Loading...
HomeMy WebLinkAboutRISK MANAGEMENTI I I I I I I I I I i I I i I I I I I Gist-brocades Food Ingredients Inc. RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY October 18, 1989 LUFT ENVIRONMENTAL CONSULTING 3701 Pegasus Drive, Suite 121 · Bakersfield, California 93308 · (805) 399-5838 October 20, 1989 Mr. Ralph Huey Hazardous Material Coordinator Bakersfield City Fire Department 2130 "G" Street Bakersfield, CA 93301 Mr. Huey: On behalf of Gist-brocades Food Ingredients Inc., Luft Environmental Consulting is submitting the enclosed Hazard and Operability Study and Offsite Consequence Analysis for your review. We would like to retrieve these documents from you upon completion of your review. If you have any questions or comments concerning the documents, please contact me at your convenience. Karl W. Luff Principal Mechanical Engineer cc: Bob Deedy Gist-brocades I I I I I I I I I I I I Gist-brocades Food Ingredients Inc. RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY Prepared By I Bakersfield, CA 93308 I I I I I Luft Environmental Consulting 3701 Pegasus Drive, Suite 121 October 18, 1989 DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY PREFACE Pursuant to Article 2, Chapter 6.95, Division 20, California Health and Safety Code (H & S Code), an Risk Management and Prevention Program must be prepared for an existing facility which handles listed acutely hazardous materials, if requested by the Administering Agency. Gist-brocades Food Ingredients Inc. utilizes chlorine as a biocide in their cooling towers. The storage capacity of the chlorination system is greater than the levels specified in Section 25536 (a), H & S Code. A Risk Management and Prevention Program (RMPP) Supporting Document has been prepared for Gist-brocades Food Ingredients Inc., based on a request letter from the Administering Agency (Bakersfield City Fire Department) in October, 1988. As a requirement of the RMPP, a Hazard and Operability Study (HazOp) was conducted. The HazOp study is the subject of this document. This package is provided for review by the Bakersfield City Fire Department only. A summary of the HazOp is provided in the RMPP Supporting Document. Gist-brocades Food Ingredients Inc./RMPPIHazOp/Oct., 1989/Page 1 Luft Environmental Consulting I I I I I I I I I I I I I I I I I I DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY The RMPP must, according to Section 25534 (d)(1), include the "results of a hazard and operability study which identifies the hazards associated with the handling of an acutely hazardous material due to operating error, equipment failure, and external events, which may present an acutely hazardous materials accident risk". A. HAZOP TECHNIQUE The HazOp technique that was used for the Gist-brocades Food Ingredients Inc. facility was a modified "guide word" approach for a Hazard and Operability Study. The basic guide word HazOp was chosen since it allows a systematic and thorough review of every part of the facility that handles AHMs. In conjunction with the guide word approach, a "what if" analysis Was incorporated into the review. Both of these approaches are described in the AlChE Guidelines for Hazard Evaluation Procedures1, which is referenced in Section 25534 (I), Chapter 6.95, Division 20, California Health and Safety Code. Other publications further describe the guide word approach, including A Guide to Hazard And Operability Studies2. Procedures for conducting the HazOp study were adapted to accommodate the simplicity of the vacuum operated chlorination system. A series of two meetings were held at the plant site. The first meeting focused on the design of the system and the equipment utilized on site. A second meeting was conducted to review the literature search on the equipment, operating practices, and safety procedures pertaining to the chlorination system. The Chlorine Institute, Inc. (Institute) provided a groat deal of literature on the handling, storing and use of chlorine gas. The Chlorine Institute, Inc., is a trade association of chlorine manufacturers, packagers, distributors and related companies. One of the objectives of the Institute is to promote occupational and environmental safety in the manufacture, transport and use of chlorine. The 1 Guidelines for Hazard Evaluation Procedures, American Institute of Chemical Engineers, New York, 1985. 2A Guide to Hazard and Operability Studies, Chemical Industry Safety and Health Council of the Chemical Industries Association, London, 1985. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 2 Luft Environmental Consulting I I I I I I I I I I I I I I I I I I Institute is a technical information center with a focus on issues involving safe design of chlorine containers, transportation of chlorine, employee health and safety and control of chlorine emergencies. Information concerning the control of emergencies included emergency response to a release and containment of the release. Additionally, some case histories of releases were supplied by the Institute. This information was used by the participants of the HazOp to help in the evaluation of Gist-brocades Food Ingredients Inc.'s chlorination system. B. PROCESS AND EQUIPMENT DESCRIPTION The chlorination system was installed in 1985 and is used as a biocide in the cooling tower system at the plant. Chlorine gas is stored on site in one ton containers. The chlorine site has two sets of trunnions designed to handle two containers. Each container can be rotated into the appropriate operating position on the trunnions. Typical operation of the system is to empty one container completely before connecting to the second container. The empty container is then shipped to be refilled at a chlorine distribution plant. When the full container returns to the plant, it is off loaded on to the vacant set of trunnions and stored until it is placed into service. Chlorine is drawn from the container through the vacuum feed regulator by a vacuum created by ejectors. The ejectors are located in a discharge branch of each of the three cooling tower water circulation pumps. Chlorine flow is controlled by two devices, an automatic chlorine flow control valve and individual rotometers. The flow through the automatic chlorine flow control valve is determined by an electrical signal from the residual chlorine analyzer. This flow is further proportioned by the rotometers associated with the individual ejectors. A schematic overview of the system is shown in Figure 1. Ton Container Ton containers are welded steel tanks with a chlorine capacity of 2,000 pounds Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 3 Luft Environmental Consulting I I I Ton Container Automatic ~ Chl°rine F~al°W R°t°m I Control Ive ~ I Vacuum Feed Residual Regulator 4-20 mA , J Chlorine I Signa: =:,,~ ~' Analyzer Typical to other Ejectors To Cooling Tower Basin To Spargers in the I I I I ! I I Ejector Cooling Tower Turbine Pump To Cooling Tower Circulation System PIPING KEY Thin Lines = Vacuum Lines Bold Lines = Pressure Lines Dashed Lines = Electronic Signal Figure I Schematic diagram of the chlorination system. (one short ton). Full containers weigh approximately 3,700 pounds. The heads of the container are convex inward and forge welded to the shell. Shell ends are crimped inward to form chimes which form a substantial gripping surface for the lifting beams that are used to load/unload the containers. These containers are constructed in accordance with the U.S. Department of Transportation's (DOT) requirements pursuant to 49 Code of Federal Regulations (CFR). I I Each container is equipped with two identical valves located near the center of one end of the vessel. These standard one ton container valves are connected to an internal eduction pipe, as shown in Figure 2. When the valves are aligned Gist-brocades Food Ingredients I nc./RMPP/HazOp/Oct., 1989/Page 4 Luft Environmental Consulting I I I I I I I I I I I I I I I I I I · vacuum chlor, ne ~.~ ~ ~ .. .. liquid ..~ -~regulat°5 vacuum chlorine liquid va,ve -~~ i j.- _~..~ tub,ng internal eduction pipe ~:~ liquid trap (heated) Figure 2 Installation of the vacuum feed regulator on the ton container. vertically, the top valve will discharge chlorine vapor. The bottom valve in this configuration would discharge liquid chlorine. A removable protective steel housing covers the container valves during the loading/unloading process and general storage of the vessel. Ton containers are also equipped with fusible metal pressure relief devices. Most containers have six fusible metal plugs threaded into them, three at each end spaced 120 degrees apart (see Fig. 3), that are designed to relieve pressure in the event that the container is exposed to high temperature. If these plugs reach a temperature of 158-165°F, the fusible metal inside the plug will yield and discharge chlorine to prevent overpressuring the container. ._._~~ Fusible Metal Fusible Metal Plug I ~ Cross Se~ion View Ton Container Container Valves Figure 3 Fusible metal pressure relief devices. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 5 Luff Environmental Consulting Chlorine Feed Vacuum Regulator Chlorine feed vacuum regulators mount directly on the one ton containers. The regulator is secured to the angle valve on the container with a yoke type clamp which utilizes a lead gasket to seal the mating flanges. These angle valves are designed specifically for chlorine service. Regulators for ton containers have a liquid trap that is electrically heated to vaporize liquid chlorine prior to the vacuum feed regulator. A schematic diagram of a vacuum feed regulator is presented in Figure 4. inlet safety valve I & valve seat I ton container container valve '1 vent '1 I I I lead ~ gasket yoke clamp liquid trap (heated) inlet filter diaphragm ! I I I vacuum tubing Figure 4 Schematic of a vacuum feed regulator valve. Pressure from within the cylinder and the closing spring keep the inlet safety valve closed until there is sufficient vacuum on the diaphragm to open the Gist-brocades Food Ingredients Inc./RMPPIHazOp/Oct., 1989/Page 6 Luft Environmental Consulting I I I I I I I I I I ! I ! I I I I I valve. If the vacuum is lost for any reason, the closing spring will drive the inlet valve closed. The closing spring material in the vacuum feed regulator at Gist-brocades is a tantalum alloy. This tantalum alloy has a corrosion resistance equivalent to glass and is chemically inert in this application. Structural properties of the alloy include excellent ductility and a tensile strength of 50,000 psi. The double diaphragm assembly is manufactured from "Halar" plastic material. The manufacturer of the regulator unconditionally guarantees the closing spring and diaphragm assemblies. At Gist-brocades, the vacuum is created when water flows through an ejector located on a branch of the discharge of the cooling tower water circulation pumps. The ejectors are discussed further in item 6 below. Residual Chlorine Analyzer The residual chlorine analyzer is an amperometric device designed to provide continuous measurement of the concentration of residual chlorine in water. A side stream of water is drawn from the discharge of the cooling tower water circulation pumps prior to the ejectors (the point of chlorination). This water sample is mixed with a small amount of pH buffer solution and reagent prior to flowing through a measurement cell that contains two dissimilar metal electrodes. As the water sample flows past the electrodes, an electrical current is generated which is directly proportional to the concentration of chlorine. Either free or total chlorine concentrations can be measured in the cell, depending on the reagent selection. Each reagent is added to a pH buffer solution to maintain a constant pH of the sample. When measuring free chlorine residual, potassium bromide is added to the buffer. The potassium bromide reacts with the free chlorine to liberate bromine in an amount equal to the free chlorine. The bromine depolarizes the measuring cell in the same manner as chlorine does and a current directly proportional to the free chlorine residual is generated. When measuring total chlorine residual, potassium iodide and sodium hydroxide are added to the buffer. This solution reacts with the free and Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 7 Luff Environmental Consulting I I I I I I I I I I I I I I I I I I combined chlorine to liberate iodine in an amount equal to the total chlorine concentration. The iodine depolarizes the measurement cell and generates a current directly proportional to the total chlorine residual in the water. The electrical signal generated by the reagents is amplified by solid-state electronics and produces an isolated 4-20 mA direct current signal via an electronic controller. This signal is the input to the automatic chlorine flow control valve, for regulation of the chlorine dosage. Automatic Chlorine Flow Control Valve An automatic chlorine flow control valve is used to regulate the flow of chlorine gas from the vacuum feed regulators. This valve is a proportioning valve that dispenses chlorine at a rate that is inversely proportional to the residual chlorine concentration in the cooling tower water. The valve receives a 4-20 mA signal from the residual chlorine analyzer controller which is directly related to the residual chlorine concentration. The basic operation principle is that the higher the electrical signal, the higher the residual chlorine, and hence, the lower the chlorine flow rate at the automatic chlorine flow control valve. The drive assembly on the valve uses an electric step motor with a hollow shaft machined to accept a lead screw. As the motor rotates one step (a revolution), the lead is threaded in or out of the shaft depending upon direction of rotation of the motor. Since the lead screw is restrained from rotating, the resulting motion is a defined linear step. The lead screw is attached to and directly drives a valve plug with identified flow characteristics. This unit can be operated in a manual mode with the position of the valve plug being set by an operator. Manual operation may take place in the event of an analyzer or power supply failure. If a power failure occurs, the position of the valve will remain at the last set point until power is resumed or the valve is manually adjusted. The demonstrated maximum chlorine flow rate through this valve in the Gist- brocades system is 50 pounds per day. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 8 Luft Environmental Consulting I I Rotorneter Flow Meters I The rotometer flow meters are connected between the chlorine piping from the automatic chlorine flow control valve and the individual ejectors to each i cooling tower pump discharge. These rotometers provide chlorine gas flow control from the common manifold to each of the ejectors as they attempt to draw chlorine from the automatic chlorine flow control valve. I Ejectors I Ejectors are used on each discharge branch from the circulating pumps to draw chlorine into the cooling tower water. Water flowing through the ejector I creates a vacuum that opens a spring loaded check valve and draws in chlorine gas, as shown in Figure 5. If there is no water flowing through the ejector, the check valve closes and prevents the pressurized water from I entering the vacuum tubing. The chlorinated water then flows through spargers in the cooling tower basin. I Chlorine Check Valve I (Vacuum Un~, ~1 Assembly Diaphragm <:[----Closing Spring I I I I Water Chlorinated ~.~ Water I I Figure 5 Cross sectional view of ejector. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 9 Luft Environmental Consulting I I I I '1 I I I I I I I I I I I I C. HAZOP REVIEW FOR GIST-BROCADES FOOD INGREDIENTS INC. A review of the chlorination equipment was used to generate the following subsystems that were investigated. Item #1: Item #2: Item #3: Item #4: Item #5: Item #6: Ton Container Vacuum Feed Regulator Residual Chlorine Analyzer Automatic Chlorine Flow Control Valve Rotometers Ejector A guide word approach was used to systematically identify the intended design and operation of the item in the system. Due to the limited amount of equipment required to operate the system, the actual guide word application was more complex than the system warranted. Therefore, deviations from the intended operation and the potential for a chlorine release from each piece of equipment were discussed on a "what if" basis. D. RESULTS OF HAZOP In order to meet the requirements of the statute, an offsite consequence analysis of the most likely hazards has to be performed for the RMPP. The term "most likely hazards" typically represents very minor releases with no public consequence. For this reason, the "worst credible events" were actually investigated for the offsite consequence analysis. These worst credible events are described briefly below. Such events could be caused by operating error, equipment failure or external events (including earthquakes). Double jeopardy events were not considered in this analysis. Release Events The HazOp team reviewed the chlorine handling, storage and operating procedures at the site. Worst case release events centered around a failure of one of the components associated with the ton container. These components include the fusible plug, the chlorine valve, and the chlorine valve packing. The chlorine valve and the fusible plug have similar orifice sizes and the rate of discharge from each is roughly equivalent. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 10 Luft Environmental Consulting I I I I I I I I I I I I I I I I I I The Chlorine Institute, Inc.'s Emergency Kit "B" is designed to stop leaks from valves, fusible plugs, and the sidewalls of ton containers. Containment of a leak is accomplished by either plugging the leak mechanically (ie. forcing a drift pin into an orifice) or by placing a cap over the leaking device. The caps are held in place by an adjustable bar that bridges across the chimes. Ton Container Case #1 - Total Release - Liquid Phase The first failure investigation was that of a catastrophic failure of the chlorine vessel. In accordance with the U.S. Environmental Protection Agency (EPA) "Technical Guidance for Hazards Analysis''3, the worst case release event assumes that the entire contents of the vessel would be released in a 10 minute period. Since this unlikely worst case release event was developed using a guidance document, no consideration was given to liquid pooling and evaporation of the chemical over time. With a maximum container capacity of 2000 pounds, the release rate would be 200 pounds per minute. This guideline developed release rate for a catastrophic failure of the vessel is similar to the rate of discharge from a total failure of a fuse plug or an uncontrolled full opening of the liquid chlorine valve. In an uncontrolled liquid discharge, the actual rate of atmospheric release would depend on the rate of release and the evaporation rate of the liquid chlorine pool. Evaporation of the chlorine pool would be proportional to the surface area of the liquid pool, the type of substrate the pool formed on, and the available heat to vaporize the chlorine. The heat that is available to vaporize the chlorine is dependent on the ambient temperature and temperature of the substrate material. For simplicity, the guideline developed rate was modeled. If a release from a fusible plug or an open valve were to occur, the release rate in this event would likely be significantly reduced within a few minutes by a very simple adjustment to the system. Since the chlorine container is stored on trunnions, the vessel could be easily rotated to place the leak point in the vapor phase of the chlorine. This simple operational task would be accomplished after operations personnel donned their personal protective equipment. 3Technical Guidance forHazards Analysis, U.S.E.P.A., FEMA, U.S. DOT, Dec., 1987. Gist-brocades Food Ingredients Inc,/RMPP/HazOp/Oct., 1989/Page 11 Luff Environmental Consulting I I I I I I I I I I I I I I i I I I Rotation of the container so the discharge is in the vapor phase could reduce the release rate by approximately 700 percent. Case #2 - Valve Packing Leak - Liquid Phase A more likely worst credible release event involves a leak that could occur in the packing of a valve or a faulty fusible metal plug. This release rate was calculated to be 20.4 pounds per minute of liquid chlorine from a 1/16 inch by 1/16 inch orifice. A release duration of 30 minutes was assumed until the leak could be contained. For this release duration, the release rate would be relatively constant. The valve packing release could be mitigated rapidly by closing the container valve. Containment of the release from the fusible plug would be made utilizing a Chlorine Institute, Inc.'s Emergency Kit "B". It should be noted that documents received from the Institute state that the Emergency Kit "B" can be installed in approximately 8 minutes. Allowing time to discover the problem, don protective clothing and respirators, and install the Kit accounted for the half hour response time. This response time is a very conservative estimate considering the availability of respirators and emergency protective clothing (rain suits and chemical resistant gloves). Even though the release was a liquid, it was assumed that no pooling of the chlorine would take place and that the entire stream would vaporize almost immediately upon release. In a like manner to Case #1 above, the release rate would likely be significantly reduced within a few minutes by rotating the container to place the leak point in the vapor phase of the chlorine. After rotation of the vessel, the Emergency Kit "B" would be installed. Case #3 - Total Release - Vapor Phase A vapor phase case similar to Case #1 above was also reviewed. This worst case release event was developed assuming a total failure of a fusible plug or mechanically shearing off of the vacuum regulator from the valve. Although unlikely, the vacuum regulator could be sheared off by an uncontrolled spreader bar used to load and off load containers. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 12 Luft Environmental Consulting I I I I Response and containment of the leak was again assumed to be 30 minutes. The initial release rate from the .34 inch diameter orifice would be very high, approximately 80 pounds in the first two minutes (40 pounds per minute). After this initial release, the release rate would become more constant at about 13 pounds per minute for approximately 10 more minutes. After a fifteen minute duration, the release rate would reduce down to approximately 8.5 pounds per minute. For the purposes of the air model, an average release rate of 14.6 pounds per minute was used. I I I I Case #4 - Valve Packing Leak - Vapor Phase In a similar event to Case #2, a vapor release event was developed for a leak in the packing of a valve or a limited failure of the fusible plug. The area of the leak was assumed to be .0039 inches (1/16 inch by 1/16 inch). This opening resulted in a calculated initial release rate of 3.0 pounds per minute. Response time was again assumed to be 30 minutes. Since the release rate was relatively Iow and the duration did not release a significant percentage of the container contents, the initial release rate was used in the air model. I Vacuum Feed Regulator I I I I I ! Chlorine feed vacuum regulators mount directly on the one ton container valves with a yoke type clamp. This clamp has a tongue and groove flange which utilize a lead gasket to seal the mating surfaces. Pressure from within the cylinder and the closing spring keep the inlet safety valve closed until there is sufficient vacuum on the diaphragm to open the valve. If the vacuum is lost for any reason, the closing spring will drive the inlet valve closed. Since the vacuum feed regulator system only delivers chlorine into an evacuated pipeline, any leak in the chlorine piping system would result in a loss of vacuum and a decrease in chlorine flow. Vacuum is created for the chlorination system when water flows through an ejector located on a branch of the discharge of the cooling tower water circulation pumps. Several modes of failure that could result in a release from the vacuum regulator were investigated by the HazOp team. Literature from the Chlorine Institute, Inc. was also used to help define parameters of potential failures. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 13 Luff Environmental Consulting I Case #5 - Vacuum Regulator Installation Release I When the chlorination system is assembled, the container valve is open and pressure exists between the valve and the vacuum regulator. The improper I installation of the regulator could result in a leak between the container valve and the vacuum regulator. Operating procedures at the Bakersfield Plant are I in place to minimize the potential for a release. Each time a regulator is installed on a container, a new lead gasket is utilized. I After installation, the container valve is briefly opened and to provide closed pressure to the regulator. A squeeze bottle of ammonium hydroxide is used to I provide ammonia vapors near the gasket to locate any leaks. If a leak is present, a localized white cloud will form when the ammonia reacts with the chlorine gas. Once the leak is located, the failed part would be replaced. I When the installation of the regulator is verified to be leak free, the container valve is opened no more than a half turn. I Release from the installation of the regulators would be very minor due to the minimal amount of chlorine available as a result of only briefly opening the i A release of this would certainly be less than the previous cases. valve. type I Case #6- Vacuum Regulator Gasket Release During steady state operations of the chlorine system, it is possible that a leak I from the gasket could occur as a result of an external force, such as an ' earthquake. The likelihood of this leak being significant is Iow due to the precision machined tongue and groove matching flanges on the container and the regulator. If a leak were to occur, the available flow area for this type of leak would be very small as a result of the design of the flanges. The flow rate of a release from this case would be less than the cases described under the "ton container" above. Additionally, the leak could be stopped by simply turning off the container valve. I I I I I Case #7 - Vacuum Feed Regulator Diaphragm Failure Failure of the diaphragm in the regulator was also investigated. This event would probably not result in a atmospheric release during normal operating Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 14 Luft Environmental Consulting I I I I I I I I I I I I I I I ! I I conditions. During normal operations, the diaphragm failure would result in a loss of vacuum and lower chlorination. This failure would be detected during the daily inspection of the system by a Iow chlorine residual in the cooling tower water. When the system is not operational, the inlet safety valve should seal off the chlorine flow. If the inlet safety valve does not completely seal, the chlorine could escape through the atmospheric vent on the regulator. The available flow area for this release would again be less than in the cases concerning the ton container. Mitigating this type of release is as simple as shutting off the container valve. Residual Chlorine Analyzer Failure of the residual chlorine analyzer could not result in a direct release of chlorine. If the analyzer failed with a "no residual" reading signal, it would tend to drive open the automatic chlorine flow control valve and sup(~r chlorinate the cooling tower water. This would be discovered in the daily inspections on the cooling tower system. Automatic Chlorine Flow Control Valve The automatic chlorine flow control valve is operated on a vacuum. Any sealing device failure would result in lower chlorine flow into the cooling tower water system and would be discovered by operations personnel during daily inspections. If the automatic chlorine flow control valve were'driven open by a faulty residual chlorine analyzer, the demonstrated maximum flow rate available would only be 50 pounds per day. This flow rate is approximately two times the normal rate and would be discovered as high chlorine residual during the daily inspections. There is no potential for a significant atmospheric release from this device. Rotometers Failure of the rotometers would result in lower chlorine flow, as in the other vacuum operated devices. There is no potential for a significant atmospheric release of chlorine. Gist-brocades Food Ingredients Inc./RMPP/HazOp/Oct., 1989/Page 15 Luft Environmental Consulting I I I I I I I I I I I I I I I I '1 I Ejector Ejector failure would result in limited or no chlorine flow. a significant atmospheric release. Gist-brocades Food Ingredients I nc./RMPP/HazOp/Oct., 1989/Page 16 There is no potential for Luft Environmental Consulting Gist-brocades Food Ingredients, Inc~ RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) SUPPORTING DOCUMENT Gist-brocades Food Ingredients, Inc. RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) SUPPORTING DOCUMENT Prepared By Luft Environmental Consulting 3701 Pegasus Drive, Suite 121 Bakersfield, CA 93308 October 16, 1989 RECORD OF REVISIONS Change Number Date of Change Date Entered Signature of Person Entering Change TABLEOFCONTENTS II. III. IV. VI. VII. IX. Xo XI. XIV. PREFACE INTRODUCTION CERTIFICATION BY QUALIFIED PERSON AND FACILITY OPERATOR ACCIDENT HISTORY FACILITY DESCRIPTION DESIGN, OPERATING AND MAINTENANCE SAFETY SYSTEMS DETECTION, MONITORING AND CONTROL SAFETY SYSTEMS AUDITING AND INSPECTION PROGRAMS RECORDKEEPING PROCEDURES PERSONNEL AND TRAINING EMERGENCY RESPONSE PROCEDURES DESCRIPTION OF THE HAZARD AND OPERABILITY STUDY OFFSITE CONSEQUENCE ANALYSIS IMPLEMENTATION OF THE RMPP APPENDIX A: RESUMES OF QUALIFIED PERSONS APPENDIX B: STATUTE WITH AMENDMENTS AND LIST OF AHMS Gist-brocades Food In~'edients, Inc./RMPP/October, 1989/Page i Luff Environmental Consulting I. PREFACE Pursuant to Article 2, Chapter 6.95, Division 20, California Health and Safety Code (H & S Code), a Risk Management and Prevention Program (RMPP) must be prepared for an existing facility which handles listed acutely hazardous materials, if requested by the Administering Agency. Gist-brocades Food Ingredients Inc. utilizes chlorine as a biocide in their cooling tower. The storage capacity of the chlorination system is greater than the levels specified in Section 25536 (a), H & S Code. An RMPP was requested by the Administering Agency for the existing chlorine system in October of 1988. The RMPP prepared for the Gist-brocades Food Ingredients Inc. facility, which is described in this Supporting Document, addresses the process and equipment associated with the handling and storage of chlorine. Section 25534 (I) of the Health & Safety Code states that "[t]he Office of Emergency Services shall adopt, and publish for distribution, guidelines for the preparation and submission of risk management and prevention programs." This document is based on a review of the following publications. The requirements of the statute (California Health and Safety Code, Division 20, Chapter 6.95, Article 2). "Risk Management and Prevention Program Guidance", by the California Office of Emergency Services, June 1989. "Guidelines for Hazard Evaluation Procedures", prepared by Batelle Columbus Division for the Center for Chemical Process Safety of the American Institute of Chemical Engineers, 1985. This document describes the elements of the RMPP and the results of the hazard and operability study and the offsite consequence analysis. The format of the RMPP Supporting Document follows, in general, the sequence of items as presented in the statute. Throughout this document, section numbers quoted are from Chapter 6.95, Division 20, California Health and Safety Code, unless stated otherwise. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page I-1 Luft Environmental Consulting I1. INTRODUCTION Gist-brocades Food Ingredients Inc.'s Bakersfield plant produces baker's yeast under the label of Eagle Baker's Yeast. The yeast is grown in batch fermentation tanks. In order to maintain the optimum growth temperature for the yeast, the fermentation tanks are cooled with water from the cooling tower system. A. ACUTELY HAZARDOUS MATERIALS (AHM) Gist-brocades Food Ingredients Inc. utilizes chlorine, which is a listed acutely hazardous material, as a biocide in the cooling tower system at the plant. This RMPP addresses in detail the processes and equipment associated with the handling and storage of chlorine gas. B. REQUIREMENTS OF THE STATUTE Assembly Bill 3777 of 1986 added Article 2 to Chapter 6.95, Division 20, of the California Health and Safety Code. A summary of Article 2 as it pertains to the Gist-brocades Food Ingredients Inc. facility is provided below. Section 25534 (a) of Article 2 states that "... the Administering Agency may require the submission of an RMPP. " The Bakersfield City Fire Department (the Administering Agency) issued a letter in October, 1988 requesting Gist-brocades Food Ingredients Inc. to submit an RMPP. m' According to Section 25532(g), a Risk Management and Prevention Program, or RMPP, "means all the administrative and operational programs of a business which are designed to prevent acutely hazardous materials accident risks, including, but not limited to, programs which include design safety of new and existing equipment, standard .operating procedures, preventive maintenance programs, operator training and accident investigation procedures, risk assessment for unit operations, or operating alternatives, emergency response planning, and internal or external audit procedures to ensure that these programs are being executed as planned." Pursuant to 25534 (d), the RMPP shall be based upon "an assessment of the processes, operations, and procedures of the business, and shall consider all of the following: Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page I1-1 Luff Environmental Consulting (1) The results of a hazard and operability study which identifies the hazards associated with the handling of an acutely hazardous matedal due to operating error, equipment failure, and external events, which may present an acutely hazardous materials accident risk. (2) For the hazards identified in the hazard and operability studies, an offsite consequence'analysis which, for the most likely hazards, assumes pessimistic air dispersion and other adverse environmental conditions." An RMPP is to contain the following elements, as specified in Section 25534 (c). "(1) A description of each accident involving acutely hazardous materials which has occurred at the business or facility within three years from the date of the request made [by the Administering Agency], together with a description of the underlying causes of the accident and the measures taken, if any, to avoid a recurrence of a similar accident. (2) A report specifying the nature, age, and condition of the equipment used to handle acutely hazardous materials at the business or facility and any schedules for testing and maintenance. (3) Design, operating, and maintenance controls which minimize the risk of an accident involving acutely hazardous materials. (4) Detection, monitoring, or 'automatic control systems to minimize potential acutely hazardous materials accident risks. (5) A schedule for implementing additional steps to be taken by the business, in response to the findings of the assessment performed, to reduce the risk of an accident involving acutely hazardous materials. These actions may include any of the following: (A) Installation of alarm, detection, monitoring, or automatic control devices. (B) Equipment modifications, repairs, or additions. (C) Changes in the operations, procedures, maintenance schedules, or facility design. (6) Auditing and inspection programs designed to allow-.the handler to confirm that the RMPP is effectively carded out. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page 11-2 Luft Environmental Consulting 8 m am 10. 11. (7) Recordkeeping procedures for the RMPP." Section 25534 (f) states that "[a] handler shall maintain all records concerning an RMPP for a period of at least five years." Pursuant to Section 25534 (g), the RMPP must "identify, by title, all personnel at the business who are responsible for carrying out the specific elements of the RMPP, and their respective responsibilities, and the RMPP shall include a detailed training program to ensure that those persons are able to implement the RMPP." Section 25534 (h) requires that "[t]he handler shall review the RMPP, and shall make necessary revisions to the RMPP at least every three years, but, in any event, within 60 days following a modification which would materially affect the handling of an acutely hazardous material." The RMPP, and any required revisions, "shall be certified as complete by a qualified person and the facility operator'', in accordance with Section 25534 (j). Section 25534 (k) states that "[e]xcept as specified in subdivision (d) of Section 25535 [item 10 below], the handler shall implement all activities and programs specified in the risk management and prevention program within one year following the certification made pursuant to subdivision (j) [item 8 above]. Implementation of the RMPP shall include carrying out all operating, maintenance, monitoring, inventory control, equipment inspection, auditing, recordkeeping, and training programs as required by the RMPP." In accordance with 25535(d), "[t]he owner or operator shall implement all programs and activities in the RMPP before operations commence, in the case of a new facility, or before any new activities involving acutely hazardous materials are taken, in the. case of a modified facility" Additionally, as added to the statute, effective January 1, 1989, Section 25534.1 requires that "[e]very RMPP . . . shall give consideration to the proximity of the facility to schools, general acute care hospitals, and long- term health care'facilities." Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page 11-3 Luft Environmental Consulting C. RMPP ELEMENTS AND REQUIREMENTS The elements and requirements of the RMPP, in accordance with the statute, are addressed in the sections of this document listed in Table I1-1. D. PROCEDURES FOR UPDATING THE RMPP As required by SeCtion 25534 (h), the RMPP will be reviewed, and revised as necessary, at least every three years. It also will be revised within 60 days following a modification which would materially affect the handling of an acutely hazardous material. Need for revision of the RMPP will be partially based on the routine inventories performed for business plan purposes. Pursuant to Section 25510, within 30 days of the receipt of a new hazardous material or a 100% or more increase in a hazardous material currently on-site, .an amendment is to be made to the inventory form. Filing of an amendment to the inventory form will trigger a review by Gist-brocades Food Ingredients Inc.'s Maintenance Engineer. The determination as to whether the RMPP requires revision will be made at that time. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page 11-4 Luft Environmental Consulting TABLE I1-1 RMPP ELEMENTS AND REQUIREMENTS AS SPECIFIED IN CHAPTER 6.95, DIVISION 20, CALIFORNIA HEALTH AND SAFETY CODE ~ection of Cha0ter 6.95 Section of RMPP Document 25532 (g) VI, VIII, X, XI, XII 25534 (c)(1) IV 25534 (c)(2) V 25534 (c)(3) VI 25534 (c)(4) VII 25534 (c)(5) XIV 25534 (c)(6) VIII 25534 (c)(7) IX 25534 (d)(1) XII 25534 (d)(2) XIII 25534 (f) IX 25534 (g) X 25534 (h) II 25534 (j) III 25534 (k) VI, VII, VIII, IX, X 25535 (d) XIV 25534.1 XIII Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page 11-5 Luff Environmental Consulting II!. CERTIFICATION BY QUALIFIED PERSON AND FACILITY OPERATOR Section 25534 (j), H & S Code requires that the RMPP shall be certified as complete by a qualified person and the facility operator. These certifications are provided below. I certify that I am qualified to attest to the validity of the hazard and operability studies performed pursuant to Section 25534, and the relationship between the corrective steps taken by the handler following the hazard and operability studies and those hazards which were identified in the studies. Additionally, I certify that this risk management and prevention program is complete. This certification is based on my understanding that the data and documents provided by Gist-brocades Food Ingredients Inc. are true and correct and that~the plans, programs, and procedures will be implemented as described. Signature Kad W. Luft. P. E. Name Signature A, Sue Luff. CHMM. R.E.A. Name principal Mechanical Engineer Title Date Prirlcipal Environmental Enoineer Title Date As.facility operator, I hereby certify that this risk management and prevention program is complete. · ? / ~.Si'gnatu re Robert Deedy Name Plant Manager Title ,/ Date Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page II1-1 Luft Environmental Consulting IV. ACCIDENT HISTORY Section 25534 (c)(1) of the statute requires that the RMPP include "[a] description of each accident involving acutely hazardous materials which has occurred at the business or facility within three years from the date of the request [for the RMPP], together with a description of the underlying causes of the accident and the measures taken, if any, to avoid a recurrence of a similar accident." A. DESCRIPTION OF ACCIDENT Gist-brocades Food Ingredients Inc. has not had an accident involving chlorine at the Bakersfield Plant. B. ACCIDENT PREVENTION MEASURES This subsection is not applicable since there have been no accidents involving acutely hazardous materials at the Bakersfield Plant. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page IV-1 Luff Environmental COnsulting V. FACILITY DESCRIPTION In accordance with Section 25534 (c)(2), the RMPP must include "[al report specifying the nature, age, and condition of the equipment used to handle acutely hazardous materials at the business or facility and any schedules for testing and maintenance." This section provides a description of the facility and equipment which handle AHMs. A. GENERAL DESCRIPTION OF FACILITY Gist-brocades Food Ingredients Inc.'s Bakersfield plant produces baker's yeast in batch fermentation 'tanks. The biological activity of the yeast growing in the molasses mixture generates heat. In order to maintain the optimum growth temperature for the yeast, the fermentation tanks are cooled with water from the cooling tower system. After the fermentation process is complete, the yeast is dried and packaged for the industrial market. B. DESCRIPTION OF AHM PROCESS AND EQUIPMENT The chlorine system was installed in 1985 and is used as a biocide in the cooling tower system at the plant. Chlorine gas is stored on site in one ton containers. The chlorine site has two sets of trunnions designed to handle two containers. Each cylinder can' be rotated into the appropriate operating position on the trunnions. Chlorine is drawn from the container through the vacuum feed regulator by a vacuum created by ejectors. The ejectors are located in a discharge branch of each of the three cooling tower water circulation pumps. Chlorine flow is controlled by two devices, an automatic chlorine flow control valve and individual rotometers. The flow through the automatic chlorine flow control valve is determined by an electrical signal from the residual chlorine analyzer. This flow is further proportioned by the rotometers associated with the individual ejectors. A schematic overview of the system is shown in Figure V-1. Each key component in the system is discussed below. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-1 Luft Environmental Consulting Ton Container Vacuum Regulator Automatic Chlorine Flow Control Valve .\ 4-20 mA Signa', ~,,', Rotometers Residual jChlorine Analyzer ]::>' I Typical to ~>.j, other Ejectors To Cooling Tower Basin To Spargers in the Tower Basin Ejector Cooling Tower Turbine Pump To Cooling Tow~er Circulation System PIPING KEY Thin Lines = Vacuum Lines Bold Lines = Pressure Lines Dashed Lines = Electronic Signal Figure V-1 Schematic diagram of the chlorination system. 1. Chlorine Ton Container Ton containers are welded steel tanks with a chlorine capacity of 2,000 pounds (one short ton). Full containers weigh approximately 3,700 pounds. The heads of the container are convex inward and forge welded to the shell. Shell ends are crimped inward to form chimes which form a substantial gripping surface for the lifting beams that are used to load/unload the containers. These containers are constructed in accordance with the U.S. Department of~Transportation's (DOT) requirements pursuant to 49 Code of Federal Regulations (CFR). ? Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-2 Luff Environmental Consulting Each container is equipped with two identical valves located near the center of one end of the vessel (Fig. V-2). These standard one ton container valves are connected to an internal eduction pipe, as shown in Fig. V-3a. When the valves are aligned vertically, the top valve will discharge chlorine vapor. The bottom valve in this configuration would discharge liquid chlorine. A removable protective steel housing covers the container valves during the loading/unloading process and general storage of the vessel. Ton containers are also equipped with fusible metal pressure relief devices. Most 'containers have six fusible metal plugs threaded into them, three at each end spaced 120 degrees apart (see Fig. V-2), that are designed to relieve pressure in the event that the container is exposed to high temperature. If these plugs reach a temperature of 158-165°F, the fusible metal inside the Plug will yield and discharge chlorine to prevent overpressuring the container. j Fusible Metal ~ Top View · Fusible Metal Plug r ~ ,~ cross section view One T. on' tiiiiiiiii~ Fusible Metal Ton Container ContaIner Figure V-2 Fusible metal pressure relief devices 2. Chlorine Feed Vacuum Regulator Chlodne feed vacuum regulators mount directly on the one ton containers. The regulator is secured to the angle valve on the container with a yoke type clamp which utilizes a lead gasket to seal the mating flanges.' These angle valves are designed specifically for chlorine service. Figure V-3a shows a vacuum feed regulator installed on a ton container. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-3 Luft 'Environmental Consulting Regulators for ton containers have a liquid trap that is electrically heated to vaporize liquid chlorine prior to the vacuum feed regulator, as shown in Figure V-3b. ~-f / ch,orine gas va,ve chlorine gas ' · ',Vent · chlorine ~_~ _~ j~ aCuU~Jar~or liquid__ ..."~ -~,,,,,,'~eg [_vacuum o ,or, e i 1.-- internal edu~ion ~~N pipe ~ · liquid trap (heated) Figure V-3a Installation of a vacuum feed regulator valve on a ton container container wall inlet safety valve & valve seat ton container valve vent lead ~ gasket yoke clamp liquid trap inlet (heated) 3> filter vacuum tubing diaphragm Figure V-3b Schematic of a vacuum feed regulator valve. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-4 Luff Environmental Consulting Pressure from within the cylinder and the closing spring keep the inlet safety valve closed until there is sufficient vacuum on the diaphragm to open the valve. If the vacuum is lost for any reason, the closing spring will drive the inlet valve closed. At Gist-brocades, the vacuum is created when water flows through an ejector located on a branch of the discharge of the cooling tower water circulation pumps. The ejectors are discussed further in item 6 below. 3. Residual Chlorine Analyzer The residual chlorine analyzer electronically provides a continuous measurement of the concentration of residual chlorine in the cooling tower water. Water is drawn from the discharge of the cooling tower water circulation pumps prior to the'point of chlorination. This water sample passes through a measurement cell in the residual chlorine analyzer that generates an electrical signal which is directly proportional to the concentration of residual chlorine. The electrical signal is processed by a controller to provide an output signal for the automatic chlorine flow control valve. 4. Automatic Chlorine Flow Control Valve An automatic chlorine flow control valve works in conjunction with the residual chlorine analyzer to regulate the flow of chlorine gas from the vacuum feed regulators. The flow control valve is a proportioning valve that dispenses chlorine inversely proportional t° the residual chlorine in the cooling tower water. 5. Rotometer Flow Meters The rotometer flow meters are connected betwben the chlorine piping from the automatic chlorine flow control valve and the individual ejectors to each cooling tower pump discharge. These rotometers provide chlodne gas flow control from the common manifold to each of the ejectors as they attempt to draw chlodne from the automatic chlorine flow control valve. 6. Ejectors Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-5 Ejectors ar'e used on each discharge branch from the circulating pumps to draw chlorine into the cooling toWer water. Water flowing through the ejector Luff Environmental Consulting creates a vacuum that opens a spring loaded check valve and draws in chlorine gas, as shown in Figure V-4. If there is no water flowing through the ejector, the check valve closes and prevents the pressurized water from entering the vacuum tubing. The chlorinated water then flows through spargers in the cooling tower basin. Chlorine Check Valve (Vacuum Line) '~II> j Assembly Diaphragm <~-- Closing Spring ---J Water ~ Chlorinated · Water 'I I Figure V-4 Cross sectional view of ejector. C. SCHEDULES FOR TESTING AND MAINTENANCE 1. Equipment Test Schedules All one ton containers are constructed in accordance with the DOT requirements pursuant to 49 CFR. These requirements specify that the one ton containers are inspected and tested by licensed professional inspectors at given intervals. DOT Regulations also address the refilling of the containers. Components of the chlorination system (other than the one ton container) are routinely tested after any maintenance or repair work is completed. The overall system pedormance is monitored daily by testing for residual chlorine in the cooling water system. 2. Maintenance Schedules Routine maintenance on the equipment is performed in accordance with the manufacturer's recommended procedures for each major piece of Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-6 Luff Environmental Consulting equipment. The equipment is visually inspected daily to see if the chlorine flow is at the proper rate. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page V-7 Luff Environmental Consulting VI. DESIGN, OPERATING AND MAINTENANCE SAFETY SYSTEMS Section 25534 (c)(3) requires that the RMPP address design, operating, and maintenance controls which minimize the risk of an accident involving acutely hazardous materials. This section of the RMPP discusses these controls. It also addresses the design safety of the existing equipment, standard operating procedures, and preventive maintenance programs, as required by Section 25532 (g); and operating and maintenance programs, as required by Section 25534 (k). A. GENERAL The chlorination system on the cooling towers features an all vacuum feed system. A fail safe vacuum regulator is mounted directly on the one ton container valve (vapor phase position). It allows chlorine to flow from the container only if sufficient vacuum is available in chlorine feed line to overcome the closing spring and container pressure. This regulator closes in the event of a loss of vacuum. According to the equipment manufacturer, this type of system is the safest and most commonly used method of handling chlorine. The basic design of the system has been utilized for more than 20 years. B. CONTROL SYSTEMS 1. Vacuum Feed Regulator Chlorine feed vacuum regulators mount directly on the angle valve on the one ton containers. The regulator is secured to the valve with a yoke type clamp which utilizes a lead gasket to seal the mating flanges. This regulator has a liquid trap that is electrically heated to vaporize liquid chlorine prior to the vacuum feed regulator. Pressure from within the,cylinder and the closing spring keep the inlet safety valve closed until there is sufficient vacuum on the diaphragm to open the valve. If the vacuum is lOst for any reason, the . closing spdng will drive the inlet valve closed. The closing spring material in the vacuum feed regulator at Gist-brocades is a tantalum alloy. This tantalum alloy has a corrosion resistance equivalent to glass and is chemically inert in this application. Structural properties of the alloy include excellent ductility and a tensile strength of 50,000 psi. The double diaphragm assembly is manufactured from "Halar" plastic material. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-1 Luft Environmental Consulting The manufacturer of the regulator unconditionally guarantees the closing spring and diaphragm assemblies. An additional feature on the vacuum feed regulator is the container empty or "no gas" indicator. This indicator will show "no gas" if a vacuum is drawn on the normally pressUrized container. 2. Automatic Chlorine Flow Control Valve An automatic chlorine flow control valve is used to regulate the flow of chlorine gas from the vacuum feed regulators. This valve is a proportioning valve that dispenses chlorine at a rate that is inversely proportional to the residual chlorine concentration in the cooling tower water. The valve receives a 4-20 mA signal from the residual chlorine analyzer controller which is directly related to the residual chlorine concentration. The basic operation principle is that the higher the electrical signal, the higher the residual chlorine, and hence, the lower the chlorine flow rate at the automatic chlorine flow control valve. The drive assembly on the valve uses an electric step motor with a hollow shaft machined to accept a lead screw. As the motor rotates one step (a revolution), the lead is threaded in or out of the shaft depending upon direction of rotation of the motor. Since the lead screw is restrained from rotating, the resulting motion is a defined linear steP. The lead screw is attached to and directly drives a valve plug with identified flow characteristics. This unit can be operated in a manual mode with the position of the valve plug being set by an operator. Manual operation may take place in the event of an analyzer or power supply failure. If a power failure occurs, the position of the valve will remain at the last set point until power is resumed or the valve is manually adjusted. The demonstrated maximum chlorine flow rate through this valve in the Gist- brocades system is 50 pounds per day. 3. Residual Chlorine Analyzer The residual chlorine analyzer is an amperometric device designed to Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-2 Luff Environmental Consulting provide continuous measurement of the concentration of residual chlorine in water. A side stream of water is drawn.from the discharge of the cooling tower water circulation pumps prior to the ejectors (the point of chlorination). This water sample is mixed with a small amount of pH buffer solution and reagent prior to flowing through a measurement cell that contains two dissimilar metal electrodes. As the water sample flows past the electrodes, an electrical current is generated which is directly proportional to the concentration of chlorine. Either free or total chlorine concentrations can be measured in the cell, depending on the reagent selection. Each reagent is added to a pH buffer solution to maintain a constant pH of the sample. When measuring free chlodne residual., potassium bromide is added to the buffer. The potassium bromide reacts with the free chlorine to liberate bromine in an amount equal to the free chlorine. The bromine depolarizes the measuring cell in the same manner as chlorine does and a current directly proportional to the free chlorine residual is generated. When measuring total chlorine residual, potassium iodide and sodium hydroxide are added to the buffer. This solution reacts with the free and combined chlorine to liberate iodine in an amount equal to the total chlorine concentration. The iodine depolarizes the measurement cell and generates a current directly proportional to the total chlorine residual'in the water. The electrical signal generated by the reagents is amplified by solid-state electronics and produces an isolated 4-20 mA direct current signal via an . electronic controller. This signal is the input to the automatic chlorine flow control valve, for regulation of the chlorine dosage. 4. Rotometers Rotometers provide a means of distributing flow from the automatic chlorine flow control valve to the individual ejectors in the cooling tower system. The chlorine flow rate to each ejector is controlled by varying the flow orifice with a threaded thumb wheel on top of the rotometer. As the thumb wheel is threaded in or out, a tapered valve plug is moved toward or away from the valve seat which varies the flow area. These gas specific rotometers have a ball float that rises in the calibrated clear plastic tube .in proportion to the volumetric flow rate through the meter. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-3 Luft Environmental Consulting This visual representation of the flow allows an operator to adjust the chlorine flow to the individual ejectors and have an immediate reading as to the exact flow rate into the cooling tower water. The system also provides a means to visually monitor the operation of the chlorination package. 5. Ejectors Ejectors are located on a side stream off of the discharge of each of the cooling tower water circulating pumps. The ejectors receive high pressure water from the pumps and. discharge into spargers in the cooling tower basins. Water flowing through the ejector creates a vacuum that pulls open the spring loaded check valve, and draws in chlorine gas from the vacuum regulator, throUgh all of the flow control devices, and into the cooling tower water. If there is no water flowing through the ejector, the spring loaded check valve closes and prevents the water from entering the vacuum tubing. C. SAFETY EQUIPMENT 1. Pressure Relief Valves Ton containers are equipped with fusible metal pressure relief devices. Most containers have six fusible metal plugs threaded into them, three at each end spaced 120 degrees apart. If these plugs reach a temperature of 158-165°F, the fusible metal inside the plug will yield and discharge chlorine to prevent over pressuring the Container. 2. Respirators Several respirators and three self contained breathing apparatus (SCBAs) are onsite for response to emergency situations. The respirators are full face canister type and are located near the chlorine containers and throughout the plant. One of the self contained breathing apparatus is equipped with an air line that is connected to a large air cylinder on a portable cart. The other two SCBA's have standard 30 minute air bottles. All of the respirators are located in separate geographical areas within the plant. These separate locations of the respirators ensure that emergency response personnel can obtain the necessary personal protective equipment during an emergency. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-4 Luft Environmental Consulting D. FIRE PROTECTION SYSTEM Hydrants/Hose Reels There are three fire hose houses (stand pipe with hose) in the immediate area of the chlorine containers. One of the. hose houses is located adjacent to the chlorine containers. The second hose house is to the east of the cooling towers near the "cleaning in place" building, while the third one is to the north near the boiler house. A fire hydrant, with a hose house, is also located to the west of the cooling towers. E. STANDARD OPERATING PROCEDURES Chlorination of the cooling tower water is a continuous process during the operation of the cooling tower water 'circulation pumps. Residual chlorine levels are checked daily by the operations personnel. As necessary, operators make minor' adjustments in the chlorine flow rates to maintain the proper chlorine residual in the water. Average chlorine consumption for the chlorination system is approximately 22 pounds 'per day. At this rate, a one ton container is consumed in about three months. The changing of the ton containers is done according to the standard procedures developed for the Bakersfield facility. A ~brief description of the procedures follows. When the active chlorine container becomes empty, the chlorination system is transferred onto the spare container. The vacuum regulator is removed from the empty container and installed with a new lead gasket on the full container. Standard procedures for checking for chlorine leaks with ammonium hydroxide are followed prior to placing the container into service. These procedures follow the guidelines of the Chlorin~ Institute, Inc. and the chlorination equipment manufacturers. The empty chlorine container is shipped to an approved chlorine refilling station. All inspection, testing, and filling are done in accordance with the U.S. Department of Transportation requirements. The full container is returned to the Bakersfield plant and placed on a set of trunnions for storage until it is needed. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-5 Luft Environmental Consulting F. PREVENTATIVE MAINTENANCE PROGRAM The chlorination system is maintained on an as needed basis. Daily tests on the residual chlorine levels and routine inspections on the.equipment provide a continuous performance baseline for the system. If the system loses performance, the necessary equipment is thoroughly inspected and properly maintained. G. OTHER OPERATING AND MAINTENANCE PROGRAMS For non-routine.tasks, job procedures and duties discussions may be conducted prior to commencing the work. These discussions ensure that the maintenance personnel understand the specific hazards of the non-routine job. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VI-6 Luff Environmental Consulting VII. DETECTION, MONITORING AND CONTROL SAFETY SYSTEMS Detection, monitoring, and automatic control syStems to minimize potential acutely hazardous materials accident risks are discussed in this section, as required by Section 25534 (c)(4) of Chapter 6.95. -Pursuant to Section 25534 (k), monitoring programs are also addressed. A. AUTOMATED MONITORING Residual Chlorine Analyzer Residual chlorine levels are continuously monitored by this analyzer. Data from the analyzer is used to automatically adjust the chlorine flow rate to maintain a proper chlorine residual in the cooling tower water. B. MANUAL MONITORING Monitoring of the chlorination system is done on a routine basis. Operations personnel physically inspect the system on a daily basis to verify the proper function and integrity of the equipment. Cooling tower 'water samples are analyzed for residual chlorine on a daily basis to insure proper operation of the chlorine system. The plant is manned 24 hours per day. During times when the normal operations staff is not available, the plant has a security service present that could report a release of chlorine, All records of maintenance, both preventative and response, on the chlorine system.will be recorded in a log book. These maintenance records will be kept for a minimum of five years. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VII-1 Luft Environmental Consulting VIII. AUDITING AND INSPECTION PROGRAMS In accordance with Section 25534 (c)(6), this section describes the auditing and inspection programs designed to allow the handler to confirm that the RMPP is effectively carried out. Also discussed are the internal and external audit procedures, pursuant to Section 25532 (g), and equipment inspection programs and auditing programs, per Section 25534 (k). A. INTERNAL AUDITS Performance tests on the chlorination system are conducted daily in .the form of residual chlorine analYses for the cooling water tower water. B. OUTSIDE AUDITS 1. OSHA Federal and California OSHA (Occupational Safety and Health Administration) may conduct periodic safety audits. Complete OSHA safety audits may result from complaints made by employees or other outside agencies. Safety audits may also arise from a serious industrial accident resulting in a lost time injury or death. 2. Bakersfield City Fire Department The Bakersfield City Fire Department (BCFD) will, as a minimum, be making annual fire safety inspections of the Gist-brocades Food Ingredients Inc. facility. The BCFD will conduct inspections at least once every three years, pursuant to Section 25537, Chapter 6.95, Division 20, of the California Health. and Safety Code. C. EQUIPMENT INSPECTIONS Gist-brocades Food Ingredients Inc. personnel pedorm periodic inspections on equipment. Maintenance is pedormed by Gist-brocades Food Ingredients Inc. personnel or contractors on an as-needed basis or as recommended by the equipment manufacturers and/or suppliers. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VIII-1 Luff Environmental Consulting D. AUDITS OF CONTRACTORS Whenever the inspection and maintenance of equipment is.contracted to a independent firm, Gist-brocades Food Ingredients Inc. personnel may audit the findings of those inspections and witness the maintenance of the equipment.. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page VIII-2 Luff Environmental Consulting IX; RECORDKEEPING PROCEDURES Pursuant to Section 25534 (c)(7), Chapter 6.95, this section of the RMPP describes the reCordkeeping procedures for the risk management and prevention program. It also discusses inventory control programs and recordkeeping programs, per Section 25534 (k). As required by Section 25534 (f), the handler shall maintain all records concerning an RMPP for a pedod of at least five years. A. RECORDKEEPING Recordkeeping associated with the RMPP is the responsibility of the Maintenance Engineer. The Maintenance Engineer will ensure that all of the appropriate records (i.e. safety training, chlorination system inspection and testing) are kept. This individual is responsible for maintaining a record of revisions to each document and ensuring that all copies are properly distributed. B. RECORD RETENTION All documents relating to the RMPP shall be maintained for a period of at least five years. Records are kept in the Maintenance Engineer's office at the Gist- brocades Food Ingredients Inc. facility. These records include the following. * Hazardous material inventories * Maintenance records * Daily operations reports * Employee training records C. INVENTORY CONTROL AND RECORDKEEPING PROGRAMS 1. Hazardous Materials Inventories As required by Chapter 6.95, Division 20, California Health~ and Safety Code, inventories of hazardous materials at the facility are submitted to the Bakersfield City Fire Department. Amendments to the inventories are · submitted as required. These inventory records are kept for a minimum of five years. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page IX-1 Luft Environmental Consulting 2. Maintenance Records Documentation of maintenance on the chlorination system shall be kept and retained for a minimum of five years. 3. Operations Records Available records (from the chlorine refill plant) pertaining to the inspection and testing of the chlorine containers shall be kept for the Bakersfield Plant. 4. Employee Training Records All records of employee training are maintained for at least five years. These files contain information on the employees trained, their respective duties, and the date and content of the training program. Gist-brocades Food Ingredients inc./RMPP/October, 1989/Page IX-2 Luff Environmental Consulting X. PERSONNEL AND TRAINING As required by Section 25534 (g), the RMPP shall identify, by title, all. personnel at the business who are responsible for carrying out the specific elements of the RMPP, and their respective responsibilities. It shall include a detailed training program to ensure that those persons are able to implement the RMPP. This section discusses the employee training program for the Gist-brocades Food Ingredients Inc.'s facility, pursuant.to Section 25534 (k) and operator training, per Section 25532 (g). A. SCOPE OF TRAINING Each individual undergoes an assessment of training needs based on previous experience/training and their job assignment at Gist-brocades Food Ingredients Inc.. Supplemental training is provided as necessary to personnel prior to the individual performing unsupervised work. Contractors work under the supervision of Gist-brocades Food Ingredients Inc. personnel. Prior to commencing the work, the contractors are thoroughly briefed on the safety requirements at the plant. B. IMPLEMENTATION OF TRAINING The personnel listed in Table X-1 are responsible for the implementation of the RMPP and for the training of the Gist-brocades Food Ingredients Inc. employees. C. PLANT OPERATIONS PERSONNEL TRAINING AlI operations personnel are specifically trained for their responsibilities within the plant. They are required to read the procedure manuals for each major piece of equipment in the plant. For reference purposes, these manuals are located in the Maintenance Engineer's office. In addition to the equipment training, the operations personnel receive training in fire prevention, fire control, accident prevention, and safety and first aid, in accordance with Gist-brocades Food Ingredients Inc. standards. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page X-1 Luft Environmental Consulting TABLE X -1 PERSONNEL RESPONSIBLE FOR IMPLEMENTING THE RMPP TITLE SPECIFIC RESPONSIBILITY plant Manager Maintenance Engineer Implementation of the entire RMPP Implementation of the entire RMPP 'Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page X-2 Luft Environmental Consulting Prior to any personnel working unsupervised in the Gist-brocades Food Ingredients Inc. facility, they must complete a detailed training program with the following elements. Overview of the process and the equipment. Plant and personnel safety program, "Hazardous Materials Communication personnel protective equipment use emergency procedures, such as response procedures, are also part of this program. including the contents of the Program" and the site specific and maintenance. Standard to a fire and evacuation * System(s)/equipment operation (as applicable to work assignment) · Description of the system/process · Process variables and control o Normal operation o Troubleshooting * Normal Start-up/Shutdown * Emergency Shutdowns * UtilitY and Auxiliary Systems MAINTENANCE TRAINING Maintenance personnel are required to read the available maintenance manuals from the manufacturer before pedorming any work on each major piece of equipment in the plant. For reference purposes, these manuals are located in the Maintenance Engineer's office. In addition to the equipment training, the maintenance personnel receive the standard training in fire prevention, fire control, accident prevention, safety and first aid. The training is in accordance with Gist-brocades Food Ingredients Inc. standards. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page X-3 Luft Environmental Consulting The maintenance personnel's training needs are met on an individual basis. Each maintenance employee receives general presentations along with scheduled training and directed work experience to meet their individual needs. E. ENVIRONMENTAL AND SAFETY CONCERNS All personnel at the Gist-brocades Food Ingredients Inc. facility are instructed in the site specific environmental and safety requirements on a periodic basis. The general subject matter for this training includes the following elements. * HaZardous materials handling * Hazard communication program Notification 'procedures and reporting requirements are discussed, as applicable to each job classification. Specified employees are instructed in, and responsible for, the notification procedures for emergency response organizations. These emploYees are responsible for implementing procedures and coordinating actions of plant personnel with the local emergency response agencies. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page X-4 Luft Environmental Consulting Xl. EMERGENCY RESPONSE PROCEDURES This section addresses the requirements for emergency response planning and accident investigation procedures, pursuant to Section 25532 (g). A. NOTIFICATION PROCEDURES In case of an emergency involving the release or threatened release of a hazardous material, 911 and 1-800-852-7550 or 1-916-427-4341 will be called. This will notify the local emergency agencies, the Bakersfield City Fire Department (station located approximately 2.13 miles away on Sorrano Ave.), and the State Office of Emergency Services as required by law. Personnel responsible for emergency response will be notified via intercom, telephone or voice contact. As necessary, employees, contractors, and other individuals within the facility will also be notified in a similar manner. B. MEDICAL ASSISTANCE PLAN The following local emergency medical facility and ambulance service will be used in the event of an emergency. Memorial Urgent Care 6501 Ming Avenue Bakersfield 397-4Oo4 Hall Ambulance 327-4111 C.. EVACUATION PLAN Immediate notification and evacuation of the business will be performed utilizing Gist-brocades Food Ingredients Inc.'s Evacuation Procedures. Evacuation drills have been conducted at the facility using the following plant evacuation procedures. In the event of a significant chlorine release, Gist-brocades Food Ingredients Inc. personnel will be instructed by the Plant Manager, or his designee, to follow the Plant's emergency evacuation plan. Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page XI-1 Luff Environmental Consulting w Employees will notify all contractors in the area via voice contact. An emergency response contractor and the neighboring businesses will be notified of the release per the Plant's evacuation procedure. All personnel will evacuate in accordance with the pre-designated plan to a safe location. 5. As necessary, proper personnel protective equipment will be worn by trained personnel for response to the emergency, ACCIDENT INVESTIGATION PLAN Each supervisor is responsible for the investigation of any accident at the Gist- brocades Food Ingredients inc, plant, The objective of the plan is to determine the cause of the accident and to initiate preventative measures to help ensure that the accident does not recur. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xl-2 Luft Environmental Consulting XlI. DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY This section of the RMPP addresses the requirement of Section 25534 (d)(1) that the RMPP include the "results of a hazard and operability study which identifies the hazards associated with the. handling of an acutely hazardous material due to operating error, equipment failure, and external events, which may present an acutely hazardous materials accident risk". Additionally, Section 25532 (g) specifies that the RMPP include programs which include risk assessment for unit operations or operating alternatives. A. HAZOP TECHNIQUE The HazOp technique that was used for the Gist-brocades Food Ingredients Inc. facility was a modified "guide word" approach for a Hazard and Operability Study. The basic guide word HazOp was chosen since it allows a systematic and thorough review of every part of the facility that handles AHMs. In conjunction with the guide word approach, a "what if" analysis was incorporated into the review. Both of these approaches are described in the AlChE Guidelines for Hazard Evaluation Procedures1, which is referenced in Section 25534 (I), Chapter 6.95, Division 20, California Health and Safety Code. Other publications further desCribe the guide word approach, including A Guide to Hazard And Operability Studies2. Procedures for conducting the HazOp study were adapted to accommodate the simplicity' of the vacuum operated chlorination system. A series of two meetings were held at the plant site. The first.meeting focused on the design of the system and the equipment utilized on site. A second meeting was conducted to review the information obtained during a literature search on the equipment, and the operating practices and safety procedures pertaining to the chlorination system. The Chlorine.Institute, Inc. (Institute) provided a great deal of literature on the handling, storing and uSe of chlorine gas. The Chlorine Institute, Inc., is a trade association of chlorine manufacturers, packagers, distributors and related 1 Guidelines for Hazard Evaluation Procedures, American Institute of Chemical Engineers, New York, 1985. 2A Guide to Hazard and Operability Studies, Chemical Industry Safety and Health Council of the Chemical Industries Association, London, 1985. Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page Xll-1 Luff Environmental Consulting companies. One of the objectives of the Institute is to promote occupatiOnal and environmental safety in the manufacture, transportation and use of chlorine. The Institute is a technical information center with a focus on issues involving safe design of chlorine containers, transportation of chlorine, employee health and safety, and control of chlorine emergencies. Information concerning the control of emergencies included emergency response to a release, containment of the release, and some case histories was supplied by.the Institute. This information was used by the participants of the HazOp to help in the evaluation of Gist-brocades Food Ingredients Inc.'s chlorination system. B. HAZOP REVIEW FOR GIST-BROCADES FOOD INGREDIENTS INC. A review of the chlorination equipment was used to generate the following subsystems that were investigated. Item #1: Item #2: Item #3: Item #4: Item #5: Item #6: Ton Container Vacuum Feed Regulator Residual Chlorine Analyzer Automatic Chlorine Flow Control Valve Rotometers Ejector A .guide word approach was used to systematically identify the intended operation of the item. Due to the limited amount of equipment required to operate the system, the actual guide' word application was more complex than the system warranted. Therefore, deviations from the intended operation and the potential for a chlorine release for each piece of equipment were discussed on a "what if" basis. C. RESULTS OF THE HAZOP REVIEW In order to meet the requirements of the statute, an offsite consequence analysis of the most likely hazards has to be performed for the RMPP. The term "most likely hazards" typically represent very minor releases with no public consequence. For this reason, the "worst credible events" were actually investigated for the offsite consequence analysis. These "worst credible events" are described briefly below. 'Such events could be caused by operating error, Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XII-2 Luft Environmental Consulting equipment failure or external events (including earthquakes). Double jeopardy events were not considered in this analysis. Ton Container Chlorine is stored at the plant in the liquid phase. Container valves are available to supply both liquid or gaseous chlorine. The HazOp team reviewed the chlorine- handling, storage and operating procedures on site. Worst case release events centered around a failure of a fusible plug, the chlorine valve, and the chlorine valve packing. The chlorine valve and the fusible plug have similar orifice sizes; and therefore, the rate of discharge from each is roughly equivalent. Case #1 - Total Release - Liquid*Phase The first failure investigation was that of a catastrophic failure of the chlorine vessel. In accordance with the U.S. Environmental Protection Agency (EPA) "Technical Guidance for Hazards Analysis''3, the worst case release event assumes that the entire contents of the vessel would be released in a 10 minute period. Since this unlikely worst case release event was developed using a guidance document, no consideration was given to liquid pooling and evaporation of the chemical over time. With a maximum container capacity of 2000 pounds, the release rate would be 200 pounds per minute. This guideline developed release rate for a catastrophic failure of the vessel is similar to the rate of discharge from a total failure of a fuse plug or an uncontrolled release from the liquid chlorine valve. In an uncontrolled liquid discharge, the actual rate of atmospheric release would depend on the rate of evaporation of the liquid chlorine pool. Evaporation of the chlorine pool would be proportional to the sudace area of the liquid pool, the type of substrate the pool formed on, and the available heat of vaporization (ambient temperature and substrate material dependent). For simplicity, the guideline developed rate was modelled. Case #2 - Valve Packing Leak - Liquid Phase A more likely worst credible release event involves a leak that could occur in the Packing of a valve or a faulty fusible metal plug. This release rate was calculated to be 20.4 pounds per rain.ute of liquid chlorine from a 1/16 inch by 3Technical Guidance for Hazards Analysis, U.S.E.P.A., FEMA, U.S. DOT, Dec., 1987. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XlI-3 Luff Environmental Consulting 1/16 inch orifice. A release duration of 30 minutes was assumed until the leak could be contained. For this release duration, the release rate would be relatively constant. The valve packing release could be mitigated rapidly by closing the container valve. Containment of the release from the fusible plug would be made utilizing a Chlodne Institute, Inc.'s Emergency Kit "B" (see Section XIV - Implementation of the RMPP). It should be noted that documents received from the Institute state that the Emergency Kit "B" can be installed in aPproximately 8 minutes. Allowing time to discover the problem, don protective clothing and respirators, and install the Kit accounted for the half hour response time. This response time is a very conservative estimate considering the availability of respirators and emergency protective clothing (rain suits and chemical resistant gloves). Even though the release was a liquid, it was assumed that no pooling of the chlorine would take place and that the entire stream would vaporize almost immediately upon release. The release rate in this event would likely be significantly reduced within a few minutes by a very simple adjustment to the system. Since the chlorine container is stored on trunnions, the vessel could be easily rotated to place the leak point in the vapor phase of the chlorine. This simple operational task would reduce the release rate by approximately 700 percent. Case #3 - Total Release - Vapor Phase A vapor phase case similar to Case #1 above was also reviewed. This worst case release event was developed assuming a total failure of a fusible plug or shearing off of the vacuum regulator from the valve. Response and containment of the leak was assumed to be 30 minutes as in Case #2. The initial release rate would be very high, approximately 80 pounds in the first two minutes. After this initial release, the release rate would become more constant at about 13 pounds per minute for approximately 10 more minutes. After a fifteen minute duration, the release rate would reduce down to approximately 8.5 pounds per minute. For the purposes of the air model, an average release rate of 14.6 pounds per minute was used. Gist-brocades Food Ingredients Inc./RMPPIOctober, 1989/Page XII-4 Luff Environmental Consulting Case #4 - Valve Packing Leak - Vapor Phase In a similar event to Case #2, a vapor release event was developed for a leak in the packing of a valve or a limited failure of the fusible plug. The area of the leak was assumed to be .0039 square inches (1/16 inch by 1/16 inch)..This opening resulted in a calculated initial release rate of 3.0 pounds per minute. Response time was again assumed to be 30 minutes, Since. the release rate was relatively Iow and the container pressure would not significantly drop due to the discharge, the initial release rate was used in. the air model. Vacuum Feed Regulator Chlorine feed vacuum regulators mount directly on the one ton container valves with a yoke type clamp. This clamp has a tongue and groove flange which utilizes a lead gasket to seal the mating surfaces. Pressure from within the cylinder and the closing' spring keep the inlet'safety valve closed until there is sufficient vacuum on the diaphragm to open the valve. If the vacuum is lost for any reason, the closing spring will drive the inlet valve closed. Since the vacuum feed regulator system only delivers chlorine into an evacuated pipeline, any leak in the chlorine piping system would result in a loss of vacuum and a decrease in chlorine flow. Vacuum is created for the chlorination system when water flows through an ejector located on a branch of the discharge of the cooling tower water circulation pumps. Several modes of failure that could result in a release from the vacuum regulator 'were investigated by the HazOp team. Literature from the Chlorine Institute, Inc. was also used to help define parameters of potential failures. Case #5 - Vacuum* Regulator Installation Release When the chlorination system is assembled, the container valve is open and pressure exists between the valve and the vacuum regulator. The improper installation of the regulator could result in a leak between the container valve and the vacuum regulator. Operating procedures at the Bakersfield Plant are in place to minimize the potential for a release, as follows. Each time a regulator is installed on a container, a new lead gasket is utilized. After installation, the container valve~is briefly opened and closed to provide pressure to the regulator. A squeeze bottle of ammonium hydroxide is used to Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xll-5 Luff Environmental Consulting provide ammonia vapoi's near the gasket to locate any leaks. If a leak is present, a localized white cloud will form when the ammonia reacts with the chlorine gas. Once the leak is located, the failed part would be replaced. When the installation of the regulator is verified to be leak free, the container valve is opened no more than a half turn. Release from the installation of the regulators Would be very minor due' to the minimal amount of chlorine available as a result of only briefly opening the valve. 'A release of this type would certainly be less than the previous cases, 'and therefore, was not modeled. Case #6-Vacuum Regulator Gasket Release During steady state operations of the chlorine system, it is possible that a leak from the .gasket could occur as a result of an external force, such as an earthquake. The likelihood of this leak being significant is Iow due to the precision machined tongue and groove matching flanges on the container and the regulator. If a leak were to occur, the available flow area for this type of leak is very small as a result of the design of the flanges. The flow rate of a release from thi~ case would be less than the cases described under the "ton container" cases above. For this reason, this event was not modeled. Additionally, the leak could be stopped by. simply turning off the container valve. Case #7 - Vacuum Feed Regulator Diaphragm Failure · Failure of the diaphragm in the regulator was also investigated. This event would probably not result in an atmospheric release during normal operating conditions; and therefore, was not modeled. During normal operations, the diaphragm failure would result in a loss of vacuum and a lower chlorination rate. This failure would be detected during the daily inspection of the system by a Iow chlorine residual in the cooling tower water. When the system is not operational, the inlet safety valve should seal off the chlorine flow. If the inlet safety valve does not completely seal, the chlorine could escape through the atmospheric vent on the regulator. The available flow area for this release would again be less than in the cases concerning the ton container. Mitigating this type of release is as simple as shuting off the container valve. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xll-6 Luft Environmental Consulting Residual Chlorine Analyzer Failure of the residual chlorine analyzer could not result in a.direct release of chlorine. If the analyzer failed with a "no residual" reading signal, it would tend to drive open the automatic chlorine flow control valve and super chlorinate .the cooling tower water. This would be discovered in the daily inspections on the cooling tower system. Automatic Chlorine Flow Control Valve The automatic chlorine flow control valve is operated on a vacuum. Any sealing device failure would result in lower chlorine flow into the cooling tower water system and would be discovered by operations personnel during daily inspections. If the automatic chlorine flow control valve were driven open by a faulty residual chlorine analyzer, the demonstrated maximum flow rate available would only be 50 pounds per day. This flow rate is approximately two times the normal rate and would be discovered as high chlorine residual during the daily inspections. There is no potential for a significant atmospheric release from this device. Rotometers Failure of the rotometers would result in lower chlorine flow as in the other vacuum operated devices. There is no potential for a significant atmospheric release. Ejector Ejector failure would result in limited or no chlorine flow. for a significant.atmospheric release of chlorine. There is no potential Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XII-7 Luff Environmental Consulting XllI. OFFSITE CONSEQUENCE ANALYSIS Section 25534 (d)(2), Division 20, California Health and Safety Code .states that "[t]he RMPP shall consider..., [f]or the hazards identified in the hazard and operability studies, an offsite consequence analysis which, for the most likely hazards, assumes pessimistic air dispersion and other adverse environmental conditions". Additionally, as added to the statute, effective January 1, 1989, Section 25534.1 requires that "[e]very RMPP... shall give consideration to the proximity of the facility to schools, general acute care hospitals, and long term health care facilities". For the release cases identified in the HazOp study, an air dispersion model was used to evaluate the offsite consequences of these events. The model and its results are discusSed in this section. A. GENERAL MODEL INFORMATION The pollutant dispersion results were generated by the CAMEO~II-ALOHA model developed by the Hazardous Materials Response Branch of the National Oceanic and Atmospheric Administration. CAMEO~II is the Computer-Aided Management of Emergency Operations program which was designed to help emergency planners and first responders both plan for, and safely handle, chemical accidents. ALOHA stands for Areal Locations of Hazardous Atmospheres and serves as a tool for estimating the movement and dispersion of an atmospheric pollutant. ALOHA is an atmospheric dispersion model which plots the distribution of a pollutant gas as a series of Gaussian distributions. The Gaussian. equation describes a bell-shaped or normal curve. Concentration distribution at ground level is calculated and the bell-shape spreads out and gets wider and flatter as the pollutant drifts downwind. The concentration to be calculated is set to a particular numerical value such as parts per million (ppm), or can be the value of a toxicological parameter such as Immediately Dangerous to Life or Health (IDLH). The model will give results using either an instantaneous source release or a continuous source release. When the release of the pollutant occurs during one very short time period, it is modeled as an instantaneous release. An instantaneous release is handled as a series of puffs. As the wind carries the puff Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xill-1 Luff Environmental Consulting away, the puff spreads out in all directions (see Figure Xlll-1). The puff of vapor becomes larger and less dense as it drifts, but the core which exceeds the concentration value given will eventually get smaller. The plume ends when the spreading has reduced the central concentration of the puff to below the value being calculated. Vapor puff Part of the puff that exceeds thresho~ntr~ Pollutant~ FIGURE XlII-1 - Top View of a Footprint from an Instantaneous Spill A continuous release occurs when a pollutant is being released over a longer period of time. A continuous release results in a concentration curve termed a footprint (see Figure XIII-2). The area inside the curve is the region that is predicted to have ground level concentrations above the limit set by the modeler. Pollutant Source ~ FIGURE XlII-2 - Footprint from a Continuous Source Spill The ALOHA air dispersion model incorporates several assumptions regarding chemical source, meterological data and terrain. The chemical spill is assumed to have occurred at ground level and all concentrations are also calculated at Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xlll-2 Luff Environmental Consulting ground level. The model is not designed to handle elevated sources. Meterological data includes atmospheric stability, wind speed and wind direction. Atmospheric stability classes are measured from A to F with class A being the most unstable and class F the most stable. Unstable conditions result ina large amount of mixing of the atmosphere, causing the pollutant chemical to prodUce a shorter threat distance but a wider plume. However, this condition also' includes more variable wind directions, resulting in 'a threat zone that may tend to meander. Stable conditions give opposite results. The stability can be influenced by both heating and mechanical stirring of the atmosphere. Heating of the surface layer of the atmosphere (daytime) leads to unstable conditions and ground cooling (nighttime) results in more. stable conditions. Classes A to C occur during daytime, classes E and F occur at night and class D may occur during day or night. Mechanical stirring is caused by the winds, with strong winds tending to cause neutral stability (classes C and D). Wind speed and wind direction are input directly by the modeler; although the air model checks that wind speeds are consistent with the stability class chosen. If the wind speed and stability class are inconsistent, no plume dispersion is calculated by the model. The turbulence caused as the wind flows over and around obstacles is included in. the 'model and termed ground roughness. The model adjusts for mixing consistent with stability class by allowing for either rural ground roughness, a terrain with few obstacles, or an urban ground roughness, characterized by many obstacles. The effects of terrain on the speed and direction of the plume are not included in the model. The model assumes that the winds are uniform throughout the plume. B. LIMITATIONS OF ALOHA DISPERSION MODEL Even with all the above mentioned factors included in the plume concentration calculations, as with all dispersion models, the ALOHA model does have several limitations. It is important to remember that the model has a nominal accuracy of a factor of 2, so a predicted concentration of 50 ppm may actually be in the range of 25 to 100 ppm. This degree of accuracy is consistent with other dispersion models. The model also does not accurately represent several conditions which are discussed below. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XIII-3 Luft Environmental Consulting A general limitation of all Gaussian plume models is that they are not accurate at Iow wind speeds (less than 1 mile per hour) or at very stable atmospheric conditions, which produce these Iow wind speeds. As mentioned earlier, the model does not include terrain steering effects caused by topography or wind shifts and the model assumes that all winds are constant throughout the plume. Wind variation can reduce the accuracy of the results when the plume travels more than a mile from the source. The model can also only give plume dispersion results at distances less than 10 miles. Another situation not handled well by the air model is concentration patchiness in the area 50 to 100 yards from the source. This area is termed "pre-Gaussian". In this vicinity, the pollutant will meander and be more patchy than the model can predict. After about 100 yards, the plume has experienced enough mixing eddies to assume a Gaussian distribution. The ALOHA model is unable to handle the initial plumes resulting from heavy gases (gases with a molecular weight greater than air) using the Gaussian equations. The release of a chemical under pressure which results in the formation of an aerosol mist will behave as a heavy gas. In a heavy gas release, the plume initially sinks and then gravity and heating disperse the plume. As the cloud is diluted and the density approaches that of air, turbulence 'is solely responsible for plume spread. When the heavy gas reaches about a one percent concentration, the plume assumes a Gaussian distribution and the model assumptions are valid. This transition usually takes place quickly (within 50 yards). Thus, there are only minor errors in the model's estimation of threat distances for small spills of chemicals with levels of concern in the parts per million range. Errors can be more significant with large spills, but the likelihood of a complete release is remote. A chemical released under pressure will always result in a variable source rate problem which cannot be handled by the model. Any continuous release must be constant and not decrease over time. A pressurized release can result in a rapid initial evaporatiOn or flash-boil which can lead to the sudden release of a large amount of chemical that cannot be properly represented by the model, The model will warn'if a flash-boil is a problem. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XIII-4 Luff Environmental Consulting C.. EVENTS MODELED FOR GIST-BROCADES For Gist-brocades Food Ingredients Inc., the chemical that was modeled is chlorine stored under pressure at ambient temperature. Chlorine gas is. greenish-yellow with a bleach-like odor. Chlorine has a molecular weight of 71 and is typically stored under sufficient, pressure to liquify it at ambient temperature. Chlorine is'shipped as a nonflammable, poisonous gas and is an irritant to the eyes, skin and lungs. Exposure of the skin to liquid chlorine will cause frostbite. Chlorine has a perceptible odor as Iow'as 0.09 - 0.2 ppm and an odor threshold of 3.5 ppm. A release of this chemical may result in the formation of an aerosol mist which, initially, behaves as a heavy gas. The dispersion modeling results are still valid, as the chlorine behaves in a normal Gaussian distribution once the concentration drops to about 10,000 ppm. For a small release, this will occur within the first 50 yards from the source. Because the chemical modeled is under pressure, the rate of release will be variable. This model limitation is overcome by assuming a known quantity of chemical is being released in a specified period of time. Since chlorine is stored at the plant in ton containers in both liquid and vapor phases, release events for each phase were' modeled. As discussed in the Sections pertaining to the HazOp study, four release cases were modeled for Gist-brocades Food Ingredients Inc. Case #1 - Total Release - Liquid Phase In accordance with the U.S. Environmental Protection Agency (EPA) "Technical Guidance for Hazards Analysis", the worst case release event assumes that the entire contents of a full one ton chlorine container would be released in a 10 minute period. With a maximum container capacity of 2000 pounds, the release rate would be 200 pounds per minute. This worst case release event was developed using the subject guidance document. No consideration was given to liquid pooling and evaporation of the chemical over time. However, a liquid release of this type would like!y develop a liquid pool. Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page XIII-5 Luff Environmental Consulting The actual' rate of atmospheric release during a liquid chlorine discharge would be dependent on-the percent of liquid that would flash to vapor upon release and the rate of evaporation of the pooled liquid. Liquid'chlodne vapor flash is a function of storage temperature. At 85 °F, approximately 20 percent of the fluid would flash to vapor immediately upon exposure to atmosPheric pressure. Evaporation of the chlorine pool would be proportional to the sudace area of the liquid pool,, the type of substrate the pool formed on, and the available heat of vaporization (ambient temperature and substrate material dependent). Due to the limitations of the model, and for simplicity, the guideline developed rate was modeled. Case #2 - Valve Packing Leak - Liquid Phase A more likely worst credible release event involves a leak that could occur in the packing of a valve or a faulty fusible metal plug..This release rate was calculated to be 20.4 pounds per minute from an estimated 1/16 inch by 1/16 inch orifice. The initial rate was assumed to be relatively constant, since this release rate represents only one percent of container capacity .per minute. Therefore, the event was modeled as a continuous release. The valve packing release could be mitigated rapidly by closing the container valve. Containment of the release from the fusible plug would be made utilizing a Chlorine Institute,. Inc.'s Emergency Kit "B" (see Section XIV - Implementation of the RMPP). It should be noted that. documents received from the Institute state that the Emergency Kit "B" can be installed in approximately 8 minutes. Allowing time to discover the problem, don protective clothing and respirators, and install the Kit accounted for a half hour response. This response time is a very conservative estimate considering the availability of respirators and emergency protective clothing (rain suits and chemical resistant gloves). Even though the release was a liquid, it was assumed that no pooling of the chlorine would take place and that the entire stream would vaporize almost immediately upon release due to the limitations of the model. A liquid pool with a surface area of approximately three hundred square feet could evaporate liquid chlodne at this rate. The release rate in this event could be significantly reduced within a few minutes by a very simple adjUstment to the system. Since the chlorine Gist-brocades Food Ingredients InC./RMPP/October, 1989/Page XIII-6 Luft Environmental Consulting container is stored on trunnions, the vessel could be easily rotated to place the leak point in the vapor phase of the chlorine. This simple operational task would reduce the release rate by approximately 700 percent.. Case #3 - Total Release - Vapor Phase A vapor phase case similar to Case #1 above was also reviewed. This worst case release event was developed assuming a total failure of a fusible plug or shearing off of the vacuum regulator from the valve. Response and containment of the leak was again assumed to take 30 minutes. The initial release rate is very high, approximately 80 pounds in the first two minutes. After this initial release, the release rate would become more constant at about 13 pounds per minute for approximately 10 more minutes. After roughly a fifteen minute duration, the release rate would drop to approximately 8.5 pounds per minute. For the purposes of the air model, an average release rate of 14.6 pounds per minute was used. Case #4 - Valve Packing Leak - Vapor Phase In a similar event to Case #2, a vapor release event was developed for a leak in the packing of a valve or a limited failure of the fusible plug. The area of the leak was assumed to be .0039 square inches (1/16 inch by 1/16 inch). This opening resulted in a calculated initial release rate of 3.0 pounds per minute. Response time was again assumed to be 30 minutes. Since the release rate was relatively Iow and a fairly short duration, a significant percentage of the container contents would not be released. For this reason, the initial release rate was assumed to be constant and was Used in the air model. In all cases, the releases were modeled for concentrations set at 25 ppm and 2.5 ppm. A value of 25 ppm is considered as Immediately Dangerous to Life or Health (IDLH)I. At 25 ppm, there is severe irritation of the eyes, ears, nose and throat; but there are no lasting effects for a short exposure of less than 30 minutes. A value of 2.5 ppm is considered by EPA to be the Level of Concern (LOC). At 2.5 ppm, it is believed that nearly all individuals could be exposed for up to one hour without developing life threatening health effects. Because the chlorine odor is perceptible at concentrations as Iow as 0.09-0.2 ppm, it is IlDLH levels are defined by NIOSH (National Institute for Occupational Safety and Health), U. S. Department of"Health and Human Services. Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page Xlll-7 Luft Environmental Consulting unlikely that a person would become unknowingly overexposed. Pursuant to Assembly Bill 3205 of 1988, "[e]very RMPP. shall give consideration to the proximity of the facility to schools, general acute care hospitals, and long-term health care facilities." Sensitive population sites considered in this offsite consequence analysis included residences, schools, and health care facilities.. Sensitive population sites identified near Gist-brocades are shown in Figure XIII-3 and listed in Table XIII-1. The impact of the chlorine dispersion was evaluated at these sites. The meteorological conditions used to determine maximum impact distances and times were based on data collected by the U.S. Weather Service at Meadows Field Airport in Bakersfield, CA. This data is summarized in Table XIII-2 and includes monthly and annual distributions of wind speed and wind direction. The lowest monthly average wind speed of 5.0 miles per hour was used in the model since lower wind speeds generally give the most pessimistic dispersion results. A wind direction prevails from the northwest nine months of the year and was used to represent the most common conditions. The area southeast of the plant which would be impacted by the prevailing northwest winds consists mainly of single-family residences and rural areas, with a few schools and. no health facilities. The wind direction will average from the east-northeast for three months of the year. The area southwest of the plant impacted by this condition is generally rural with a few housing developments and no existing schools or health facilities. Sensitive population sites directly north and to the northwest of the plant would be impacted on the rare occasion when winds are from the south or southeast. This condition is usually caused by storm activity which results in good mixing of the air and rapid dispersal of any plume which may occur. Most of the ALOHA limitations discussed previously have minimal effects on the Gist-brocades modeling results. The ALOHA limitation regarding terrain steedng effects and Wind shifts caused by topography is not a problem in the area being modeled because the terrain is generally flat and unobstructed by large hills or valleys. The ground roughness of the area is determined to be rural terrain. Effects of concentration patchiness within the first 50 to 100 yards from the source will be minimal because there are no sensitive populations within this pre- Gaussian area. Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page Xlll-8 Luft Environmental Consulting GIST-BROCADES PLANT [~ STOCKDALE HWY c3 ~ w z n" CSUB n-3:: - ~- 0 u~. I=: ~ m ~h ~o ~ . '~~ ~ GIST-SRO :ADES WHITE LN · m X PACHECO RD > PANAMA LN HOSKING AVE LEGEND ~ ~ - School 1~ - Nearest Residence [~] -Fire Station ~1 - Proposed School ~ - Hospital I Mile = 1 Inch FIGURE XlII-3 - Map showing the sensitive population sites near the Gist-Brocades plant on District Blvd. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xlll-9 Luft Environmental Consulting TABLE XlII-1 GIST-BROCADES PLANT SENSITIVE POPULATION SITES SITE APPROXIMATE DISTANCE FROM PLANT (Miles) CRITICAL WIND DIRECTION (Wind From) Nearest Residence 0.38 SOuth or Northwest Nearest Hospital 0.37 South Nearest School 1.07 Southwest Nearest Fire Station 2.13 Southeast Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page Xlll-10 Luff Environmental Consulting TABLE Xlll-2 METEROLOGICAL DATA from Meadows Field Airport MONTH WIND SPEED WIND DIRECTION (mph) (from)* Jan 5.2 NW Feb 5.8 ENE Mar 6.5 NW Apr 7.1 NW May 7.9 NW Jun 7.9 NW Jul 7.2 NW Aug 6.8 NW Sep 6.2 WNW Oct 5.5 NW Nov 5.1 ENE Dec 5.0 ENE Annual Average 6.4 NW * Critical Wind Direction indicates direction of origin of the wind. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XII1-11 Luft Environmental Consulting D. RESULTS OF THE ALOHA DISPERSION MODELING The results of the ALOHA plume dispersion modeling are summarized in Tables XlII-3 and Xlll-4. Tables XlII-3 and XlII-4 give the total plume dispersion distance and the total dispersion time at vadous stability classes for each release event. In the unlikely worst case-event, Case #1, a complete release of 2000 pounds of chlorine in 10 minutes, the plume of 2.5 ppm would reach the nearest residences under the worst case (pessimistic) meteorological conditions with a wind from the northwest. At concentrations of 25 ppm for the same case, the plume would also impact the nearest residences. For Case #2, the accidental release of 20.4 pounds per minute of liquid chlorine for 30 minutes, the dispersion distance of the chlorine plume would be reduced significantly for both the 2.5 and 25 ppm concentrations. Total dispersion times for the plumes would be 24 and 6 minutes respectively. Both plumes would impact only the nearest residences southeast of the plant. With the worst case vapor phase release, Case #3, the release of 14.6 pounds per minute of chlorine over 30 minutes was modeled. Plumes of 2.5 and 25 ppm concentrations could reach the nearest residences. This Case was modeled assuming the most pessimistic meteorological conditions with a wind from the northwest. The total dispersion time for the 2.5 ppm concentration was 1'9 minutes; while, the time in the 25 ppm case was 5 minutes. Case #4, a vapor phase worst credible release event assumed an accidental release of 3.0 pounds per minute of chlorine 'for 30 minutes. In this event, the dispersion distances of the chlorine are reduced significantly for both the 2.5 and 25 ppm plumes. The 2.5 ppm plume would impact only the nearest residences. Total dispersion time for the 25 ppm case was 7 minutes. The 25 ppm plume, with a total dispersion time of 2 minute, would not impact any sensitive population sites. In order to show maximum plume dispersion (maximum impact), a stability class of "F" was used in the ALOHA model for Cases 1 through 4. Stability class "F" is indicative of stable atmospheric conditions on a very cool night with weak winds. However, a stability class of "A" is also very common to the Bakersfield area during the summer mgnths. Stability class "A" is indicative of unstable Gist-brocades Food Ingredients IncJRMPP/October, 1989/Page XII1-12 Luft Environmental Consulting fi) o o CL I Al:ILI- J~lll-~ '" GIST-BROCADES PLANT ~ PLUME DISPERSION RESULTS for CHLORINE- LIQUID PHASE AMOUNT OF PLUME WIND WIND STABILITY CLASSES CHLORINE CONCENTRATION SPEED DIRECTION A B C D E F RELEASED (PPM) (MPH) (FROM) UNSTABLE UNSTABLE NEUTRAL NEUTRAL STABLE STABLE DAY DAY DAY DAY OR NIGH-[ NIGHT ~ NIGHT 20.4 LBS/MIN 2.5 5.00 NORTHWES-[ 286 YDS* 434 YDS 662 YDS .63 MILES 1.01 MILES 2.01 MILES CONTINUOUS 2 MIN** 3 MIN 5 MIN 8 MIN 12 MIN 24 MIN 20.4 LBS/MIN 25 5.00 NORTHWES-I' 90 YDS 136 YDS 203 YDS .17 MILES .27 MILES .48 MILES 'CONTINUOUS 37 SEC 56 SEC 1 MIN 2 MIN 3 MIN 6 MIN 200 LBS/MIN 2.5 5.00 NORTHWEST .52 MILES .79 MILES 1.29 MILES 2.80 MILES 5.68 MILES >10 MILES CONTINUOUS 6 MIN 9 MIN 15 MIN 34 MIN 1HR 8 MIN -- 200 LBS/MIN 25 5.00 NORTHWEST 283 YDS 429 YDS 656 YDS .62 MILES 1.00 MILES 1.98 MILES CONTINUOUS ' 2 MIN 3 MIN 4 MIN 7 MIN 12 MIN 24 MIN BOLD: DATA USED FOR PLUME DISPERSION MODEL. *: TOTAL DISPERSION DISTANCE OF PLUME. **: TOTAL DISPERSION TIME OF PLUME. <. CL tl) 0 0 TABLE Xlll-4 GIST-BROCADES PLANT PLUME DISPERSION RESULTS for CHLORINE - VAPOR PHASE AMOUNT OF PLUME WIND WIND STABILITY CLASSES CHLORINE CONCENTRATIOI~ SPEED DIRECTION A B C D E F RELEASED (PPM) (MPH) (FROM) UNSTABLE UNSTABLE NEUTRAL NEUTRAL. STABLE STABLE DAY DAY DAY DAY OR NIGHT NIGHT NIGHT 3.0 LBS/MIN 2.5 5.00 NORTHWEST109 YDS* 165 YDS 247 YDS .21 MILES .33 MILES .60 MILES, CONTINUOUS 45 SEC** I MIN 2 MIN 3 MIN 4 MIN 7 MIN 3.0 LBS/MIN 25 5.00 NORTHWEST 34 YDS 52 YDS 77 YDS 108 YDS 175 YDS .17 MILES CONTINUOUS 14 SEC 21 SEC 32 SEC 44 SEC I MIN 2 MIN 14.6 LBS/MIN 2.5 5.00 NORTHWEST 241 YDS 366YDS 557YDS .52 MILES .82 MILES 1.59 MILES CONTINUOUS 2 MIN 3 MIN 4 MIN 6 MIN 10 MIN 19 MIN 14.6 LBS/MIN -25 5.00 NORTHWES'I' 76 YDS 115 YDS 172 YDS .14 MILES .23 MILES .40 MILES CONTINUOUS '31 SEC 47 SEC I MIN 2 MIN 3 MIN 5 MIN BOLD: DATA USED FOR PLUME DISPERSION MODEL. *: TOTAL DISPERSION DISTANCE OF PLUME. **: TOTAL DISPERSION TIME OF PLUME. c) 0 r' atmospheric conditions characterized by strong sunshine and weak winds. Tables XIII-3 and XIII-4 show that when using "A" stability, in the event of a complete release, the 2.5 ppm plume would travel a maximum.distance of 0.52 miles. Under those conditions, the worst case release would only impact the nearest residences around the plant site. The plumes of any other releases would not reach even the closest sensitive population sites. Offsite consequences of events were reviewed that addressed both liquid and vapor phase releases. The worst credible event in each case showed that exposure· levels to the sensitive population sites would probably not represent life threatening conditions or irreVersible health effects. Gist-brocades Food Ingredients Inc./RMPP/Oct0ber, 1989/Page XII1-15 Luft Environmental Consulting XlV. IMPLEMENTATION OF'THE RMPP Section 25534 (c)(5) of the statute requires that the RMPP include "[a] schedule for implementing additional steps to be taken by the business, in response to the findings of the assessment performed.., to reduce the risk of an accident involving acutely hazardous materials. These actions may include any of the following: (A) Installation of alarm, detection, monitoring, or automatic control devices. (B) . Equipment modifications, repairs, or additions. (C) Changes in the operations, procedures, maintenance schedules, or facility design." Additionally, Section 25534 (k) states that, for an existing facility, "the handler shall implement all activities and programs specified in the RMPP within one year following the certification. Implementation of the RMPP shall include carrying out all operating, maintenance, monitoring, inventory control, equipment inspection, auditing, recordkeeping, and training programs as required by the RMPP." In the case of Gist-brocades Food Ingredients INC., as discussed in this RMPP Supporting Document, the facility utilizes chlorine as a biocide for the cooling tower water. The facility currently has written operating procedures that pertain to the storage, handling, and operation of the chlorination system. The following implementation plan for the RMPP pertains to the chlorination system equipment. A. ADDITIONAL'STEPS REQUIRED Gist-brocades Food Ingredients Inc. has committed to improving the storage and handling procedures associated with the chlorination system. The present system has the ton containers resting on trunnion supports for both storage and operating conditions. One of the mitigation measures to be undertaken by Gist- brocades Food Ingredients Inc. is the installation of a tie-down system for the containers. This tie-down system will immobilize the containers in the event of an earthquake. The system will also be equipped with a quick release device that will afford repositioning the vessel on the trunnions in the event of a liquid release. In the~ event of a release, Gist-brocades Food Ingredients Inc. has also taken delivery of a Chlorine Institute, Inc. Emergency Kit "B" to better respond to the Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XIV-1 Luft Environmental Consulting situation. Training on the installation of the Kit was part of the purchase package. Self contained breathing apparatus (SCBA) and protective clothing are already at the facility to augment the response of the on-site personnel. An operational procedure was also changed to help mitigate one of the release events. During the loading and off loading of containers, the chlorine container valve on the active system will be closed. This will'prevent the release of chlodne if the vacuum regulator is sheared off by the container hoist spreader bar. B. IMPLEMENTATION SCHEDULE All activities and programs specified in this RMPP shall be implemented within one year of the certification of the RMPP. The emergency response equipment has already been received at the Bakersfield Plant. Gist-brocades Food Ingredients Inc. will complete the work on the tie-down system for the containers within three months of receipt of all of the materials. Gist-brocades Food Ingredients Inc. is committed to implementation of the requirements of the RMPP as soon as possible. Gist-brocades Food Ingredients Inc./RMPP/October, 1989/Page XIV-2 Luft Environmental Consulting Gist-brocades 'Food Ingredients Inc. RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) ADDENDUM FOR ANHYDROUS AMMONIA Prepared By Luft Environmental Consulting 3701 Pegasus Drive, Suite 121 Bakersfield, CA 93308 March 1992 Ila. INTRODUCTION Gist-brocades Food Ingredients Inc.'s Bakersfield plant produces baker's yeast under the label of Eagle Baker's Yeast. The yeast is grown in batch fermentation tanks which are ferti.lized with phosphoric acid, sulfuric acid and ammonium hydroxide. In order to maintain the optimum .growth te, mperature for the yeast, the fermentation tanks are cooled with water from the cooling tower system. Chlorine, which is a listed acutely hazardous material, is utilized as a biocide in the cooling tower system at the plant. A Risk Management and Prevention Program (RMPP) for the chlorine system was prepared at the request of the Bakersfield City Fire Department and was certified as complete on October 19, 1989. In March of 1992, Gist-brocades Food Ingredients Inc. will be adding an. anhydrous ammonia storage tank to the facility. Anhydrous ammonia is also a listed acutely hazardous material. This addendum to Gist-brocades Food Ingredients Inc.'s RMPP addresses the administrative and operational programs associated with the anhydrous ammonia storage system which are designed to prevent acutely hazardous materials accident risks. A. ACUTELY HAZARDOUS MATERIALS (AHM) As an alternative to purchasing ammonium hydroxide, Gist-brocades Food Ingredients Inc. is installing an anhydrous ammonia storage tank and process equipment that will enable them to convert anhydrous ammonia to ammonium hydroxide. Since anhydrous ammonia is a listed acutely hazardous' material, an RMPP must be prepared if requested by the Administering Agency. The Bakersfield City Fire Department has requested an RMPP associated with the anhydrous ammonia system. This addendum to the original RMPP addresses in detail the processes and equipment associated with the I'iandling and storage of anhydrous ammonia. B. REQUIREMENTS OF THE STATUTE Assembly Bill 3777 of 1986 added Article 2 to Chapter 6.95, Division 20, of the California Health and Safety Code. A summary of Article 2 as it pertains to the Gist-brocades Food Ingredients Inc. facility is provided in Section II of the original RMPP. This addendum contains additions to the appropriate sections of the original RMPP which are affected by the addition of the ammonia system. Glst-brocadesJRMPP Addendum/Feb., 1992 Page Ila-1 Luft Environmental Consulting As of January 1992, section 25534(d)(2) requires the offsite consequence analysis to include "a clearly prepared map noting the location of the facility which shows the populations considered . . . and the zones .of vulnerability including the levels of expected exposure in each zone" for the hazards identified during the hazard and operability study. This RMPP addendum complies with.the requirements of Article 2, Chapter 6.95, Division 20 of the California Health and Safety Code as of January-1992. Gist-brocades/RM PP Addendum/Feb., 1992 Page Ila-2 Luff Environmental Consulting III. CERTIFICATION BY QUALIFIED PERSON AND FACILITY OPERATOR Section 25534 (j), H & S Code requires that the RMPP shall be certified as complete by a qualified person and the facility operator. These certifications are provided below. I certify that I am qualified to attest to the validity of the hazard and operability studies performed pursuant to Section 25534, and the relationship between the corrective steps taken by the handler following the hazard and operability studies and those hazards which were identified in the studies. Additionally, I certify that this risk management and prevention program is complete. This certification is based on my understanding that the data and documents provided by Gist-brocades Food Ingredients Inc. are true and correct and that the plans, programs, and procedures will be implemented as described. Signature Karl W. Luff. P.E.. R.E.A. Name Principal Mechanical Engineer Title Date As facility operator, I hereby certify that this risk management and prevention program is complete and will be implemented. / ~.~ Plant Manager Signature 'Frtle Robert Deedy -~//'///..~ ~.-- Name Date Gist-brocades/RMPP Addendum/Feb., 1992 Page Ilia-1 Luft Environmental Consulting IVa. ACCIDENT HISTORY Section 25534 (c)(1) of the statute requires that the RMPP include "[a] description of each accident involving acutely hazardous materials which has occurred at the business or facility within three years from the date of the request [for the RMPP], together with a description of the underlying causes of the accident and the measures taken, if any, to avoid a recurrence of a similar accident." A. DESCRIPTION OF ACCIDENT Gist-brocades Food Ingredients Inc. has not had an accident involVing chlorine or anhydrous ammonia at the Bakersfield Plant. B. ACCIDENT PREVENTION MEASURES This subsection is not applicable since there have been no accidents involving acutely hazardous materials at the Bakersfield Plant. Gist-brocades/RMPP Addendum/Feb., 1992 Page IVa-1 Luft Environmental Consulting Va. FACILITY DESCRIPTION In accordance with Section 25534 (c)(2), the RMPP must include "[a] report specifYing the nature, age, and condition of the equipment used to handle acutely hazardous materials at the business or facility and any schedules for testing and maintenance." This section provides.a description of the facility and equipment which handle anhydrous ammonia. A. GENERAL DESCRIPTION OF FACILITY Gist-brocades FOod Ingredients Inc.'s Bakersfield plant produces baker's yeast. The yeast is grown in a batch process in fermentation tanks. During this batch process, the yeast solution is fertilized in the fermentation tanks with phosphoric acid, sulfuric acid and ammonium hydroxide. After the fermentation process is · complete, the yeast is dried and/or packaged for the industrial market. B. DESCRIPTION OF AHM PROCESS AND EQUIPMENT Gist-brocades Food Ingredients Inc. is currently purchasing ammonium hydroxide at a concentration of approximately 25 percent ammonia. The anhydrous ammonia storage tank and ammonia conversion system are being installed to provide an economical method of making ammonium hydroxide (aqua ammonia) on site. Anhydrous ammonia will be stored on site in a. 15,325 gallon (water capacity) pressur_e vessel. Actual maximum storage capacity for anhydrous ammonia is 13,256 gallons. Ammonium hydroxide will be made by mixing approximately 7 gallons per minute (gpm) of liquid anhydrous ammonia and approximately 15 gpm of water in the ammonia conversion system. Anhydrous ammonia is fed from the storage vessel to the ammonia conversion system via a one inch welded stainless steel pipeline. Water supply for the conversion system will be from the existing water softening system located at the plant. The ammonia conversion system consists of a reaction chamber and a heat exchanger. Inside the reaction chamber, the anhydrous ammonia is added to water through a jet type mixer to make ammonium hydroxide. Since this reaction is exothermic, the ammonium hydroxide will be cooled in a shell and tube heat Gist-brocades/RMPP Addendum/Feb., 1992 Page Va-1 Luff Environmental Consulting exchanger using cooling water from the existing cooling tower. After cooling, the ammonium hydroxide will flow into an existing 20,000 gallon ammonium hydroxide storage tank. The existing ammonium hydroxide distribution system will be utilized to fertilize the fermentation tanks. J Liquid ammonia flow to the conversion system will be controlled by a pneumatic control valve. This pneumatic control valve closes upon loss of air pressure. The air supply to this control valve will be governed by a fail closed (no air flow) electric solenoid valve. Anhydrous ammonia flow into the reaction chamber will be approximately 7 gpm. The ammonia flow will be controlled by two devices; a pressure regulator and a restricted fixed orifice plate. Soft water flow to the reaction chamber will be regulated by a programmable logic controller (PLC) which will modulate a flow control valve. During processing, the PLC will monitor the ammonium hydroxide reaction temperature and the position of the flow control valve. If the ammonium hydroxide temperature leaving the reaction chamber is too high, the PLC will open the soft water control valve. Opening the soft water valve reduces the ammonium hydroxide strength and lowers the heat of reaction in the chamber. Conversely, reducing the flow of soft water into the reaction chamber increases the' ammonium hydroxide concentration and the reaction temperature. Therefore, the PLC can maintain the proper concentration of ammonium hydroxide by maintaining the appropriate reaction temperature. A schematic overview of the system is shown in Figure Va-1. Each key component in the system is discussed below. 1. Anhydrous Ammonia Vessel The anhydrous ammonia storage container is a Welded steel pressure vessel capable of holding 15,325 gallons of water (liquid ammonia capacity of 13,256 gallons, or 86.5% of water capacity). The vessel was designed and constructed in 1974 in accordance with Section VIII, Division 1 of the ASME Boiler and Pressure Vessel Code. A nameplate on the vessel identifies the maximum allowable working pressure (265 psig)and temperature (118°F) as required by the ASME code. The vessel is filled through a dedicated connection designed for a standard two inch ammonia hose. Vapors displaced during the filling operation are returned to the delivery truck via a second dedicated connection and hose. Gist-brocadesJRMPP Addendum/Feb., 1992 Page Va-2 Luff Environmental Consulting Both the liquid and vapor lines are equipped with excess flow valves. The liquid line is also equipped with a back check valve. Both of theSe delivery lines are connected to a bulkhead that is capable of withstanding a 2000 pound pull in any direction. All of the isolation valves, including the back check valve, are protected from any pulling force by this bulkhead. The ammonia storage tank is equipped with four new pressure relief valves that are set at 265 psig and will operate if the pressure inside the vessel approaches the maximum allowable working pressure. These pressure relief valves are required to flow a minimum of 6835 cubic feet per minute in the event of a fire. The tank is .firmly attached to a concrete foundation' which was designed and installed for seismic Zone 4 conditions according the Uniform Building Code. Excess flow valves are also provided on the ammonia storage tank. The excess flow valve on the liquid ammonia line is set at 225 gpm. The vapor return line (fill connection) is equipped with a 95 gpm excess flow valve. Additionally, there is a 24 gpm excess flow valve installed on the one inch liquid line from the anhydrous tank to the ammonia conversion system. All of' these valves will close at their respective set points and prevent the uncontrolled release of ammonia if the piping downstream of the excess flow valve ruptures. 2. Automatic Ammonia Flow Control Valve A pneumatic-flow control valve regulates the ammonia flow to the reaction chamber. This valve is a pressurize to open (fail closed) control valve. Air pressure to the control valve is governed by an electric solenoid valve. The electric solenoid valve is a fail close valve. Commands to open or close the valve originate from the programmable logic controller (PLC) if the system is in the automatic mode. The ammonia flow valve can also be opened manually. In order to open the ammonia control valve manually, the operator must satisfy two conditions. First, the operator must use a key to turn the keyed switch from auto to manual. Second, water flow to the reaction chamber must be established. The water flow is verified by the PLC using a water flow indicator switch. This "water flow permissive" prevents the flow of anhydrous ammonia to the reaction chamber if the soft water supply is not present to complete the reaction to Gist-brocades/RMPP Addendum/Feb., 1992 Page Va-3 Luft Environmental Consulting ammonium hydroxide. 3. Reaction Chamber The reaction chamber is simply a mixing vessel where the liquid ammonia and water are mixed to produce ammonium hydroxide in a process that releases heat. Liquid' ammonia is introduced into the chamber through a jet type nozzle. Soft water flow into the reaction chamber is regulated by a variable position control valve. This' control valve regulates the concentration of the ammonium hydroxide by varying the soft water flow rate into the reaction chamber which has a fixed anhydrous ammonia flow rate. Reducing the soft water flow rate increases the ammonia concentration in the ammonium hydroxide. Similarly, increasing the soft water flow rate reduces the ammonia concentration in the final product. Since'this reaction is exothermic, the relative strength of the ammonium hydroxide can also be determined by monitoring the temperature of the ammonium hydroxide as it leaves the reaction chamber. The higher the reaction temperature, the greater the ammonium hydroxide concentration in the solution. Similarly, a Iow 'reaction temperature will indicate a weak ammonium hydroxide solution.. By monitoring the temperature of the ammonium hydroxide leaving the reactor chamber, the PLC can modulate the water control valve to produce the desired reaction temperature and the desired, concentration of ammonium hYdroxide. 4. Aqua Ammonia Heat Exchanger A shell and tube heat exchanger is located downstream of the reaction chamber to cool the ammonium hydroxide before it is fed into the ammonium hydroxide storage tank. Cooling water supplied to the shell and tube heat exchanger is provided by the existing cooling tower system. The PLC operates a variable position flow control valve to regulate the flow of cooling water to the heat exchanger. 5. Aqua Ammonia Storage Tank The aqua ammonia storage tank provides up to 20,000 gallons of storage Gist-brocades/RMPP Addendum/Feb., 1992 Page Va-4 Luft Environmental Consulting capacity for aqua ammonia created by the conversion system. The storage tank functions as a reservoir for aqua ammonia that is slowly metered to the fermentation tanks for optimum yeast growth. C. SCHEDULES FOR TESTING AND MAINTENANCE 1. Equipment Test Schedules The anhydrous ammonia pressure vessel was' originally designed and constructed in 1974 according to ASME pressure vessel codes. The vessel was inspected, pressure tested, and checked for leaks 'by California Tank Lines before delivery to Gist-brocades Food Ingredients, inc. Components of the ammonia conversion system are routinely tested after any maintenance or repair work is completed. Additionally, the overall system performance is monitored online during each batch by testing the' specific gravity of the ammonium hydroxide as it leaves the shell and tube heat exchanger. 2. Maintenance Schedules Routine maintenance on the equipment is performed in accordance with the manufacturer's recommended procedures for each major piece of equipment. The equipment is visually inspected daily for leaks from valves and fittings and signs of corrosion. Gist-brocades/RMPP Addendum/Feb., 1992 Page Va-5 Luff Environmental Consulting Via. DESIGN, OPERATING AND MAINTENANCE SAFETY :SYSTEMS Section 25534 (c)(3) requires that the RMPP address design, operating, and maintenance controls which minimize the risk of an accident involving acutely hazardous materials. This section of the RMPP discusses these controls. It also addresses the design safety of the existing equipment, standard operating procedures, and preventive maintenance' programs, as required by Section 25532 (g); and operating and maintenance programs, as required by Section 25534 (k). A. GENERAL The ammonia conversion system is monitored and controlled by an Allen BradleyTM programmable logic controller (PLC). The PLC controls process equipment and provides alarm functions for process' parameters that exceed normal operating limits. These process parameters include temperature, flow, and fluid level. If the alarm conditions are not corrected, resulting in the process parameters returning to normal operating limits, the PLC will activate the shutdown sequence. The shutdown sequence includes closing the anhydrous ammonia control valve and purging the reactor vessel and heat exchanger with water. Upon completion of the water purge, the soft water flow to the conversion system will cease along with the cooling water flow. B. CONTROL SYSTEMS 1. Excess Flow Control The anhydrous ammonia storage tank is equipped with excess flow valves. Each excess flow valve is set to operate at a given maximum flow rate. If this maximum flow rate is exceeded for any reason, the excess flow valve will operate instantaneously to shut off the flow of ammonia. These valves are designed to automatically operate in the event of a hose or pipe fitting failure downstream of the excess flow valve. An exCess flow Valve is a normally open valve which allows fluid flow in both directions.. The valve is held open by the force of the valve spring shown in Figure.Via-1. Although fluid can flow in both directions, the valve will operate and. shUt off flow in only one direction. Fluid will pass through the Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-1 Luft Environmental Consulting excess flow valve (in the direction of the flow arrow) at any rate of flow up to and including the specified maximUm flow rate of the valve. As the flow through the valve increases, the fluid drag on the valve head increases until the specified maximum flow rate is reached. If the specified maximum flow rate is exceeded, the fluid drag on the valve head will exceed the force of the spring and th'e valve head will close against the valve seat shutting off the flow. Direction of FI Figure Via-1 Excess Flow Valve Once the excess flow valve operates, the upstream pressure from the tank will keep the valve closed until the pressure across the valve head can be equalized. Since the excess flow valve is designed with a very small orifice in the valve head, the pressure across the valve head will equalize when 'the downstream line failure is isolated. As the differential pressure across 'the valve head decreases, the closing force on the valve head drops. When the closing pressure drops below the valve spring pressure, the excess flow valve will open. There are three excess flow valves in the anhydrous ammonia storage system. Two of these excess flow valves are in liquid service and one is in vapor service. A three inch excess flow valve set at 225 gallons' per minute (gpm) is threaded directly into the ammonia storage tank. This valve prevents the uncontrolled release of the contents of the ammonia storage tank in the event of a catastrophic liquid line rupture downstream of the excess flow valve. The second liquid line excess flow valve is a one inch valve, set at 24 gpm, installed in the liquid ammonia pipeline that supplies the ammonia conversion system. This valve was installed to prevent an Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-2 Luff Environmental Consulting uncontrolled release from the ammonia supply line and ammonia conversion system in the event of a catastrophic failure in the pipeline or reaction chamber. The excess flow valve that is installed in the ammonia vapor return line is set at 95 gpm. This flow rate is equivalent to approximately 119 cubic feet per minute (at 70°F) or about 51 pounds of ammonia per minute. A schematic diagram of the excess flow valve arrangement is shown in Figure Via-2. Vapor Return Line Excess Flow Valv Bulkhead ¢<1 , i ,'. Liquid Ammonia Line ._.~ 1 Excess Flow Valve - 3 inch Liquid Ammonia Line To Ammonia Excess Flow Valve- 1' inch Conversion S~stem Figure Via-2 Excess Flow Valve Arrangement 2. Ammonia Feed Valve The ammonia feed valve is a pneumatically driven control valve located in the line between the ammonia storage vessel and the reaction chamber. This valve is a fail closed valve and must be pressurized to o~pen. An electric solenoid valve controls the air flow to the pneumatic control valve. If the electric power or instrument air is lost for any reason, the closing springs on the respective valves will drive the valves closed. The PLC based control system ensures that this ammonia valve will only open if soft water is flowing into the reaction chamber. This "water flow permissive" prevents anhydrous ammonia flow if the soft water supply is not functioning properly. Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-3 Luft Environmental Consulting Ammonia Vapor Return Line -1 NH3 TANK Excess Flow Valves ~ Liquid Ammonia Line Excess Flow Valve Pressure Reducing Valve Set at 120 psig AQUA VENT TANK Set at 25 psig To Fermentation Towers Figure Va-1 The ammonia flow valve can also be opened manually. In .order to operate the ammonia control valve manually, the operator must use a key to turn the keyed switch on the control panel from auto to manual. The water.flow permissive must also be satisfied in order to open the solenoid valve. In either the automatic or manual mode, the maximum liquid ammonia flow rate is approximately 7 gallons per minute. This flow rate is governed by a pressure reducing valve and a restricted orifice plate that are installed prior to the pneumatic control valve. 3. Soft Water Flow Control Valve A flow control valve is used to regulate the flow of soft water into the reaction chamber. Since the anhydrous ammonia flow rate is fixed, this control valve regulates the concentration of the ammonium hydroxide by varying the volume of water available for the conversion process. The PLC modulates the water flow based on the ammonium hydroxide reaction temperature at the exit of the mixing chamber. This temperature is monitored by resistance temperature detectors (RTDs) that produce an electrical signal that is transmitted to the PLC. The PLC determines if the reaction temperature is within the normal operating parameters and generates the appropriate output signal for control of the soft water valve. The output signal from the PLC is an electrical Current ranging from 4 to 20 milliamps. Since the soft water control valve is pneumatically operated, the electrical signal from the PLC must be converted to a pneumatic signal. This is accomplished via a current to pneumatic transducer (I to P transducer). The I to P transducer receives the 4 to 20 milliamp signal from the PLC and converts this to a 3 to 15 psig pneumatic signal for'modulating the control valve. Prior to reaching the control valve, this pneumatic signal passes through a automatic/manual pneumatic controller. In the automatic mode, this device allows the pneumatic signal to pass through unchanged. In the manual mode, the BLC pneumatic signal is blocked and a separate pneumatic signal from the instrument air sYstem is used to control the soft water valve. This pneumatic signal from the instrument air system is manually adjusted by an air pressure regulator mounted in the controller. The pneumatic Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-4 Luff Environmental Consulting controller allows the operator to make adjustments to the ammonia conversion system in the manual mode. It is important to note that all alarm and shutdown systems are active in this manual mode. 4. Temperature Sensors Temperature sensors located throughout the ammonia conversion system monitor fluid temperatures. Temperatures downstream of the reaction chamber and heat exchanger are monitored to ensure proper operation of the system. A temperature that exceeds a set maximum or falls below a set minimum will initiate a shutdown sequence for the' entire system. As mentioned above, several process temperatures are measured with RTDs. RTDs operate on the principle that electrical resistance changes with temperature in electrical conductors. The resistive value of the element is directly proportional to the temperature of the media surrounding the RTD. The RTD resistive element is typically enclosed in a metal or ceramic casing. In the application for Gist-brocades, the RTDs will be installed in thermowells in the process piping. The control panel is equipped with digital displays for the reaction chamber temperature and the heat exchanger temperature. In addition to the RTDs, there are also several temperature gauges installed along the process piping for confirmation readings by the operator. 5. Cooling Water Control Valve Cooling water flow for the shell and tube heat exchanger is regulated by a pneumatic control valve. The control system for this valve is virtually identical to the soft water control valve discussed in item #3 above. An RTD provides temperature signals to the PLC, which signals the I to P transducer for pneumatic manipulation of the cooling water control valve. 6. Level Indicators There are three level indicators in the ammonia conversion system. Two of the level indicators are dedicated to the anhydrous ammonia tank. A mechanical displacement type gauge was supplied with the ammonia tank. In addition to this gauge, a differential pressure type gauge was added to the system to provide an electdc signal back to the yeast process control room. This differential pressure gauge also has a local display near the tank. Both of these gauges on the anhydrous tank will be used to verify the tank level Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-5 Luft Environmental Consulting for ordering purposes. There is also a differential pressure type gauge on the ammonium hydroxide tank. This level indicator will send a 4-20 milliamp electrical signal to the PLC and provides for a digital display located on the control panel that shows tank level. The PLC will initiate alarm functions for the ammonium hydroxide tank for Iow level, high level and high-high level. The high-high level will initiate an immediate shutdown of the ammonia conversion system. 7. Flow Indicators There are two flow indicating switches installed in the ammonia conversion system, one floTM indicator is installed in the soft water supply line. This indicator signals the PLC when a minimum of 12 gallons per minute of soft. water is flowing to the reaction chamber. This is the water flow permissive signal that allows the PLC to open the anhydrous ammonia control valve. The second flow indicator installed in the system is the ammonia flow indicator. This indicator is installed in the one inch liquid ammonia line in case ammonia flow is established when the system is not in the conversion process. If a tenth of a gallon per minute of ammonia flows through this indicator, the PLC will set off a local ammonia flow alarm. C. SAFETY EQUIPMENT 1. Pressure Relief Valves The anhydrous ammonia vessel is equipped with four new pressure relief valves, set at 265 psig, located on two 2" relief valve manifolds. This relief system is capable of flowing the minimum requirement-of 6835 cubic feet per minute in the event of a fire. The existing ammonium hydroxide tank is also equipped with a pressure relief valve set at 25 psig. The ammonium hydroxide tank is also equipped with a back pressure regulating valve that will vent the tank vapors during the filling process. All displaced tank vapors will be bubbled through a 800 gallon tank filled with water. When sufficient ammonia accumulates in the vent tank, the contents of the tank will be' pumped into the ammonium hydroxide tank. The vent tar~k will then be refilled with soft water. Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-6 Luft Environmental Consulting 2. Respirators Several 'respirators and three self contained breathing apparatus (SCBAs) are on site for response to emergency situations. The respirators are full face canister type and are located throughout the plant. One of the self contained breathing apparatus is equipped with an air line that is connected to a large air cylinder on a portable cart. The' other two SCBAs have standard 30 minute air bottles. All of the respirators are located in separate geographical areas within the plant. These separate locations of the respirators ensure that emergency response personnel can obtain the necessary personal protective equipment during an emergency. 3. Emergency Showers Emergency showers equipped with eye wash stations are located close to the anhydrous ammonia tank, ammonium .hydroxide tank, and throughout the facility. D. FIRE PROTECTION SYSTEM Hydrants/Hose Reels There are three fire hose houses (stand pipe with hose) in the immediate area of the ammonia conversion system. One of the hose houses is. located adjacent to the chlorine containers. The second hose house is to the' east of the cooling towers near the "cleaning in place" building, while the 'third one is to the north near the boiler house. A fire hydrant, with a hose house, is also' located to the west of the cooling towers. E. STANDARD OPERATING PROCEDURES Ammonium hydroxide will be produced in a batch process that will be initiated and monitored by a trained operator. With the keyed selector switch in the automatic mode and the power switch in the on position, the operator will simply push the start up sequence button. This will initiate the automatic batch process. Once the PLC verifies operational status, the PLC will start the soft water flow to the reaction chamber and cooling water to the heat exchanger. After a short time delay, the PLC will open the anhydrous ammonia valve to begin the conversion process. Since the conversion process is exothermic, the PLC will hold process Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-7 Luft Environmental Consulting control until the conversion system's temperature stabilizes. After these short time delays, the PLC will control the ammonia conversion process based primarily on reaction temperature. Plant operators will periodically check the soft water quality prior to the inlet to the reaction chamber. During the conversion process, specific gravity of the ammonium hydroxide will be tested with a hydrometer to ensure that the ammonia concentration is within specification. If the ammonia concentration is not within specification, the plant operator may manually adjust the soft water flow to the reaction chamber via the automatic/manual pneumatic controller. During manual operation of the system, all alarms and shutdown sequences will remain active. F. PREVENTATIVE MAINTENANCE PROGRAM The ammonia system is maintained on an as needed basis. Daily checks on the ammonium hydroxide levels, specific gravity tests, and routine inspections on the equipment provide a continuous performance baseline for the system. If the system loses performance, the necessary equipment will be removed from service, thoroughly inspected and appropriately maintained. G. OTHER OPERATING AND MAINTENANCE PROGRAMS In addition to the standard operating and maintenance procedures, "tailboard briefings" or "job procedure discussions" will be conducted prior to commencing non-routine tasks. This procedure ensures that the operation and maintenance personnel under'Stand the job specific potential hazards and how to react if a hazard develops. Gist-brocades/RMPP Addendum/Feb., 1992 Page Via-8 Luff Environmental ConsuLting Vlla. DETECTION, MONITORING AND CONTROL SAFETY SYSTEMS Detection, monitoring, and automatic control systems to minimize potential acutely hazardous materials accident risks are discussed in this section, as required by Section 25534 (c)(4) of Chapter 6.95. Pursuant to Section 25534 (k), monitoring programs are also addressed. A. AUTOMATED CONTROL 1. Programmable Logic Controller All process parameters for the ammonia conversion system are monitored and controlled by an Allen Bradley SLC 500 programmable logic controller (PLC). The process parameters controlled by. the PLC include reaction temperature and heat exchanger temperatures. Control of these parameters is accomplished through flow control valves installed in the soft water supply line and the cooling water supply line. B. AUTOMATED MONITORING 1. Reaction Temperature The temperature of the ammonium hydroxide leaving the reaction chamber is continuously monitored, via a RTD, during the operation of the ammonia conversion SYstem. Reaction temperatures that exceed normal operating parameters will activate local alarms. If these alarms are not cleared by operations personnel, the PLC will initiate a system shutdown. 2. Heat Exchanger Temperature Monitoring of the ammonium hydroxide temperature leaving the heat exchanger is virtually identical to the reaction chamber temperature monitoring. The PLC monitors the RTD in the ammonium hydroxide line leaving the heat exchanger and modulates the cooling water control valve to obtain the desired temperature. Gist-brocades/RMPP Addendum/Feb., 1992 Page Vlla-1 Luff Environmental Consulting 3. Soft Water Flow Indicator A flow indicating switch, set at 12 gpm, is installed in the soft water supply line. This indicator vedfies that at least 12 gpm of soft water is flowing before the PLC allows the anhydrous ammonia'control valve to open. This water flow permissive ensures that there is sufficient soft water flow into the reaction chamber to complete the conversion to ammonium hydroxide. By verifying this soft water flow rate, the likelihood of having anhydrous ammonia flow through the system unreacted is greatly reduced. 4. Anhydrous Ammonia Flow Indicator The one inch anhydrous ammonia line is also equipped with a flow indicating switch. This indicator is installed as a safeguard to detect anhydrous ammonia flow when the system is idle. If the conversion system is not in service and the anhydrous ammonia flow rate reaches 0.1 gpm, a local alarm will be activated. Early detection of anhydrous ammonia flow when the system is idle should enable operations personnel to respond quickly to an equipment malfunction. 5. Anhydrous Ammonia Tank Level The anhydrous ammonia tank is equipped with a differential pressure gauge that is installed to determine the level of liquid anhydrous ammonia. This gauge sends an electrical signal to the PLC and a digital display on the local control panel. 6. Ammonium Hydroxide Tank Level The ammonium hydroxide tank level is monitored exactly the same way as the anhydrous ammonia tank. A differential pressure gauge sends an electrical signal, which is proportional to the tank level, to the PLC and a digital display-on the local control panel. B. MANUAL MONITORING Monitoring of the ammonia system is done on a routine basis. Operations personnel physically inspect the system on a daily basis to verify the prolser function and integrity of the equipment. Gist-brocades/RMPP Addendum/Feb., 1992 Page Vlla-2 Luff Environmental Consulting 1. Hydrometer Ammonium hydroxide samples are analyzed for ammonia concentration on a daily basis to insure proper operation of the ammonia system. A sample of ammonium hydroxide is diverted to the hydrometer which measures the specific gravity of the solution. After measuring the specific gravity, the sample is discharged into the 800 gallon vent tank. 2. Temperature There are several manual temperature gauges located on the process piping for local verification of temperatures. 3. Pressure There are several manual pressure gauges located on the process piping for local verification of pressures. 4. Tank Levels A mechanical level indicator is installed on the anhydrous ammonia tank. This gauge is used to verify'the differential pressure gauge which also monitors the liquid level in the tank. The ammonium hydroxide tank is equipped with a sight glass to veri,fy liquid level in the tank. Gist-brocades/RMPP Addendum/Feb., 1992 Page Vlla-3 Luft Environmental Consulting Xila. DESCRIPTION AND RESULTS OF THE HAZARD AND OPERABILITY STUDY This section addresses the requirement that the RMPP be based upon an assessment of the processes,' operations, and procedures of the business, and shall consider the "results of a hazard and operability study which identifies the hazards associated with the handling of an acutely hazardous material due to operating error, equipment failure, and external events, which may present an acutely hazardous materials accident risk." [§ 25534 (d)] Additionally, Section 25532 (g) specifies that the RMPP include programs which include risk assessment for unit operations or operating alternatives, in accordance with these requirements, a Hazard and Operability (HazOp) study was conducted. This section summarizes the results of the study. .A. HAZOP TECHNIQUE The HazOp technique that was used for the Gist-brocades Food Ingredients Inc. facility was a modified "guide word" approach for a Hazard and Operability Study. The basic guide word HazOp was chosen since it allows a systematic and thorough review of every part of the facility that handles AHMs. However, due to the simplicity of the ammonia conversion system, a "what if" analysis was incorporated into the guide word approach. Both of these approaches are described in the AIChE Guidelines for Hazard Evaluation Procedures1, which is referenced in Section 25534 (I), Chapter 6.95, Division 20, H & S Code. Other publications further describe the guide 'word .approach, including A Guide to Hazard And Operability Studies2. B. PROCESS AND EQUIPMENT DESCRIPTION Gist-brocades Food Ingredients Inc. will produce ammonium hydroxide on site by mixing approximately 7 gallons per minute (gpm) of liquid anhydrous ammonia and approximately 15 gpm of water in the ammonia conversion system. Anhydrous ammonia is fed from the 13,256 gallons (ammonia capacity) storage tank to the ammonia conversion system via a one inch welded stainless steel 1 Guidelines for Hazard Evaluation Procedures, Amedcan Institute of Chemical Engineers, New York, 1985. 2A Guide to Hazard and Operability Studies, Chemical Industry Safety and Health Council of the Chemical Industries Association, London, 1985. Gist-brocades/RMPP Addendum/Feb., 1992 Page Xlla-1 Luff Environmental Consulting pipeline. Water supply for the conversion system comes from the existing water softening system located at the plant. The ammonia conversion system consists of a reaction chamber and a heat exchanger. Inside the reaction chamber, the anhydrous ammonia is added to the soft water through a jet type mixer to make ammonium hydroxide. Since this reaction is exothermic, the ammonium hydroxide will be cooled in a shell and tube heat exchanger using cooling water from the existing cooling tower. After cooling, the ammonium hydroxide will flow into an existing 20,000 gallon ammonium hydroxide storage tank. The existing ammonium hydroxide distribution system will be utilized to fertilize the fermentation tanks. Liquid ammonia fioWto the conversion system will be controlled by a pneumatic control valve. This pneumatic control valve closes upon loss of air pressure. The air supply to this control valve will be governed by a fail closed electric solenoid valve. The ammonia flow will be limited to approximately 7 gpm by two devices; a pressure regulator-and a fixed orifice plate. Soft water flow to the reaction chamber will be regulated by a programmable logic controller (PLC) which will modulate a flow control valve. During processing, the PLC will monitor the ammonium hydroxide reaCtion temperature and the position of the flow control valve. If the ammonium hydroxide temperature leaving the reaction chamber is too high, the PLC will open the soft water control valve. Opening the soft water valve reduces the ammonium hydroxide strength and lowers the heat of reaction in the chamber... Conversely, reducing the flow of soft water into the reaction chamber increases the ammonium hydr(~xide concentration and the reaction temperature. Therefore, the PLC can maintain the proper concentration of ammonium hydroxide by maintaining the appropriate reaction temperature. C. HAZOP REVIEW FOR GIST-BROCADES FOOD INGREDIENTS INC. A review of the ammonia conversion system was used to generate the following subsystems that were investigated. Item #1: Item #2: Item #3: Item #4: Fill Connections for the Anhydrous Ammonia Storage Tank Anhydrous Ammonia Storage Tank Liquid Anhydrous' Ammonia Line Reaction Chamber Gist-brocades/RMPP Addendum/Feb., 1992 Page Xlla-2 Luft Environmental Consulting A modified guide word approach was used to systematically identify the intended design and operation of the item in the system. 'Deviations from the intended operation and the potential for an anhydrou~ ammonia release from each piece of equipment were discussed on a "what if" basis. D. RESULTS OF HAZOP In order to meet the requirements of the statute, a HazOp study was conducted to identify the hazards associated with the handling of an acutely hazardous material. The potential hazardous events determined by the Haz©p review are described briefly.below. These events could be caused by operating error, equipment failure or external events (including earthquakes). From these events, the worst credible ammonia releaSe was generated for input into the atmospheric dispersion model and offsite consequence analysis. Discussion of the offsite consequence analysis, based on the results of the worst credible release events, is provided in Section XIIla. 1..Uncontrolled Tank Filling There are two methods of dispensing liquid ammonia from the delivery truck to the storage tank. In the first method, the vendor truck is equipped with an ammonia compressor which takes ammonia vapor from the client tank, compresses it, and pressurizes the vendor tanker with the compressed vapor. This pressure differential bet~veen the vendor tanker and the client's storage vessel enables the vendor to discharge the ammonia at a high flow rate. The normal flow rate for the ammonia vendor using this delivery method is about 90 gallons per minute (gpm). This is the delivery system that Gist-brocades Food Ingredients Inc. is currently Planning to use. The second method of transferring liquid ammonia to the storage tank is by a liquid ammonia pump on the ammonia tanker. The typical delivery rate for this type of system is approximately 75 gpm. In both delivery systems, the vendor's tanker is equipped with excess flow valves that will shut.off immediately if normal delivery rates are exceeded. These valves are designed to automatically operate in the event of a hose or pipe fitting failure.downstream of the excess flow valve. A discussion on the Gist-brocades/RMPP Addendum/Feb., 1992 Page Xlla-3 Luff Environmental Consulting operation of an excess flow valve was provided in Section Via. There are three excess flow valves in the anhydrous ammonia storage system. Two of these excess flow valves are in liquid service and one is in vapor service. A three inch excess flow valve set at' 225 gallons per minute (gpm) is threaded directly into the ammonia storage tank. This valve prevents the uncontrolled release of the contents of the ammonia storage tank in the event of a catastrophic liquid line rupture downstream 'of the excess flow valve. The second liquid line excess flow valve is a one inch valve, set at 24 gpm, installed in the liquid ammonia pipeline that supplies the ammonia conversion system. This valve was installed to prevent an uncontrolled release from the ammonia supply line and ammonia conversion system in the event of a catastrophic failure in the pipeline or reaction chamber. A schematic diagram of the excess flow valve arrangement is shown in Figure Via-2. The excess flow valve that is installed in the ammonia vapor return line is set at 95 gpm. This flow rate is equivalent to approximately 119 cubic feet per minute (at 70°F) or 54 pounds of ammonia per minute. A release during the tank filling operation could occur if the tank operator could not shut the delivery system down for any reason, and the ammonia delivery tanker had sufficient product to ovedill the storage tank. This event could happen if the ammonia delivery truck operator becomes incapacitated for any reason, such as suffering a heart attack during the 'ammonia delivery. When the st. orage tank completely fills, the liquid ammonia will flow into the storage tank vapor return line and attempt to flow to the tanker. However, the excess flow valve on the vapor return line is set at 95 gpm of liquid ammonia. Since the discharge rate from the ammonia tanker is typically around 90 gpm, the excess flow valve on the vapor return line could close immediately and stop the flow out of the storage tank. If the ammonia is delivered via liquid pump, an immediate release would not occur since the typical delivery pump does not generate sufficient pressure to operate the pressure relief valve on the storage tank. When the vapor line excess flow valve operates, the liquid ammonia pump would bypass the flow back to the tanker. A release may occur a little later due to the thermal expansion of the liquid ammonia forcing the storage tank pressure relief Gist~-brocades/RMPP Addendum/Feb., 1992 Page Xlla-4 Luft Environmental Consulting valve to operate. The release rate due to thermal expansion of the liquid ammonia would be less than the release rate of hose failure event discussed in #3 below. If the ammonia is delivered utilizing an ammonia compressor on the tanker, the possibility of an immediate release exists. The compressor on the ammonia tanker can generate enough pressure to lift the pressure relief valves on the storage tank if the vapor return line excess flow valve closed due to an overfill situation. Under these circumstances, a release equivalent to the delivery rate of the ammonia tanker, approximately 90 gallons per minute (7.6 pounds per second) could develop. However, since the tankers carrY approximately 7,900· gallons of ammonia and the system capacity is 13,256 gallons, it would be impossible to overfill the tank if the order is placed when the ammonia storage capacity is below 5000 gallons. In order to ensure that this type of release does not occur, ammonia will only be ordered when the tank contains less than 4000 gallons. 2. Storage Tank Line Failure As discussed previously, the anhydrous ammonia storage tank is equipped with three excess flow valves. There is a three inch excess flow valve set at 225 gpm located where the liquid line is threaded into the storage tank. This valve prevents an uncontrolled release through the liquid ammonia deliverY line. Additionally, there is a check valve installed in the liquid ammonia delivery line to prevent the reverse flow of ammonia during a delivery. Since this check valve is installed in the short section of piping between the delivery bulkhead and the excess flow valve, the likelihood of a significant release from the 2 inch liquid line is very Iow. A one inch excess flow valve, set at 24 gpm, is installed at the beginning of the one inch liquid ammonia supply line that feeds the conversion system. The maximum release from a rupture of this one inch liquid ammonia line would be limited to the excess flow valve rating. The third excess flow valve in the ammonia conversion system is installed where the vapor return line connects to the anhydrous ammonia storage tank. This excess flow valve is set for 95 gpm. This flow rate is equivalent to Gist-brocades/RMPP Addendum/Feb., 1992 Page Xlla-5 Luft Environmental Consulting approximately 119 cubic feet per minute (at 70°F) or about 51 pounds of ammonia per minute. All of the above release events would require the failure of schedule 80 piping. Failure of the schedule 80 piping is unlikely since these short piping sections are secured to a substantial bulkhead and the pipelines are protected with pressure relief valves. Additionally, the maximum ammonia release rates from a pipeline failure would be below the hose failure event discussed in event #3 below. 3. Hose Failure With the mitigation measures taken into account in event #1 and event #2 above, the worst credible release event would involve a failure in the two inch ammonia loading hose. Two types of hOse failures were investigated; a complete failure of a hose and a partial failure of a hose. If the hose completely failed, the excess flow valve on the del!very truck would close immediately. It was assumed that the excess flow valve would operate properly. In this release event, only the contents of the loading hose would be released. A worst case release event would be that of partial failure of the hose. In this case, a failure of the hose near the fitting was reviewed. The failure was assumed to occur near the hose end fitting, where the pressure stresses are compounded by the mechanical connection stresses. For a worst case release eveht, the resultant flow through the excess flow valve on the delivery truck would be just below the operating set point of the valve. Since the excess flow valve settings vary slightly between ammonia delivery companies, it was assumed that the operating set point was equivalent to the excess flow valve on the storage tank. It should be noted that the anhydrous ammonia storage tank at Gist-brocades was previously a cargo carrier and was built to United States Department of Transportation specification MC 331. For the ammonia release calculations, a worst case release rate of 200 gpm was assumed. This rate is approximately 90 percent of the excesS~ flow Gist--brocades/RMPP Addendum/Feb., 1992 Page Xlia-6 Luft Environmental Consulting valve setting. If the liquid ammonia delivery pressure is approximately 145 psig (30 psig greater than normal storage pressure), about 20.7 percent of the liquid ammonia would flash to the vapor phase as'soon as it was released to the atmosphere. Therefore, the worst case release was calculated as 3.44 pounds per second of vapor and 13.16 pounds per second of liquid ammonia. Since the ammonia delivery trucks are equipped with manually-operated emergency shut. down valves, two versions of this release event were reviewed. In the first case, the ammonia truck operator or a plant operator would respond to the release and operate the emergency shut down valve.. Response to the hose failure and shut down of the delivery system by one of these operators was assumed to occur within 10 seconds. The net release from this 10 second release event would be 34.4 pounds (3.44 pounds per second for 10 seconds) of ammonia vapor and 131.6 pounds (13.16 pounds per second for 10 seconds) of liquid ammonia. The 34.4 pounds of ammonia vapor represents an instantaneous release to the atmosphere while the 131.6 pounds of liquid ammonia must evaporate over a period of time. In the second case, it was assumed that the truck driver and/or operator. could be overcome by ammonia vapors and may not be able to respond immediately to the ammonia release. This case would result in a "continuous" release of 16.6 pounds per second of ammonia. Gist-brocades Food IngredientS Inc has several Self Contained Breathing Apparatus (SCBA) on site for emergency response. As such, plant personnel would be able to don the SCBAs and operate the emergency shut down valve within a few minutes. Therefore, a continuous release of 16.6 pounds per second was modeled as a worst case event. 4. Fitting Leak. The available leak flow area for all valves and threaded fittings on the ammonia system would be less than the area of the modeled hose failure. Although a valve packing is one of the most likely release events, the smaller flow area would result in a release well below the hose failure event. Gist-brocades/RMPP Addendum/Feb., 1992 Page Xlla-7 Luft Environmental Consulting Xilla. OFFSITE CONSEQUENCE ANALYSIS Section 25534 (d)(2), Division 20, California Health and Safety Code states' that "[t]he RMPP shall consider.. [f]or the hazards identified in the hazard and operability studies, an offsite consequenCe analysis which, for the most likely hazards, assumes pessimistic air dispersion and other adverse environmental conditions." Additionally, Section 25534.1 requires that "[e]very RMPP . . . shall give consideration to the proximity of the facility to schools, residential areas, general acute care hospitals, long-term health care facilities, and child day care facilities." Effective January 1, 1992, Section 25534(d)(2) was revised and now.requires that the offsite consequence analysis include "a clearly prepared map noting the location of the facility which shows the po. pulations considered.., and the Zones of vulnerability, including the levels of expected exposure in each zone for the hazards identified in the hazard and operability study." This section addresses these requirements. The HazOp study generated release events for the ammonia system based on the design of the facility, potential operator error and external events, such as an earthquake. These release events were reviewed to determine whether there was a high likelihood of occurrence or a significant offsite consequence if the release were to occur. The releases associated with a high likelihood of occurrence were very Iow release rates resulting in an insignificant offsite consequence. For the purpose of emergency response planning, a worst credible release scenario was identified in the HazOp study. For this release, an air dispersion model was performed utilizing adverse (pesSimistic) and average (prevailing wind and common warm weather parameters) meteorological conditions. Both sets of meteorological data were used to provide a basis for evaluating the offsite consequences of the release events. The air dispersion model and its results are discussed in this document. A. GENERAL MODEL INFORMATION The pollutant dispersion results were generated by the CAMEO~3.0-ALOHA 5.0 model developed by the Hazardous Materials Response Branch of the National Oceanic and Atmospheric Administration (NOAA). CAMEOTM is the Computer- ''Aided Management of Emergency Operations program which was designed to help emergency planners and first responders both plan for, and safely handle, Gist-brocades/RMPP Addendum/Feb., 1992 Page Xllla-1 'Luff Environmental Consulting chemical accidents. ALOHA stands for Areal Locations Of Hazardous Atmospheres and serves as a tool for estimating the movement and dispersion of an atmospheric po!lutant. NOAA recently updated ALOHA to include both a Gaussian dispersion model and'a heavy gas model. ALOHA has'the ability to recommend whether the-Gaussian or heavy gas model is appropriate for a chemical release occurring during the selected meteorological conditions. The modeler can select either model to run. The Gaussian dispersion model plots the distribution of a pollutant gas from a series of Gaussian equations as described by Turner i'n the "Workbook of Atmospheric Dispersion Estimates''1. The Gaussian equation describes a bell- shaped or normal curve. Concentration distribution at ground level is calculated and the bell-shape spreads out and gets wider and flatter as the pollutant drifts downwind. The heavy gas dispersion model uses calculations found in the widely accepted DEGADIS2 model. The model has been simplified to provide quicker results during emergency use, but is still accurate to within ten percent of the original DEGADIS model. In both models, the pollutant concentration to be calculated is set to a particular numerical value such as parts per million (ppm), or can be the value of a toxicological parameter such as those set in the Emergency Response Planning Guidelines issued by the American Industrial Hygiene Association3. The models will' give results using either an instantaneous source release, when the release of the pollutant occurs dudng one short time, pedod; or a continuous source release, when the pollutant is being released over a longer period of time. Both releases result in a. pollutant concentration curve termed a footprint (see Figure XIIla-1). The area inside the curve is the region that is predicted to have ground level concentrations above the limit set by the modeler. 1Turner, D. Bruce, 1974. Workbook of Atmospheric Dispersion Estimates. National Technical Information Service, Springfield, Virginia. 2Spicer, Tom and Jerry Havens, 1989. Users Guide for the DEGADIS 2.1 Dense Gas Dispersion Model, EPA-450/4.-89-019. U.Si EPA, Cincinnati, ~hio. 3American Industrial Hygiene Association, October, 1988: Emergency RespenSe Planning Guidelines. AIHA ERPG Committee, 475 Wolf Ledges Parkway, Akron, OH 44311. Gist-brocades/RMPP Addendum/Feb., 1992 Page Xllla-2 Luff Environmental Consulting Pollutant Source FIGURE XIIla-1 - Footprint from a Continuous or Instantaneous Source Spill The ALOHA model incorporates several assumptions regarding chemical source, meteorological data and terrain. In the heavy gas model, the chemical spill is assumed to have occurred at ground level. The Gaussian model has the option of using elevated sources. For both models, all concentrations are calculated at ground level. Meteorological data includes atmospheric stability, wind speed and wind direction. Atmospheric stability classes are identified by the letters A through F with class A being the most unstable and class F the most stable. Unstable conditions result in a large amount of mixing of the atmosphere, causing the pollutant chemical to produce a shorter threat distance but a wider plume. However, this condition also includes more vadable wind directions, resulting in a threat zone that may tend to meander. Stable conditions give a longer, narrower plume that tends to travel in a single direction. The stability can be influenced by both heating and mechanical stirring of the atmosphere. Heating of the sudace layer of the atmosphere leads to unstable conditions and ground cooling results in more stable conditions. Mechanical stirring is caused by the winds, with strong winds tending to cause neutral stability (classes C and D). Wind speed and wind direction are input directly by the modeler; although the air model Checks that wind speeds are consistent with the stability class chosen. If the wind speed and stability class are inconsistent, no plume dispersion is calculated. The turbulence caused as the wind flows over and aroUnd obstacle~ is included in the model and termed ground roughness. The model adjusts for mixing Gist-brocades/RMPP Addendum/Feb., 1992 Page Xllla-3 'Luff Environmental Consulting consistent with stability class by allowing for either rural ground roughness, a terrain with few obstacles, or an urban ground roughness, characterized by many obstacles. The effects of terrain on the speed and direction of the plume are not included in the model. The model assumes that the winds are uniform throughout the plume. B. LIMITATIONS OF ALOHA DISPERSION MODEL As with all dispersion models, the ALOHA model does have several limitations even with all the above mentioned factors included in the plume concentration calculations. It is imPortant to remember that the model has a nominal accuracy of a factor of 2, so a predicted concentration of 50 ppm may actually be in the range of 25 to 100 ppm. This degree of accuracy is consistent with other dispersion models. The model also does not accurately represent several conditions which are discussed below. As mentioned earlier, tl~e ALOHA model does not include terrain steering effects caused by topography or wind shifts and the model assumes that all winds are constant throughout the plume. Wind variation can reduce the accuracy of the results when the plume travels more than a mile from t~he source. The model will only calculate plume dispersion results at distances less than 6.4 miles (10 km). Also, dispersion models are not accurate at Iow wind speeds (less than 1 mile per hour) or at very.stable atmospheric conditions, which produce these Iow wind speeds. ALOHA does not allow a wind speed of less than one meter per second (2.237 miles per hour). Another situation not handled well by ALOHA (or air models in general) is concentration patchiness in the area 50 to 100 yards from the source. In this vicinity, the pollutant may meander and be more patchy than the model can predict. After about 100 yards, a plume will have experienced enough mixing eddies to reduce the-irregular concentrations within the plume. At this point, the model will predict the pollutant concentrations within the accuracy of the model. C.. MODEL PARAMETERS FOR GIST-BROCADES For the release cases identified in the HazOp study, an air dispersion model was Gist-brocades/RMPP Addendum/Feb., 1992 Page Xllla-4 'Luff Environmental Consulting used to evaluate the offsite consequences of the release events. The air dispersion modeling was pedormed utilizing .both adverse and average meteorological conditions. Also, modeling was performed using urban and rural terrain roughness factors as the facility is surrounded by industrial buildings to the north and open fields to the south. By applying both sets of data to the release events, the modeling results provide qualitative' information to help evaluate the offsite consequence associated with the worst credible release events. Anhydrous ammonia is a gas composed of three parts hydrogen and one part nitrogen (NH3). It has a molecular weight of 17.03 and is lighter than air. When stored under sUfficient pressure at ambient temperature, ammonia is liquefied. Ammonia exists in both liquid and vapor phases in the storage tank. Anhydrous ammonia is shipped as a nonflammable gas and is an irdtant to the eyes, skin and mucous membranes. Ammonia has a perceptible odor as Iow as 5 ppm and is readily detectable at 10 ppm. A release of anhydrous ammonia vapors under pressure may result in the formation.of an aerosol mist which behaves as a heavy gas (heavier than air). However, the ammonia behaves in a normal Gaussian distribution once the concentrations in air drop to about 10,000 ppm. For a small release, this will .occur within the first 50 yards from the source. If anhydrous ammonia is released as a liquid, a portion of the liquid will flash to the vapor phase immediately. The amount of' liquid ammonia that will flash to the vapor phase is dependent on the storage pressure prior to the release. For each release of liquid ammonia, the percentage of ammonia flashing to the vapor phase was calculated. The balance of the release would form a liquid pool that would evaporate over a period of time depending on ambient conditions. The air dispersion model calculates the evaporation release rate based on the input meteorological parameters. Therefore, a release of liquid anhydrous ammonia would generate two vapor dispersion models; one for the initial flash of the release and the second for the evaporation of the liquid pool. Results were modeled for concentrations set at 50 ppm and 500 ppm as specified in the "Outline of RMPP Requirements"? prepared by the Bakersfield City Fire Department Hazardous Materials Division. The 50 ppm value is tine tenth of the Immediately Dangerous to Life and Health concentration published by the EPA. Gist-brOCades/RMPP Addendum/Feb., 1992 Page Xllla-5 ' 'Luff Environmental Consulting An ambient concentration of ammonia of 500 ppm is considered as Immediately Dangerous to Life or Health (IDLH). At 500 ppm, there is severe irritation of the eyes, nose and throat, but there are no irreversible health effects to healthy adults for a short exposure of less than 30 minutes. Since ammonia has a perceptible odor as Iow as 5 ppm, it is unlikely that a person would become unknowingly overexposed.4 The meteorological conditions used to determine maximum impact distances and times were based on data collected by the U.S. Weather Service at Meadow Fields Airport in Bakersfield, California. The data in Table XIIla-1 summariZes the meteorological data for both average and adverse conditions. On an annual basis, the predominant meteorological conditions are with a wind direction from the north-northwest at 6.4 miles per hour. A chemical release during these average atmospheric conditions represents the most likely offsite consequence resulting from the release. For modeling purposes, however, winds from the south at 5.0 miles per hour were also considered because they represent'the worst case or "pessimistic" atmospheric conditions. Although these weather conditions occur less than 3 percent of the time, modeling releases With these adverse weather conditions would present an offsite consequence that would potentially have the greatest affect on sensitive populations. The Gist-brocades facility is located on District Blvd. in a commercial/light industrial area. Pursuant to Section 25534, California Health and Safety Code, "[e]very RMPP... shall give consideration to the proximity of the facility to schools, residential areas, general acute care hospitals, Iong-te[m health care facilities, and child~day care facilities." Sensitive population sites considered in this offsite consequence analysis included residences, schools, emergency and health care facilities. Figure XIIla-2 shows an overview of the sensitive population sites near the plant. The distances between the Gist-brocades facility and the sensitive populations are summarized in Table XIIla-2. The nearest residences are approximately a third of a mile north and, also, a third of a mile to the southeast of the Gist-brocades facility. There is a hospital to the north of the Gist-brocades facility, approximatelY a quarter of a mile away. The nearest school is roughly 1.07 miles to the northeast of the plant. A Bakersfield City Fire Department station 4American Industrial Hygiene Association, October. 1988: Emergency Response Planning · Guidelines. AIHA ERPG Committee, 475 Woff Ledges Parkway, Akron, OH 44311. Gist-brocades/RMPP Addendum/Feb., 1992 Page Xilla-6 Luff Environmental Consulting is located 2.13 miles northwest of the Gist-brocades facility. There are adjacent industrial sites one tenth of a mile to the north, east and west of the facility. Most of the ALOHA limitations discussed previously have minimal effects on the model results. The ALOHA limitation regarding terrain steering effects and wind shifts caused by topography is not a problem in the area being modeled because the terrain is generally flat and unobstructed by large hills or valleys. The ground roughness of the area is determined to be a combination of rural and urban terrain. The effects of concentration patchiness within the first 50 to 100 yards from the source will be minimal because there are no resident populations within this pre-Gaussian area. D. RESULTS OF THE ALOHA DISPERSION MODELING During the HazOp study, one release event was identified as the worst credible release that warranted air dispersion modeling. This release was due to the failure of the liquid ammonia delivery hose dudng a delivery. As stated in Section XII, the worst case release was calculated as 3.44 pounds per second of vapor and 13.16 pounds per second of liquid ammonia (16.6 pounds per second total). Air dispersion modeling was performed for the delivery hose failure with two different release durations. In the first scenario, an immediate response from the personnel witnessing the tank loading was assumed. The duration of the release was assumed to be ten seconds.. In the second scenario, a continuous release was assumed. The model runs indicated that the plume is fully developed in less than one minute. The release associated with the delivery hose failure has the potential to occur on an infrequent basis when ammonia is delivered to the plant to recharge the ammonia system. Only dudng ammonia delivery could the delivery hose failure release event develop. As mentioned earlier, a release of liquid anhydrous ammonia would result in a portion of the liquid flashing to the vapor phase immediately, depending on the storage pressure prior'to the release. For the liquid ammonia release cases that were modeled, the percentage of ammonia flashing to the vapor phase was calculated and modeled as a vapor release. The balance of the release would form a liquid pool that would evaporate over a period of"iime depending on ambient conditions and the size of the pool. The air dispersion model calculates Gist-brocadeslRMPP Addendum/Feb., 1992 Page Xllla-7 Luff Environmental Consulting the evaporation release rate based on the input meteorological 'parameters and .the surface area of the pool. In general, the larger the surface area of the spill the greater the evaporative release rate. This situation would generate a larger plume in the offsite consequence analysis, but the duration of exposure would be significantly shorter. Conversely, the smaller the sudace..area of the spill, the smaller the dispersion plume and the longer the duration of exposure. For the hose rupture release events, the assumed ammonia pool surface area was 150 square feet for the 10 second release event and 1000 square feet for the-continuous release event. The results of the ALOHA plume dispersion modeling are summarized in Tables Xllla-3 through XIIla-6. Under all meteorological conditions modelled, ALOHA predicted that the concentration of the plume would drop'to less than 10,000 ppm within 50 yards from the facility. Since the nearest sensitive receptor is more than 100 yards from the facility, the Gaussian model was used to evaluate the offsite consequences from a potential release. ALOHA recommended the use of the Gaussian model for these release events. Table Xllla-3 gives the plume distance for the 10 second release event using the Gaussian model and average and adverse meteorological conditions with urban terrain roughness factors. The model indicates that the 50 ppm plume would travel the furthest, 233 yards under adverse weather conditions. Since the nearest sensitive receptor is over 400 yards away, the 50 ppm and 500 ppm plumes from this release event would not reach any of the sensitive receptors. Nearby businesses to the north, east and west of Gist-brocades could be exposed to slightly higher concentrations in the unlikely event of a release. The plume distances for a continuous release using the Gaussian model and average and adverse meteorological conditions with urban terrain roughness factors are listed in Table Xllla-4. As shown in Figure Xllla-3, the 50 ppm plume from a continuous release under adverse weather conditions would travel over 500 yards and reach the nearest sensitive receptor. The air dispersion model results indicate that the level of exposure at this receptor site would be below 100 ppm. An exposure of 100 ppm is well below the Emergency Response Planning Guidelines .Level 2 (ERPG-2) concentration of 200 ppm at which "it is believed Gist-brocades/RMPP Addendum/Feb., 1992 Page Xllla-8 Luft Environmental Consulting that nearly all individuals could be exposed for up to one hour without experiencing or developing irreversible or other serious health effects or symptoms which could impair an individual's ability to take protective action''5 Since all release events would last less than one hour, there should be no significant adverse health effects on the sensitive receptors associated with these releases. Figure XIIla-4 shows the 500 ppm plume from a continuous release using the Gaussian model and adverse meteorological conditions with urban terrain roughness. The 500 ppm plume would not reach any sensitive recePtors. Tables Xllla-5 and XIIla-6 give the total plume distance travelled for the Gaussian models using average and adverse meteorological conditions with rural terrain roughness factors. These plumes travel in a southerly.direction as shown in Figures XIIla-3 and XIlla-4. The modeling shows that the 50 ppm plume may travel as much as 800 yards away from the facility under adverse weather conditions. Since the nearest sensitive receptor is approximately 600 yards away, the receptors to the south of the facility would be exposed to the 50 ppm plume. Air dispersion modeling results indicate that this exposure would be under 100 ppm. Business sites less.than 600 yards from the Gist-brocades facility could be exposed to slightly higher concentrations. Due to the differences in terrain factors, the zones of vulnerability are represented as semi-circles. It should be noted that changing the wind direction and speed affects atmospheric stability in the air diSpersion model substantially. Such changes in the atmospheric conditions would most likely result in shorter plume distances than those shown in the Figures. In all of the release events modeled above, the greatest impact to offsite receptors is associated with the flashing vapor release, instead of the evaporation of the pool. The evaporative pool releases, although longer in duration, result in a lower release rate than the vapor release. As such, evaporative pool releases disperse much quicker and do not travel as far as the vapor releases. At the release rate modeled, the entire contents of a full ammonia tanker truck (7900 gallons) would be released in approximately 40 minutes in the unlikely. 5American Industrial Hygiene Association, Apdl, 1988: Emergency Response Planning Guidelines. AIHA ERPG Committee, 475 Wolf Ledges Parkway, Akron, OH 44311. Gist-brocades/RM P P Addendum/Feb., 1992 Page XIIla-9 Luff Environmental Consulting event that attempts to stoP the release are not successful.' In emergency situations, Gist-brocades personnel would respond to a release in less than 10 minutes. The fire department would respond within 10 minutes, and the fire department hazardous materials team would respond within 30 minutes. It is unlikely that all of the reSponding personnel would be completely unsuccessful in attempting to reduce the release and its impact. Therefore, sensitive population exposure from the vapor release would occur only for the few minutes that Gist-brocades or emergency personnel take to stop the release. In most cases, exposure could be expected to. last less than 10 minutes. Since the local fire department response time is 10 minutes or less, a water spray could be applied to the release to substantially reduce the offsite impact within 10 minutes. In summary, the modeling shows that the sensitive populations near the facility Could be exposed to Iow concentrations of ammonia in a worst credible release event. However, there probably would not be any irreversible adverse health effects to.these sensitive populations due to the Iow ammonia concentrations in the plumes and the relative short exposure times for the sensitive receptors. The non-sensitive receptors adjacent to the Gist-brocades facility could be exposed to higher concentrations for a relatively short exposure time, however, there probably would not be any irreversible health effects. Gist-brocades,/RMPP Addendum/Feb,, 1992 Page Xllla-10 Luff Environmental Consulting TABLE Xllla-1 MEADOWS FIELD METEOROLOGICAL DATA MONTHLY SUMMARY MONTH WIND SPEED WIND DIRECTION (MPH) (FROM) JAN 5.20 NW FEB 5.80 ENE MAR 6.50 NW APR 7.10 NW MAY 7.9O NW JUN 7.90 NW JUL 7.20 NW AUG 6.60 NW SEP 6.20 WNW OCT 5.50 NW NOV 5.10 ENE DEC 5.00 ENE AVERAGE 6.40 NW GIST-BROCADES FOOD INGREDIENTS, INC. [~ MING AVE ~ WHITE LANE []_., ~IST-BROCADES DISTRICT BLVD '~ ~ ~ I','''' '''''''''''' ''''' ' I I I I I I I I I I ~ PACHECO RO3 LEGEND q' & -sohoo, ~ N 1 Inch = 0.5 Miles I I I ' Railroad ~--] - Nearest Residence - Hospital r~ - Fire station FIGURE Xllla-2 - Map shows the sensitive populatjon sites near the Gist-Brocades plant. TABLE Xilla-2 GIST-BROCADES PLANT SENSITIVE POPULATION SITES SITE Nearest Hospital Nearest Residence DISTANCE FROM PLANT (Miles) 0.24 CRITICAL WIND DIRECTION (Wind From) South South or Northwest Nearest School 1.07 Southwest Nearest Fire Station 2.13 Southeast TABLE Xllla-3 GIST-BROCADES FOOD INGREDIENTS, INC. PLUME DISPERSION RESULTS FOR AMMONIA DELIVERY HOSE RUPTURE VAPOR RELEASE OF 3.44 LBS/SEC AMMONIA-FOR 10 SECONDS LIQUID RELEASE OF 13.16 LBS/SEC OF AMMONIA-FOR 10 SECONDS ATMOSPHERIC MODELING CONDITIONS ATMOSPHERIC MODELING RESULTS (URBAN TERRAIN) CONDITION AVERAGE ADVERSE PLUME DISTANCE, OF PLUME TRAVEL CONCENTRATION VAPOR RELEASE LIQUID RELEASE (PPM) GAUSSIAN GAUSSIAN TEMPERATURE 95°F 40°F MODEL MODEL AVERAGE HUMIDITY 25% 50% CONDITIONS 6 minute evaporation release CLOUD COVER 10% 70% 50 159 yards 140 yards WIND SPEED 6.4 MPH 5.0 MPH 500 51 yards 44 yards WIND DIRECTION NW ENE ...... ADVERSE RELEASE DATE Summer Winter CONDITIONS 7 minute evaporation release RELEASE TIME 4:00 P.M. 8:00 A.M. -50 233 yards 177 yards . STABILITY CLASS B-Unstable C-Unstab/Neut 500 74yards 55 yards TABLE XIIla-4 GIST-BROCADES FOOD INGREDIENTS, INC. PLUME DISPERSION RESULTS FOR AMMONIA DELIVERY HOSE RUPTURE VAPOR RELEASE OF 3.44 LBS/SEC AMMONIA LIQUID RELEASE OF 13.16 LBS/SEC OF AMMONIA ATMOSPHERIC MODELING CONDITIONS ATMOSPHERIC MODELING RESULTS (URBAN TERRAIN) CONDITION AVERAGE ADV. ERSE PLUME DISTANCE OF PLUME TRAVEL ~ CONCENTRATION VAPOR RELEASE LIQUID RELEASE (PPM) GAUSSIAN GAUSSIAN TEMPERATURE 95°F 40°F MODEL MODEL AVERAGE HUMIDITY 25% 50% CONDITIONS CLOUD COVER 10% 70% 50 375 yards 349 yards WIND SPEED 6.4 MPH 5.0 MPH 500 124 yards 114 yards WIND DIRECTION NW ENE ADVERSE RELEASE DATE Summer Winter CONDITIONS RELEASE TIME 4:00 P.M. 8:00 A.M. 50 549 yards 460 yards STABILITY CLASS B-Unstable C-urJstab/Neut 500 181 yards 143 yards TABLE Xllla-5 GIST-BROCADES FOOD INGREDIENTS, INC. PLUME DISPERSION RESULTS FOR AMMONIA DELIVERY HOSE RUPTURE VAPOR RELEASE OF 3.44 LBS/SEC AMMONIA-FOR 10 SECONDS LIQUID RELEASE OF 13.16 LBS/SEC OF AMMONIA-FOR 10 SECONDS ATMOSPHERIC MODELING CONDITIONS ATMOSPHERIC MODELING RESULTS (RURAL TERRAIN) CONDITION AVERAGE ADVERSE PLUME DISTANCE OF PLUME. TRAVEL CONCENTRATION VAPOR RELEASE LIQUID RELEASE (PPM) GAUSSIAN GAUSSIAN TEMPERATURE 95°F 40°F MODEL MODEL AVERAGE ; HUMIDITY 25% 50% CONDITIONS 6 minute evaporation release CLOUD COVER 10% 70% 50 233 yards 205 yards WIND SPEED 6.4 MPH 5.0 MPH 500 73 yards 64 yards WIND DIRECTION NW ENE ADVERSE RELEASE DATE Summer Winter CONDITIONS 6 mir~ute evaporation release RELEASE TIME 4:00 P.M. 8:00 A.M. 50 375 yards 284 yards STABILITY CLASS B-Unstable C-Unstab/Neut 500 117 yards 88 yards TABLE Xllla-6 GIST-BROCADES FOOD INGREDIENTS, INC. PLUME DISPERSION RESULTS FOR AMMONIA DELIVERY HOSE RUPTURE VAPOR RELEASE OF 3.44 LBS/SEC AMMONIA LIQUID RELEASE OF 13.16 LBS/SEC OF AMMONIA ATMOSPHERIC MODELING CONDITIONS ATMOSPHERIC MODELING RESULTS (RURAL TERRAIN) CONDITION AVERAGE ADVERSE PLUME DISTANCE OF PLUME TRAVEL CONCENTRATION, VAPOR RELEASE LIQUID RELEASE (PPM) GAUSSIAN GAUSSIAN TEMPERATURE 95°F 40°F MODEL MODEL AVERAGE HUMIDITY 25% 50% CONDITIONS ~', ~t f',~ CLOUD COVER 10% 70% 50 564 yards 530 yards WIND SPEED 6.4 MPH '5.0 MPH 500 180 yards 166 yards WIND DIRECTION NW ENE ADVERSE RELEASE DATE Summer Winter CONDITIONS RELEASE TIME 4:00 P.M. 8:00 A.M. 50 801 yards 745 yards STABILITY CLASS B-Unstable C-Unstab/Neut 500 290 yards 231 yards GIST-BROCADES FOOD INGF{EDIENTS, INC. DISPERSION OF AMMONIA RELEASE FROM DELIVERY HOSE FAILURE 3,44 {bs/sec VAPOR-13,16 lbs/sec LIQUID CONTINUOUS RELEASE ADVERSE WEATHER CONDITIQNS 50 PPM PLUME CONCENTRATION  MING AVE rn ~ ~ WH~E ~NE ~ ~IST BROCADES DISTRICT BL P L ANT 1 Inch = 0.5 Miles ~t - School I I- Raltroad LEGEND ~' - Nearest Residence ~ - Hospital ~- Fire Station FIGURE Xllla-3 ~Z~JZ~: 3.44 lbs/sec Ammonia Li(~uid Source: 13.16 lbs/sec Ammonia Total Dis0ersion Distance for 50 _eom Plume: .~;}~ZLE~t: 801 yards-rural LiQuid Release: (shaded plume) 745 yards~ral 460 yards-urban - Map shows the impact on the sensitive population sites from subject ammonia release case. /~0verse Weather Conditions Modelled Stabili _ty Class: C Wind Speed: 5.0 mph Wind Direction: From the ENE-rural From the South-urban ~: 40°F LEC/GIST-B ROCADES/2-92 GIST-BROCADES FOOD INGREDIENTS, ~NC. DISPERSION OF AMMONIA REI~EASE FROM DELIVERY HOSE FAILURE 3,44 lbs/sec VAPOR-13.16 lbs/sec LIQUID CONTINUOUS RELEASE ADVERSE WEATHER CONDITIONS $00 PPM PLUME CONCENTRATION ~ MING AV~ ~ WHITE ~NE DISTRICT BLVD P LA NT I1'''' '''''11~ III I~ ~ ~ ' ~ ~ ~ ~ PAOHEGO I Inch = 0.5 Miles School III-Railroad LEGEND 1~' - Nearest Residence r~ - Hospital r~- Fire station FIGURE Xllla-4 - Map shows the impact on the sensitive population sites from subject ammonia release case. _~ZQ~E~: 3.44 lbs/sec Ammonia ~: 13.16 lbs/sec Ammonia Total DisDersion Distance for 500 Dom Plume: _'~: 290 yards-rureJ 181 yards-urban ~: (shaded plume) 231 yards-ruraJ 143 yards-urban Adverse Weather Conditions Modelled Stability_ Classt C Wind Soeed: 5.0 mph Wind Direction: From the ENE-ruraJ From the South-urban Temperature: 40°F LEC/GIST-BROCADES/2-92 XIVa. IMPLEMENTATION OF THE RMPP Section 25534 (c)(5) of the statute requires that the RMPP include "[a] schedule for implementing additional steps to be taken by the business, in response to the findings of the assessment performed.., to reduce the risk of an accident involving acutely hazardous materials. These actions may include any of the following: (A) Installation of alarm, detection, monitoring, or automatic control devices. (B) Equipment modifications, repairs, or additions. (C) Changes in the operations, procedures, maintenance schedules, or facility design." Additionally, Section 25534 (k) states that, for an existing facility, "the handler shall implement all activities and programs specified in the RMPP within one year following the certification. Implementation of the RMPP shall include carrying out all operating, maintenance, monitoring, inventory control, equipment inspection, auditing, record keeping, and training programs as required by the RMPP." In the case of Gist-brocades Food Ingredients Inc., as discussed in this RMPP Supporting Document, the facility utilizes anhydrous ammonia in the production of aqua ammonia. The facility currently has wdtten operating procedures that pertain to the storage, handling, and operation of the ammonia system. The following implementation plan for the RMPP pertains to the ammonia system equipment. A. ADDITIONAL STEPS REQUIRED Gist-brocades Food Ingredients Inc. has committed to safe storage and handling procedures associated with the ammonia system. As the ammonia system is new, mitigation measures were incorporated during the design and construction phases. Therefore, no additional steps are required by Gist-brocades Food Ingredients Inc. to implement the RMPP. B. IMPLEMENTATION i SCHEDULE All activities and programs specified in this RMPP shall be implemented before anhydrous ammonia is brought on site. Manual monitoring of the aqua ammonia storage tank on a daily basis has already been implemented to familiarize operating personnel with the procedure and establish a routine. Gist-brocades/RM PPAddendum/Feb., 1992 Page XlVa-1 Luft Environmental Consulting Gist-brocades. Food Ingredients Inc. RISK MANAGEMENT AND PREVENTION PROGRAM (RMPP) ADDENDUM FOR SULFURIC ACID Prepared By Luft Environmental Consulting 3701 Pegasus Drive, Suite 121 Bakersfield, CA 93308 February 1993 CERTIFICATION BY QUALIFIED PERSON AND FACILITY OPERATOR Section 25534 (j), H & S Code requires that the RMPP shall be certified as complete by a qualified person and the facility operator. These certifications are provided below. I certify that I am qualified to attest to the validity of the hazard and operability studies performed pursuant to Section 25534, and the relationship between the corrective steps taken by Gist-brocades. Food Ingredients Inc. following the hazard and operability studies and those hazards which were identified in the studies. Additionally, I certify that'this risk management and prevention program for sulfuric acid is complete per the requirements of the Bakersfield City Fire Department. This certification is based on my understanding that the data and documents provided by Gist-brocades Food Ingredients Inc. are true and correct and that the plans, programs, and procedures will be implemented as described. Signature Principal Mechanical Engineer Title Karl W. Luff, R.E.A. Name Date As facility operator, I hereby certify that this risk management and prevention program is complete and will be implemented. Signature . / Robert Deedy Name Plant Manager Title Date Gist-brocades/Sulfuric Acid/RMPP Page 1 Luff Environmental Consulting GIST-BROCADES FOOD INGREDIENTS, INC. SULFURIC ACID SYSTEM RMPP A. GENERAL DESCRIPTION OF FACILITY Gist-brocades Food Ingredients Inc.'s Bakersfield plant produces baker's yeast. The yeast is grown in a batch process in fermentation tanks. During this batch process, the yeast solution is fertilized in the fermentation tanks with phosphoric acid, sulfuric acid (93%), and ammOnium hydroxide. After the fermentation process is complete, the yeast is dried and/or packaged for the industrial market. B. DESCRIPTION OF' SULFURIC ACID PROCESS AND EQUIPMENT The' sulfuric acid. system has been maintained in .excellent condition since it began service in 1985. The system consists of a 700 gallon storage tank, a 20 gallon dosing tank, and associated transfer and distribution piping as shown in Figure 1. Material of construction for the sulfuric acid system is 316L stainless steel. All tanks and pipelines are welded. The 700 gallon .storage tank is an atmospheric pressure tank. The dosing tank is a pressure vessel that is equipped with a pressure relief valve set at 35 psig. This pressure relief valve relieves back to the acid storage tank. The sulfuric acid equipment is inspected for leaks and general condition on a weekly basis. Maintenance is performed on the system as needed. The dosing tank is pressure tested as required by the American Society of Mechanical Engineers (ASME) pressure vessel code. Sulfuric acid is delivered to the storage tank by the vendor using a pneumatically driven diaphragm type pump. After the delivery is complete, the acid remaining in the fill line is blown into the storage tank using compressed air from the plant air system. The storage tank overflow line is piped directly to.the plant wastewater system. Since the yeast production process is a batch process, sulfuric acid is added to the yeast solution in a batch process. To accomplish this, sulfuric acid is gravity fed from the storage tank to the dosi,ng tank when needed. Once the dosing tank is isolated (valved out) from the storage tank, compressed air from the plant air system is used to force the acid from the dosing .tank into the acid distribution piping. The dosing tank is normally empty and is filled only when sulfuric acid is to be fed to the fermenters or the yeast storage tanks. Gist-brocades/Sulfuric Acid/RMPP Page 2 Luft Environmental Consulting YEAST STORAGE · .. TANKS . FERMENTATION TANKS · STORAGE / ':" ~' ', ~ TANK DOSXN6I SULFURIC ,~,CI O SYJ'FE~I FIGURE 1 Phosphoric acid is also used during the yeast production process. Normally, the phosphoric acid is delivered to the process through its own distribution piping. However, phosphoric acid can be delivered to the sulfuric acid dosing tank for distribution to the fermenters or the yeast storage tanks. Phosphoric acid (37%) and sulfuric acid (93%) are compatible compounds. A minor amount of heat may be generated if the two 'acids are mixed due to the relative strengths (water content) of the acids. C. ACCIDENT HISTORY Gist-brocades Food Ingredients Inc. has not had an accident involving sulfuric acid at the Bakersfield Plant within the last three years. D. RECORDKEEPING PROCEDURES The sulfuric acid storage tank level is checked once per week at a minimum. Written weekly and monthly inventories are kept that include a record of the storage tank level. An order for sulfuric acid is placed when the inventory indicates that the tank level is below the reorder point. All documents relating to the RMPP are maintained for a minimum of five years. E, PERSONNEL AND TRAINING All oPerations personnel are specifically trained for their responsibilities within the plant. In addition to the equipment training, the operations personnel receive training in fire prevention, fire control, accident prevention, and .safety and first aid, in accordance with Gist-brocades Food Ingredients Inc. standards. All personnel at the plant are instructed in the site specific environmental and safety requirements on a periodic basis. The general subject matter for this training includes the following elements: * Hazardcus materials handling * Hazard communication program * Emergency response and evacuation procedures Annual safety training includes first aid procedures for individuals exposed to acids used at the facility. Emergency evacuation routes and the locations of possible hazards are included in the safety program. The Plant Manager and Maintenance Engineer are responsible for the Gist-brocades/Sulfuric Acid/RMPP Page3 Luff Environmental Consulting implementation of the RMPP and for the training of the Gist-brocades Food Ingredients Inc. employees. F. EMERGENCY RESPONSE PROCEDURES In the event of a large volume acid spill, the acid would be allowed to drain into the wastewater system. Depending on the nature of the spill, plant personnel may don the appropriate personnel protective equipment and attempt to stop the spill. A large volume spill would be permitted to drain into the wastewater system. Following the acid spill, plant personnel would wash any residual acid into the wastewater system using copious amounts of water. G. HAZARD ANALySis The HazOp study generated release events for the sulfuric acid system based on the design of the system, potential operator error, and external events, such as an earthquake. The HazOp study indicated that the possible release events were a hose failure during acid delivery and a tank or pipeline failure as the result of an earthquake. These release .events were reviewed to determine whether there was a high likelihood of occurrence orca significant consequence if the release were to occur. None of the release events had a high probability of occurrence. The worst credible release would involve a large volume aCid spill onto the concrete and into the wastewater system. The maximum volume of acid that could be released from the sulfuric acid system is 700 gallons from the storage tank. Due to the Iow vapor pressure of sulfuric acid, a significant airborne release from a spill of the acid is very unlikely. 1. ACID SPILL ON CONCRETE The worst case release'from the sulfuric acid system involves an external event, such as an earthquake, that causes a catastrophic failure of the main sulfuric acid storage tank. If the release were to. occur shortly after an acid delivery, a maximum of 700 gallons of sulfuric acid could be released. The area around the acid storage tank is sloped so that any liquid spill would flow toward a storm drain and into the wastewater system. The sulfuric acid storage tank is mounted on a concrete pad, and is surrounded by concrete and rubber coated concrete. In the event of a release, the acid would move in a fairly narrow stream toward the storm drain and would react with any concrete in its path. This chemical reaction would release hydrogen gas. In order to ignite the Gist-brocades/Sulfuric Acid/RMPP Page4 Luft Environmental Consulting hydrogen, the hydrogen concentration in the air at the combustion source would have to exceed the lower explosive limit (LEL) of 4% (40,000 ppm). Discussion with.a sulfuric acid manufacturer1 indicates that the rate of evolution of hydrogen is not expected to be sufficient to support combustion. Since the combustion sources are located approximately 80 feet from the acid tank and the acid spill would flow away' from the combustion sources due to drainage grading, an explosion and/or fire is very unlikely. Additionally, there are no confined spaces above the sulfuric acid system to trap the hydrogen gas. In this release event, it is possible that employees would exit the building on the south side through a door located near the sulfuric acid storage tank. These workers could be inadvertently exposed to the acid as it drained from the tank to the wastewater system. Injury could result from individuals walking through the acid by splashing the acid on clothing and exposed areas of skin. However, there are emergency eyewash/shower stations located near the acid storage tank and the acid fill connections. Use of these emergency eyewash/shower stations would minimize the injuries to exposed workers. 2. ACID SPILL IN WATER Sulfuric acid could be released to the wastewater system as a result of a catastrophic failure of the storage tank or as the result of overfilling the storage tank. The compartmentalized vendor trucks carry a maximum of 650 gallons of sUlfuric acid. Based on normal reorder quantities, overfilling the storage tank could release a maximum of 60 gallons of acid. to the wastewater system. A catastrophic failure of the storage tank could release 700 gallons of acid to the wastewater system. Since the mixing of sulfuric acid and water produces heat, the possibility of an offsite consequence if the acid enters the Wastewater system was investigated. If an acid spill reached the wastewater system, the acid would react with the water in the wastewater system. This reaction would evolve heat. Temperatures in excess of 400° F could be obtained depending on the rate the acid is added to the water and the relative amounts of water and acid that are present.2 Engineering calculations also indicate that temperatures above 325° F (163° C) are possible as shown in Figure 2. Since the wastewater system contains a minimum of 5,600 gallons of water 1jerry Mitchell, Cargill Inc., Telephone communication on December 17, 1992. 2jerry Mitchell, Cargill Inc., Telephone communication on December 17, 1992. Gist-brocades/Sulfuric Acid/RMPP Page 5 Luff Environmental Consulting TEMPERATURE OF SULFURIC ACID ADDED TO WATER 350 T T 2oo250300 150 100 50 0 0 1 2 3 4 5 6 RATIO OF ACID TO WATER Gist-brocades/Sulfuric Acid RMPP Figure 2 Luft Environmental Consulting at all times, the highest ratio of acid to water is 0.13 which corresponds to a temperature of approximately 125° F, as shown in the. figure. At that temperature, no steam or acid mists are expected to form. Also, the vapor pressure of a 93% sulfuric acid solution is relatively Iow, even at 'elevated temperatures (see Attachment 1)3. If sulfuric acid (100%) is heated to temperatures above 572° F (300° C), it begins to decompose and release sulfur oxides. Further heating and higher temperatures accelerate the decomposition process. As stated above, the acid used at the Gist-brocades facility is a 93% solution. The reduced concentration of sulfuric acid lowers the vapor pressure of the sulfur oxides substantially. The reduced vapor pressure substantially raises the temperature required for the evolution of sulfur oxides from the acid. Since there are no combustible materials present near the storage tank, heating of the tank due to a nearby fire is extremely unlikely. It is, therefore, unlikely that the storage tank could be heated sufficiently to cause decomposition of the acid. H. OFFSITE CONSEQUENCE ANALYSIS The hazard analysis indicated that a worst case release of sulfuric acid could release 700 gallons of sulfuric acid to the wastewater system. Large volumes of water are contained in the wastewater system, so the heat produced by this reaction would not be sufficient to cause decomposition of the acid or produce steam that could carry sulfuric acid into the air. 'Since the sulfuric acid. would not be carried offsite, no offsite consequences are~ expected. Emergency eyewash/shower stations are available for onsite personnel who may be exposed to the acid. 3perry's Chemical Engineer's Handbook, 6th ed., McGraw-Hill Book Company, New York, NY 1984 Gist-brocades/Sulfuric Acid/RMPP Page 6' Luff Environmental Consulting Attachment #1 Sulfuric Acid Vapor Pressure 3-68 PHYSICAL AND CHEMICAL DATA TABLE 3-14b Sulfuric Acid Parlial Pressure, bar, over Aqueous Sulfuric Acid' I0 20 179E-0~'~ 50 594 E - 08 40 lalE-O7 50 .513E-07 70 .-%~E -(~ ~ .7~E-(~ ~ .175E--~5 I I0 .752E .152~-~ .416E-~ .~E-~ 210 .I~E-~ .~E-~ .~E-~ 2~ .I~E-OI 270 .I~E-01 ~ .~E--OI ~ .~SE-OI ~ ~6E-OI 310 ~ .~OE-OI ~ .~E-OI ~0 .116 ~ .1~ I~.0 0 ~E--~ .~E--~ ~E--~ .~E--~ 70 .7~E--~ 110 1~ .HOE--~ ~IE~ ~4E--~ ~IE--~ .I~E--~ ~ .H4E--01 ~ .I~E-01 2~ ~E-01 ~ ~I4E-OI ~ .~E--01 ~0 ~4E--01 ~ .76~-01 ~ .t~ ~ .l~ 310 .1~ ~ .213 ~ .2~ ~0 ~ .417 ~ ~m~l ~i~g ~. Uni~ty ~ ~ ~ke~. VAPOR PRESSURES OF SOLUTIONS 3-6q TABLE 3-15 Total Pressure, bar, of Aqueous Sulfuric Acid Solutions' 'C- O 4O 50 60 TO ~0 gO I~0 II0 130 140 ITO 180 ~00 ~10 ~0 ~0 ~4O ~0 ~0 ~0 3i0 330 ~4O 10.0 .223E-OI .4o4E--01 .?08£-01 .117 .189 .296 .449 .c:fo4 1.349 1.868 2.324 4.404 9.0{13 11.289 16.841 20.264 28.561 :~k494 4,s.l~xs $1.72~ 59~015 ;.7~495 ' g4.Tos 200 :](I (! 40 O 448E -02 909E-02 174E-Ol 319E-01 .5.58E -- OI .939E--01 32~E- ~ 670E - 02 13()E -OI .241[-01 .4271[- Ol 725E--01 107E--OI .205E--01 .373E-01 .649E--01 .109 .175 ,275 .417 .617 1.2~8 1.740 [361 3;149 4.132 6.810 8.571 10.~ 13.107 15.951 19.225 22.960 27.188 81.989 37.24O 43.11`8 49.590 56.68f 64.400 72.7,81 81.816 91.518 101.894 112.947 · 152 .2.39 .3&5 542 .786 I.II3 I.,~4 2.101 2.810 3.697 4.793 6.127 7.731 · 9 640 11.~7 14.505 17.529 24.927 54.334 45.984 52.715 60.081 . ~.101 ' 76.792 ~.172 107.O4,3 119 18~ .290 434 .634 .904 1.264 2.,~3 3.090 4.G31 5.1~ 6.~ 10`245 11576 15.~7 18.414 ~.~2 ~.7~ 41.514 47.~ ~.~9 6~ 70.~7 ~.~8 ~,~O 1~.~7 92.0 94.0 9~.0 .109E--O6 .~48E--06 .102E--O5 ~9E--~ .70~E--05 .169E--'O4 ,380E--04 .813E--04 .IIOE--02 A27E ~,~9E--02 .866E--02' .I~E--OI .~E--OI ,312E--01 .4~E--OI .~E--O! .944E--O1 .132 .247 .~1 .439 .575 .744 ~E~ .221E--~5 ~E--~ .147E--04 .~44E--O4 ,759E--04 .159E--03 .519E~03 .612E--03 .346E--02 .576E--02 .1~9E--02 .14OE--OI ..~0E--OI. ~47E--0! .7~7'E--01 .1~7 .379 .ffoO 1,099 1.3~ .416E--07 .136E--06 .400E--~ .113E--05 .293E--~5 ~712E--~ .164E--O4 .742E--~ .14~E'--O3 .2~E--O3 ~21E--03 .161E~02 .270E--02 .4411[--02 .703E--O2 .IIOE--OI .167E--01 .25OE-O1 .300E--O1 ...~OE--OI .752E--01 .14.5 .197 .2~4 .~1 .460 ,597 19:1E -0~ 4~$E- 0'2 180E- 02 151~-OI 710£-OZ 272E -01 131E-01 470[ -01 .232E-01 .7~2E-Ol . 395E-01 126 .E$1E--OI . I SK, j .104 298 161 .9~.3 519 1.70~ 1.020 2.289 1.392 3,021 1.870 3.9,30 2.475 5.045 6.397 4.169 8.020 5.312 9.948 6.696 12.217 8.354 14.~64 10.322 17.929 12.641 21.452 15.351 2~.472 18.496 30·0~0 ·22.122 ~5.168 26.275 40.926 31.004 47.347 54.470 42.~J8 62.338 49.168 70.990 56.727 80.466 65.130 90.806 74.437 70 O 7.5 o 800 850 207E-(~ 467E-03 .201E-02 .387E--02 .716E-~ .127E-01 .217E-01 .~E-01 .578E-01 .~E-0I .1~ .~1 .~1 1.1~ 1.5~ ~017 2.~ 3.~ 4.~1 8.~9 15.272 ~1.~14 ~.812 41.~ 47.7~ .388E 811E-0~ .162£-~2 .309E - 0'2 .~7E-OI .170E-01 .~IE-OI .452E--01 .7~E--OI .478 . 2.715 5.481 6.~1 [~7 10`147 14.~5 17.4~ ~.611 2~ t~ ~.1~ ' 490E-~ 253E-0B .531E-03 .106E-O2 376£ -02 668E-02 · I ISE--OI .192E--01 .312E-01 .493E-01 AlS .170 .246 350 .489 6.73 .~13 1.610 2.O98 2.701 3.439 4.332 6.402 6·673 8,170 9.916 11,939 14.2~4 16.9f6 19.920 97.0 96.0 .235E--00 .117E--07 .774E-00 .391E--00 .235£--06 .1211[--06 .6.59E--06 .344E-06 .f73E-05 .914£--06 .425E--05 .228E--05 .987E--O5 .538E--05 ~tSE--~ .12OE--~ ,458E--04 .257E--04 .~tE--~ ~4E--~ .178E--03 .3321[--03 .598E-03 .1~1[-~ .626E--03 .I~E--~ .100E--~ .2~E--02 .4711[--02 .2931[--O2 .741E--02 .466E--~2 .1141[--01 .7261[--02 .172E-01 .1111[-01 .2~5E-01 .166E--Ot .371E-01 .2451[-0l .531E--01 .354E--O1 .749E-01 .5~E--OI .104 .TI0E-0I .1~ .~$E--OI .193 .135 ~ .t53 .341 .245 .446 .324 .343E - 05 952E -0,5 245E-~ ,589E-(~ .28~E- O3 .58-1E- O3 .38~E- 02 ~666E--02 .183E-01 .461E-01 .6831[-01 .147 .209 .292 ,402 .544 .726 · 1.242 3.886 4.740 5.732 6.876 &185 9.672 99.0 99.5 100`0 .479E-06 .I~6E~ .52~E -00 .155E--0~ .~E--~ .109E--~ ~E~ .~E -~ .1~--~ ~IE-~ .~E -~ .~E-~ .I~E-~ ~E~ .478E-O2 .II~E-01 .I~E-OI ~E-0I .~E--01 .~E-OI .Ill .1~ ·215 .313E--06 .113E--00 .373E--00 .114E--06 .323E--06 .861E--06 · .216E--(~; .514E--~ .117E--04 .2.531[--04 .5Z7E--04 .206E--03 .387E--03 .70~E--03 .12~E--02 .219E-02 -372E--02 .619E-02 .101£--01 .161E-01 .25~E--Of .392E--01 .596E--01 .8~5E-01 .132 .19~ .279 .5~2 .76~E--06 .261E--00 .812E--~ ~E~ .I.~E--~ ~E~ .856E--~5 .184E--04 .~lE--~ ,I~E~ ~E~ .I~E--~ ~E~ ~E--~ ~IE--~ .I~E--OI .I~E-OI ~E--0t .417E-01 5~-01 .~E--OI .115 .1~ .213 .~E--O8 .12,4E--07 .4~E--O7 .141E--06 .42~E--06 .120E--~ .~19E--O5 .19~E--04 .441E--O4 ~E--~ .414E--0~ .~14E--03 .I~E--~ .237E--02 .516E--02 .905E--02 .IS5£--01 ~oOE--OI .4~E--OI .687E- 01 .l~J .169 .258 .389 .577 .846 1.225 1.7,51 I I 300 2.2t$1 1.761 1.211 ,767 J 578 J .425 ,379 .399 .785 2.476 ilo 2.'2'91 .~199 1.524 .977 [ .74! J .353 .4~ .5,.~ I,O&5 3.4~ 5~0/:> ~17 2.723 1.901 1.234 .944 .713 .649 .714 1.486 4,800 830 4.153 3.347 2.353 1.545 I. 191 .911 .840 .944 340 5.011 4.0~4 2.8~1 1.919 ~ .491 I. 1.5~ 1.008 I.~39 2.718 8.957 850 6.006 4.949 3.523 2.36~ 1.852 1.456 1.374 1.614 3.631 12.009 ij ;Ven~.~. ,.,[~.Ro~nso~ N~u~en. and Cmit~o. AI(..TaE me~in~. Anaheim. Calif.. 1982; and p~i~rale oommunicalioa from Pro(. The, od~¢ Ve~m~ule~ Chemical Eagi~ng Dcp~.. mv~-m), d Cahfocn~