Document 2Yaj42GNDmGkXKdnB56O0NzR

INTER-OFFICE CORRESPONDENCE To. Date. Memo NAME June 10, 1981 LOCATION From. Subject _ Mr. J. A. Karr: name location VCSA MEETING - BERMUDA - OCT. 23-24, 1980 Copy to: Hooker Chemical PLASTICS DIVISION GENERAL: The Association now includes a total of 44 companies, 20 of them from outside the U.S. The total is up 5 from the addition of new entrants, among them XCI and Fills. At the 1980 meeting, then, there were more companies and fewer people than at last year's meeting. Next year's meeting place -- Hilton Head, S. C., Sept. 24-25, 1981 TABLE OP CONTENTS OF. FOLLOWING WRITE-UP Page Page Page Page Page Page Page Page Page Pag Page Page Page Page Page Page 1 2 3 3 4 4 4 6 6 6 6 7 7 7 8 8 VCM Monitoring VCM Emission Control and Redundancy VCM Absorption in EDC (Ethylene Dichloride) Respiratory Protection; plus Attachment () Preparation for Vessel Entry Industrial Risk Insurance, plus Attachment(5) Regulatory Update, U.S. European Regulatory Update, plus AttachmentQi-T) High Pressure Cleaning System Vent Relief Systems; plus Attachment (c)' VCM Monitoring System Improvements for Safety VCM Safety Audit Electrical Accident Power Loss and VCM Release VCM Charge Pump Fracture; plus Attachment(5) True Confessions; Great American disaster on Page 1. VCM Monitoring. Ethyl Corp.. Baton Rouge, Henry Smith: The Company started with a Miran system primarily for OSHA, but also for the EPA VCM Standard. Because of interferences from moisture, ammonia, etc., they switched to a chromatograph system. They have four 10 point units (15 min. cycle); one in each of two poly areas and two VCM production areas. A computer produces an 8 hour shift report, a daily report and a monthly report. They formerly used polyethylene tubing, but found Teflon best. Sensing points are set about 53$ feet off the floor level. They have sample points up and down wind of each plant area, but none at the fence line. They check each sensing point by bag sampling every 3-6 months. MCC-14] Results: Poly areas - average 0.7 ppm; 93% 0-1 ppm; 6% 1-5 ppm; 6T8% 5-25 ppm; 0.2% above 25 ppm. VCM Production - average 0.2 ppm; 98.7% 0-1 ppm; 1.2% 1-5 ppm; 0.1% 5-25 ppm; a few each month above 25 ppm; two blips above 25 ppm starts their Leak Detection/Prevention program in action. Control Rooms - avg. 0.7 ppm; 1.6 ppm max. R^fl'Ltor Lntry - avg. 9.7 ppm; range 0.4-60 ppm; mean 2 ppm VCM Loading - avg. 1.4 ppm; range 0.1-2.7 ppm; mean 0.9 ppm Compressor, Ar^or, - avg. 1.4 ppm; range 0.2-5*1 ^8 J>pm -2- 1. VCM Monitoring:, Ethyl Corp.. Baton Rouge, Henry Smith (cont'd) The alarm levels are the same as in our plants. They plan no changes in the system. Meeting discussion: Some companies present say that Teflon has a VCM memory. Ethyl has not established this. Pour attendees said they monitor at the fence line and that about 90% of the time levels are below 1 ppm. Six plants have back up monitors. There was a consensus agreement that, wherever possible, all companies would exert pressure against the requirement for daily calibrations. 2. VCM Emission Control and Redundancy. Tenneco, Texas, Tom Moore: This plant uses the Shin-Etsu process with large reactors like Hooker's Addis plant. They have two computers backed up by a battery system. Instrument electrical power is backed up by a second battery bank. The plant also has a standby diesel generator. They incorporated several key design features from their experience of the Flemington explosion some years ago. For reactor charging they use three devices; an oval gearmeter with tank level gage, an oval gearmeter with DP orifice meter and a turbine meter with tank load cell. The reactor stopper system is like that at Addis. They have dual relief valves on each reactor and capability to relieve overpressure to atmosphere or to a flare and either automatically or manually. The have nuclear devices, but are relocating these, since they did not always insure reactors were empty. They have two incinerators for emergencies only for the recovery system vent. They are underdesigned and must be used together. However, they are correcting this fault so that a spare will be available. Their incinerator stacks are high enough so that HC1 discharge is no problem. A smoke problem was worked out in agree ment with the Government. Tenneco cautions against overreacting in providing redundancy to an existing plant. They consider preventive maintenance to be vital. They put great importance on regular simulation testing. They insist that their Engineering Dept, follow up to insure that the systems operate as they were designed to do. Their operator training stresses minimizing errors and not so much prevention of errors. They have rewritten their computer program three times during their expansion. occ 2938 ,, t; _ 3. VCM Absorption in EDO, Mr. Swinkel of AKZO, Netherlands: Most of the VCM produced at a VCM plant near Rotterdam is loaded onto ships. The plant made 524,000 tons of VCM in 1979. At rated capacities the three former Firestone FVC plants consume just over 300,000 tons. Vent gas containing VCM comes from the ship loading, some RR car loading, storage areas, VCM purification and EDC purification. Total vent gas flow is 10-50 kg/hour. VCM emissions are regulated by an agency equivalent to US EPA. The VCM emissions have been reduced to less than 40 ppm by use of the EDC absorption tower. They consider the absorption system to be highly reliable. About 1000 cu. meters/hr. (580 CFM) is fed into the bottom of a tower, 0.6 meter diam., 5 meters high (2 ft. diam., 16 ft. high). The tower is filled with 1" 0 rings. The EDC is at about 0F. 4. Respiratory Protection, Glen Schaaf, Goodrich: This Company makes the operator responsible for keeping his respirator clean. The Company provides the means to do this. One tool is an automatic dishwasher (home type) with the drying cycle blocked out. Concerning beards, Goodrich requires a man to shave or transfer to a no-mask area, if his beard results in his failing the mask fit test. Asked what they do if there is no other job available, G. Schaaf could not answer. One company was told by the Government that it had to provide a job in such a case. One company was cited on noncompliance of bearded men. Ten plants require clean shaves - no beards. During one company's plant inspection, the OSHA inspector said he wasn't required to wear a mask. Goodrich tests bearded men twice per year and others once per year. They allow only the positive pressure respirator for bearded men. Fourteen companies use two types of respirators, ten use three types and one uses only one type, the full face mask. Fourteen companies do qualitative testing of masks while three do quantitative testing. The latter is required for the Lead Standard and the test equipment costs $10,000-$20,000. Only a few companies do the qualitative fit test in the plant processing areas. Results differ there from the standard test. Freon is approved for the fit test. OCC 2939 -4 4. Respiratory Protection, Glen Schaaf, Goodrich (cont'd): Only two companies enforce the use of air paks when responding to a fire alarm. This is required in the OSHA manual. No company sends anyone into an explosive cloud of VCM - except for life rescue. In general in emergencies the problem is to keep people out. In the four pages of Attachment at the end of this write-up additions] information of the Goodrich program is shown. 5. Preparation for Vessel Entry. Glen Schaaf, Goodrich: This Company considers this the most hazardous job in its plants. They average about 180 entries per month. However, "vessel" has the broad meaning and includes any confined area. Goodrich has one man always standing by outside the vessel. He NEVER goes in. They have a hoist to lift out the man inside. After locking out power equipment, it is required on their check list that the start switch .be tried. Every 3 minutes the man inside must pull a cord to silence an alarm that sounds every 3 minutes. An air-line respirator is required for poly vessel entry and any other, if only a double block and bleed is used. They have not set a limit of VCM concentration in the air of vessels to be entered. No company only closes valves to "blank" a vessel; eleven use double block and bleed; 22 companies blank or break lines; five use radios inside vessels; 19 require lifelines. Luncheon Speaker, Industrial Risk Insurers, John Davenport: A write-up of this address is attached as Attachment^ 6. Regulatory Update, John Barr, Air Products: OSHA: The carcinogen policy issued Jan., 1980, is in litigation but was ruled a standard by the 5th Circuit Court. The candidate list of 250 suspected carcinogens was issued in Aug., 1980 and is in two parts; 200 of EPA concern; 150 OSHA. Priority lists are due next. The benzene standard was overturned and the coke oven case was withdrawn. On the lead standard, a man removed from an exposure area for health reasons, will have his pay rate retained for 18 months, at least. On May 23 a rule was issued that records of exposure and health data must be retained for 30 years and be available to anyone. This is in litigation by the La. Chem. Assoc. OCC 2940 -5- 6. Regulatory Update, John Barr, Air Products (cont'd): OSHA (cont'd): The Fire Brigade Standard was issued Sept. 12, 1980. A brigade is not a requirement. OSHA did not follow proper procedure in promulgating rules on walkaround pay. The court remanded this. OSHA will try again. Labeling proposal is past due. This would require an operator to be supplied with a "map" of valves, pipes, etc. in his area. Because OSHA inspectors have been refused entry in several instances, they will now arrive with warrants. At a NIOSH/OSHA symposium in March, 1980, an Edinburgh study was re viewed. There were unclear indications of long term adverse effects from FVC dust inhalation. The regulatory groups have assigned a project officer on this potential problem indicating that the possibility exists that the issue may be reopened. EPA: Hearings were held on EPA's carcinogen policy in March, 1980. A suspect chemicals list was released, but this will be revised in view of the testimony of the Scientific Advisory Board. The body was critical of the quality of the data used to include Toluene, TCE, Acrylonitrile, Methyl Chloride, Perchloroethylene and Methylene Chloride saying that there were no justifiable bases for the selection. The issues are still open. EPA has contracted with TRW for a five-year study to determine the need and bases for revision of the VCM Standard. The outcome of the carcinogen policy issues will bear on this. Emergency Releases will be a major issue. A General Fugitive Emissions regulation for loading and storing all organic chemicals will be forthcoming soon. This will be similar to that of the VCM Standard. EPA is in total disarray on the Water Quality Criteria and no further action on this is foreseen. The National Cancer Institute has found that phthalates and adipates are carcinogenic in rodents. John Barr conducted a poll on Emergency Discharges for the last year: 8 companies had less than 2 discharges 9 companies had 3-6 discharges 3 companies had 7-10 discharges one company had over 10 discharges (no company reports such discharges to OSHA unless personnel are exposed) OCC 2941 -6- 7. European Regulatory Update, Herbert Deuschl, Wacker-Chemie: This is summarized in Attachment 8. High Pressure Cleaning; System, Varren Clark, Dow: This was a discourse of two accidents related to their high pressure system. The first concerned a severe steam burn suffered by a man when a valved off pump was left running. Afterward a T connection was installed on pump discharge to drain the line. The second was more serious in that an operator lost one eye when he dropped the lance and the fail safe did not work. After the accident Dow reevaluated their system. From this they set new standards: 1. No wand will be less than 4- ft. long. 2. Maximum hose length between lance and bypass valve will be 25 ft. 3. All flow must stop in 0.432 seconds when a lance is dropped to the floor from a height of 3 feet. 4. The nonhazardous level is when there is zero pressure on the lance. 5. A three-way dump valve must be installed on all lances; this is to be actuated electrically from a micro-switch on the handle of the gun. Dow uses the Par Tek nozzle. 9. Vent Relief Systems. Glen Schaaf, Goodrich: The results of a survey conducted prior to the meeting in Bermuda are given in Attachment(c)at the end of this writeup. 10. VON Monitoring System Improvements for Safety, Mike Rodnicki. Shell: In their plant in 1980 they had 880 OSHA alarms from 8500 samples taken per month: 600 alarms - above 0.5 ppm VCM 240 alarms - above 1.0 ppm VCM 40 alarms - above 5-0 ppm VCM In that year from 1200 samples per month they had 575 EPA alarms: 500 alarms - above 0.25 ppm VCM 50 alarms - above .50 ppm VCM 25 alarms - above 5 ppm VCM OCC 2942 -7 - 10. VCM Monitoring System Improvementsfor Safety, Mike Rodnicki, Shell (cont'd): In their Leak Detection and Prevention Program, leak checks are per formed by one "sniffer" and one maintenance man. They have a drawing of each piece of equipment and leak locations are marked on the drawing. They designate which leaks are to be repaired at once and which can wait until later. 11. YCM Safety Audit, J. Cizek, Diamond Shamrock: Their audit team was made up of one man each from departments of Safety, Engineering, Environmental Compliance, Loss Prevention, Personne and Production. The audit covered VCM handling up to reactor charging: storage, pipe lines, vapor cloud formation, emergency procedures, fire protection. The audit covered 3 days and started with an initial conference with all concerned present. For each area a "what if" study was made. Then each area was inspected and the people in the area were consulted. The problem items in each area were delineated and a report was sub mitted to Management with recommendations for action. The findings showed that the greatest need was to update Emergency Procedures. 12. Electrical Accident, Glen Schaaf. Goodrich: In one plant a fused type breaker was wrongly installed in a panel for nonfused types. The only difference between the two was the number stamped on each - size, shape and appearance were identical. The fused breaker rested slightly deeper in and contacted the metal of the panel box. On the power start, the breaker shorted with a large flash and fire. An Engineer at the spot was burned and lost three months work recovering The breaker box and panel were practically destroyed. 13* Power Loss and VCM Release, J, Zimmerman, Diamond Shamrock: At the Deer Park, Texas plant they had 12 Emergency Relief Valve Dis charges from 6 reactors and associated equipment. In Mar., 1980, a maintenance error relating to a transformer caused loss of agitation. The automatic stopper addition system was down for repair. They used manual stopper addition, but this was ineffectual. There was insufficie: room for blowdown. About 4-2,000# of VCM was discharged. All personnel were evacuated. No fire or explosion resulted. They received a notice of violation from the State, but have heard nothing from either EPA or OSHA. OCC 2943 -8 - . Power Loss and VCM Release, J. Zimmerman, Diamond Shamrock (cont'd): In future similar emergencies, they intend to manually vent to try to get enough agitation for manual stopper addition to be effective. 14. VCN Charge Pump Fracture, J. Miller, Esso~Canada: Attachment at the end of this writeup presents a case history of this fracture. The cause was operation with closed suction and discharge valves. No peroxides were later found nor any evidence of auto ignition temperature of which is 878F. They concluded that over 1000 PSIG was achieved. Tests showed that the Durco impeller raises VCM temperature 68F in one minute. In group discussion, Great American reported a similar failure of a cast iron pump. Goodrich reported two: a stainless steel pump "blew up" with slurry trapped in it; another fragmented holding flush water. In the latter they concluded pressure was over 1200 PSIG. 15* True Confessions: A. Great American Chemical Co.. Ray Asti, Fitchburg, Mass; EXPLOSION & FIRE On a Friday at 9:00 p.m. in Sept., 1980, an operator in a building of three reactors noticed that the reactor flush water was not draining out the bottom. He went below to open the 6" reactor drop valve and the 4" drain valve (to atmosphere) on the cross in the drop line below the reactor drop valve. In error he opened the drop valve of a live reactor. At first a little water drained out. Thinking he could open what seemed to be a partial plug from above, he turned to leave and at this time smelled VCM. As he reached the door the live batch blew out. When the Foreman heard the blow out, he went at once to the lower level. Looking in the open door he saw a cloud of VCM vapor streaming past over his head. At the base of the 4" drain valve he saw a grapefruit sized blue glowing ball. The ball expanded in size and then the vapor cloud ignited. He felt no heat nor wind force and, miraculously, he was not physically harmed in any way. The wall on an adjoining building, across the alley about ten feet away, was charred from the blast except for an area roughly equivalent to the superimposition of his form, somewhat enlarged, on the surface. Other clouds of VCM traveled through other exits in various directions. Some burned and others did not in a random pattern. In a group of trees about 150 feet away, one tree would be totally scorched while another next to the first was untouched. OCC 2944 -Q_ 15. True Confessions (cont'd): A. Great American Chemical Co., Ray Asti, Fitchburg, Mass.: (cont'd) EXPLOSION & FIRE (cont'd) The VCM vapor blovm by the fans up the old, high stack (6 ft. diameter, constructed of heavy tiles, adapted for ventilation) overflowed the top and flowed down the sides onto the roof. When the cloud ignited, the stack was demolished and 90 feet of it came down on the roof. The roof then collapsed on top of the reactors. There was much building structural damage and piping, wire conduits, etc. were turned into spaghetti, but there was little serious damage to major process equipment or vessels. After the worst was over, numerous small flares from minor leaks in the reactor room had to be put out. This was a nightmare working under the collapsed roof by only the light of the leak flares. The 110 volt power was, of course, shut off. The reactor charge vessel was full but was secured without harm. At the start alarms went off at 20% of LEL and others at 4-0%. The latter activated the sprinklers. At 20% of LEL the ventilation fans automatically went to high speed. On hearing the blow out the tank farm operator blocked off transfer lines. Other operators in buildings stopped all transfers of materials and other process operations. Reactor operators stoppered all batches within % hour. Neither agitation nor cooling water flow was lost during the crisis. One operator suffered first, second and third degree burns and another, first and second degree burns. There was no experience of 'HC1 fume irritation because the fire took the vapors upward very fast. Total VCM lost was about 15,000#. The reactor was of 3750 gals, size with an initial charge of 10,000# VCM. It had had only a few batches prior to the blow out. The operator had one year's experience on the job, but was considered very competent. It is not known if a polymer plug or a faulty valve was the cause of the temporary delay before the blow out. Within 15 minutes from the blast, Ray Asti was on the site. EPA came in one hour later with protective clothing and instruments. Their readings at the fence were all less than 0.5 ppm VCM. ORHA came in, but has not been heard from since. The Fire Dept, (municipal) had to be called in despite their being automatically alarmed when the sprinklers were activated. OCC 2945 10 True Confessions (cont'd): A. Great American Chemical Co., Ray Asti, Fitchburg, Mass, (cont'd): EXPLOSION & FIRE (cont'd): There was much initial adverse publicity in the news media. The Government attitudes to Great American's intent to rebuild was favorable. When they do rebuild, they will position sprinkler heads at the bottoms of reactors. This is from their belief that the glowing blue ball was static electricity. B. Keysor-Century Corp.; Saugus, Calif.: FIRE: Maintenance had changed a reactor drop valve to a Jamesbury wafer type, glass filled, teflon lined. They had not followed the correct procedure and, therefore, the valve would not fully close. The operator checked the valve and thought it was closed. When he charged the vessel, VCM vapor traveled 50 ft. away to a nonexplosion proof area where an electric switch ignited the cloud first outside the reactor building. This two story building, 150 ft. x AO ft., houses all but the reactors. The reactors are out in the open with no deluge system. The alarms sounded but very shortly the rupture disc blew on a VCM charge tank. A very short time later some relief valves of reactors blew. A few residents nearby were evacuated and road blocks were put up in a 2J4 mile radius. The local Fire Dept, would not enter the plant. There was still fire in the reactor charging building 1% hours later and the last fire was not out until ten hours later. C. Borden Chemical Co., Ray Jepsen, Mass.: 1. Gasket failure, reactor top entering agitator pedestal: On a 5200 gals, vessel the gasket was of 18" diameter and glass lined with a teflon envelope. They concluded its failure was from a loss of structural integrity, although part of the problem was that the bolts were not tightened sufficiently. The gasket had been in service for nine years. They had noticed some corrosive attack after five years. They will now regasket reactors every five years. A poll of the audience showed that no company had a gasket check or scheduled replacement program. OCC 2946 11 15. True Confessions (cont'd): C. Borden Chemical Co., Ray Jensen, Hass.: 2. Gasket failure, reactor foot bearing: Pfaudler foot (step) bearings had been removed from 5200 gals, reactors, but were still on the 5700 gals, reactors. The bearing gaskets were teflon, a replacement for the formerly used lead sheath asbestos type. However, because the bearing surface of the teflon was only 50% of the lead type they had changed back to the latter. The first one installed failed releasing about 5000# of VCM from a cold charge at 34- PSIG. The flange dropped between There was no lead left on about 180 of the gasket circumference. There was no fire nor other crisis. They have no explanation for the failure. A second batch was charged after installation of a new lead gasket. However, the bolts were retorqued after normal installation. They could take 1/8 turn to 1 turn on various bolts around the circle. Batch cycle was normal. A third batch was also normal. Before this charge the bolts were retorqued without changing gaskets. All other reactors were normal. Their Illiopolis plant uses the lead type without problems. Pfaudler recommends teflon and Borden will change to this type. D. PYC Plant, Identity Unknown: A 105 300W rail car of VCM was being unloaded. As the cover of a magnetic gage was being unscrewed, it blew off said the gage rod blew out to a point 35 feet away. An operator was hit by an edge of the cover. They have no explanation. The speaker asked for a safe procedure for removing this type cover. A VCM producer in the audience reported that one of his plants had a similar case while loading a car. They found it to come from a weld leak. E. British Company, Identity Unknown: They had a fatality from nitrogen displacement of air in a vessel, despite the nitrogen system being color coded and having different fittings. OCC 2947 Zrr- AtiimxplH'ric convent ration of rinttl chloride S. F, 6ooc>r tu4 Required apparatus. Artm, (i) Unknown, or above 3.600 ppm....................................Open-circuit, self-contained breathing apparatus. pressure demand type, with full facepiece. (ii) Not over 3.600 ppm ___ (a) Combination type C supplied air respirator. pressure demand type, with full or half facepiece, and auxiliary self-contained air supply; or -- .... (b) Combination type C, supplied air respirator continuous flow type, with full or half facepiece and auxiliary self-contained air - supply.___________ ;_____________________ (iii) Not over 1,000 ppm___ Type C, supplied air respirator continuous flow type, with,full or half facepiece, helmet or hood.__________________________ (iv) Not over 100 ppm......... (a) Combination type L supplied air respirator demand type, with full facepiece, and aux iliary self-contained air supply; or (b) Open-circuit self-contained breathing ap paratus with full facepiece, in demand mode; or (c) Type C supplied air respirator, demand type, t with full facepiece. i t (v) Not over 25 ppm........... (a) A powered air-purifying respirator with hood, helmet, full or half facepiece, and a canister which provides a service life of at least 4 hours for concentrations of vinyl chloride up to 25 ppm, or (b) Gas mask, front- or back-mounted canister which provides a service life of at least 4 hours for concentrations of vinyl chloride up to 25 ppm. (vi) Not over 10 ppm........... (a) Combination type C supplied-air respirator, demand type, with half facepiece, and aux iliary self-contained air supply; or (b) Type C supplied-air respirator, demand type, with half facepiece; or (c) 4ny chemical cartridge respirator with an organic vapor cartridge which provides a service life of at least 1 hour for concentra tions of vinyl chloride up to 10 ppm. OCC 2948 BFGOODRICH CHEMICAL GROUP Attacm. (a) -Atmospheric Concentration of Vinyl Chloride Required Apparatus Unknown, or above 1,000 ppm,,,................ ............... Open-circuit, self-contained breathing apparatus, PRESSURE DEMAND TYPE, WITH FULL FACEPIECE, Not over 1,000 ppm AND ALL LOWER LEVELS......... .. TYPE C, SUPPLIED AIR RESPIRATOR CONTINUOUS FLOW TYPE, WITH FULL OR HALF FACEPIECE, HELMET OR HOOD. OCC 2949 B.fr (JOObRicw A-rrACrf, (a) c=L3 aN oHE 0? ? t Quantitative Test OCC 2950 RESULTS OF THtS EXAMINATION INDICATE: EXAMINEE NAME PHYSICIAN SIGNATURE DATE DATE THERE IS THERE IS NOT THIS PERSON IS THIS PERSON IS NOT REASON FOR EXCLUDING THIS PERSON S CONTINUED EXPOSURE TO LEVELS WHICH ARE PERMISSIBLE UNDER OSHA S STANDARD OR BFGOODRICH MEDICAL POLICIES. APPROVED FOR THE USE OF PROTECTIVE EQUIPMENT AND RESPIRATORS. O' EUROPEAN REGULATIONS COVERING VCM ENVIRONMENT Federal Republic of Germany France Italy Norway Netherlands United Kingdom Sweden 3 3 kg/hour/source or 150 mg/m /source 0.3 rag/m^ for 99% of measures taken in inhabit* zones authorisation of regional mines offices project: monitoring in few factories emission of up to 60 kg VCM/hour as a one year average at a production of 100.000 tons PVC/ye; From June 30, 1982 on: reduction from 60 to 20 kilos VCM/hour * > 3 authorisation of regional authorities monitoring by environment inspector 0.5 kilos/ton S-PVC 10 ppb 13.0 kilos/ton P-PVC OCC 2952 EUROPEAN REGULATIONS COVERING VCM Armca.(S-x WORKER PROTECTION Belgium Denmark Federal Republic of Germany Finland France Italy Norway Netherlands United Kingdom Spain Sweden * Switzerland EEC regulation project (not yet valid): 5 ppm/week, action level 15 ppm for all factories 1 ppm/8 hours for producers existing factories: 5 ppm/hour, 15 ppm/hour, max.40 factories to be built + processors: 2 ppm/6 ppm/hour, max. 40 5 ppm/8 hours (personnel), max. 15 ppm/15 minutes existing factories: 5 ppm/week, action level 15 ppm new factories: 1 ppm/week, action level 5 ppm EEC: project ( not yet valid): 3 ppm/technical long term limit/year, action level ISppm 1 ppm VCM, mean value for 8 hours shift 10 ppm max. (personnel) Code of practice project, not yet valid: ceiling value 30 ppm and TWA lO ppm/8 hours EEC: project (not yet valid): 3 ppm/technical long-term limit/year, action level 15 ppm authorities require: 1 ppm/8 hours (personnel), max. 5 ppm/15 minutes (personnel) 2 ppm/8 hours (average MAK) project (not yet valid): 3 ppm/technical long-term limit/year, action level 15 ppm OCC 2953 EUROPEAN REGULATIONS COVERING VCM ATTACH. FOOD PACKAGING Belgium Denmark Finland Federal Republic of Germany France Italy Norway Netherlands United Kingdom Sweden Switzerland Spain w- EEC regulation Migration max. 0.05 mg/kg project (not yet valid): 1 ppm max. in packaging, migration max. 0.05 mg/kg max. 0.05 mg/kg VCM in foodstuff 1 ppm max. in packaging material project (not yet valid): migration max.0.05 mgAg | migration max. 0.05 mg/kg on simulants no government regulation. Industry uses BGA: max. 1 ppm in packaging material max. 0.01 mg/kg in foodstuff project (not yet valid): migration max.0.05 mg/6dnf no regulation migration max. 0.01 mg/kg migration max. O.Ol.mg/kg max. 1 ppm in food packaging material migration max. 0.05 mg/kg project (not yet valid) :*max. 1 ppm in packaging material migration n^ax. 0.01 mgAg occ 2954 Attach-. VENT RELIEF SYSTEMS 1. RESULTS'OF SURVEY 2. METHODS OF RELIEF SYSTEM SIZING 3. QUESTIONS & DISCUSSION Glen D. Schaaf Oct., 1980 OCC 2955 SUMMARY OF SURVEY ATTACH. OVERPRESSURE PROTECTION PRIMARY RELIEF SYSTEMS - MOST PVC REACTORS ARE FITTED WITH RUPTURE DISK - RELIEF VALVE COMBINATIONS - ALL PRIMARY RELIEF SYSTEMS DISCHAR6E TO THE ATMOSPHERE - MANY REACTORS HAVE MULTIPLE RELIEVING SYSTEMS - ALMOST 100% OF THE POLYMERIZERS HAVE MANUAL VENTS - SOME MANUAL VENTS DISCHARGE TO THE ATMOSPHERE - GAS HOLDERS - RECOVERY - ONLY A FEW COMPANIES HAD BLOCK VALVES PRECEEDING THE RUPTURE DISK OR RELIEF VALVE -3' ) SIZING OF RELIEF SYSTEMS - TOTAL RELIEF AREA AVAILABILITY (CALCULATED ON NOZZLE SIZE) VARIED FROM A MIN, OF 1.57 TO A MAXIMUM OF 12.56 SQUARE INCHES PER 1,000 GAL. OF REACTOR CAPACITY - SMALL REACTORS 1,000 - 5,000 GALS. = 2 - 6 SQ. IN. 1,000 GAL. - LARGE REACTORS 6,000 - 35,000 GALS. 3 - 12.56 SQ. IN, 1,000 GAL. occ 2956 /Vttaok (c) III. OVERPRESSURE PREVENTION - MANY REACTORS HAVE A RELIEF SYSTEM THAT CAN BE ACTIVATED EITHER AUTOMATICALLY OR MANUALLY - A FEW OF THESE SYSTEMS DISCHARGE TO THE ATMOSPHERE - BUT MANY DISCHARGE TO GAS HOLDERS - SOME TO RECOVERY AND SOME TO CLOSED TANKS - ALMOST 100% HAVE A SHORTSTOP' SYSTEM - MOST FEEL IT IS EFFECTIVE. SHORTSTOP CAN BE INJECTED EITHER MANUALLY OR AUTOMATICALLY - ALL HAVE RECOVERY SYSTEMS AVAILABLE FOR RUN-AWAY BATCHES - MOST HAVE SOME METHOD OF DISCHARGING LIQUID TO PREVENT FAILURES DUE TO HYDROSTATIC CONDITIONS IV. i METHODS TO HANDLE A RUNAWAY BATCH - COOLING - ADD WATER - OPERATOR INVOLVEMENT - ADD SHORTSTOP - VENT TO RECOVERY - GAS HOLDER - OR EMPTY VESSEL - MANUALLY VENT REACTOR - USE OF RUPTURE DISK & RELIEF VALVE TO VENT TO ATMOSPHERE OCC 2957 attack, V. SIZING OF RELIEF SYSTEMS - ASHE CODE FOR UNFIRED PRESSURE VESSELS - 100% VCH - FIRE CONDITION - REACTION EXOTHERH AND NO COOLING - VCM VAPORIZATION RATE REQUIRED TO REMOVE HEAT OF REACTION AT RELIEVING TEMPERATURE TIMES TWO, VAPOR FLOW WITH NO BACK PRESSURE - FACTORY MUTUAL DATA SHEET EMERGENCY VENTING OF VESSELS - BOYLES PAPER - MONSANTO FORMULA MAX, RATE OF REACTION AT TEMPERATURE EQUIVALENT TO PRESSURE OF RELIEF SYSTEM - CORRECT REACTION KINETICS FOR THE TYPE OF PVC PRODUCED COMBINED WITH A VALID EQUATION FOR HANDLING A MULTIPHASE FLOW THROUGH THE RELIEVING SYSTEM - SIMULATE LOSS OF COOLING, DETERMINE RELIEF REQUIRED CONSIDERING REACTOR KINETICS AND MULTIPHASE RELIEF FLOW, SIZE RELIEF SYSTEMS TO KEEP MAX. PRESSURE BELOW ASME CODE LIMITS. t OCC 2958 ATTACH. @ Rupture Disk Nominal Size (in.) Flow Area (in.^) 3 7.393 4 12.73 6 28.89 ` 8 50.03 Relief Valve* Nominal Size (in.) Flow Area 3X4 1.287 2.853 4X6 6.38 6X8 6 X 10 11.05 16.00 8 X 10 26.00 * Largest orifice size for that size valve. I 0CC 2959 COMPANY AND LOCATION 1. 2. 3. 4.' 5. 6. 7. 8. 9. 10. POLY SIZE GAL 1,1.00 2,500 3,300 4,300 16,500 2,000 4,000 4,000 6,000 2,400 3,500 4,000 6,000 24,000 24,000 20,000 PVCPOLY SIZE RELIEF AREA CDMPARISION RATING PSIG RELIEF NOZZLE SIZE NUMBER 200 2" 1 200-400 4" .1 250 4" 3 300 4" 1 240-300 180 180 10" 3" 3" 1 3. 3 200 4" 200 4" 1 1 200 1.85-200 4" 4" 1 1 200 4" 220 4" 200 18" 180 16" 260 16" 1 1 1 1. 1 TOTAL AREA AVAILABLE 3.14 12.56 12.56 12.56 78.5 7.07 7.07 12.56 12.56 12.56 12.56 1.2.56 12.56 254.34 200.96 200.96 RELIEF AREA PER 1000 GAL 2.85 5.02 3.81 2.9 4.75 3.53 1.77 3.64 2.09 5.2 3.5 3.14 2.09 10.59 8.37 10.05 OCC 2960 PVC POLY SIZE RELIEF AREA OOMPARISION COMPANY AND LOCATION n. 12. 13. 14. 15. 16. 17. POTY SIZE GAL 36,000 35,000 2,200 5.000 7.000 20,000 3,500 7,400 4,200 5,700 4,500 3,700 5,200 17 ,000 20,000 RATING PSIG 172 185 200 225 200 225 300 250 200 200 200 200 200 165 200 RELIEF NOZZLE SIZE 8" 8" 6" 4" 6" 6" 8" 6" 6" 4" 4" 3" 4" 6" 8" x 12" i 6 8" NUMBER 2 2 2 3 1 1 2 3. 1 1 1 1 1 1 2 2 1 Note: Total relief area is based cn full nozzle size for comparison and does not include the reduction of area if a relief valve is used. Q 2376E/wp/4 o o TOTAL AREA AVAILABLE 100.48 357 12.56 28.26 28.26 100.48 28.26 28.26 12.56 12.56 7.07 12.56 28.26 213.52 107.36 - RELIEF AREA PER 1000 GAL 2.87 4.49 5.71 5.65 4.04 5.02 4.07 3.82 2.99 2.20 1.57 3.39 5.43 12.56 5.34 'tOV-U-V