HARVIN REACTION TECHNOLOGY, INC. - Executive Summary |
|
The Company: Harvin Reaction Technology is a small company that produces specialized polyols for use in rigid urethane foams. The site has been in operation under various owners since 1970 and polyols have been produced since 1976. Harvin has operated this business since September 1997. Covered Processes: Two parallel batch units are on site in which various starting materials are used in conjunction with catalyst and propylene oxide (PO) in a chemical reaction involving ring opening of the epoxide group from PO. The PO adds to active hydrogens on the starting materials yielding propylene ether groups with terminal hydroxyls. The chemical reaction evolves substantial heat which is controlled by deliberate maintenance of unreacted PO as a minor weight component in the reaction charge. Excess heat is removed by external circulation of cooling water. Nitrogen is used as a blanket in the storage of PO, in the reactors themselves, and as a blanket over products limiting the possibility of fire or explosion. Explosion proof equipment is used in the area around the reactors. PO is fed into the reactors slowly by displacing it by modest nitrogen pressure. This automatically limits the maximum reactor pressure against which a PO addition can be made. There are in addition both pressure and temperature interlocks to shut off flow when predetermined setpoints are reached. Sensors to detect any release of PO and sound emergency alarms inside the plant have been installed. Remote emergency stop switches can shut down all flow of PO. New Reactor R3: A new reactor has been added in 1999 which should be operational in July. This reactor features computerized controllers to achieve improved operational safety over the older process due to interlocks which cannot be overridden except under very limited circumstances by operations management. The features on this reactor will make it easier to operate and control. Old Reactor R1: The old reactor h as less extensive, non-computerized interlocks and controls. This reactor is scheduled to be taken out of service in August 1999 for extensive nondestructive testing to assure its full integrity. It will be recertified and refitted with upgraded sensors and controllers to enhance its ease of control and operational safety similar to the new R3 reactor by June 2000. Emergency Planning: An Emergency Plan has been developed that interlocks with the existing Emergency Plan for Greensboro/Guilford County (http://www.ci.greensboro.nc.us/ema). This plan is coordinated by Greensboro/Guilford County Emergency Management (the LEPC) and involves fire, police, health department, and if needed state and federal participation. Under this plan local industries provide notification of a problem using the 911 system; the first responder(Greensboro Fire Department) will establish a chain of command wherein the ranking officer becomes the Incident Commander; all notification of the public mus t be done through an Information Officer under the Incident Commander. Our responsibility is to promptly notify, maintain our personnel at a safe distance awaiting the respnders, and provide all possible information and assistance to the Incident Commander and all subsidiaries. A weatherproof "disaster box" has been acquired which will contain all pertinent information inluding MSDS's with medical treatment information, who-to-call list, plant diagram, etc. at a safe place away from the area where PO is handled. To facilitate any emergency response, we have had walk-through visits to familiarize the firemen at the closest station and especially their hazmat team. The Greensboro/Guilford County Plan specifically enjoins local industries from direct notification of TV and radio stations to prevent uncontrolled panic and requires them to coordinate with local authorities. Most Likely Scenerio: In preparing the worst case and alternate case scenerios we followed EPA guidelines an d treated PO as a toxic material. We believe that any significant release would more likely result in a fire and/or explosion than in a toxic release because PO is very flammable. Using RMP*Comp for the most similar flammable that we could identify (Cyclopropane CAS# 75-19-4 which IS NOT PRESENT AT THIS SITE) we calculate a distance to the overpressure endpoint for total loss of all PO on site which is 0.081 as large as our offical toxic worst case and which is even smaller than the official toxic alternate case scenerio. The portion of the facility where PO is handled therefor has been fitted with appropriate explosion proof equipment. |