Colorite Polymers - Executive Summary |
|
Introduction Colorite Polymers, a division of Tekni-Plex, Inc., owns and operates a polyvinyl chloride manufacturing plant in Burlington, NJ. This facility is subject to U.S. EPA regulations governing Accidental Release Prevention (ARP) Requirements: Risk Management Programs under Section 112(r) of the Clean Air Act (40 CFR Part 68). This Risk Management Plan (RMP) has been developed in accordance with the requirements specified under 40 CFR Part 68, Subpart G. The RMP report certifies that Colorite polymers has instituted a Risk Management Program at the Burlington Facility that is in compliance with U.S. EPA ARP requirements. The RMP includes an Executive Summary and Data Elements following the format published by U.S. EPA. In addition to identifying the applicable corporate policies and risk management systems, this RMP identifies a set of worst case and alternative release scenarios, the potential off-site consequences of those releases, and the facility's five-year accidenta l release history. (This facility has had no accidental release events that would require reporting under these new federal rules). This RMP also certifies that prevention and emergency response programs are in place so as to minimize risks to both workers and the potentially affected public. Accident Release Prevention Program and Emergency Response Policy It is the policy of Colorite Polymers management to implement the requirements of this Risk Management Program (RMP) in accordance with the USEPA regulations under 40 CFR Part 68. The objective is to minimize the risk of a release of a hazardous material; but if a release occurs, to minimize the potential impact to facility employees, the public and the environment. This objective is being accomplished by utilizing good operating practices and procedures, providing appropriate training to all employees, and coordinating response activities, as necessary, with the local emergency response providers. Colorite's management is co mmitted to providing the resources necessary to implement this policy. Further, the Colorite Burlington Facility recognizes that it has a duty to protect its employees and neighbors from the effects of all dangerous chemicals. Colorite Polymers has developed a management system to implement and maintain compliance with the ARP Program, which is coordinated with its already-established OSHA Process Safety Management and plant emergency response programs. At the Colorite's Burlington Facility, the Plant General Manager has primary responsibility for ensuring overall implementation and compliance with these programs. Other individual employees such as the Technical Director, Safety and Training Manager are responsible for the technical implementation of individual sections of these programs. Facility Description The Burlington South Plant is a polyvinyl chloride (PVC) manufacturing facility in the Specialty Vinyls Resin Division of Techni-Plex. The facility began operation in 196 2 as Cary Chemical. Tenneco acquired it in 1966 and by OxyChem in 1986. In 1995 Colorite Polymers (Pure Tec Inc., parent company), bought the facility. In 1998 Tekni-Plex took ownership through a merger. The plant is located at 35 Beverly Road in the township and county of Burlington. Of the total plant property of 139 acres, approximately 40 acres are used for industrial activities and are fenced. The Delaware River is north of the facility. Public Service Electric property is north and east of the facility. Residential area is to the east, south and west, with the closest residents about 500 feet north and 1,000 feet south. The distance to the closest industrial neighbor property is approximately 400 feet to the west of the main VAM storage tank. There are 125 site personnel at Colorite. Production operations are conducted on a seven-day week, 24-hour day basis. Plant security is provided by a 24-hour contract guard service. The site has supporting offices, laboratory, m aintenance, warehouse and utilities. PVC resins are manufactured by two interconnected process systems. They both rely on the same permanent connections to the storage facilities that contain the raw materials for both of the processes. Therefore for the purposes of the RMP program, Colorite considers this storage area and each of the main manufacturing systems as subparts of the same overall PVC manufacturing process. Operations subject to the EPA RMP Rule include the storage of raw materials, weighing tanks, process reactors, batch mixing tanks, VCM stripping and recovery from PVC. Manufacturing PVC involves both the dispersion process (rated capacity 50 mm #/yr.) and the suspension process (rated capacity 70 mm #/yr.). The dispersion process uses licensed Wacker technology in fourteen 3,000 to 3,800 gallon column reactors. The suspension process uses eight 5,000-gallon tank reactors. The primary raw material, vinyl chloride monomer (VCM), is received in railcars and stored i n four 80,000-gallon 'bullet' tanks. The co-polymerizing material is exclusively vinyl acetate monomer (VAM), which is received by truck and stored in one 30,000-gallon vertical fixed-roof tank. Both raw materials contain traces of inhibitor to prevent premature polymerization. Product resin is shipped by bulk in railcars and trucks and in bags. Program Level Identification The EPA Risk Management Regulation identifies three levels of requirements defined as "programs". EPA recognizes that some regulated processes would not pose an off-site hazard in the event of an accidental release. Such processes are classified as Program 1. Program 1 is applicable to any process for which it can be demonstrated that the public would not be adversely affected by a release. Thus, to qualify for Program 1, a facility with a regulated toxic chemical needs to meet two criteria. 7 no release over the most recent 5-year period has resulted in off-site injury or environmental damage; and 7 Disp ersion modeling demonstrates that a worst-case release (as defined by the regulation) will not result in concentrations at public receptors that exceed the toxic or thermal or overpressure endpoints specified by the U.S. EPA. The Colorite Plant has not had an accident or release involving either of its regulated substances, vinyl chloride monomer, or vinyl acetate monomer, over the last 5 years. However, modeling calculations of the possible consequences of the "worst-case release", as defined by the regulation, predict the possibility of thermal effects of VCM or airborne concentrations of VAM at public receptors that could exceed the toxic endpoint identified in the rule. Therefore, neither of these two materials is eligible for Program 1. Program 2 and 3 are applicable to processes that do not qualify for Program 1. Program 2 applies to any process that is ineligible for Program 1 and is not subject to Program 3. Program 3 applies to all processes, such as those for VCM at Colorit e, that are subject to the Occupational Safety and Health Administration (OSHA) Process Safety Management Standard (PSM). Much of the basic foundation of the Colorite Accident Prevention Program is adopted directly from the compliance program for the PSM standard for VCM. These OSHA requirements are directly consistent with the Program Level 3 RMP Program requirements under 40 CFR Part 68. For VCM, therefore, Program 3-related information is included in the Data Elements section of this Risk Management Plan. VAM is not directly subject to the OSHA PSM Rule, and therefore was not previously the specific subject of the existing Colorite Accident Release Prevention program. This material is, however, utilized in many of the same processes and most of the same equipment systems as VCM, because Colorite also makes a number of copolymer products which require a blend of both of these basic chemicals in the formulation. Therefore, VAM is classified as qualifying at the Program 2 level und er 40 CFR part 68, and the Data Elements section of the Risk Management plan related to VAM are addressed in a manner that is consistent with this classification. Worst-Case and Alternative-Release Scenarios The RMP regulations require that each facility identify worst-case and alternative case release scenarios. For the purposes of developing and maintaining adequate Risk Management Plans, the EPA has defined in its governing rules and guidance a series of modeling methods and assumptions, which are to be used as administrative guides for planning purposes. In order to standardize and simplify the many factors that can potentially occur in an accidental release situation, some of these assumptions may not take into account the available preventive measures or mitigation methods that could diminish or even eliminate the implied risks that are suggested by "worst-case" analyses. For that reason, both the results for the standardized "worse-cases" defined by the EPA methods and a set of alternative cases. The latter set of alternative cases are believed by the technical staff and management of the Colorite Burlington Plant to more realistically represent situations that may possibly, but rarely, occur within the lifetime of their facility. EPA has defined a worst-case release as the instantaneous release of the entire contents of the largest vessel or connected piping that contains a regulated substance. For either flammable liquids or vapors, such as VCM, or potentially toxic gases or vapors, such as VAM stored at ambient temperatures, it is assumed that the release occurs over a 10-minute period. The release event is then evaluated using modeling techniques and/or reference tables to define the distance to a specified endpoint (concentration of toxic vapors, or overpressure or thermal load due to rapid combustion). The distance to the endpoint is affected by several factors including molecular weight, volatility, heat of combustion, and physical setting (ur ban or rural). For potentially toxic chemicals, EPA has established "toxic endpoints" based on the American Industrial Hygiene Association (ACGIH) Emergency Response Planning Guideline, Level 2 (ERPG-2), which protects individuals from health-threatening or escape-impairing injury. The ERPG-2 is defined as: "The maximum airborne concentration below which it is believed that nearly all individuals could be exposed for up to 1 hr without experiencing or developing irreversible or other serious health effects or symptoms which could impair an individual's ability to take protective action." Given that the worst-case is a 10-minute release duration with a 10-minute concentration averaging time, the actual hazard zone is typically overestimated by selecting the ERPG-2 value as the criterion because it id defined as applicable to a one-hour period of exposure. Within such a time period there are a number of emergency measures that can reduce or eliminate exposures to levels this high for an interval much shorter than an hour. For flammables, like VCM, the endpoint is based upon the EPA-required assumption that a vapor cloud forms instantly from the released liquid and burns with explosive force producing a shock wave. The endpoint of interest is the distance at which the overpressure produced is equal to 1psi above ambient pressure, or the thermal effects of an associated fire producing a thermal load at a nearby receptor of 5 kw/m2. The first of these effects would be sufficient to produce structural damage that could lead to injuries or fatalities; the second could produce significant burns. Worst-Case Releases The SLAB (Dense Gas Dispersion) model was used to simulate the worst-case release scenario. The SLAB model is a publicly available model that simulates contaminant movement for heavier-than-air gases such as concentrated VCM vapor or VAM vapor for either instantaneous or continuous near-ground-level releases. For VAM, which is subject to the RMP rule due to its moderate toxicity, rather than its flammability, the worst case is a 30,000-gal spill into the 40 x 40-ft. dike surrounding the main storage tank. In this case, due to its higher boiling point temperature, the EPA specifications for the "worst case" analysis method allows appropriate corrections to be made for the beneficial effects of the dike surrounding the storage tanking. The EPA's version of the SLAB dispersion model was utilized to evaluate the potential consequence area for this worst case scenario. SLAB is a publicly available model which readily simulates the initial heavier-than-air vapor formed by the evaporation of VAM from the diked pool. In the immediate vicinity of the Colorite Storage tanks and the polymer manufacturing buildings, the aerodynamic surface roughness is close to 'urban' in nature, due to the many large buildings, tanks and other structures at the site. At greater distances, the surface roughness is a mixture of both urban and rural charact eristics. Therefore, for the required worst-case scenario analysis by the SLAB model, a 40-cm surface roughness parameter value was utilized. The modeling results demonstrate that the presence of the dike around the storage tank limits the distance to the endpoint concentration, 75ppm (260 mg/m3), to a maximum of 0.37 mi. The overall worst-case scenario for the Colorite Burlington Plant involves VCM, which is included in the RMP program as a Program 3 regulated substance due to its high flammability characteristics. The EPA-specified "worst case" consists of a 10-minute gaseous ground-level release of 80,000 gal of stored VCM, without regard for any benefits of storage container diking. For flammable materials that are gases at normal temperatures and pressures, the EPA "worst case" does not assume any credit for the presence of the dike. The vapor is formed immediately and assumed to disperse under atmospheric conditions of F stability and 1.5 m/sec wind speed, characteristic of nighttime releases. (A gaseous release of this magnitude over such a short time-span is physically impossible because the amount of heat energy required to so rapidly vaporize the liquid VCM would not be available. Even if the storage tank were to breach, some liquid VCM would spill to the ground in a boiling pool. Furthermore, if a hole were to form in the vapor space of the storage tank, the flashing vapor in the tank would cause the tank to auto-refrigerate. This would gradually reduce the release rate with time). The selection of the worst-case release for VCM should, therefore, be viewed as a regulatory necessity, rather than a realistic representation of a worst-case release event. For this EPA-defined "worst case" release, consequence modeling was performed with the AIRTOX Flammable Jet Model. The AIRTOX models are a proprietary set of hazards assessment models developed by ENSR and officially recognized as appropriate for RMP use in the U.S. EPA's Offsite Consequence A ssessment Guidance (OCAG), as well as in listings for similar applications in the states of New Jersey, California, and Texas. Modeling results indicate that the maximum distance to the EPA-specified endpoint for a vapor cloud explosion, which could occur from such a release scenario, is approximately 0.51 miles. Alternate Releases Alternative releases are intended to represent release scenarios that have a greater likelihood of occurrence than a worst-case release. Alternative releases do not necessarily represent the types of releases that the PSM hazards analysis and/or accident history indicate would be most frequent, but rather a release that is somewhat more likely than the worst-case release and that generally still has the potential to affect off-site receptors. In accordance with the EPA's OCAG, a single alternative release scenario is reported for each regulated substance. Through its conduct of Process Hazards analyses for VCM and VAM, the Colorite Plant performed a t horough review of the facility utilizing engineering plans, operational experience, and maintenance records, and the results of the OSHA PSM review in order to determine potential alternative release scenarios that would result in the greatest toxic endpoint distance. Each scenario was evaluated and the scenarios most likely to occur and to result in off-site impacts were selected for modeling. In order to determine the most conservative off-site consequence estimates, the Colorite Plant has chosen to include in the Risk Management Plan the alternative release scenarios that resulted in the greatest toxic endpoint distance. In accordance with the RMP rule, alternative releases are modeled under typical (rather than worst-case) dispersion conditions. The EPA OCAG default dispersion conditions are neutral atmosphere, with dispersion neither enhanced nor limited (D stability and 3 m/sec wind speed). Unlike the worst-case release (for which an instantaneous spill or 10-minute ground-lev el gas release is assumed), alternative scenarios can account for the actual release configuration, and account for both active and passive mitigation. For the VAM alternative case, review of the PHA results led to the selection of a leak of 420 lb/min from a one-inch drain valve into an area with no diking. Although preventive maintenance and routine inspection procedures make it unlikely that this event will ever occur, it is possible. It is assumed that it takes at least ten minutes to discover and stop this leak condition. In that time the liquid is assumed to spread out into a 1-cm deep pool and rapidly evaporate. This spill of VAM is assumed to be onto an undiked area, somewhat separated from the storage tanks, it was conservatively assumed that the initial roughness would be nearly equivalent to "rural". Therefore, a rural set of conditions assuming little roughness was utilized with the EPA SCREEN dispersion model. which treated the spilled pool as a thin volume source. The evaporation rate was computed from the PAVE model developed by the American Petroleum Institute (API) in both this case and the worst case for VAM. If the incident continued for more than ten minutes there would be little effect on the distance predicted for potential exposures to the EPA-specified toxic endpoint concentration, due to the short time of transport to that distance compared with the spill duration assumed. In this alternative case scenario, the vapors are predicted to reach the endpoint concentration of 75 ppm for VAM at a distance of 0.26 mi. This distance does extend beyond the Colorite property boundary. The closest neighbor to the Colorite facility is, however, another industrial facility and both a warning system at that facility and an additional public notification plan for neighbors has been established Although the hypothetical worst case scenario for VCM had the largest potential area for affecting the plant's neighbors, a careful review of realisti c alternative cases could only identify one other scenario which might produce any of the specified effects beyond the plant boundary. That alternative release scenario for VCM was a separation in the 3-inch diameter portion of the line transferring VCM from one of the rail cars to the main storage tank, above an undiked area. It is conservatively assumed that the greatest release rate would be determined by the vapor pressure of a nearly full system, although this is about 50% greater that the normal average rate of transfer in this line. Normally it takes about 5 hours to transfer some 22,000 gal of VCM out of each rail car. Normally Colorite Plant staff would be able to respond to such an event and stop the flow prior to emptying the tank, because the facility maintains an external sensor and alarm system that would alert facility personnel to the tank leak. In addition a water spray mitigation system is present and would be expected to reduce the possibility of ignition and th e potential for any impact. However, it was conservatively assumed that up to 17,800 lb of a 2-phase liquid/gas mixture would spray out in a jet aimed toward the nearest property boundary. This release is further assumed to behave as a potentially flammable jet, which completely vaporizes, finds a source of ignition and produces a thermal impact, as well as a shock wave. The thermal endpoint of interest, 5 kw/m2 would occur at a greater distance than the EPA-specified benchmark of 1 psi overpressure for the associated shock wave. However, the maximum distance to the thermal effect endpoint from this unlikely, but possible, event is only about 230 ft. from the initiating event. Depending on the exact source location, this maximum distance for any significant effect would likely be limited to onsite areas. If the source event were located at the end of the plant property closest to the western boundary, it is possible that this effect distance would extend onto the neighboring in dustrial property. It is for that reason that there already exists a flammable vapors monitoring system and alarm system which would provide early warning to those neighbors. Accidental Release Prevention Program The Colorite facility is governed by a set of OSHA and USEPA regulations that require planning and facility activities intended to prevent a release of hazardous material, or if a release inadvertently occurs, to minimize the consequences of a release to the employees of the facility, the public and to the environment. These regulations include: 7 40 CFR Part 68, Accidental Release Prevention 7 40 CFR Part 112, Spill Prevention, Control and Countermeasure 7 40 CFR Part 264, Hazardous Waste Contingency Plan 7 29 CFR Part 119, Process Safety Management The VCM portion of the storage suspension and dispersion processes are subject to the Risk Management Regulation as well as the Occupational Safety and Health Administration (OSHA) Process Safety Management (PSM) Standard. T he fundamental Colorite Accident Prevention Program elements are adopted directly from the compliance program for the PSM standard, consistent with the Program Level 3 RMP Program requirements under 40 CFR Part 68. The RMP Program 3 Prevention Program elements are: 7 Employee Participation 7 Process Safety Information 7 Process Hazard Analysis (PHA) 7 Standard Operating Procedures 7 Training 7 Contractors 7 Management of Change (MOC) 7 Pre-Startup Safety Reviews (PSSR) 7 Mechanical Integrity 7 Safe Work Practices 7 Incident Investigation 7 Compliance Audits Five Year Accident History The Colorite Plant has not had an incident in the past five years involving either VCM or VAM systems that has caused any of the following: 7 On-site deaths, injuries, significant property damage, or 7 Offsite deaths, injuries, property damage, environmental damage, evacuations or sheltering in place. Emergency Response Program Colorite has personnel trained in emergency response at the facility 24 h ours per day, seven days per week. These personnel receive annual training on emergency procedures and response techniques. Response activities are coordinated with the local fire department to ensure the appropriate level of response. For both worst cases and for the more realistic alternative cases, the included Data Elements section identifies the potential size of the population and the types of receptor areas which could be affected by each of these events. Population estimates within the predicted zones are based on the 1990 Federal Census, and estimated by applying EPA's Landview III and MARPLOT programs. Types of public receptors are identified using comprehensive maps of the local area (DeLorme Street Atlas USA, Version 6.) and the regional maps published by the U.S. Geological Service. For areas adjacent to industrial facilities having large amounts of undeveloped open space surrounding them, such as Colorite, these population estimates are often somewhat exaggerated. They also represent the total circular area, which might be affected under any set of wind conditions, rather than the much smaller area, which represents the area that a single release may affect for a particular set of wind conditions. However, they do indicate the size of the area, which should be made aware of local, and plant emergency plans, which have been developed for their protection. Colorite has developed, implemented and coordinated its emergency plan with local safety officials to ensure such notification and protection. Planned Changes for Continued Improvement of Safety Colorite Polymers, under the recent RMP program, as well as its existing PSM and earlier SARA Title III Community Right-to-Know Act compliance programs has organized its management system to effectively address all hazards and potential risks. Both the advanced planning aspects of process design, operating procedures, and emergency preparedness, and the operational elements of system maintenance, s afe operating practices and ongoing personnel training are necessary to support a continual improvement in facility safety. These programs are thoroughly documented so that information about the safe handling of chemicals present at the facility is available to all employees, and can be readily interpreted by emergency response team staff and the Incident Commander. This is especially important when questions arise from public safety officials regarding potential risks to the community. Colorite's safety program incorporates continuous improvement through use of audits, inspections, and on-going evaluations of in-plant safety and program effectiveness. The information gathered is evaluated and improvements that are identified are incorporated into the program. All of these features of the RMP and the PSM program at the Colorite Plant lead to operations that are safe today, but will be even safer tomorrow. |