Document gEJxd6QjadL8xo6k2vd68ogaJ

Ms. Rebecca W. Hanmer Regional Administrator Region IV United States Environmental Protection Agency 345 Court!and Street, N.E. Atlanta, GA 30308 Dear Ms. Hamner: A leak through a double rupture disc assembly on one of the polymerization reactors was detected at approximately 12:45 p.m. on April 1, 1980. Pursuant to the National Emission Standard for Vinyl Chloride (40 CFR, part 61, subpart F, paragraph 61.65a) this incident is being reported to your office. TIY& vessel involved was polymerization reactor D-500. The leak occurred thru a crack in the top rupture disc of a double rupture disc assembly on the reactor. The rupture discs in the assembly were rated at 200 psig at 200F. The reactor design pressure is 200 psig at 200F. Reactor batch D-500 - 4684 was charged at 11:05 a.m. on April 1, 1980. The charge was normal and polymerization of the batch was initiated. At 12:45 p.m. a vinyl area supervisor thought he heard an air leak. Upon investigation he determined the noise was coming from a leak through the double rupture disc assembly. The supervisor notified the control room of the situation and instructed the panel operator to kill the reactor. The reactor was killed at 12:50 p.m. The reactor temperature was 136F at a pressure of 123 psig at the time the reaction was killed. Immediately recovery was started on the reactor and full cooling was applied to the reactor. The pressure and temperature of the reactor began dropping rapidly as the recovery of the reactor was started. At approximately 1:40 p.m. reactor D-300 became empty. The reactor was evacuated and D-500 reactor equalized with it at 1:50 p.m. The pressure in the reactor continued to drop as the recovery operation proceeded. At 5:10 p.m. the reactor pressure was reduced to less than atmospheric and the leak thru the rupture disc stopped. During the last half-hour of the recovery, steam injected into the reactor to strip the vinyl chloride from the resin increased the temperature of the reactor to the point where the vapor pressure of water in the reactor became sign ificant. Thus the vinyl chloride concentration of the leak decreased during this period. A copy of the reactor pressure and temperature strip chart is attached. The vent stack on the double disc assembly is 50 feet above grade. DTH 000115607 Ms. Rebecca W. Hanmer April 9, 1980 Page 2 i No personnel were exposed to vinyl chloride from the leak. The weather conditions from the plant weather station during the leak were as follows: Barometric Pressure - 29.88 in. Hg. Temperature - 70 F Wind Direction - variable from south and southwest Wind Speed - variable 0 to 5 miles per hour The amount of vinyl chloride from the leak is estimated to be 283 pounds. This is based on orifice calculations using the wetted perimeter of the crack and calculating the flow rate of the leak using the average pressure and temperature of the reactor over ten minute intervals. A copy of the calculation is attached. New improved rupture disc holders and discs were ordered for the reactors in December 1979. This type rupture disc holder had previously been purchased and installed on D-300 reactor. Since this holder did have improvements over those in use at the time, they were ordered for all the other reactors. These new holders and discs were received in the plant in mid March of this year. Two new type holders and discs were installed on D-500 reactor before the next batch was charged. These new holders are the improved type that the disc is put in the holders and pretorqued in the shop. The holders with the disc in place is then taken to the field and installed on the reactor. Pretorquing in the shop minimizes the chance of a premature disc failure due to improper torquing on the disc. This improved type holder and disc is being installed on all the polymerization reactors. In addition a feasibility study will be conducted to determine if the double rupture disc assembly can be replaced with a single disc and a relief valve. The reporting requirements of the standard are summarized below: Source: Reactor D-500 Nature: Leak through crack in top disc of double rupture disc assembly. Cause: Premature failure of rupture disc. Date: April 1, 1980 Time: 12:45 p.m. DTH 000115608 Ms. Rebecca Hanmer April 9, 1980 Page 3 Approximate Discharge: 283 pounds of vinyl chloride Method of Determining Discharge: Orifice calculation Actions Taken To Prevent Discharge: 1. Reaction Killed 2. Batch recovered 3. Batch equalized with another reactor Measures Taken To Prevent Future Di scharges: 1. New type rupture disc holders and rupture discs installed in the reactor. 2. A feasibility study will be conducted to determine if the double rupture disc assembly can be replaced with a single disc and a relief valve. If there are any questions on this matter, please contact me at (601) 369-8111, Ext. 239. Sincerely, R. A. Frohreich Chief Process Engineer cjt Attachments c: Wayne Anderson - Mississippi Bureau of Pollution Control be: CLM, CRM, JEB, MPB, SJV, JJH DTH 000115609