Document NG2GX7g7E4XNXdy5np3Va0dRg

VINYL CHLORIDE SAFETY ASSOCIATION MEETING Program Mlmitea of February 10-11. 1971 meeting. Menmhla. Teen. 1. Plant Safety Design J. T. Barr, Alreo Chemicals All aspects of control room design were discussed. Location is certainly one of the most controversial items, A poll indicated 16 hare con trol rooms, 11 do not. There vere 13 near or adjacent to the poly room and 5 remote with distances varying from 60*300 feet. Two vere of blast proof con struction, 13 vere pressurized by a remote source of air. The actual strength of a blast proof vail is subject to considerable variation. One such vail moved by the force of the explosion and interrupted the power source. In another explosion a section of vail was projected through the reactor room, laboratory and into a nearby building. Suggested correction for that one was reinforcing rods on the side where personnel protection is needed. Pressure in the protected control rooms was 0.1 in. water, most used standard electrical fittings. One had a sensor in the duct for the remot air. Some have delta pressure alarms or other flow indicator alarms. Windows between the control room and poly room - 2 spoke against them, one had none but pot one in during an expansion reasoning that the operators were more interested and more conscious of reactors and conditions with visual contact. Question: Are groups on reactor floor separate from those in the central room? Answer: 1 - Yes, Many * No. One newly constructed plant designed its control room 300 feetsway with TV and radio contact. Fireproofing far structures varied somewhat. Concrete eront around steel is common. Factory Mutual reportedly specifies 2*3 in. ffcom glass held by SS clamps covered by grout. They also accept reinforced concrete structural toll. Valves. Teflon. Where there is metal to metal seal, teflon is all right but ground contacts most be maintained. Open versus dosed buildings. Seemed to conclude that in the south open buildings are fine but in the north, not so. Flexible lines; 8 use. Copper; only 1 company uses monel. 2. Emergency Procedures * R. S. Allen, Conoco ISA, Johnson-Wllliams and 3 others now make continuous monitoring sensors. Normal activation is 25 and 50$ LEL far high speed from operation end deluge respectively. Factory Mutual would not accept a pressure fog system seemingly because they had no experience with it. Steam sponge in the air S /ventilation at 25$ LEL will help raise humidity fast. Deluge heads can be sheared off in an explosion and render the system ineffective. Fixes can be piped from the area with valving to isolate damaged sections. Conoco*s new plant has 8 headers leading to the reactor area from a pill box across the road. Activation can be manual or automatic or combined. Electrical rooms are remote. N provision i mad for water run off. Personnel safety can be increased by providing guide rojea for people caughuia heavy smoke or blinded. In an eirarpeney. hit the deluge Mid rat out. OCC 2614 -z- Batches can be terminated on power failure or excess exotherm. Seven have auxiliary power, one has total auxiliary power* Ten have drills on emergency shutdowns and 7 tag valves to be sure all were checked* Be sure to have bead count after ordering evacuation* Boob threats. Five have procedures. The Any will assist in setting procedures and handling the bomb. There were 7 reported bomb threats, seven are planning bomb scare procedures. Normally the police and firemen voa51 hunt for bombs but may provide deactivation service when located* Recycle VC. One company stares at -10 C. some add inhibitor and then distill to purify. Should use recycle as quickly as possible to prevent peroxide and polymer problems* 3. Disposal of Vented Vapors - Q. B. Speed, Escambia Rupture discs and relief valves. None had rupture discs only andnone had a valve only to protect tank contents. One had a relief valve only. Thirteen had a rupture disc and relief valve. Three had rupture discs in series. One had rupture discs in parallel. Blowdown tanks. Three have uaagltated tank outside the building and can put isoprene (or other) stopper in to kill batch. Tanks nnrmny hold 2 batches. 4. Review of Recent Accident at Diamond - E. A. Ross, Diamond The Deer park, Texas plant started in 1953 and has open construction. New employees are trained to appreciate the hazards of vinyl chloride even to a witness of burning VC and the fumes given off. The accident record was good first lost time accident wae a ladder Incident in 1959 * received Best Safety Awards in 1956, 1957 and I960 and accumulated 1,000,000 man hours accident free record. The accident wae reviewed essentially as it was written. A large manhead gasket gave way and an explosion occurred ffcora the vapors released. A banlcn shirt caught fire and was responsible for unnecessary burns. The Insurance people credit static for the Initiation of the explosion but two other possible sources, a light fixture and a gas burner for the dryer, could have been responsible. Heat melted vires and power vas lost, discs and valves blew on 5 reactors but na fire came tram these vapors released to the roof* The large manhoods of the remaining reactors of similar design have been welded (certified) to the reactors. Agitation will be cut up to bo removed end split hub type used. A caution was Issued to consider the possibility of vinyl vapors trapped in the shaft and not cut with a torch* An outside ladder will be installed in the dryer tovero Glass lined pipe has been removed since static can build up and cause sparking, a deluge system Is being Installed* Rings vers welded to tbs smaller heads to provide better seals for the Lenape heads. Improvements are being mad to the lighting system. A Loss Prevention Committee has been formed. All gaskets will ba changed annually at least. 5. Review of Accidents at Other F7C Plants. A. Atlantic Tubing - E. T. Biehl. Pantasots OCC 2615 3 A reactor manhaad gasket was reused and not properly centered - a very alls part formed the seal on one edge* The reactor wouldn't pressure test but was charged anyway end blew oat* Static is the suspected cause although an electrical short is possible* The roof concrete lifted end fell bending relief valves and piping* Reactors are 2200 gal* B* Tenneeo - C. Q. Thompson, Tenneeo The release of vapors was purely accidental caused ty an operator reaching out to grab something and grabbing the handle of the drop valve on a recently charged reactor* The source of ignition was approximately 150 feet fram the source of vapors* A sketch shoved the burn areas and buildings in volved and slides showed damage and protective measures during the suooeeding days* All valves now require a wheel geared for several turns to open to prevent' l a "quick full open" accident. All personnel are required to wear long sleeves with collar buttoned* long pants* hard toes and hard bat* Having the body fully covered could have prevented much of the body bums sustained* Fan operation 1b this* Normal * Winter - 15 changes/hour Summer - 30 changes/hour High Speed - Vinter - 30 ohanges/bour Simmer - 60 changes/hour Some of the complications of "outside" help such as volunteer fire department people* state police and the press were outlined* Consider a fire control van from which operations can be directed. Panel Discussion* Fog vs* Deluge* Mr* Schwab of Allied Chemical discussed their deluge system. The normal air change is 10 per hour * at 20% LEX the fans automatically go to 20 changes per hour* At 40% LEX the deluge comes on* Coverage is *35 e*11*"1" per sq* ft* per minute* Fung) pressure is 125 psi with one electric and one dlesd. pump with capacity of 2000 gpm* They are adding a 2 million gnn reservoir for additional capacity* To test the sensors they prepare a bomb with the desired percentage of vinyl end air and check each point. Bach sensing point monitors continuously as contrasted to the system at Atlantic Tubing which monitored each point in a sequence which required 8*10 minutes to complete the cycle* Grinnel- Multimatic valves are used* Mr* Dowell of Goodrich discussed their fog system. The sole purpose f the fog is to increase the humidity to prevent the possibility of stati spark. Target is 80-90% humidity* The units are operated at 80 psi* ere about 7-8 feet above the floor and disperse 1.1 gpaio Coverage per unit is ID f t diameter area. The folly of using black iron pipe was pointed out but galvanized pipe has proven satisfactory. At 21 changes per hour the humidity was raised 'sufficiently in l-g- minutes* At 6 changes per hour the humidity was raised from 53$ to 85% in one minute* OCC 2616 - 4- Additlonal eomnants trm the floor - one installation can deluge throe buildings in 12 hours at 10,000 gput and also provide steam injection at the inlet fans* One company has been requested by insurance to go away froa the HAD system and install deluge* One fin was caused by venting from a dipstick to cod down a sample bomb but the bomb vaant groundwd. 6* Safety Procedures with High Pressure Water Cleaning Systems - M* G* Glenn, General Tiro In this installation the high pressure pump is Installed in the utility building (and considered a utility) which is about 70 feet from the reactor building* The pump is a Manton Gaulia 3600 KL4 which was surplus equip* matt and available* Horse power is 50 at 1200 RPW (cut from 100 HP originally)* Relief is set for 3500 psl The lance has 6-800 psi rating, socket welded - L used Vogt unions, certified welder* Orbit'valves at 6000 psi are used* The gun Is rated at 6000 psi with other models available* Capacity at 3000 pal is 26 gpm* The hose is attached to the header reaming down the reactor row and is approximately 8 feet long and can be quickly moved from one looation to another. The major failing in this set up is the dome area which ean*t be adequately cleaned* Nozzles have limited life* Sc<c > Other comments - can shield couplings for personnel protection. One company pressure tests hose every 3 months (nine companies do this). Hose lengths were up to 25 feet* Must go into reactor anyway. Glass etching at high pressures is no problem. Deionised water has caused sparks up to 30 inches* If an abrasive is put in the water, steel or concrete can be cut (as well as the nosale)0 5* Gasketing Material * E. A* Rose, Diamond (Note that the timbers follow the proposed program and not the sequence of presentations*) Gasketing was discussed and samples shown* Two people are using polypropylene gaskets with the Lenape head* Polypropylene will expand 20*30 times and still be a good gasket* Only the filled gaskets should be used on manhoods and one suggests only the corrugated ring should be used* 9* Spark Sources * J* T. Barr, Aire Spark sources are numerous with some hot enough to cause an explosion and others not hot enough* Sparks from metal contact (tools) and personnel are not supposed to fas hot enough, but moving equipment can generate enough heat* Non polar liquids passing through a polar medium or visa versa may areata a hot spark* fa one case a monomer pump seal failed* It was velved off and let run and a flash occurred* Possibly cutting off ths pump and then valving off might have prevented the spark, electric fittings were discussed along with motor control centers and use of remote air for pressurizing. The sealing coo* pound used to seel conduits was soluble in VC and VC was found in the conduits. Personnel training, smoking provision, etc. were discussed* Various organizations have films for training including Bp Mines* Ethyl Corp, and Caromody Trainer (?X OCC 2617 5 All aynthetic clothing shouldn't bo allowed. One company supplies cotton clothes to all employees. One nan was burned from flash caused by cloth static spark. Same fluids are so prone to buildup of potential that flows must be restricted despite all efforts to ground (e.g. kerosene). Fiberglass pipe or glass lined pipe cannot be adequately grounded. Vinyl acetate is difficult to ground totally - run dip leg to below surface to present potential fTon falling through the air0 The opinion was expressed that vinyl vapor escaping from a reactor leak won't create enough static to spark but the foaming brings to the leak and this will create the potential to spark. On snail leaks the vinyl nay be diluted before it reaches a sensing head and not trigger the high speed fan or deluge but near the leak there nay be explosive conditions. Seven plants use lightening rods or arrestors to prevent lightening In the Thompson disaster* there were three explosions* oca due to static generated in the dryer dust collector. During all the conotion the purification column was broken and vinyl spilled. After a period of tine when nothing happened four men re-entered the area presumably from curiosity and were killed when the vapors ignited. In other words* four of the five killed were unnecessary. In a Spanish plant an operator had overcharged a reactor* opened to relieve into the knockout vessel above* valved the so vessel off and the vinyl and catalyst reacted rapidly enough to blow cut a 1391b. section of the go tank and hurl it 600 feet away. 12. {falsa policies - G. D. Schaaf, Goodrich (The choice of this last discussion was by voice vote from the floor.) There is an article in the Plant Engineering magazine, November 12, 1970, issue on noise. The Welsh Healy Act was described and various deeibl (db) level significance pointed out. Throshhbld of pain is 140 db* a compressor (net defined) may be 90-100 db, a room with people may be 50 db. Ear protection is provided where a loss of 5 db is evident In a year's time. If more loss is sustained the men is removed from that area. Either ear plugs or muffs can be usod. END OCC 2618