 12,000 lb. ignited at furnace 150 ft. away. 0/4 . .Failure of carbon steel line due to hydrogen attack. Slight area explosion after release of hydrocar bons in cyclic hydroformer. 0/2 , .Crack in tank caused tank failure releasing 665 lb. of gaseous hydrogen to atmosphere. Cloud ex ploded with an apparent center 30 ft. above grade and an energy release equivalent of 20 to 40 .14/7 lb. of TNT. . .Extensive damage. No other details available. 18 18 1/15 . .Line failure at 2,000 Ib./sq. in. led to ethylene re lease with subsequent vapor cloud explosion. Code Year Location Appendix B. Vapor cloud survey, no pressures. Type of Process Material Released Actual Property Loss, $ million Property Lost Corrected to 1976. $ million Killed/ Injured. Available Data 20 ,, .1943 . .California 36 .. .1945 . .New Jersey .. -- . .Refinery 26 .. .1954 . .Term, , .Pesticides 43 ...1957 . .Quebec, Canada ., -- 46 .. .1958 - .Michigan 45 .,. 1958 . .California , .Refinery , .Refinery 50 .. .1965 . .Texas 71 ,, ,1966 . .France . .Refinery .Butane .Crude Oil .Butane Butane .Oil Froth .Butane ,03 Hydro carbons CEP September 1977 0,10 0,25 3.2 0,5 9 7 7 . 0.68 ,. 7.1 . 1.1 . 19.4 7 7 *> .. 1/7 .. 1/7 2/18 7 ..45/7 , .Bottom connection on 4,300 gal- tank failed creating 1/4 mi. diameter cloud. ..Failure of cast iron valve on 31.000 bbl. tank created cloud which was igmted at highway 209 ft. away. . .Vapor cloud formed from unknown leak source. Vinyl chloride cylinders ruptured from overheat ing creating loud noisea. . .Overfilling of storage sphere created 1,200 ft. diam eter cloud. Storage spheres subsequently rup tured causing some explosion damage. , .Overpressure of 21,000 gal. tank caused failure of tank. Cloud ignited 3$0 ft from source. ..Waterin 50,000 bbl. oil tank caused frothing and tank overflow. Amount of vapor formation from froth not known. . .Removal of wrong blind flange led to spill of un known quantity which was ignited by heater 60 ft. from spill point. , .Frozen open water draw*off valve on bottom of storage sphere resulted in cloud which was ignited by vehicle 300 ft, away. Sphere subsequently ruptured. Continued on pact' 62 61 COLORITE 007410 Appendix B. Vapor cloud survey, no overpressures, (continued) Code Year Location Type of Process Material Released Actual Property Lobs, $ million Property Loss Corrected to 1976, $ million Killed/ Injured Available Data 41 .. .1966 ..Pennsylvania 72 ., .1968 . .Texas 27 .. ,1968 . .Louisiana 58 .. .1969 . .England 35 .. .1972 . .Texas 14 .. .1972 . .Montana 2 .. .1974 ..Louisiana M.. .1976 . .Texas 16 .. .1976 . .Puerto Rico . .Refinery . .Isopropyl Alcohol ..VCM Plant . Cyclohexane Oxidation . .Alkylation . .Ethylene -- . .Cumene, Benzene, Propane . .C3 Hydrocarbons . .2 & acid mixture . .Vinyl Chloride . 0,83 .. 7 3.6 1.48 .None . 1/2 . . None of 6 in. pump discharge pipe. Cloud was wind affected. .Cloud formed when 10 in, spool piece split longi tudinally was ignited at furnace 180 ft. away. ...Casing joint or seal failure on pump created cloud which was ignited at furnace 100 ft. from point of failure. Cloud less than 3 ft high in most areas. . .Cyclohexane 7 7 . 2/23 A low cloud formed from 4 tons, which was released at lOOlb./sq. in. and 30Q'F. from an improperly blinded pipe. Ignition was by a diesel engine which overspeeded located 100 ft. from spill ..Crude Oil point. Cloud covered an area 250 ft. x 150 ft. ? 7 . 1/2 .Split in 6 in. pipeline ran for 4-1/2 hr, before igni tion. Scorched area 1,800 ft. long. About 332,000 gal. of material contained 98% C5 or lighter hy . .Butane -2.3 drocarbons at 530 lb /&q. in, gauge was released. 3.3 .1/1 .Leak from partly open 6 in. flange was ignited at furnace 100 ft. away. Resulting Fire caused vessel . .Propylene .2.6 to rupture which created minor overpressures. 3.1 . 0/0 . Power failure led to failure of expansion joint re leasing about 10,000 gal. prior to ignition at furnaces. One deluge system between spill point . .Natural Gas Liquids ..C5 Hydrocarbons .0 05 0.05 . 1/4 and ignition source tripped by steam leak prior to ignition. . .Crosscountry pipeline failure led to cloud which .. 7 .. ? ..1/2 ignited and destroyed farmhouse. Tank overpressure led to release of vapor or mist which was 10 to 25 ft. high. Ignition was at a fur nace 450 ft. from the tank. Scorched area was Code Year Location Appendix C. Vapor cloud survey, nc Type of Process Material Released Actual Property Loss, $ million Property Loss Corrected to 1976. $ million Killed/ Injured Available Data 34 ,. .1951 , .Illinois 67 ,, .1967 , .Texas 5 .. .1970 . .Louisiana 10 .. ,1971 . .Florid* 21 ....1972 . .Delaware 69 .. .1975 . .Louisiana 73 .1975 . .Louisiana ? . .Butane - ? . VCM Plant . .Cyclohexane Oxidation . .Ethylene -- ..Vinyl Chloride & .. Ethylene Dichloride " . .Cyclohexane -,PVCi>lnt , .Vinyl Chloride . .Synthetic Rubber -- . .Butadiene . .Propane -- -- .. -- -- ~' -- -- ,-- ,-- * -- -. .-- about 800 ft. diameter. Cloud was consumed in 10 to 30 sec. ...9,500 gal. of butane released from ruptured bottom discharge pipe on tank due to freezing of col lected water. Excess flow valve failed to operate. Pilot on 16 in. flare faded allowing release of 4,000 lb. in a 100 ft. deep cloud. Wind was 12-15 mi./hr. Cloud of 150 ft diameter formed when packing on ball valve blew out. Waterspray manually ac tuated. Wind blew cloud away from possible ignition sources. Failure of a 4 in. weld connection to an 8 in. pipe released 75,000 )b. of material. Cloud was 2,000 ft. long, 1,200 ft. wide and 100 ft. high. Material reached and burned in boilers but did not ignite. A number of other possible ignition sources were passed without ignition. Numerous waterspray systems were actuated manually. Failure of 3 in. gasket on reactor top released 1,800 gal. Waterspray was actuated by vapor de tectors. .. Approximately 1,000 gal. of material discharged as a gas from a hole in 2 in. nozzle on the top of a storage tank. Waterspray tripped manually. Overpressure caused split in 6 in. underground line from storage well releasing 600,000 lb. Cloud was 30 to 40 ft. high and vapor was noted 1 mi. down wind. Literture cited in Appendices. Code No. 1. Adcock, C. T. and J. D. Weldon, "Vapor Release and Explosion/' in 62 "Loss Prevention," Vol. 1, pp- 70-74, a CEP Technical Manual, AlChE, New York (1967). Private Communication. 2. Saia, S. A., "Vapor Clouds and Fires in a Light Hydrocarbon Plant," Chem Eng. Prog, 72 (U) 56-61 (1976). 3. Private Communication. CEP September 1977 COLORITE 007411 4. Klock, JohnC., "Fire in Ethyl Chloride Reactor at Baton Rouge Plant of Ethyl Corporation," American Petroleum Institute Meeting, Phila delphia (October 1965). 5. Private Communication, 6. National Fire Protection Association Fire Records Dept., Fire J,, 62 (3)8(1968). Private Commun cation. 7. Kevil, C. G-, "Flaie Line Rupture in an Ethyiene Plant," in "Safety in Air and Ammonia Plants," Vol. 9, a CEP Technical Manual, AIChE, New York (l%7). 8. Private Communication. 9. Chemical Marketing Reports, (October 4,1976). Dafter, Ray, "Flixborough --A Failure of Safety Management," The Financial Times (May 13, 1975), Dafter, Ray, "Nypro Must Take Some of Blame," (Flixborough Disaster Report) The financial Times, pp. 14-16 (May 13,1975). Dafter, Ray, "The Process Industries," The Financial Times, p. 20, (April 30,1975). Dafter, Ray, "Touchstones for Chemical Plant Safety," The Financial Times, p. 24 (February 13, 1976), Dept, of Employment, "The Flixborough Disaster --Report of the Court of Inquiry," Her Majesty's Stationary Office, London (1975). "Experts Differ on Possible Cause of Flixborough Explosion," Fire International, 45, pp. 18-23. Kletz, T. A., "The Flixborough Cyclohexane Disaster," in "Loss Pre vention," Vol. 9, pp. 106-113, a CEP Technical Manual, AlChE, New York (1975). McDonald, J., "Nypro Errors Caused Factory Blast, Union QC Claims," The Financial Times (February 19,1975). "A Review of the Flixborough Disaster and its Implications to the Insurance and Reinsurance Market," The Mercantile and General Reinsurance Co., Ltd., London. Taylor, H, D-, "Flixborough, The Implications for Management," Keith Shipton Developments Ltd., London (1975), Tucker, D. M., "The Explosion and Fire at NypTO (UK) Ltd., Flix borough, on 1 June 1974," paper presented at Interfire, London (July 30, 1975). Warner, F.f "Flixborough Disaster," Chem. Eng. Prog., 71 (9) 77-84, (1975). Private Communication. 10. "Design Deficiency?," National Safety Council Safety Newsletter, Chemical Section (June, 1975). Private Communication. 11. Private Communication. 13. Ministry of Social Affairs and Public Health, "Report on the Cause of the Explosion at Shell Perm*," State Publishing House, The Hague, The Netherlands (1968). "Shell Refinery Fire," Fine J,, 62. (5) 110-112(1968), 14. Private Communication. 15. Private Communication. 16. Private Communication. 17 Private Communication. 18. Private Communication. 19. Private Communication. 20. "LP Gas Firesand Explosions,"NFPA Quarterly, 46, (1) (July, 1952). 21. Private Communication. 22. Furiow, W., "Fire, Explosion at Chemical Company Injure Six Workers," The Houston Post (September 16,1971). Private Communication. 23. "Hazardous Materials Railroad Accident in the Alton and Southern Gateway Yard in Eaat SL Louis, Illinois, Jan. 22, 1972," Railroad Accident Report NTSB-RAR-73-1, National Transportation Safety Board, Washington, D.C. (1973). 24. Burgess, D. S. and M. G. Zebetakis, "Detonation of a Flammable Cloud Following a Propane Pipeline Break--The December 9, 1970, Explosion in Port Hudaon, Mo.," 17. S. Bureau of Mines, Report of Investigations 7752, Pittsburgh (1973). 25. Official reports of the International investigation Committee on the cause of the July 28.1948, tank car explosion at BASF, Ludwigshafen, West Germany. 26. Allan, R., "Fire, Explosion Rake Chemical Plant Here" The Commercial Appeal, pp. 1-2, Memphis, Tenn, (February 3,1954), 27. Private Communication. 32. Private Communication. 33. Private Communication. 34. "LPGas Firesand Explosions," NFPA Quarterly; 46 (1) (July 1952). 35. "Pipeline Accident Report--Exxon Pipe Line Company Crude Oil Ex plosion, St. Hearne, Texas, May 14, 1972." National Transportation Safety Board, Report No. NTSB-PAR-73-2, Washington, D.C. (1973). 36 Rugbee. P., "Fire Protection Developments in 1945." NFPA Quarterly, 39, (3) 148 (January, 1946), 37. "Railroad and Ship Fires," NFPA Quarterly, 48 (3) 315 (Jan. 1955), 38. "Bi-monthly Fire Record," FireJ., 69 (1) 54 (January, 1975), 40. "Corrosion Causes Multimillion Dollar Refinery Fire," National Safety Council Newsletter, Chemical Section. CEP September 1977 Private Communication, 41. McCarty, J. J.< "Petrochemical Fire Threatens Giant Relinerv," Fire Engineering, pp. 103-104 (August, 1966). 42. "LargestIndustrial Fire Lossesof 1955," NFPA Quarterly 49 (4) 309 (April, 1956). /fLP Gas Fires and Explosions," NFPA Quarterly, 55 (*>) 186 (Oc tober, 1961), 43. "LP Gas Fires and Explosions," NFPA Quarterly, 55 (2) 186 (October, 1961), 44. "LPGas Fires and Explosions," NFPA Quarterly, 55 (2) 187 (Oc tober, 1961). 45 Woodworth, Miles, "Oil Froth Fire at Signal Hill Refinery" NFPA Quarterly, 52, (2) 85-98 (October. 1958). 46. "Large Loss Fires of 1958." NFPA Quarterly, 52 (4) 320 (April, 1959). 47. "Large Loss of Life Fires in 1959," NFPA Quarterly, 54 (1) 35-36 (July, 1960). 48. Walls, W., "LP-Gas Tank Truck Accident and Fire, Berlin, N.Y.." NFPA Quarterly, 57,(1)9-14 (July, 1963). 49. Private Communication. 50. Private Communication, 51. Private Communication. 52. Private Communication. 54. "Railroad Accident Report--Hazardous Materials Accident at the Southern Pacific Transportation Company's Englewood Yard, Houston, Texas, September 21, 1974," National Transportation Safety Board, Report No. NTSB-RAR-75-7, Washington, D.C. (1975). Private Communication. 57. Private Communication, 58. a. "Wilton Explosion -- Diesel Engine Danger," Chemistry in Britain, p. 3(1970) b. "Diesel Engine Dangers," Chem Age, p. 40 (December 12, 1969) c, "More Dangerous Diesel," Chem, Age, p. 11 (January* 7,1970). 59. "Coordinated Attack Limits Post-Blast Damage," Fire Command, 42 (7) 14-17 (July. 1975). "Railroad Accident Report --Hazardous Materials Accident m the Railroad Yard of the Norfolk and Western Railway at Decatur, Illinois. July 19, 1974," National Transportation Safety Board, Report No. NTSB-RAR-75-4, Washington, D.C, (1975). Private Communication. 60. Private Communication. 61. Private Communication. 62. "Accidents in the Hydrocarbon Spiegel, December, 1969. Processing Industry," Schaden Bradford, W. J. and T. L. Culbertson, "Design of Control Houses to Withstand Explosive Forces," in "Loss Prevention," 1, 28-30, a CEP Technical Manual, AIChE, New York (1967). Private Communication. 63. Reider, R., H. J. Otway, and H. T. Knight, "An Unconfined, Large- Volume Hydrogen/Air Explosion," Pyrodynamtcs. 2, 249-261 (1965); paper presented at 53rd National Meeting of American Insti tute of Chemical Engineers, Pittsburgh, May 1964. 64. "Mapco Pipeline Bursts in Texas, Killing One," H<j(/ Street J.. February 27,1976. 65. Toland, J., "Ships in the Sky," Holt, New York, (1957). pp. 65-76. 67. Private Communication. 68. "Propane Gas Explosion Rattles Florida Town," Houston Post, January 5,1974. 69. Private Communication. 70. Private Communication. 71. Private Communication. 72. Private Communication. 73. Private Communication. 74. Private Communication. 76. Troyan, J. E. and R. Y. Levine, "Ethylene Oxide Explosion at Doe Run," in "Losa Prevention," Vol. 2, pp. 125-130, a CEP Technical Manual, AIChE, New York (1968). Private Communication, 77, Private Communication. 79. Private Communication. J. A. Davenport, who earned his B S Ch E trom Virginia Polytechnic Institute, joined the Factory Insurance Assn, in 1963 as an inspector m the Philadelphia area. In 1971, he <*a- promoted to chemical engineering supervisor and mo\ed to the Eastern regional office in Hartlord, Conn, His present duties include research m chemical processes, the fire hazards ot plastic*, and related subjects. 63 COLORITE 007412