Kimberly-Clark Everett Mill - Executive Summary |
|
REVISED SECTION 112(r) RISK MANAGEMENT PLAN KIMBERLY-CLARK TISSUE COMPANY EVERETT, WASHINGTON EXECUTIVE SUMMARY Introduction This revised Risk Management Plan is submitted by the Kimberly-Clark Tissue Company, Everett, Washington to comply with Section 112(r) of the Clean Air Act Amendments of 1990. This Executive Summary is provided to demonstrate Kimberly-Clark's continuing commitment to run an operation in Everett that is safe both to its employees and the surrounding community. Elimination of Elemental Chlorine from the Bleaching Operation and Startup of a New Chlorine Dioxide Bleach Plant This revised RMP is submitted because Kimberly-Clark is discontinuing use of elemental chlorine at the mill and is starting up a new chlorine dioxide bleaching system, which uses dilute aqueous chlorine dioxide solution. The chlorine dioxide is subject to EPA's RMP program. This re-submitted RMP describes the prevention program used for the new chlorine dioxide system, and adds infor mation on the modeling of hypothetical accidental releases of chlorine dioxide. Facility Description and Regulated Substances The Kimberly-Clark (K-C) Everett mill consists of a 500 air dried metric ton ammonium based sulfite pulp mill, a bleach plant, two pulp drying machines, and five paper machines. The pulp mill was built in the 1930s and was owned by the Soundview Pulp Company until the early 1950s, when it was purchased by Scott Paper Company. Scott Paper added paper machines and owned the operation until it merged with Kimberly-Clark Corporation in 1995. In the sulfite pulping operation, softwood and hardwood chips are cooked in batch digesters utilizing ammonium bisulfite cooking acid prepared on site. The cooking separates the pulp fibers from the binding material (lignin) which holds them together. To prepare the cooking acid, sulfur dioxide gas (SO2) is mixed with an aqueous solution of ammonia (NH3). The sulfur dioxide is primarily generated by burning elementa l sulfur, but is also reclaimed from the recovery boiler flue gas. Alternatively, it is obtained by the purchase of pressurized SO2 gas. This purchased "liquid SO2" is brought in by truck and stored in two 100-ton tanks located inside the pulp digester building. This quantity of SO2 is above the 5,000 pound threshold for the Risk Management Program; consequently, it is subject to the RMP regulations. The ammonia is brought to the site in anhydrous form via 80 ton rail cars. This quantity exceeds the 10,000 pound RMP threshold for anhydrous ammonia; consequently, it is subject to the RMP regulations. The ammonia is pumped from the rail cars and diluted with water prior to storage in a 150,000 gallon tank. The aqueous ammonia is stored at approximately 20% strength; this quantity stored exceeds the 20,000 pound RMP threshold for aqueous ammonia, so aqueous ammonia is also subject to the RMP regulations. The ammonia solution is employed on site not only to prepare the cooking ac id, but also for pollution control: it is used as a reactant in the sulfite recovery boiler to remove nitrogen oxides from air emissions, and it is used in the mill's secondary wastewater treatment plant to provide the nutrients needed for proper operation. It is also used to buffer concentrated spent sulfite liquor prior to storage. Pulp leaving the digesters is transported to three counter-current pressure washers; weak spent sulfite liquor from the washers is transported to multiple effect evaporators where it is concentrated. The concentrated sulfite liquor is then burned in the sulfite recovery boiler to generate steam. After washing, the sulfite pulp is bleached. The original bleaching system, described in the plan submitted to EPA in June 1999, utilized three bleaching stages: chlorination, caustic extraction with oxygen, and hypochlorite. The new bleaching system also has three stages: chlorine dioxide, caustic extraction with oxygen and peroxide, and chlorine dioxide. Unlike chlorine, which was transported to the site by barge and/or rail car and historically stored in a 720 ton tank, the chlorine dioxide solution will be manufactured at the mill via the Solvay process. In this process, sodium chlorate, sulfuric acid, and methanol are mixed in a primary reactor through which air is bubbled. The methanol reacts with the sodium chlorate in this acidic environment to produce gaseous chlorine dioxide (ClO2). The primary reactor underflows to a smaller secondary reactor where additional methanol and bubbled air are added to drive the reaction to its conclusion. The chlorine dioxide gas is collected in the vapor space of the reactor vessel and transferred to an absorption tower. In this tower chilled water is introduced to absorb the chlorine dioxide, forming a solution containing 8 to 14 grams per liter (approximately 1% ClO2). The chlorine dioxide solution is then transferred to storage tanks prior to use in the bleaching process. Two 50,000 gallon tanks of 1% chlorine dioxide solution have been constructed. Each tank will contain up to 5,000 pounds of chlorine dioxide. As this quantity exceeds the 1,000 pound RMP reporting threshold, chlorine dioxide is subject to the RMP rule. Timing The chlorine dioxide generator will start-up in late July 2000. From late July through September, the chlorine dioxide will be introduced into the first bleaching stage along with elemental chlorine to test out the generator and the effect of ClO2 on the bleaching process and pulp quality. Hypochlorite bleach liquor will still be used in the third bleaching stage. In early October 2000 the mill will be shut down for a ten day conversion period. After this shutdown, chlorine will no longer be used at the site, and bleaching will be as described above utilizing chlorine dioxide, caustic extraction, and chlorine dioxide bleaching stages. This RMP Plan is written to describe the process as it will be constituted in October 2000 and the reafter. At that time, the RMP chemicals on site will be sulfur dioxide, ammonia, and chlorine dioxide. The worst case release scenario will be a discharge of sulfur dioxide. During the short transition period (late July through September 2000) the new chemical (chlorine dioxide) will be on site, and so will the chemical being replaced (elemental chlorine). During this 2 1/2 month period, the worst case release scenario will still be a release of chlorine, as described in the RMP submitted in June 1999. Information on chlorine releases (worst case and alternative release scenarios) from the June 1999 submission are attached as an appendix to this revised RMP submittal. These appendices will no longer be relevant and can be discarded after October 2000. Kimberly-Clark's Safety Program We at Kimberly-Clark view safety as the most important aspect of our operations. Consequently, not only what we do, but also how we do it is of utmost importance; this world-class safety philoso phy extends to all facets of our operations. The mill's chemical safety record speaks for itself: over the past five years, there have been no deaths, injuries, or significant property damage on site, or known off site deaths, injuries, evacuations, sheltering in place, or environmental damage caused by accidental releases from RMP covered processes. The facility employs a full time Safety Manager and three environmental professionals. An environmental engineer monitors the Process Safety and Risk Management programs, ensuring the elements of these programs as listed below are properly executed. In addition, the mill maintains an emergency response (HazMat) team staffed with trained individuals; this team is available 24 hours a day. The HazMat team conducts quarterly drills to respond to simulated chemical accidents. A plant health and safety committee, composed of representatives from both the salaried and hourly work force, meets regularly to review all aspects of safety at the site. Finally, the company's insurance carrier, Factory Mutual, conducts annual audits which include evaluation of the safety of process equipment. For equipment and operations associated with its chlorine, sulfur dioxide, and anhydrous ammonia systems, the Kimberly-Clark Everett Mill is subject to OSHA's Process Safety Management (PSM) regulations and to the analogous program codified in EPA's RMP requirements. The new chlorine dioxide system is also covered by PSM and RMP requirements. While not covered by PSM, the aqueous ammonia system is also covered by the EPA program. Both rules require the following stringent activities which are practiced at the Everett site to minimize the potential for an accidental release: ? Review of the design of all equipment and controls to ensure proper lay out and installation. ? Updating of standard operating procedures to include specific information on safety procedures. All procedures must be reviewed and certified annually. ? Initial safety training and 3-year refresher training for all operators and maintenance staff. ? Procedures to ensure that all contractors receive the same safety training that Kimberly-Clark provides for its own employees. ? Regular inspection of all equipment, monitoring systems and controls, with stringent documentation of all inspections. ? Prompt corrective action for any non-conforming items identified by the regular inspections. ? Rigorous safety reviews conducted prior to system startup if any equipment or operations are modified. ? Stringent investigation of releases of any size and any "near-miss" incidents that might have led to a release. ? Periodic evaluation of the safety records of all outside contractors who work on the regulated systems. ? Development of an effective emergency response program. ? Implementation of an employee participation plan to ensure that all plant-wide staff are aware of these programs, and are actively consulted regarding safe ty issues. ? Independent audits of the entire PSM program and RMP program every three years. Safety is provided not only by the above procedures, but also by actual hardware installed on each chemical system: ? The liquid sulfur dioxide system is equipped with automatic shutoffs in case of an emergency. Any system overpressure would cause SO2 to vent back into the SO2 collection header. The header is vented back to the absorption tower, part of the acid making system. Any excess SO2 would react to form sulfite cooking acid. ? The anhydrous ammonia system is equipped with pressure relief valves, automatic shutoff devices, temperature regulators, video monitoring, and leak detection devices. Any overpressure would normally vent to the aqueous ammonia tank. ? The aqueous ammonia system is equipped with emergency shutoffs, pressure control, and leak detection devices. Overpressure normally vents from the tank via a stack equipped with a scrubber which converts any gas bac k into ammonia solution. ? The new chlorine dioxide reactor is a state-of-the-art installation equipped with fail-safe controls, which will shut down the system should an upset ever occur. Either high temperature or high pressure in the generator off-gas will activate the interlocking system. As a result: 1) The chemical feeds will stop. 2) The dilution air will be diverted from the bottom of the reactor to the gas phase. This will stop the stripping of chlorine dioxide, will purge the ClO2 and decomposition products from the reactor, and will cool the gas space and gas lines. 3) The generator cannot be restarted until the gas temperature falls below the critical set point. ? The mill's process control system is used to control the ClO2 process. To control the generation of chlorine dioxide and attain the desired yield, the mill ensures that: 1) The chemical and air flow controls are accurately calibrated and frequently checked. 2) The chemical feed streams are set and control led by the process control system at the recommended flows for the desired production rate. 3) The process streams are accurately analyzed at the appropriate frequency. 4) The operators are knowledgeable in the recommended operating parameters and adhere to standard operating procedures. 5) The raw materials are purchased to the recommended quality specifications and contamination of the raw material is prevented or removed prior to entering the generator. ? For the safe storage of ClO2 in the two 50,000 gallon tanks, the following design features have been incorporated: 1) High temperature of the chilled water will shut down the ClO2 generating system. 2) The two storage tanks are vented with sweep air back to the absorption tower. 3) Tanks are designed with a high aspect ratio (small diameter). 4) Tanks have adequate safety hatch relief areas, similar to the safety hatch on the reaction vessel. 5) Chlorine dioxide solution will be introduced into the tanks so that solution d oes not agitate the tank contents. 6) A floating roof is installed in each tank to reduce the amount of vapor space above the solution. Emergency Response Procedures Kimberly-Clark uses its Emergency Response Guidebook to provide step-by-step procedures for emergency response in the unlikely event of an accidental release. The key elements of the emergency preparedness program are as follows: ? All plant staff are trained in the specific elements of the program. ? A team of engineers, supervisors and operators are trained, certified and equipped for hazardous materials (HazMat) emergency operations to stop uncontrolled releases. ? The plant uses a combination of audible alarms and a plant loudspeaker system to alert the staff of a potential accident and to conduct in-plant communications. ? In the event of a large release the facility would immediately contact a telephone call list that includes the Everett Fire Department (911) and the Snohomish County Local Emergency Planning Committee (LEPC). K-C has provided the county "911" system with an Emergency Alert System (EAS) whereby citizens can be advised of a problem and given directions via "Emergency Broadcast System" interruption of radio and television transmissions. Steps To Improve Safety Kimberly-Clark has completed the following steps to reduce the likelihood of accidental releases, and to minimize the impacts of releases that could conceivably occur. ? Elimination of Elemental Chlorine Bleaching. Kimberly-Clark is installing a new chlorine dioxide bleach plant, replacing the elemental chlorine bleaching system. This will eliminate the import, storage, and use of compressed liquid chlorine at the mill. The dilute aqueous chlorine dioxide solution will be essentially used as it is generated, keeping the quantity stored low and greatly reducing risks to the surrounding community. ? Improved Traffic Barriers. Kimberly-Clark evaluated potential traffic accidents that could damag e piping containing aqueous ammonia solution, and installed new traffic barriers at some locations. After installation of the barriers, there is no realistic likelihood of a traffic accident causing damage to process piping at the ammonia system. ? Enclosure of Liquid Sulfur Dioxide Tanks. The two liquid SO2 tanks are located inside the pulp digester building. Historically the end of the building containing the tanks was open to the atmosphere. Large building doors were refurbished and are now kept shut thus enclosing the tanks, which reduces significantly the distance any release would travel. ? Community Cooperation. Based on the history of the plant, a significant accidental release is only a remote possibility. However, since such a release could impact industries, commercial areas, and residential areas, Kimberly-Clark is working with the Snohomish County Local Emergency Planning Committee to minimize any potential impact. As mentioned above, K-C worked with the Coun ty Department of Emergency Management (DEM) to upgrade their existing Emergency Alert System (EAS) so it now provides "911" service 24-hours a day, 7 days a week. This should greatly reduce community risk should a major chemical release or other catastrophe occur. Hypothetical Accidental Release Scenarios The Risk Management Plan must assess the downwind impacts of hypothetical uncontrolled accidental releases. EPA requires facilities to model the distance that a plume of released gas would travel before it disperses to an ambient concentration equal to the "Toxic Endpoint Concentration." The Toxic Endpoint Concentrations for various compounds are specified by EPA, and are generally concentrations that would cause no lasting physical harm but could interfere with the ability of people to leave the area. The Toxic Endpoint Concentrations for the RMP chemicals at the facility are: 200 parts per million (ppm) of ammonia, 1 ppm of chlorine dioxide, and 3 ppm of sulfur dioxide (th e elemental chlorine endpoint is also 3 ppm). Kimberly-Clark conducted safety reviews with plant operators, engineers, and safety managers to evaluate a wide range of hypothetical accidents that could cause releases of ammonia, chlorine dioxide, or sulfur dioxide. In accordance with EPA's rule, two general types of hypothetical accidental release scenarios were developed: ? The "Administrative Worst-Case Release" that arbitrarily assumes the entire contents of the largest container of chemical is released to the atmosphere in 10 minutes. Kimberly-Clark is unaware of any conceivable event that could actually cause such a catastrophic release at the Everett mill. ? "Alternate Release Scenarios," which are releases that the safety review teams concluded have a realistic (but small) chance of actually occurring at the mill. These hypothetical releases generally consist of flange leaks, temporary process upsets, and breakage to pipes or tanks. Worst-Case Release Scenario: Sulf ur Dioxide For the Worst-Case Release, the entire contents of a 100-ton tank containing pressurized liquid sulfur dioxide was assumed to be released as a gas in 10 minutes. Kimberly-Clark is unaware of any conceivable accident that could such a large release. The sulfur dioxide storage tanks are indoors. Equations from EPA's "RMP Guidance for Wastewater Treatment Plants" were used to estimate the release rate. Based on EPA guidance it was assumed that the building would partially contain the gaseous release, so of the 100 tons released inside the building only 34 tons would be lost to the outside air in the first 10 minutes. Lookup tables in EPA's RMP guidance documents were used to estimate the distance that the sulfur dioxide plume would travel. It is estimated that the plume would travel 8.8 miles before it dispersed to EPA's toxic endpoint concentration of 3 ppm. Alternate Release Scenario: Sulfur Dioxide Kimberly-Clark's safety review team selected the following hypoth etical accident as the Alternate Release Scenario: a large earthquake breaks a 420-foot long pipe connecting the SO2 storage tanks to the processing plant. The entire contents of the pipe (590 pounds of anhydrous liquid) spills onto the ground, forms a liquid pool that immediately cools to its boiling point temperature (-14 degrees F), and forms a puddle of SO2 "ice". It is estimated that 60 percent of the solid puddle would evaporate in 20 minutes to form a gas cloud that could blow downwind. The resulting gaseous SO2 release is 17.5 pounds per minute for the first 20 minutes. After that time the remaining SO2 would volatilize slowly. Of the 590 pounds of spilled liquid, only 350 pounds would be released as a gas in the first 20 minutes. The 20-minute release period used for this calculation corresponds to the averaging time for the ERPG-2 concentration that EPA used as the basis for the Toxic Endpoint. To calculate the impact of this hypothetical release, Kimberly-Clark reviewe d historical wind speed and direction patterns utilizing data from a weather station operated near the mill by the Puget Sound Air Pollution Control Agency (PSAPCA). The station's data indicates that the prevailing wind direction is toward the west, and the median wind speed is 1.8 meters/second. Although EPA suggests that facilities may use a relatively high wind speed of 3.0 meters/second for modeling the Alternate Release Scenario, Kimberly-Clark elected to use a reduced wind speed even though it results in downwind impacts that are more severe than if EPA's suggested wind speed was used. To account for the measured low wind speeds, Kimberly-Clark modeled the downwind impacts by averaging two values taken from EPA's "RMP-COMP" model: the value for the Worst-Case Release (1.5 meters/second wind speed and F stability), and the value for the Alternate Release (3.0 meters/second and D stability). The Alternate Release Scenario sulfur dioxide plume was modeled to travel 1,300 feet before it dispersed to the 3 ppm Toxic Endpoint Concentration. The prevailing wind direction at the Everett mill is toward the west, which would normally blow the hypothetical release out into Everett Harbor. However, it is possible that the wind could blow the plume toward populated areas. Such an event could affect some residential areas, commercial areas, and industrial facilities adjacent to the mill. Therefore, as described in the section entitled "Emergency Response Program," Kimberly-Clark is working with the LEPC to facilitate prompt notification of those nearby facilities in the unlikely event of an actual release. Alternate Release Scenario: Aqueous Chlorine Dioxide Solution Chilled, aqueous chlorine dioxide is stored in two above-ground tanks inside a secondary containment dike. Kimberly-Clark's safety review team selected the following hypothetical accident as the Alternate Release Scenario: 1,000 gallons of solution spills from a broken fitting into the secondar y containment. After the spill is stopped, the solution accumulated in the secondary containment is manually drained to the mill's closed sewer system and routed to the mill's wastewater treatment plant. EPA's WATER8 emission model was used to calculate the amount of chlorine dioxide that would volatilize as the solution spilled into the containment dike. WATER8 was also used to calculate the amount of chlorine dioxide that would volatilize from the sewer system and the wastewater treatment plant. Of the 84 pounds of chlorine dioxide in the original 1,000-gallon spill, an estimated 21 pounds would volatilize and disperse downwind in the first 10 minutes. EPA's SLAB dense-gas dispersion model was used to estimate the distance that the chlorine dioxide would blow downwind before it dispersed to EPA's toxic endpoint concentration of 1 ppm. The SLAB model was used because it models dense gas releases and to take advantage of the model's transient release capabilities. The EPA de faults for wind speed of 3 meters/second and D stability were used in the model. SLAB estimated a distance of 0.14 miles to the toxic endpoint, which barely extends to the nearest plant boundary. Alternate Release Scenario: Anhydrous Ammonia Kimberly-Clark's safety review team evaluated a wide range of hypothetical events, and selected the following scenario to represent releases of anhydrous ammonia. Anhydrous ammonia is imported to the facility in rail cars. A gasket on the unloading rack between the rail car and Kimberly-Clark's process line develops a small leak (assumed to be a 1/8-inch wide split). The leaking gasket is exposed to liquid ammonia at a pressure of 120 psig, which results in a release of ammonia droplets. An ammonia monitor at the unloading station would detect the leak and automatically alert the control room, which would dispatch the mill's HazMat team to the unloading area to repair the leak. The TANKLEAK emission module of the AIRTOX model was used to c alculate the ammonia emission rate. Based on an assumed 1/8-inch hole size and a system pressure of 120 psig the calculated release rate is 2.5 gallons per minute (13 pounds per minute). As described previously, PSAPCA's nearby weather station indicates that the historical median wind speed at the mill is 1.8 meters/second. To account for that low wind speed, the downwind ammonia impact caused by the hypothetical 13 lbs/minute release was calculated by averaging two values from EPA's "RMP*Comp" dispersion model: the value for the Worst-Case Release (1.5 meters/second wind speed and F stability), and the value for the Alternate Release (3.0 meters/second and D stability). The Alternate Release Scenario ammonia plume was modeled to travel 800 feet before it dispersed to the 200 ppm Toxic Endpoint Concentration. The prevailing wind direction at the Everett mill is toward the west, which would normally blow the hypothetical release out into Everett Harbor. However, it is possible th at the wind could blow the plume toward populated areas. Such an event could affect some residential areas, commercial areas, and industrial facilities adjacent to the mill. Therefore, as described in the section entitled "Emergency Response Program," Kimberly-Clark is working with the LEPC to facilitate prompt notification of those nearby facilities in the unlikely event of an actual release. Alternate Release Scenario: Aqueous Ammonia Solution Aqueous ammonia is manufactured onsite by blending anhydrous ammonia with water to a solution strength of 20%, cooling the solution in a heat exchanger, and storing the solution in a large storage tank. A gasket on the heat exchanger processing 20% ammonia solution develops a small leak (assumed to be 1/8 inch diameter). Equations from EPA's "RMP Offsite Consequence Analysis Guidance" were used to estimate the leakage flow rate. At the 25-30 psig working pressure, the 1/8-inch gasket leak would cause a leakage flow rate of 2 gallons per minute of aqueous solution (4 pounds per minute of ammonia). As a conservative step it was assumed that all of the ammonia in the spilled solution would volatilize immediately to form a gas cloud. Ammonia detectors in the area would detect the leak and alert the control room, which would shut off the feed pumps to the system and stop the leak. The total duration of the leak would be less than 15 minutes, after which the ammonia emissions would be limited to relatively small amounts of evaporation from the accumulated pool of spilled solution. As described previously, PSAPCA's nearby weather station indicates that the historical median wind speed at the mill is 1.8 meters/second. To account for that low wind speed, the downwind ammonia impact caused by the hypothetical 4 lbs/minute release was calculated by averaging two values from EPA's "RMP*Comp" dispersion model: the value for the Worst-Case Release (1.5 meters/second wind speed and F stability), and the value for the Alternat e Release (3.0 meters/second and D stability). EPA's "RMP*Comp" dispersion model was used to estimate the downwind distance the ammonia release could travel before it dispersed to EPA's toxic endpoint concentration of 200 ppm. Based on "Urban" surface roughness to account for tall buildings near the release site, the "RMP*Comp" model predicts that the ammonia cloud would travel approximately 500 feet before dispersing to the 200 ppm toxic endpoint concentration. The modeled distance to the toxic endpoint concentration is less than the distance to the facility boundary. APPENDIX As discussed above, chlorine dioxide will be introduced to the Everett mill in late July, but elemental chlorine will not be totally removed from the site until October. In the interim period, the current chlorine bleaching system will continue to operate. Thus for a 2 1/2 month period, the worst case release scenario for the site is represented by a release of chlorine, not sulfur dioxide. Data for t he worst case and alternative release scenarios for chlorine as reported in June 1999 are repeated below. This information will no longer be relevant after October 2000. Worst Case Release Scenario for Chlorine Anhydrous liquid chlorine (chlorine gas that is stored as a liquid under pressure at ambient temperature) is imported to the site by 90 tons rail car. The Administrative Worst-Case Release Scenario assumes that the entire 90 tons of chlorine in a car are emitted as a gas cloud in 10 minutes, during a period of exceptionally calm winds and stagnant atmospheric conditions (1.5 meter/second wind speed and "F stability") that would result in minimal dispersion of the gas cloud as it blew downwind. A speaker at a technical conference from a southern mill recently put the worst case release scenario into perspective. He calculated that a discharge of the entire contents of a 90 ton chlorine rail car in 10 minutes would require blasting a 3 foot diameter hole into the car. S ince the car is stationary when on the plant site, it is difficult to conceive what sort of force short of an asteroid collision could cause such a cavity. EPA's scenario contemplating full evaporation in 10 minutes is equally extraordinary. The thermodynamic properties of anhydrous chlorine indicate that such a large instantaneous gas release is not possible. If the 90 tons of liquid chlorine was somehow discharged from the tank or car it would spill onto the ground and immediately cool itself until it formed a puddle of ice, which would take much longer than 10 minutes to evaporate into a gas cloud. To evaporate the chlorine in 10 minutes would require a heat source equivalent to a small boiler, but no such heat source is present in the vicinity of the chlorine storage area. Finally, this imaginary heat source would create convection currents, increasing the height of the plume. But EPA's modeling directive assumes the gas cloud remains at ground level. Thus the worst case re lease scenario violates the laws of physics and provides little useful information. Nevertheless, the RMP rule dictates that the Worst-Case Scenario assumes the release of 90 tons of gaseous chlorine, so Kimberly-Clark used the most recent "RMP*Comp" computer model to estimate the downwind impacts for the 90 ton chlorine release. The model was set to use "Urban" surface roughness conditions to account for urban areas, rolling terrain, and coniferous forest in the Everett vicinity. RMP*Comp indicates that the gas cloud would travel 14 miles before it dispersed to the 3 ppm Toxic Endpoint Concentration. Alternate Release Scenario for Chlorine Kimberly-Clark's safety review team selected the following hypothetical accident as the Alternate Release Scenario: a large earthquake breaks a 65-foot long pipe connecting the rail car to the bleach plant. The scenario assumes that liquid chlorine flows from the broken pipe for 2 minutes before the bleach plant operator activates remot e emergency shutoffs, after which time the chlorine drains by gravity onto the ground. A total of 28 gallons of liquid chlorine would spill onto the ground, form a pool, immediately cool to its boiling point temperature (-29 degrees F), and form a puddle of solid chlorine "ice". It is estimated that 60 percent of the solid puddle would evaporate in 20 minutes to form a gas cloud that could blow downwind. Kimberly-Clark reviewed historical wind speed and direction patterns from a weather station operated near the mill by the Puget Sound Air Pollution Control Agency (PSAPCA). The station showed that the prevailing wind direction is toward the west (blowing from the mill away from populated areas), and the median wind speed is 1.8 meters/second. Although EPA suggests that facilities may use a relatively high wind speed of 3.0 meters/second for modeling the Alternate Release Scenario, Kimberly-Clark elected to use a reduced wind speed even though it results in downwind impacts tha t are more severe than if EPA's suggested wind speed was used. To account for the measured low wind speeds, Kimberly-Clark modeled the downwind impacts by averaging two values taken from EPA's "RMP-COMP" model: the value for the Worst-Case Release (1.5 meters/second wind speed and F stability), and the value for the Alternate Release (3.0 meters/second and D stability). The Alternate Release Scenario chlorine plume was modeled to travel 1,100 feet before it dispersed to the 3 ppm Toxic Endpoint Concentration. The prevailing wind direction at the Everett mill is toward the west, which would normally blow the hypothetical release out into Everett Harbor. However, it is possible that the wind could blow the plume toward populated areas. Such an event could affect some residential areas, commercial areas, and industrial facilities adjacent to the mill. Therefore, as described in the section entitled "Emergency Response Program," Kimberly-Clark is working with the LEPC to facilitate p rompt notification of those nearby facilities in the unlikely event of an actual release. |