APPENDIX K

PIPELINE RISK ANALYSIS

Pipeline Risk Analysis

Mountain House Specific Plan III San Joaquin County, California

June 2004

J House Environmental

Site Assessment Soil & Water Remediation Safety Risk Analysis

Pipeline Risk Analysis

Mountain House Specific Plan III San Joaquin County, California

June 4, 2004

Prepared for:

EDAW, Inc. 2022 J Street

Sacramento, California 95814

Prepared by:

J House Environmental 220 Hidden Creek Drive Auburn, California 95603

TABLE OF CONTENTS Page

INTRODUCTION ..........................................................................................................................1

Purpose and Scope ...................................................................................................................1 Report Organization...................................................................................................................1

SETTING ......................................................................................................................................3

Study Area Development Plans.................................................................................................3 Pipeline Construction Specifications and Operating Parameters ..............................................3 Pipeline Operation, Maintenance, and Safety Procedures ........................................................4 Historic Pipeline Incidents..........................................................................................................5

RISK ANALYSIS ...........................................................................................................................7

Event Identification ....................................................................................................................7 Probability/Frequency Analysis..................................................................................................8 Consequence Analysis ............................................................................................................10 Estimated Individual Risk.........................................................................................................12

RISK MANAGEMENT.................................................................................................................14

ACCEPTABLE LEVEL OF INDIVIDUAL RISK ...........................................................................17

CONCLUSIONS AND RECOMMENDATIONS...........................................................................19

REFERENCES ...........................................................................................................................21

FIGURES

Figure 1.............................................................................................................. Site Location Map

Figure 2............................................................................................................................ Site Plan

Figure 3 .......................................................................................................... Pipeline Risk Zones

TABLES

Table 1 ........................................................................................................Pipeline Specifications

Table 2 .................................................. Estimated Individual Risk for Land Uses Along Pipelines

Table 3 ........................................................................ Individual Risk Values for Various Hazards

Table 4 .................................................. Recommended Setbacks for Land Uses Along Pipelines

APPENDICES

APPENDIX A .....................................Information Provided by Pacific Gas and Electric Company

APPENDIX B ................................................. Information Provided by Coates Field Service, Inc.

APPENDIX C ............................................ Information Provided by California State Fire Marshall

APPENDIX D .......................... Quantitative Risk Analysis Performed by Quest Consultants, Inc.

INTRODUCTION

This report presents the results of the pipeline risk analysis performed by J House Environmental for Mountain House Specific Plan III (SP III). The study area consists of approximately 900 acres located within the Mountain House Master Plan area in southwest San Joaquin County, California (Figure 1). The study area includes approximately 758 of the 812 acres that comprise Specific Plan III, and an approximately 142-acre parcel located adjacent to the southeast boundaries of SP III. The existing Grant Line Village residential neighborhood, located in the northwest portion of SP III, is not included in the study area (Figure 2).

Two natural gas pipelines and a crude oil pipeline are located in a utility corridor that traverses the study area (Figure 2). Natural gas pipelines are also located within the Mountain House Parkway and Von Sosten Road right-of-ways, immediately east of the study area (Figure 2). The natural gas pipelines are owned and operated by Pacific Gas and Electric Company (PG&E). The crude oil pipeline is owned and operated by Chevron Pipe Line Company (CPL).

Purpose and Scope

The purpose of the risk analysis is to identify potential safety hazards associated with the natural gas pipelines and crude oil pipeline and to estimate risks associated with development in proximity to the pipelines. Recommendations for development setbacks for land uses planned along the pipeline alignments are provided based on the risk analysis.

The risk analysis is based on information obtained from PG&E, CPL, and the California State Fire Marshall, Pipeline Safety Program (CSFM) regarding construction specifications, operating parameters, and inspection and maintenance procedures for the subject pipelines. Potential risks associated with pipeline leak and rupture incidents are estimated based on: 1) an identification of events that could lead to pipeline failure; 2) an assessment of the probability or frequency of these events occurring; and 3) an evaluation of the consequences that could result from a pipeline failure. Industry literature and statistics provide a basis for the event identification and probability analyses. Computer modeling and engineering calculations are used to complete the consequence analysis and to estimate the annual individual risk of fatality at varying distances from the subject pipelines.

Report Organization

The remainder of this report is organized into the following sections:

• Setting;

• Risk Analysis;

• Risk Management;

• Acceptable Level of Individual Risk;

1

• Conclusions and Recommendations; and

• References.

Supporting documentation is included in attached appendices.

2

SETTING

This section presents a description of the study area, construction specifications and operating parameters for the subject pipelines, and information regarding PG&E’s and CPL’s operation, maintenance, and safety procedures. Results of a records search for historic pipeline incidents in the vicinity of the study area are also presented in this section.

Study Area Development Plans

The SP III Initial Study (EDAW, 2003) and the SP III Land Use Plan (EDAW, 2004) present a concept plan that includes residential, office/commercial/light industrial, elementary school/community facility, Delta Community College, and open space/recreational development within the study area (Figure 2). The specific development layout for each use area has not yet been determined, and drainage and grading plans for the study area have not been finalized.

Pipeline Construction Specifications and Operating Parameters

Two utility corridors that contain underground pipelines are located in and around the study area. For purposes of discussion in this report, the utility corridors are identified as follows (Figure 2):

• Corridor #1: Combined PG&E/CPL corridor that contains two natural gas pipelines (L401 and L002) and one crude oil pipeline (CSFM 0499).

• Corridor #2: PG&E corridor along Mountain House Parkway and Von Sosten Road. Contains natural gas pipeline L162A, which is fed by L401 from a point south of I-205, and continues east along Von Sosten Road. Contains natural gas pipeline L176, which is fed by L162A and extends along Mountain House Parkway north of Von Sosten Road.

Table 1 presents a summary of available information on construction specifications and operating parameters for the subject pipelines.

According to Mr. Greg Parker, Risk Management Technical Specialist with PG&E, pipeline L401 is a main natural gas transmission line for California. This 36-inch diameter line operates at a pressure of 1040 pounds per square inch gage (psig), which represents a hoop stress of 59.96 percent (%) of the specified minimum yield strength (SMYS). L401 is part of the PG&E Backbone Gas Transmission System that runs from Oregon to Arizona (California Energy Commission, 2001). Pipeline L401 currently has a Class 2 location designation. A Class 2 location is: 1) any class location unit (an area that extends 220 yards on either side of the centerline of any continuous one mile length of pipeline) that has more than 10 but fewer than 46 buildings intended for human occupancy (Code of Federal Regulations, Title 49 [CFR 49] Part 192.5).

3

Pipelines L002, L162A, and L176 are also part of PG&E’s California natural gas transmission system. L002 is a 26-inch diameter line that operates at 890 psig, which represents a hoop stress of 59.96% SMYS. L162A is a 10.75-inch line that operates at 365 psig (24.85% SMYS) and L176 is a 6.625-inch diameter line that operates at 365 psig (21.44% SMYS). Line L162A currently has a Class 2 location designation, as described above for L401. Lines L002 and L176 currently have a Class 1 location designation. A Class 1 location is: 1) an offshore area; or 2) any class location unit that has 10 or fewer buildings intended for human occupancy (CFR 49 Part 192.5).

According to Mr. Parker, L401, L002, L162A, and L176 are constructed of welded steel and have cathodic protection systems. A facsimile from Mr. Parker that presents information on the subject pipelines is presented in Appendix A.

Mr. Larry Whitehead and Mr. Ernie Browning of CPL were contacted to obtain details regarding the crude oil pipeline that traverses the study area. The request for information was forwarded to Mr. Mark Zahn of Coates Field Service, Inc. (Coates), contract right of way agent for CPL. According to Coates, the KLM (Kettleman-Los Medanos) crude oil pipeline was installed in 1945 and has a diameter of 18-inches. The pipeline is constructed of welded carbon steel. Coates has indicated that additional details regarding construction and operation of the CPL crude oil pipeline are company confidential and are not available for release to the public. The information provided by Coates is presented in Appendix B.

Mr. Bob Gorham, Supervising Pipeline Safety Engineer with CSFM, was contacted for additional public information on the CPL pipeline. CSFM regulates the safety of intrastate hazardous liquid transportation pipelines and acts as an agent of the Federal Department of Transportation, Office of Pipeline Safety (OPS) concerning the inspection of interstate pipelines. According to CSFM, the subject pipeline, CSFM# 0499 KLMR, was installed between 1942 and 1946 and extends from Los Banos to Los Medanos. The 18-inch diameter pipeline has an impressed current cathodic protection system. The maximum operating pressure for this pipeline is reported as 920 psig. Information provided by Mr. Gorham is presented in Appendix C.

Pipeline Operation, Maintenance, and Safety Procedures

The PG&E natural gas pipelines and the CPL crude oil pipeline are constructed, operated, and maintained in accordance with state and federal regulations set forth in CFR 49 Parts 190, 191, 192, and 195; California Public Utilities Commission General Order No. 112-E; and the California Pipeline Safety Act of 1981 (California Government Code). Requirements and procedures established in these regulations to safeguard health, property, and public welfare include the following:

• Procedural manuals for operation, maintenance, and emergencies, including an emergency contingency plan, are maintained;

4

• Public information outreach programs, including public awareness education and pipeline marking that includes emergency telephone numbers, are implemented;

• Pressure tests are performed for all new installations and specified spacing for main-line valves is adhered to;

• Corrosion protection measures, i.e. cathodic protection, pipeline coating, and annual reads on corrosion potentials, are implemented on all pipelines;

• Annual, semi-annual, and/or quarterly pipelines inspections are performed, including annual valve maintenance;

• Patrolling for evidence of pipeline damage or any condition which may impact continued safe operation is periodically conducted;

• Pipeline integrity testing is periodically conducted;

• Pipelines are installed with a minimum 30- to 36-inch cover and are constructed using modern weld design techniques; and

• Any excavation activities near pipelines may only be conducted 48 hours after Underground Services Alert (USA) has been notified.

Historic Pipeline Incidents

PG&E records indicate that there have been no pipeline leaks or ruptures in the vicinity of the study area since the time of installation of the subject natural gas pipelines. The OPS natural gas transmission incident databases for 1970 to mid-1984, mid-1984 to 2001, and 2002 to present (as of 11/7/03) were reviewed. No pipeline incidents were recorded within San Joaquin County for PG&E natural gas pipelines L401, L002, L162A, and L176.

On December 4, 2003, crude oil pipeline CSFM 0499 was accidentally struck by a tractor working on farmland in the southeast portion of the study area. Approximately 21,000 gallons (500 barrels) of crude oil were reportedly released due to this incident. No injuries were reported. Additional details regarding the extent of soil and/or groundwater impact resulting from this release were not available at the time of preparation of this report.

CSFM records (as of 11/17/03) indicate that two incidents have occurred along the subject crude oil pipeline, one at a location approximately 13 miles from the study area and the second at a location approximately 98 miles from the study area. Both incidents were pinhole leaks that resulted from external corrosion. The leaks were reported on December 5, 1997 and involved the release of 4 barrels and 12 barrels, respectively, of crude oil.

5

The OPS hazardous liquid incident databases for pre-1986, 1986 to January 2002, and January 2002 to present (as of 11/7/03) were reviewed. No pipeline incidents were recorded for the CPL crude oil pipeline in the immediate vicinity of the study area.

The National Response Center (NRC) is the federal point of contact for reporting oil and chemical spills. A NRC database search (12/1/03) showed no reportable oil spills associated with the CPL crude oil pipeline in the immediate vicinity of the study area.

The California Department of Toxic Substances Control (DTSC) hazardous waste and substances site list (Cortese List), a database of hazardous materials release sites, was reviewed (11/7/03) and showed no indication of releases from the CPL pipeline in the immediate vicinity of the study area. Database searches conducted as part of the Phase I Environmental Site Assessments for properties that comprise SP III (Teixeira Property [Wallace Kuhl & Associates, 2003a], Muela Property [Wallace Kuhl & Associates, 2003b], Tuso Property [Wallace Kuhl & Associates, 2003c], and Proposed Delta College [Levine Fricke, 2001]) and for the parcel located adjacent to the southeast boundaries of SP III (Kleinfelder, Inc., 2002a) did not identify any petroleum product releases from the CPL pipeline.

6

RISK ANALYSIS

This section presents the risk analysis for the subject pipelines. Since the pipelines do not pose a safety hazard unless their structural integrity is compromised, resulting in a release to the environment, the first step in this risk analysis is to identify events that could lead to a pipeline leak or rupture. In the second step, the probability or frequency of such events occurring is assessed. Consequences that could result from a pipeline leak or rupture are then evaluated and the annual individual risk (expressed as probability of being exposed to a fatal hazard over a one-year period) at varying distances from the subject pipelines is estimated.

Event Identification

Four types of events are generally recognized as the main causes of pipeline leak and/or rupture:

• Third Party Dig-ins;

• Corrosion and Deterioration;

• Weld or Material Defects; and

• Ground Movement.

Third party dig-ins can result from activities that are not associated with pipeline construction and maintenance. Third party dig-ins are generally associated with development or reconstruction projects (i.e., subsurface digging with a backhoe or exploratory soil borings).

Pipeline corrosion and deterioration can occur both internally and externally. There are a number of possible causes of corrosion and deterioration. The presence of carbon dioxide and water in natural gas is generally the main reason for internal corrosion of natural gas pipelines. External corrosion or deterioration of natural gas and crude oil pipelines is generally the result of direct contact of the pipeline material with soils, water, and/or air.

Weld or material defects can weaken pipeline structures and result in leaks and/or ruptures. Improper material selection, pipeline design and construction, or quality control can lead to potential weld and material defects that can compromise the pipeline integrity.

Ground movement can compromise the structural integrity of a pipeline, resulting in leaks or ruptures. Underground pipelines are most sensitive to ground movement associated with fault rupture, liquefaction, and landslides.

7

Probability/Frequency Analysis

The pipeline failure probability rates used in the risk analysis are based on OPS incident data, as described in the quantitative risk analysis performed by Quest Consultants Inc., presented in Appendix D. The assumed failure rates are as follows:

• 1.5 X 10-3 failures per mile per year for 6-inch to 12-inch diameter natural gas transmission pipelines; rate applies to L162A and L176.

• 5.6 X 10-4 failures per mile per year for 24-inch to 28-inch diameter natural gas transmission pipelines; rate applies to L002.

• 2.0 X 10-4 failures per mile per year 30-inch to 36-inch diameter natural gas transmission pipelines; rate applies to L401.

• 1.2 X 10-3 failures per mile per year for liquid pipelines; rate applies to CSFM 0499.

The probability of a pipeline incident occurring within or adjacent to the study area is related to the probability of occurrence of the four types of events described earlier. A qualitative assessment of the potential for each of these events to occur has been conducted. The qualitative assessment ranks the likelihood of an event occurring as very low, low, moderate, high, or very high. Based on results of the qualitative assessment, the OPS incident rates presented above and used in the quantitative risk analysis are considered conservative as applied to the study area.

Third Party Dig-Ins

The potential for third party dig-ins to occur is related to the amount of construction being performed in the immediate vicinity of a pipeline. The study area is planned for development as part of the Mountain House Master Plan, with build-out projected to occur over several years. As with all construction projects of this nature, work will be conducted by licensed contractors and, as required by law, USA will be contacted prior to any excavation activities. Additional precautionary measures that are being implemented as part of the Mountain House Master Plan development include close coordination between project engineers and pipeline operators, pothole mapping to confirm pipeline locations within the PG&E/CPL easement that traverses the study area, and adherence to special restrictions regarding grading, construction, landscaping, and load limitations within the pipeline easement. The potential for third party dig-ins to occur is considered moderate.

Corrosion and Deterioration

The potential for pipeline corrosion and deterioration to occur is related to pipeline material type, the age of the pipeline, and corrosion preventative measures (i.e., cathodic protection and/or

8

protective coatings). The 36-inch diameter PG&E natural gas pipeline (L401) was installed in 1993; the other three natural gas pipelines (L002, L162A, and L176) were installed between 1967 and 1973. The PG&E pipelines all have cathodic protection systems. The 18-inch diameter CPL crude oil line (CSFM 0499) was constructed in 1945 and has an impressed current cathodic protection system. Routine maintenance and inspection of the subject pipelines by PG&E and CPL have not identified any concerns with respect to corrosion or deterioration. The potential for a compromise in the structural integrity of the subject pipelines to occur due to corrosion or deterioration is considered low to moderate.

Weld or Material Defects

The potential for weld or material defects to occur is related to the use of insufficiently qualified operators (welders) and/or defectively manufactured materials. PG&E has indicated that design and construction of all gas transmission facilities is regulated by the CPUC General Order 112-E and CFR 49 Part 192. Routine maintenance and inspection of the subject pipelines by PG&E has not identified any concerns with respect to weld or material defects. Operation of the CPL crude oil pipeline is regulated by CFR 49 Part 195 and the California Pipeline Safety Act of 1981 (California Government Code). Routine maintenance and inspection of the subject pipelines by PG&E and CPL have not identified any concerns with respect to weld or material defects. The potential for a compromise in the structural integrity of the subject pipelines to occur due to weld or material defects is considered low to moderate.

Ground Movement

The potential for ground movement to occur in the area of the subject pipelines is related to the potential for surface fault rupture, seismic shaking, liquefaction, and/or landsliding. The geologic setting in the study area provides a basis for assessment of the probability of ground movement to occur.

The study area is not located within a currently-designated Alquist-Priolo Earthquake Fault Zone (Hart and Bryant, 1997). These zones are defined by the State of California, Department of Conservation, Division of Mines and Geology (DMG) to identify areas at risk from surface fault rupture. No active faults are located in the immediate vicinity of the study area; the closest known active fault is the Greenville fault, located approximately 8 miles to the southwest (Baseline Environmental Consulting, 1994; Kleinfelder, Inc., 2002b). The potential for surface fault rupture to occur in the vicinity of the subject pipelines is considered to be very low.

The study area is located in a region of moderate to high seismicity, due to proximity to the San Andreas and Great Valley fault systems. A number of active and potentially active faults are located within approximately 30 miles of the study area (Baseline Environmental Consulting, 1994; Kleinfelder, Inc., 2002b). A probabilistic seismic analysis was conducted for the southeast portion of the study area by Kleinfelder, Inc. (Kleinfelder, Inc. 2002b). Results of this analysis indicate a peak horizontal ground acceleration of 0.50g (10 percent probability of being

9

exceeded within a 50-year period). The regional probabilistic seismic map prepared by DMG for the state (Petersen, et.al., 1996) shows that the study area lies within an area of peak ground acceleration of 0.40g to 0.50g. Overall, the potential for seismic shaking to occur in the vicinity of the subject pipelines is considered to be moderate to high.

The Preliminary Geotechnical Services Report, Mountain House Business Park, Mountain House California (Kleinfelder, 2002b), prepared for the southeast portion of the study area, indicates that the potential for liquefaction to occur is remote, due to the presence of dense soils and the absence of shallow groundwater. Based on the presence of clayey soils throughout the study area (Baseline Environmental Consulting, 1994), the potential for liquefaction to occur in the vicinity of the subject pipelines is considered very low.

The study area is located in a relatively flat topographic region. Therefore, the potential for landsliding to occur in the vicinity of the subject pipelines is considered very low.

The overall potential for a compromise in the structural integrity of the subject pipelines to occur due to ground movement is considered low. The pipelines are located in an area of moderate to high seismicity. However, the potential for surface fault rupture is very low and there is little or no potential for liquefaction or landsliding to occur.

Consequence Analysis

The primary hazard associated with a release from the subject natural gas and crude oil pipelines is flammability. A secondary concern associated with a release from the crude oil pipeline is contamination of environmental media (soil, surface water, etc.). An assessment of each of these hazards is presented in this section.

Flammability

Consequences associated with a release of flammable material from the subject pipelines have been evaluated using computer modeling performed by Quest Consultants Inc. A description of the modeling approach, assumptions, and results is presented in Appendix D.

The consequence analysis evaluates a series of failure scenarios defined by different release conditions and various resulting hazards. The release conditions that were evaluated include a ¼ -inch diameter hole, a 2-inch diameter hole, and full pipeline rupture. The resulting hazards that were assessed include exposure to heat radiation from a torch fire or a pool fire, exposure to a flash fire, and exposure to explosion overpressure. A set of hazard zones for each failure scenario have been developed using the CANARY by Quest computer software hazards analysis package, which takes into account release conditions, ambient weather conditions, effects of local terrain, and mixture thermodynamics.

10

The predicted hazard zones represent areas where injuries or fatalities could occur. However, the probability of occurrence of an event that would create such a hazard zone is not incorporated into the consequence analysis. The estimation of individual risk, presented in the next section, combines the likelihood of each specific release scenario occurring with the associated consequences, to predict the probability of an individual being exposed to a fatal hazard over a one-year period.

As described in the quantitative risk analysis presented in Appendix D, the dominant hazard is exposure to heat radiation from torch fires. The analysis estimates that the maximum downwind hazard zone distance for a torch fire resulting from full rupture of the 36-inch diameter natural gas pipeline L401 is 1,496 feet with immediate ignition and 912 feet with delayed ignition. Maximum torch fire hazard zones are smaller for the smaller diameter natural gas pipelines and for the ¼-inch hole and 2-inch hole release scenarios, rather than the full pipeline rupture scenario (Table 3-10, Appendix D).

The maximum downwind hazard zone distance for a pool fire resulting from full rupture of the 18-inch diameter crude oil line CSFM 0499 is 129 feet with immediate ignition and 217 feet with delayed ignition. Maximum hazard zones are smaller for the ¼-inch and 2-inch hole release scenarios, rather than the full pipeline rupture scenario (Table 3-10, Appendix D).

The maximum downwind hazard zone distance for a flash fire resulting from full rupture of the 36-inch diameter natural gas pipeline L401 is 1,198 feet. The flash fire hazard zones are smaller for the smaller diameter natural gas pipelines and for the leak and puncture, rather than full pipeline rupture, scenarios (Table 3-10, Appendix D).

As described in Appendix D, releases from the subject natural gas pipelines would have little potential to create significant overpressures. The peak overpressure predicted in the quantitative risk analysis is well below the threshold for injury to humans. Therefore, hazard zones have not been predicted for explosion overpressure following ignition of released natural gas.

Contamination of Environmental Media

In the event of leak or rupture of the crude oil pipeline, released product would seep through surrounding soil and could result in degradation of surface water or groundwater resources. Remediation of impacted soil, surface water, and groundwater could be required to protect human health and the environment.

In the case of a relatively small release, implementation of spill response measures may be sufficient to remove impacted soil and prevent migration of released liquids. In the event of a larger release, active remediation of the impacted media (soil, surface water, and/or groundwater) may be required. The type of site remediation that would be needed would

11

depend on the volume of material released and the distribution of the released material in the environmental media.

Estimated Individual Risk

Annual individual risk (expressed as the probability of being exposed to a fatal hazard over a one-year period) is a function of the probability of a pipeline incident occurring and the consequences that could result from a pipeline incident. Risk transects (measurements of individual risk as a function of distance from a pipeline) were constructed for the subject pipelines as part of the quantitative risk analysis modeling (Figures 6-2, 6-3, and 6-4, Appendix D). The transects illustrate how the risk decreases as the distance from the pipeline increases. Individual risk contours constructed for the system of pipelines in the SP III area illustrate the combination of risk due to multiple pipelines in the different utility corridors (Figure 6-5, Appendix D).

As shown by the risk transects and risk contours, there are no locations where an annual individual risk level of 1.0 X 10-5 (one chance in one hundred thousand of being exposed to a fatal hazard over a one-year period) is reached. As described in Appendix D, individual risk levels drop to negligible values at approximately 1,500 feet from the combined PG&E/CPL corridor, 500 feet from PG&E natural gas pipeline L162A, and 300 feet from PG&E natural gas pipeline L176. These distances roughly correspond to the maximum downwind hazard zone distance for each pipeline. Beyond these distances, the risk of being fatally affected by a release from the subject pipelines is zero.

The level of risk shown by the risk transects and risk contours is the risk of lethal exposure to any of the hazards associated with the various release scenarios modeled for the subject pipelines. For example, the annual individual risk level of 10-6 shown on the contour map (Figure 6-5, Appendix D) represents one chance in a million of being exposed to a fatal hazard from any of the possible release scenarios associated with any of the subject pipelines over a one-year period. The risk transects and risk contours presented in Appendix D are based on an individual being present at a specific location for 24 hours a day, 365 days per year. As described in Appendix D, individual risk levels would be proportionately lower for areas subject to mobile populations, where individuals would only be present for a portion of a day.

The SP III concept plan incorporates a variety of land uses, including use areas subject to mobile populations. Individual risk values set forth in Appendix D have been proportionately adjusted for planned land use, based on the following occupancy factors:

• Residential use – 100% occupancy; an individual would be present 8,736 hours during a year.

• Office/commercial/light industrial use - 30% occupancy; an individual would be present 2,600 hours during a year, which reflects an average of 50 hours per week.

12

• Delta Community College - 30% occupancy; an individual would be present 2,600 hours during a year, which reflects an average of 50 hours per week.

• Open space/recreational use - 15% occupancy; an individual would be present 1,300 hours per year, which reflects an average of 25 hours per week.

Table 2 shows the estimated individual risk for the planned land uses at selected distances from the subject pipelines. As shown, the highest annual individual risk is for residential use in proximity to the combined PG&E and CPL utility corridor. The highest estimated annual individual risks associated with PG&E lines L162A and L176 are 5.3 X 10-7 and 3.7 X 10-7, respectively, for office/commercial/light industrial use.

13

RISK MANAGEMENT

Risk management measures are intended to: 1) reduce the probability of occurrence of an event that could result in a pipeline failure and 2) mitigate the consequences that could result if pipeline failure were to occur due to such an event. PG&E and CPL have a number of risk management measures in place to accomplish these goals. Additional measures have been incorporated into the Mountain House Master Plan (MHMP) and Specific Plan III requirements to minimize risks associated with development in proximity to underground pipelines.

The matrix table presented below highlights measures intended to reduce the probability of occurrence of the key events associated with pipeline failure. Risk management measures that are implemented to mitigate the consequences of a pipeline incident are also shown in the matrix.

Main Causes of Pipeline Failure

Risk Management Measures Third Party

Dig-ins Corrosion and Deterioration

Ground Movement

Weld or Material Defects

Design, construction, operation, and maintenance in accordance with CFR, CPUC, and California Government Code requirements.

PG&E CPL

PG&E CPL

Cathodic protection monitoring, annual leak surveys, pipeline patrolling and inspection, and periodic pressure tests.

PG&E CPL

PG&E CPL

Line marking, public education, participation in USA, and minimum cover.

PG&E CPL

MHMP

Pothole mapping of pipeline locations and restrictions regarding grading, construction, landscaping, and load limitations within pipeline easements.

PG&E CPL

MHMP

Development and maintenance of emergency planning documents, spill response plans, and training programs to address emergencies.

PG&E CPL

MHMP

PG&E CPL

MHMP

PG&E CPL

MHMP

PG&E CPL

MHMP

14

The amount of ground cover over pipelines is mandated by the US Department of Transportation, Office of Pipeline Safety. Ground cover requirements are set forth in the Code of Federal Regulations (49CFR192 for gas pipelines; 49CFR195 for hazardous liquid pipelines).

The PG&E natural gas transmission pipelines are subject to the following cover requirements:

--------------------------------------------------------------------------------------------------------------------------- Normal Consolidated Location soil rock --------------------------------------------------------------------------------------------------------------------------- ___Inches (millimeters)____ Class 1 locations................................................... 30 (762) 18 (457) Class 2, 3, and 4 locations.................................... 36 (914) 24 (610) Drainage ditches of public roads and railroad crossings.............................................................. 36 (914) 24 (610) ---------------------------------------------------------------------------------------------------------------------------

The Chevron Pipeline Company crude oil transmission line is subject to the following cover requirements:

---------------------------------------------------------------------------------------------------------------------------- Cover inches (millimeters) ------------------------------------------------- Location Normal Rock excavation excavation * ---------------------------------------------------------------------------------------------------------------------------- Industrial, commercial, and residential areas......... 36 (914) 30 (762) Crossings of inland bodies of water with a width of at least 100 ft (30 mm) from high water mark to high water mark................................................. 48 (1219) 18 (457) Drainage ditches at public roads and railroads….. 36 (914) 36 (914) Deepwater port safety zone................................... 48 (1219) 24 (610) Gulf of Mexico and its inlets and other offshore areas under water less than 12 ft (3.7 m) deep as measured from the mean low tide.................... 36 (914) 18 (457) Any other area....................................................... 30 (762) 18 (457) ----------------------------------------------------------------------------------------------------------------------------

* Rock excavation is any excavation that requires blasting or removal by equivalent means.

If it appears that existing cover at the project site does not meet these requirements, the pipeline operator should be contacted to determine the reason for the variance and to confirm that it has been approved by governmental oversight agencies.

As noted earlier in this report, the class location designations for the subject natural gas pipelines are Class 1 and Class 2. With build-out of the Mountain House Master Plan, a change in the class

15

location designations to Class 3 will be required. In accordance with CFR 49 Part 192.609 and 192.611, PG&E will be required to conduct a technical study to support the change in class location designations for pipelines L401 and L002, since they operate at a hoop stress greater than 40% SMYS. If results of the technical study indicate that existing pipeline construction and operating pressure cannot support a change to Class 3 designation, a reduction in operating pressure or pipeline construction upgrade could be required. This would reduce the estimated level of individual risk at varying distances from the pipeline. As indicated above, Class 3 locations require a greater amount of cover than Class 2 locations. The need for any additional cover in the project area should be addressed as part of the PG&E technical study that will be required to support the change in class location designations for pipelines L401 and L002.

16

ACCEPTABLE LEVEL OF INDIVIDUAL RISK

An acceptable level of individual risk for hazards associated with underground pipelines has not been established by the State of California or the federal government for new development projects such as the Mountain House Master Plan. Standards that have been proposed and used by various governmental agencies worldwide generally consider individual risk levels below 1 x 10–6 (one-in-a-million) acceptable (Cornwell, John B. et al, 1997). Individual risk levels greater than 1 x 10–5 (one-in-one hundred thousand) are generally considered unacceptable. Most planning and facility siting studies adopt a “gray zone” or negotiable area between the unacceptable and acceptable individual risk levels. A comparison of these standards with individual risk values for various hazards is presented in Table 3.

In selecting individual risk criteria for the Mountain House Master Plan, the local community’s tolerance for risk needs to be taken into consideration. The acceptability of the risk involved and the benefits of the proposed development should be weighed and evaluated by the various stakeholders. San Joaquin County has indicated that an approach that involves designation of three risk zones would be acceptable for SP III, as follows:

• No build zone – specified land use not allowed.

• Hazard notification zone – specified land use allowed with disclosure of potential risk to property owner.

• No constraint zone – specified land use allowed with no conditions or constraints.

Based on generally accepted criteria described earlier in this section, it is suggested that the no constraint zone be defined as locations where the annual individual risk is lower than 1 x 10–6. As shown in Table 2, the estimated annual individual risk associated with the subject pipelines is lower than 1 x 10–6 for all of the planned land uses except residential. Therefore, it is recommended that all non-residential use areas be designated as no constraint zones.

The estimated annual individual risk for residential land use exceeds 1 x 10–6 in proximity to the combined PG&E and CPL corridor (natural gas lines L401 and L002; crude oil line CSFM 0499) (Table 2). Risk transects and risk contours presented in Appendix D show that at distances greater than 249 feet from pipelines located within this corridor, the estimated annual individual risk for residential land use drops to below 1 x 10–6. It is recommended that residential use areas beyond this 249-foot setback distance be designated as no constraint zones.

Planned residential land use areas located closer than 249 feet from pipelines within the combined corridor should be designated as hazard notification zones or as no build zones. The estimated annual individual risk in these areas falls within the “gray zone” defined by risk values greater than 1 x 10–6 but lower than 1 x 10–5 (Table 3). Selection of an individual risk threshold

17

value to define the no build versus hazard notification zones in the SP III area is subjective and should be based on community risk tolerance and risk acceptability. San Joaquin County has indicated that a conservative threshold value, which would minimize potential risk to area residents, should be applied to the SP III development area. The County has indicated that a threshold value of 2 X 10–6, which is consistent with the lower 11% of the generally accepted “gray zone”, would be acceptable to define the no build versus hazard notification zones.

As shown in Table 4, the estimated annual individual risk associated with the natural gas pipelines and the crude oil pipeline located within the combined PG&E and CPL corridor exceeds 2.0 X 10–6 for a distance of 68 feet from the pipelines. A no build zone for residential use would therefore be designated for areas within this 68-foot setback. Residential use areas located between 68 feet and 249 feet from the nearest pipeline would be designated as hazard notification zones. As described earlier, residential use areas beyond the 249-foot setback would be designated as no constraint zones.

The residential setback distances are based on the cumulative estimated annual individual risk associated with the three pipelines located within the combined PG&E and CPL corridor, rather than the risk associated with any one specific pipeline. Due to the proximity of the three pipelines, the cumulative annual individual risk derived by the quantitative risk analysis modeling is based on the pipelines being co-located at the centerline of the utility corridor. This approach is reasonable given the degree of uncertainty inherent in the modeling. To ensure that setback distances are adequate to meet the desired risk threshold levels, measured distances should be perpendicular from the nearest pipeline within the corridor (rather than from the centerline of the corridor) to the closest structure intended for residential occupancy, including attached porches and balconies. Outdoor structures such as pools, fences, patios, and decks could be allowed within the no build zone, since occupancy would be less than 100%. Detached garage and attached garage could be permitted within the no build zone with conditions imposed that would prevent future use or conversion for any type of residential occupancy, including bonus rooms, granny flats, or home offices.

18

CONCLUSIONS AND RECOMMENDATIONS

The pipeline risk analysis for Mountain House Specific Plan III indicates that the estimated annual individual risk at the planned office/commercial/light industrial, Delta Community College, and open space/recreation land use areas is less than 1 x 10-6. It is recommended that these land use areas be designated no constraint zones, since the level of estimated risk is within the range that is generally considered acceptable.

The estimated annual individual risk for residential land use exceeds 1 x 10–6 in proximity to the combined PG&E and CPL corridor (natural gas lines L401 and L002; crude oil line CSFM 0499). At distances greater than 249 feet from pipelines located within this corridor, the estimated annual individual risk for residential land use drops to below 1 x 10–6. It is recommended that residential use areas beyond this 249-foot setback distance be designated as no constraint zones.

Based on a conservative threshold value 2 X 10–6, a no build zone for residential use would be designated for areas within 68 feet of the nearest pipeline within the combined PG&E and CPL corridor and a hazard notification zone would be established for residential use areas between 68 feet and 249 feet of the nearest pipeline within the corridor. The setback distances should be measured perpendicular from the nearest pipeline within the combined PG&E and CPL corridor to the closest structure intended for residential occupancy, including attached porches and balconies.

Risk management measures currently in place by PG&E and CPL appear adequate to minimize the potential for occurrence of an event that could result in pipeline failure. Additional measures have been incorporated into the Mountain House Master Plan and Specific Plan III requirements to minimize risks associated with development in proximity to underground pipelines.

To provide an added degree of risk management, J House Environmental recommends that any evacuation plans, public health and safety plans, or emergency response training plans that are developed for SP III identify the presence of the subject pipelines and take into consideration procedures that could be implemented to reduce risks associated with pipeline failure. Site-specific risk management measures could include:

• Identifying evacuation routes that direct the public away from the pipelines;

• Maintaining an emergency contact list with phone numbers of local police, fire departments, and pipeline operators; and

• Providing special pipeline incident training and equipment to local emergency responders.

19

It is recommended that site specific risk management measures for SP III be developed in coordination with pipeline operators, County officials, and the Mountain House Community Services District.

20

REFERENCES

Auburn Journal, Sunday, December 7, 2003.

Baseline Environmental Consulting, Final Environmental Impact Report, Mountain House Master Plan and Specific Plan I, September 1994.

California Energy Commission, Natural Gas Infrastructure Issues, October 2001.

California Government Code, Title 5, Division 1, Part 1, Chapter 5.5 The Elder California Pipeline Safety Act of 1981, Section 51010-51019.1.

California Public Utilities Commission, General Order 112-E.

California State Fire Marshall, Hazardous Liquid Pipeline Risk Assessment, March 1993.

California State Fire Marshall, Pipeline Safety Program, Personal communication between Bob Gorham, Supervising Pipeline Safety Engineer of SFM and Jackie House of J House Environmental.

Chevron Pipe Line Company, Personal communication between Ernie Browning of CPL and Jackie House of J House Environmental.

Chevron Pipe Line Company, Personal communication between Larry Whitehead of CPL and Jackie House of J House Environmental.

Coates Field Service, Inc., Written communication between Mark Zahn of Coates and Jackie House of J House Environmental.

Code of Federal Regulations, Title 49, Parts 190, 191, 192, 195.

Cornwell, John B. and Marx, Jeffrey D., Quest Consultants Inc., Application of Quantitative Risk Analysis to Code-Required Siting Studies Involving Hazardous Material Transportation Routes, (undated).

Cornwell, John B. and Meyer, Mark M., Quest Consultants Inc., Risk Acceptance Criteria or “How Safe is Safe Enough?”, October 13, 1997.

Cornwell, John B. and Martinsen, William E., Quest Consultants Inc., Uncertainties in Pipeline Risk Analysis, September, 1990.

EDAW, Mountain House Specific Plan III Initial Study, October 13, 2003.

EDAW, Mountain House Specific Plan III & EIR Land Use Plan, April 23, 2004

21

Federal Department of Transportation, Office of Pipeline Safety, http://ops.dot.gov/stats

Federal Emergency Management Agency, U.S. Department of Transportation, U.S. Environmental Protection Agency, Handbook of Chemical Hazard Analysis Procedures, October 1990.

Hart, Earl W. and Bryant, Williams A., Fault Rupture Hazard Zones in California, California Division of Mines and Geology Special Publication 42, 1997.

Kleinfelder, Inc., Phase I Environmental Site Assessment, Mountain House Parkway, Mountain House, California, November 12, 2002 (2002a).

Kleinfelder, Inc., Preliminary Geotechnical Services Report, Mountain House Business Park, Mountain House, California, November 6, 2002 (2002b).

Levine Fricke, Phase I Environmental Site Assessment, Proposed Delta College at Mountain House Site, San Joaquin County, California, June 29, 2001.

Major Industrial Accidents Council of Canada, Land Use Planning With Respect to Pipelines, A Guideline for Local Authorities, Developers and Pipeline Operators, September, 1999.

Pacific Gas and Electric Company, Personal communication and written communication between Greg Parker, Risk Management Technical Specialist of PG&E and Jackie House of J House Environmental.

Pacific Gas and Electric Company, Personal communication between Thomas Crawford of PG&E and Jackie House of J House Environmental.

Petersen, Mark D., et. al., Probabilistic Seismic Hazard Assessment for the State of California, California Division of Mines and Geology, Open-File Report 96-08, 1996.

Wallace Kuhl & Associates, Inc., Environmental Site Assessment, Mountain House Teixeira Property, June 2, 2003, (2003a).

Wallace Kuhl & Associates, Inc., Environmental Site Assessment, Mountain House Muela Property, June 2, 2003, (2003b).

Wallace Kuhl & Associates, Inc., Environmental Site Assessment, Mountain House Tuso Property, June 2, 2003 (2003c).

22

FIGURES

TABLES

TABLE 1

PIPELINE SPECIFICATIONS

MOUNTAIN HOUSE SPECIFIC PLAN III PIPELINE RISK ANALYSIS

Pipeline ID Type Date of Installation

Diameter (inches)

Pipe Wall Thickness (inches)

Operating Pressure

(psig)

Distance to Nearest Shutoff Valves

(feet) Utility Corridor #1

L401 Natural GasTransmission

1993 36 0.446 1040 6000’ north3000’ south

L002 Natural GasTransmission

1972 26 0.322 890 6000’ north3000’ south

CSFM 0499 Crude Oil 1945 18 NA 920 maximum 13 miles upstream 5 miles downstream

Utility Corridor #2 L162A Natural Gas

Transmission 1967 10.75 0.1880 365 7000’ northeast

4000’ south L176 Natural Gas

Transmission/ Gathering

1973 6.625 0.1880 365 5000’ south26,000’ north

NA - information not available; company confidential psig - pounds per square inch gage

TABLE 2 ESTIMATED INDIVIDUAL RISK FOR LAND USES ALONG PIPELINES

MOUNTAIN HOUSE SPECIFIC PLAN III PIPELINE RISK ANALYSIS

Combined PG&E and CPL Corridor L401, L002, CSFM 0499

PG&E Natural Gas Pipeline L162A

PG&E Natural Gas Pipeline L176

Planned Land Use

At

pipeline

At 100 feet

from pipeline

At 200 feet

from pipeline

At 300 feet

from pipeline

At

pipeline

At 50 feet from

pipeline

At 100 feet

from pipeline

At

pipeline

At 50 feet from

pipeline

At 100 feet

from pipeline

Residential

2.8 X 10-6

1.6 X 10-6

1.1 X 10-6

8.5 X 10-7

NA

NA

NA

NA

NA

NA

Office/

Commercial/ Light Industrial

8.4 X 10-7

4.9 X 10-7

3.4 X 10-7

2.6 X 10-7

5.3 X 10-7

4.5 X 10-7

3.6 X 10-7

3.7 X 10-7

2.5 X 10-7

1.5 X 10-7

Delta

Community College

NA

4.9 X 10-7

3.4 X 10-7

2.6 X 10-7

NA

NA

NA

NA

NA

NA

Open Space/ Recreation

4.2 X 10-7

2.4 X 10-7

1.7 X 10-7

1.3 X 10-7

NA

NA

NA

NA

NA

NA

Individual risk expressed as probability of being exposed to a fatal hazard over a one-year period.

Distances are in feet perpendicular to pipeline.

Assumed land use occupancy factors: residential 100%, office/commercial/light industrial 30%, Delta Community College 30%, open space/recreation 15%.

NA – Not applicable. Specified land use not planned at given location.

TABLE 3

INDIVIDUAL RISK VALUES FOR VARIOUS HAZARDS

MOUNTAIN HOUSE SPECIFIC PLAN III PIPELINE RISK ANALYSIS

Cause of Fatality Annual Individual Risk

Acceptance Criteria

Heart Disease 3.2 X 10-3

Cancer 1.9 X 10-3

Pneumonia 2.8 X 10-4

Motor Vehicles 1.9 X 10-4

Diabetes

1.5 X 10-4

Falls

5.0 X 10-5

Drowning

2.2 X 10-5

Fires, Burns

2.1 X 10-5

Greater than 1.0 X 10-5

Generally Unacceptable Level of Public Risk

Air Travel

9.0 X 10-6

Electrocution

6.0 X 10-6

Railways

4.0 X 10-6

Excessive Cold

4.0 X 10-6

Between 1.0 X 10-5 and 1.0 X 10-6 “Gray Zone” – Negotiable Level of Public Risk

Natural Disaster (tornado, flood, earthquake, etc.)

9.0 X 10-7

Excessive Heat

9.0 X 10-7

Lightning

4.0 X 10-7

Less than 1.0 X 10-6

Generally Acceptable Level of Public Risk

TABLE 4 RECOMMENDED SETBACKS FOR LAND USES ALONG PIPELINES

MOUNTAIN HOUSE SPECIFIC PLAN III PIPELINE RISK ANALYSIS

Combined PG&E and CPL Corridor L401, L002, CSFM 0499

PG&E Natural Gas Pipeline L162A

PG&E Natural Gas Pipeline L176

Planned Land Use

No Build Zone

Hazard Disclosure

Zone

No Constraint

Zone

No Build Zone

Hazard Disclosure

Zone

No Constraint

Zone

No Build Zone

Hazard Disclosure

Zone

No Constraint

Zone

Residential

Within 68’

68’ – 249’

Beyond 249’

NA

NA

NA

NA

NA

NA

Office/

Commercial/ Light

Industrial

None

None

All locations

None

None

All locations

None

None

All

locations

Delta

Community College

None

None

All locations

NA

NA

NA

NA

NA

NA

Open Space/ Recreation

None

None

All locations

NA

NA

NA

None

None

All

locations

Setbacks are distances in feet perpendicular to pipeline. No Build Zone based on annual individual risk of fatality exceeding 2.0 X 10-6. Hazard Disclosure Zone based on annual individual risk of fatality between 1.0 X 10-6 and 2.0 X 10-6. No Constraint Zone based on annual individual risk of fatality less than 1.0 X 10-6. None – Individual risk threshold for indicated zone not reached; no setback pertains to specified land use. NA – Not applicable. Specified land use not planned at given location

APPENDIX A

INFORMATION PROVIDED BY PACIFIC GAS AND ELECTRIC COMPANY

APPENDIX B

INFORMATION PROVIDED BY COATES FIELD SERVICE, INC.

APPENDIX C

INFORMATION PROVIDED BY CALIFORNIA STATE FIRE MARSHALL

APPENDIX D

QUANTITATIVE RISK ANALYSIS PERFORMED BY QUEST CONSULTANTS, INC.

QUEST

MOUNTAIN HOUSE PIPELINE RISK ANALYSIS

Prepared ForJ House Environmental220 Hidden Creek Drive

Auburn, California 95603

Prepared ByQuest Consultants Inc.908 26th Avenue, N.W.Post Office Box 721387

Norman, Oklahoma 73070-8069Telephone: 405-329-7475Telecopy: 405-329-7734

December 31, 200303-12-6501

QUEST-i-

MOUNTAIN HOUSE PIPELINE RISK ANALYSIS

Table of Contents

Page1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Hazards Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Failure Scenario Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Failure Frequency Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.4 Hazard Zone Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.5 Risk Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Properties of the Pipelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.1 Property and Pipeline Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Pipeline Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.3 Gas Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4 Meteorological Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 Potential Hazards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1 Release Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.2 Effects of Exposure to Flash Fires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.2.1 Flammable Gas Exposure Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2.2 Dispersion Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2.3 Dispersion Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.3 Effects of Exposure to Thermal Radiation from Fires . . . . . . . . . . . . . . . . . . . . . . 73.3.1 Radiant Flux Exposure Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.3.2 Fire Radiation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.3.3 Radiant Hazard Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.4 Effects of Explosion Overpressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.5 Summary of Consequence Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 Accident Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184.1 Gas Transmission Pipeline Failure Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.2 Liquids Pipeline Failure Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214.3 Hazardous Events Following Flammable Gas Releases . . . . . . . . . . . . . . . . . . . . 21

5 Risk Analysis Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

6 Risk Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256.1 Hazard Footprints and Vulnerability Zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256.2 Individual Risk Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

6.2.1 Risk Transects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266.2.2 Individual Risk Contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296.2.3 Individual Risk Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

6.3 Study Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Appendix A CANARY by Quest® Model Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1

QUEST-ii-

List of Figures

Figure Page 3-1 Release Rate vs. Time for Rupture of 36-inch Natural Gas Pipeline . . . . . . . . . . . . . . . . . 5 3-2 Fire Radiation Probit Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3-3 Overpressure Probit Relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4-1 Example Event Tree for a Flammable Gas Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

5-1 Wind Speed/Atmospheric Stability Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6-1 Maximum Flash Fire Hazard Corridor and Hazard Footprint from Line 401 . . . . . . . . . . . 26 6-2 Risk Transect for Pipeline Corridor Containing Lines 401, 002, and 0499 . . . . . . . . . . . . 27 6-3 Risk Transect for Line 162A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 6-4 Risk Transect for Line 176 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 6-5 Individual Risk Contours in the area of the Mountain House Development . . . . . . . . . . . 30

QUEST-iii-

List of Tables

Table Page 2-1 Pipeline Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2-2 Typical Natural Gas Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3-1 Crude Oil Pool Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3-2 Flammable Gas Dispersion Results for a Rupture of Line 401 (Horizontal Release) . . . . . 8 3-3 Flammable Gas Dispersion Results for a Puncture of Line 401 (Horizontal Release) . . . . 9 3-4 Flammable Gas Dispersion Results for a Leak from Line 401 (Horizontal Release) . . . . . 10 3-5 Hazardous Radiation Levels for Various Exposure Times . . . . . . . . . . . . . . . . . . . . . . . . . 11 3-6 Torch Fire Radiation Results for a Rupture of Line 401 (Horizontal Release) . . . . . . . . . 13 3-7 Torch Fire Radiation Results for a Puncture of Line 401 (Horizontal Release) . . . . . . . . . 14 3-8 Torch Fire Radiation Results for a Leak from Line 401 (Horizontal Release) . . . . . . . . . . 15 3-9 Hazardous Overpressure Levels from Probit Relationship . . . . . . . . . . . . . . . . . . . . . . . . . 17 3-10 Largest Hazard Distances for Releases from Pipelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

6-1 Approximate Distances from Pipeline or Pipeline Corridor to Individual Risk Levels . . . 27

QUEST-1-

MOUNTAIN HOUSE PIPELINE RISK ANALYSIS

1.0 INTRODUCTION

Quest Consultants was retained by J House Environmental to perform a quantitative risk analysis (QRA) forseveral pipelines in the vicinity of the Mountain House Specific Plan III Development near Tracy, California.The methodology for the risk analysis study followed rigorous, internationally accepted guidelines. Thefollowing pipelines were included in the study.

• Line 401 - high pressure natural gas• Line 002 - high pressure natural gas• Line 162A - natural gas• Line 176 - natural gas• Line CSFM 0499 - crude oil

The objective of the study was to compute the level of risk the pipelines would pose to the area within theproposed development. The study was divided into four primary tasks. First, determine potential releasesthat could result in significant hazardous conditions along the pipeline corridors. Second, calculate the con-sequences (hazard zones) associated with each potential release. Third, for each potential release identified,derive a frequency (or probability) of release. Fourth, using consistent, accepted methodology, combinepotential release consequences with the release frequencies to arrive at a measure of the “risk” the systemwould pose to the public.

1.1 Hazards Identification

The potential hazards associated with the pipelines around the proposed Mountain House Development arecommon to similar gas transmission and crude oil pipelines worldwide, and are a function of the materialbeing transported, pipeline physical properties, procedures used for operating and maintaining the equipment,and hazard detection and mitigation systems in place. The hazards that are likely to exist are identified bythe physical and chemical properties of the gas or liquid, and the pipeline conditions. For pipelines handlingflammable gases, the common hazards are:

• torch fires• flash fires• vapor cloud explosions

For pipelines handling crude oil, the common hazard is a pool fire.

1.2 Failure Scenario Definition

Potential pipeline release scenarios are determined from a combination of past history of releases from similarfacilities, project-specific information, and engineering analysis by system safety engineers.

This step in the analysis defines the various release scenarios, and sets the conditions for each release. Therelease conditions include:

QUEST-2-

• fluid composition, temperature, and pressure• pipeline diameter and length• normal flow rate, release rate, and release duration• location and orientation of the release

1.3 Failure Frequency Definition

The frequency with which a given release scenario is predicted to occur can be estimated by using acombination of:

• historical experience• failure rate data on similar types of equipment• service factors• engineering judgment

For single component failures (e.g., pipe rupture), the failure frequency can be determined from industrialfailure rate data bases.

1.4 Hazard Zone Analysis

The release conditions (e.g., pressure, composition, temperature, hole size, inventory, etc.) from the failurescenario definitions are processed, using the best available hazard quantification technology, to produce a setof hazard zones for each scenario. The CANARY by Quest® computer software hazards analysis packageis used to produce profiles for the hazards associated with the failure scenario. The models that are usedaccount for:

• release conditions• ambient weather conditions (wind speed, air temperature, humidity, atmospheric stability)• effects of the local terrain (diking, vegetation)• mixture thermodynamics

1.5 Risk Quantification

The methodology used in this study follows established techniques and has been successfully employed inseveral QRA studies that have undergone regulatory review in countries worldwide.

The result of the analysis is a prediction of the risk posed by the pipelines. Risk may be expressed in severalforms (e.g., individual risk contours, average individual risk, societal risk, etc.). For this analysis, the focuswas on the prediction of risk transects and individual risk contours.

2.0 PROPERTIES OF THE PIPELINES

2.1 Property and Pipeline Description

The area proposed for development is located on the western edge of San Joaquin County, California. Theproperty is bordered by Interstate Highway 205 to the south and Mountain House Parkway to the east. Grant

QUEST-3-

Line Road forms part of the northern border, while the Delta Mendota Canal and other properties form thewestern border.

The pipelines in the area of the proposed Mountain House Development are located in two pipeline corridors.The first corridor crosses through the property of the proposed development from northwest to southeast.This easement contains two high pressure natural gas lines (401 and 002) and a crude oil pipeline (Line 0499).The second corridor begins near the southeast corner of the property, on the southern side of I-205. Fromthis location, Line 162A (which is fed by Line 401) extends northward along the eastern border of theproperty for approximately 3/4 mile, where it turns to the east. From this turning point, Line 176 (which isfed by Line 162A) extends northward past the property.

2.2 Pipeline Physical Properties

The potential hazards that could affect the Mountain House Development are functions of the five pipelinesinvolved in the study. Each pipeline presents a distinct hazard, based on line size (pipe diameter), line length,pressure, temperature, and normal mass flow rate through that section. Table 2-1 lists the operating propertiesof each pipeline. The temperature of the fluid in each pipeline was assumed to be 60°F.

Table 2-1Pipeline Properties

Pipeline Designation Outside Pipeline Diameter(inches) Pipe Wall Thickness Operating Pressure

(psig)

Line 401 36 0.446 1,040

Line 002 26 0.322 890

Line 162A 10.75 0.188 365

Line 176 6.625 0.188 365

Line 0499 - crude 18 0.562* 735**

* Standard schedule 40 pipe assumed. **Operating pressure assumed to be 80% of MAOP.

2.3 Gas Composition

The composition of the gas is identical for all natural gas pipelines in the analysis. A typical natural gascomposition was assumed for this study, and is given in Table 2-2.

Table 2-2Typical Natural Gas Composition

Component Gas Composition(mole %)

Methane 97.5

Ethane 1.2

Propane 0.3

Nitrogen 1

QUEST-4-

2.4 Meteorological Data

Weather data from the area near the Mountain House Development were applied to all pipelines. The data,obtained from the National Climatic Data Center in Asheville, North Carolina, reflect the relative probabilityof wind speed, direction, and Pasquill-Gifford atmospheric stability for the area.

The following atmospheric conditions were applied to all accident scenarios. They represent the average con-ditions for the area.

Air temperature = 70°FRelative humidity = 70%

3.0 POTENTIAL HAZARDS

Quest reviewed the pipeline specifications in order to determine credible hazardous events that have thepotential to occur. As a result of this review, the following potential releases of flammable gas or crude oilwere selected for evaluation. None of the pipelines transport any acutely toxic components.

(1) Rupture: Full rupture of the pipeline, resulting in rapid depressurization of the line. This isconsidered the maximum credible release that might occur along the pipeline.

(2) Puncture: A 2-inch hole in one of the pipelines, as a result of material defect or puncture.(3) Leak: A 1/4-inch hole in one of the pipelines, to simulate a corrosion hole in the pipeline.

For each of the above releases, several release orientations were considered.

• Vertically upward (represents 50% of all release orientations)• 45° (represents 25% of all release orientations)• Horizontal (represents 25% of all release orientations)

The release scenarios described above define the range of hazardous credible releases that might occur. Eachof these releases may create one or more of the following hazards.

• Exposure to heat radiation from a torch fire.• Exposure to a flash fire (release of pipeline fluid that forms a flammable vapor cloud which is subse-

quently ignited).• Exposure to explosion overpressure following the release and ignition of pipeline gas into a confined

or congested area.• Exposure to heat radiation from a pool fire.

3.1 Release Characteristics

The calculation of pipeline release rates in this study was accomplished with the release model contained inthe CANARY by Quest® modeling package. This model predicts the time-varying flow of vapor, aerosol,and liquids (as appropriate) following a breach of the pipe. While calculating the release rates and pressuredrop along the pipeline, the model accounts for multiphase thermodynamic behavior, including two-phaseflow due to flashing in the pipe and varying vapor/aerosol production.

QUEST-5-

Figure 3-1Release Rate vs. Time for Rupture of 36-inch Natural Gas Pipeline

Once a pipeline rupture occurs, the rate of mass leaving the pipeline drops dramatically after the ruptureoccurs as the pipeline deinventories. An example of this behavior is presented in Figure 3-1, for a ruptureof line 401, a 36-inch natural gas pipeline. Within ten seconds after the rupture occurs, the escaping massrate has dropped to half the initial rate; after three minutes the release rate stabilizes and drops much moreslowly. This behavior is due to the large initial pressure drop that occurs as gas leaves the pipe.

Since it is the high mass rate within the first few moments of a pipeline rupture that establishes the maximumextent of the flammable vapor cloud or fire, the gas trailing out of the pipe at later times will only pose ahazard closer to the rupture point. For punctures (2-inch hole) of a pipeline, the release rate begins muchlower than that for a rupture, and the pipeline depressurization behavior lasts considerably longer. Thus, thelargest hazards from a puncture are characterized by the first several minutes of the release, rather than thefirst few seconds. Pipeline leaks (1/4-inch hole) are characterized by a constant release rate (i.e., the volumein the pipeline is large enough that there is no significant depressurization until long after the release begins.)

For releases of crude oil, the behavior is similar to the gas pipelines, but with a faster decay in the release rate(depressurization of a liquid vs. a vapor). The hazards associated with crude oil are directly proportional tothe size of the liquid pool following a release. This analysis assumed that there were no local terrain effectsthat would alter or limit the formation of a liquid pool (i.e., the pool spreads as a circle at the point of release).Because the terrain features at any one point along the pipeline are unknown, this simplifying assumption isnecessary for modeling the pool size, and is typically a conservative approach. Table 3-1 presents some ofthe liquid pool information from this study. Please note that the release rates presented in Table 3-1 are theexpected rates of liquid falling to the ground; some portions of the released material may be released as anaerosol mist due to the pipeline pressure.

QUEST-6-

Table 3-1Crude Oil Pool Properties

Release TypeApproximate LiquidRelease Rate (gpm)

Approximate Volumein Accumulated Pool

(gallons)Maximum PredictedPool Diameter (feet)

Rupture 5,960 92,350 73

Puncture 925 33,400 47

Leak 6 435 9

3.2 Effects of Exposure to Flash Fires

The physiological effects of fires on humans depend on the rate at which heat is transferred from the fire tothe person, and the time the person is exposed to the fire. Even short-term exposure to high heat flux levelsmay be fatal. This situation could occur to persons wearing ordinary clothes who are inside the flammablevapor cloud, defined by the lower flammable limit (LFL), when it is ignited. Persons located outside theflammable cloud when it is ignited will be exposed to much lower heat flux levels and may be able to takeaction to protect themselves.

3.2.1 Flammable Gas Exposure Limits

The lower flammable limit (LFL) of the gas is defined by the gas composition. In the gas, there are flam-mable materials (methane, ethane, etc.) and inert materials (nitrogen). The lower flammable limit for the gasin this study is 4.99 mole percent. When performing flammable hazard zone calculations, the extent of theflammable zone is often defined by a gas concentration limit equal to the lower flammable limit (LFL). Itis consistently assumed that all persons within the LFL zone are killed if the flammable gas ignites, whilepersons outside this zone are unaffected.

3.2.2 Dispersion Analysis

When performing a site-specific risk analysis, the ability to accurately model the release, dilution, and dis-persion of gas is important if an accurate assessment of potential risk to the public is to be attained. For thisreason, Quest uses a modeling package, CANARY by Quest, that contains a set of complex models that cal-culate release conditions, initial dilution of the vapor (dependent upon the release characteristics), and thesubsequent dispersion of the vapor introduced into the atmosphere. The models contain algorithms thataccount for thermodynamics, mixture behavior, transient release rates, gas cloud density relative to air, initialvelocity of the released gas, and heat transfer effects from the surrounding atmosphere and the substrate. Therelease and dispersion models contained in the QuestFOCUS package (the predecessor to CANARY byQuest) were reviewed in a United States Environmental Protection Agency (EPA) sponsored study [TRC,1991] and an American Petroleum Institute (API) study [Hanna, Strimaitis, and Chang, 1991]. In bothstudies, the QuestFOCUS software was evaluated on technical merit (appropriateness of models for specificapplications) and on model predictions for specific releases. One conclusion drawn by both studies was thatthe dispersion software tended to overpredict the extent of the gas cloud travel, thus resulting in too large acloud when compared to the test data (i.e., a conservative approach).

QUEST-7-

A study prepared for the Minerals Management Service [Chang, et al.,1998] reviewed models for use inmodeling routine and accidental releases of flammable and toxic gases. CANARY by Quest received thehighest possible ranking in the science and credibility areas. In addition, the report recommends CANARYby Quest for use when evaluating toxic and flammable gas releases. Specific models (e.g., SLAB) containedin the CANARY by Quest software package have also been extensively reviewed. Technical descriptionsof the CANARY models used in this study are presented in Appendix A.

3.2.3 Dispersion Analysis Results

To demonstrate the range of calculations performed in the hazards analysis, consider the following examplecase. The chosen scenario is a release from Line 401, a 36-inch natural gas pipeline. Pipeline conditions are60ºF and 1,040 psig. Table 3-2 presents the downwind distances, at grade, corresponding to the lowerflammable limit following a full rupture of the pipeline. Initial conditions are a horizontal release at grade.Tables 3-3 and 3-4 present the downwind dispersion distances for the flammable (LFL) hazard, for thepuncture (2-inch) and leak (1/4-inch) releases, respectively, associated with Line 401. These tables displayall possible wind speed and atmospheric stability combinations that exist in the area of the development, andthe hazard distance that corresponds to each one.

Because of the buoyant nature of methane, many of the releases modeled in this study create very small orshortened ground-level flammable hazards. Releases at 45° and 90° (vertical) have much shorter flammablehazard distances than those given in Tables 3-2 through 3-4.

3.3 Effects of Exposure to Thermal Radiation from Fires

In the event of a continuous fire (torch or pool fire) following the release of flammable material, the radiationlevels necessary to cause injury to the public must be defined as a function of exposure time. People who areexposed to radiant hazards are aware of the hazard and know in which direction to move in a very shortperiod of time. Work sponsored by the U.S. Coast Guard [Tsao and Perry, 1979] developed a probit relation-ship between exposure time and incident heat flux. The probit takes the form:

=Pr ( )4 / 312.8 2.56 ln t I− + i i

where: = exposure time, sect= effective radiation intensity, kW/m2I

Table 3-5 presents the probit results for several exposure times applicable to pipeline releases that are ignited.The mortality rates and corresponding radiation levels are listed. The form of the radiation probit equationfor different exposure times is presented in Figure 3-2.

3.3.1 Radiant Flux Exposure Limits

The exposure time for radiant hazards is short due to people’s natural instinct to flee from an obvious hazard.For this work, an exposure time of 30 seconds was used in the probit relationship, although it is reasonableto assume that most people will find protection in much less time than 30 seconds. The effect of using thisexposure time is to increase the size of the hazard zones since longer exposure times result in lower tolerableflux levels. This conservative assumption leads to an overprediction of the size of the radiant hazard zones.

QUEST-8-

Table 3-2Flammable Gas Dispersion Results

for a Rupture of Line 401 (Horizontal Release)

Maximum downwind distances at grade (flash fire)

C (mole fraction) 4.99 mole% (LFL)

Downwind Distance in Feet to Hazard Level

11.32 m/s wind speed 581 869

10.36 m/s wind speed 604 896

7.20 m/s wind speed 705 994

4.63 m/s wind speed 630 840 1,099 1,198

2.83 m/s wind speed 584 787 991 1,175 1,188 1,125

1.03 m/s wind speed 932 1,106 1,188 1,165 1,079

Astability

Bstability

Cstability

Dstability

Estability

Fstability

Note: wind speed/stability combinations which normally occur are enclosed by the heavy line.

QUEST-9-

Table 3-3Flammable Gas Dispersion Results

for a Puncture of Line 401 (Horizontal Release)

Maximum downwind distances at grade (flash fire)

C (mole fraction) 4.99 mole% (LFL)

Downwind Distance in Feet to Hazard Level

11.32 m/s wind speed 42 55

10.36 m/s wind speed 43 55

7.20 m/s wind speed 49 57

4.63 m/s wind speed 46 55 57 55

2.83 m/s wind speed 45 54 57 56 53 51

1.03 m/s wind speed 57 57 55 53 51

Astability

Bstability

Cstability

Dstability

Estability

Fstability

Note: wind speed/stability combinations which normally occur are enclosed by the heavy line.

QUEST-10-

Table 3-4Flammable Gas Dispersion Results

for a Leak from Line 401 (Horizontal Release)

Maximum downwind distances at grade (flash fire)

C (mole fraction) 4.99 mole% (LFL)

Downwind Distance in Feet to Hazard Level

11.32 m/s wind speed 5 6

10.36 m/s wind speed 5 6

7.20 m/s wind speed 6 6

4.63 m/s wind speed 6 6 6 6

2.83 m/s wind speed 5 6 7 6 6 5

1.03 m/s wind speed 7 6 6 5 5

Astability

Bstability

Cstability

Dstability

Estability

Fstability

Note: wind speed/stability combinations which normally occur are enclosed by the heavy line.

QUEST-11-

Figure 3-2Fire Radiation Probit Relations

Table 3-5Hazardous Radiation Levels for Various Exposure Times

Exposure Time(sec) Probit Value Mortality Rate*

(percent)Incident Radiation Flux

(kW/m2) (Btu/(hr ft2)i

52.675.007.33

15099

27.87 55.17109.20

8,83317,48534,610

152.675.007.33

15099

12.22 24.20 47.39

3,873 7,67015,178

302.675.007.33

15099

7.27 14.39 28.47

2,304 4,560 9,023

602.675.007.33

15099

4.32 8.55 16.93

1,369 2,709 5,365

*Percent of population fatally affected.

QUEST-12-

3.3.2 Fire Radiation Analysis

CANARY by Quest also contains models for torch fire and pool fire radiation. These models account forimpoundment configuration (pool), gas flow rate (torch), material composition, target height relative to theflame, target distance from the flame, atmospheric attenuation (includes humidity), wind speed, and atmo-spheric temperature. They are based on information in the public domain (published literature) and have beenvalidated with experimental data. These models are described in Appendix A.

This analysis considered two types of torch fires: immediate and delayed. Immediate torch fires occur whenthe exiting gas is ignited upon release. The fire in this situation will be the largest possible size, for a givenset of initial conditions (hole size, temperature, pressure, etc). As the release progresses, the flow rate dropsas the system depressurizes. If ignition occurs later in the release, the resulting (delayed) torch fire will besmaller in size, and thus have lesser consequences.

Two types of pool fires were also considered: immediate and delayed. In this case, the immediate pool fire(immediate ignition) represents a smaller hazard because the pool has not grown to its full size. Delayedignition pool fires assume that the pool has grown to its maximum size before being ignited.

3.3.3 Radiant Hazard Results

Table 3-6 contains the maximum downwind distances to the three radiation flux levels of interest (7.27 kW/m2, 14.39 kW/m2, and 28.47 kW/m2 for 30-second exposures), for six wind speeds evaluated, for a horizontalrelease following full rupture of Line 401. The fire radiation calculations are not sensitive to atmosphericstability conditions, but are influenced by wind speed. For near-horizontal releases, higher winds have littleeffect on the flames. (For torch fires that are oriented at an angle or vertically, higher winds bend the flamedownwind toward the target, resulting in larger hazard distances.) The puncture (2-inch) results for horizontalreleases are presented in Table 3-7; the leak (1/4-inch) results for horizontal releases are given in Table 3-8.Torch fire hazard distances following 45° and vertical releases have shorter hazard distances than a com-parable horizontal release.

3.4 Effects of Explosion Overpressure

The physiological effects of explosion overpressures depend on the peak overpressure that reaches the person.Exposure to high overpressure levels may be fatal. Persons located outside the flammable cloud when itexplodes will be exposed to lower overpressure levels than persons within the flammable cloud. If the personis far enough from the edge of the exploding cloud, the overpressure is incapable of causing fatal injuries.

The hazardous area associated with the overpressure created by an explosion of a vapor cloud is dependenton the mass of flammable gas within the flammable range (i.e., between the lower and upper flammablelimits, LFL and UFL respectively) and its heat of combustion. The mass of flammable gas between the LFLand UFL is determined in the vapor dispersion analysis.

In the event of an ignition and explosion of flammable gas, the overpressure levels necessary to cause injuryto a person are often defined as a function of peak overpressure. Unlike potential fire hazards, persons whoare exposed to explosion overpressures have no time to react or take shelter; thus, time does not enter intothe hazard relationship. The Health and Safety Executive [HSE, 1991] has produced a probit relationshipbased on peak overpressure. The probit takes the form:

QUEST-13-

Table 3-6Torch Fire Radiation Results

for a Rupture of Line 401 (Horizontal Release)

Maximum downwind distances at grade (torch fire)

R lowR middleR high

7.27 kW/m2

14.39 kW/m2

28.47 kW/m2

Downwind Distance in Feet to Hazard Level

Immediate IgnitionUpon Release

Delayed IgnitionFollowing Release

11.32 m/s wind speed1,4961,3331,318

911858777

10.36 m/s wind speed1,4961,3331,318

911858777

7.20 m/s wind speed1,4961,3331,318

912858776

4.63 m/s wind speed1,4961,3331,318

912857775

2.83 m/s wind speed1,4961,3331,318

912857773

1.03 m/s wind speed1,4961,3331,318

912858766

All stabilities All stabilities

QUEST-14-

Table 3-7Torch Fire Radiation Results

for a Puncture of Line 401 (Horizontal Release)

Maximum downwind distances at grade (torch fire)

R lowR middleR high

7.27 kW/m2

14.39 kW/m2

28.47 kW/m2

Downwind Distance in Feet to Hazard Level

Immediate IgnitionUpon Release

Delayed IgnitionFollowing Release

11.32 m/s wind speed170150148

170149148

10.36 m/s wind speed170150148

170149148

7.20 m/s wind speed170150148

170149148

4.63 m/s wind speed170150148

170149148

2.83 m/s wind speed170150148

170149148

1.03 m/s wind speed170150148

170149148

All stabilities All stabilities

QUEST-15-

Table 3-8Torch Fire Radiation Results

for a Leak from Line 401 (Horizontal Release)

Maximum downwind distances at grade (torch fire)

R lowR middleR high

7.27 kW/m2

14.39 kW/m2

28.47 kW/m2

Downwind Distance in Feet to Hazard Level

Immediate IgnitionUpon Release

Delayed IgnitionFollowing Release

11.32 m/s wind speed323231

323231

10.36 m/s wind speed323231

323231

7.20 m/s wind speed323231

323231

4.63 m/s wind speed323231

323231

2.83 m/s wind speed323231

323231

1.03 m/s wind speed323231

323231

All stabilities All stabilities

QUEST-16-

=Pr ( )1.47 1.37 ln p+ i

where: = peak overpressure, psigp

Table 3-9 presents the probit results for 1, 50, and 99% fatalities. The graphical form of the explosion probitequation is presented in Figure 3-3. This probit relation defines the explosion overpressures used in theanalysis. Unlike dispersion and radiation, there is no exposure duration required for the probit relationship.Once the explosion is assumed to occur, the probit endpoints define the distances at which persons can befatally affected.

The damaging effects of a vapor cloud explosion (VCE) depend on the peak overpressure, or blast wave, thatreaches a person or a given structure (and the type of construction of that structure). The peak overpressurecan be predicted by a vapor cloud explosion model that takes into account the properties and quantity of theflammable material. VCE calculations in the current analysis were made with the Baker-Strehlow model con-tained in the CANARY by Quest hazard modeling package (see Appendix A). The Baker-Strehlow modelis based on the premise that the strength of the blast wave generated by a VCE is dependent on the reactivityof the flammable gas involved; the presence (or absence) of structures, such as walls or ceilings, that partiallyconfine the vapor cloud; and the spatial density of obstructions within the flammable cloud [Baker, et al.,1994, 1998]. This model reflects the results of several international research programs on vapor cloudexplosions, which show that the strength of the blast wave generated by a VCE increases as the degree of con-finement and/or obstruction of the cloud increases. The following quotations illustrate this point.

“Both in two- and three-dimensional geometries, a continuous accelerating flame wasobserved in the presence of repeated obstacles. A positive feedback mechanism between theflame front and a disturbed flow field generated by the flame is responsible for this. Thedisturbances in the flow field mainly concern flow velocity gradients. Without repeatedobstacles, the flame front velocities reached are low both in two-dimensional and three-dimensional geometry.” [van Wingerden and Zeeuwen, 1983] (Tests conducted by TNOin the Netherlands.)

“The current understanding of vapor cloud explosions involving natural gas is that combus-tion only of that part of the cloud which engulfs a severely congested region, formed byrepeated obstacles, will contribute to the generation of pressure.” [Johnson, Sutton, andWickens, 1991] (Tests conducted by British Gas in the United Kingdom.)

Researchers who have studied case histories of accidental vapor cloud explosions have reached similar con-clusions.

“It is a necessary condition that obstacles or other forms of semi-confinement are presentwithin the explosive region at the moment of ignition in order to generate an explosion.”[Wiekema, 1984]

“A common feature of vapor cloud explosions is that they have all involved ignition of vaporclouds, at least part of which have engulfed regions of repeated obstacles.” [Harris andWickens, 1989]

QUEST-17-

Figure 3-3Overpressure Probit Relation

Table 3-9Hazardous Overpressure Levels from Probit Relationship

Probit Value Mortality Rate(percent)

Peak Overpressure

(psig) (kPag)

2.67 1 2.4 16.55

5.00 50 13.1 90.73

7.63 99 72.0 496.83

In an area without confinement or obstacles, normal hydrocarbon deflagrations do not produce substantialoverpressures. This arises from the relationship between flame speed and initial overpressure, and theobservation that confinement or obstacles cause acceleration of the flame front. Unconfined flammableclouds have a low flame speed and, without any physical structures to cause flame acceleration, significantoverpressures are not generated. Experimental analysis with light hydrocarbon vapor clouds in open areas[Hirst and Eyre, 1982; van Wingerden and Zeeuwen, 1983; Johnson, Sutton, and Wickens, 1991] and histori-cal data (as well as engineering analysis) show this to be true. For example, Hirst and Eyre [1982] make thefollowing comments about the overpressures generated following ignition of vapor clouds in an open area.

“On the evidence of the trials performed at Maplin Sands, the deflagration [explosion] oftruly unconfined flat clouds of natural gas or propane does not constitute a blast hazard.”

“The maximum overpressure measured in any of the tests was approximately 1 mbar [1 mil-libar, or 0.0145 psig]. Pressures of this magnitude, which are consistent with the low flamespeeds and flat clouds, are far too low to cause damage either to a ship or its crew. Indeed,

QUEST-18-

they are lower than the overpressures experienced in everyday life; for example, the over-pressure created inside a car when the door is slammed is typically a few millibars.”

For the releases in this study, any potential vapor cloud explosion has little potential to create significantoverpressures. This is due to three main reasons:

• Much of the area surrounding the pipelines is flat, open terrain, and each pipeline corridor will havea cleared right-of-way. There is essentially no place that flammable vapors can collect with anydegree of confinement or congestion.

• Methane, the primary ingredient in natural gas, is classified as a low reactivity fuel.• Releases of natural gas are lighter than air. For vertical and 45° releases, there is only a small area

of flammability near grade level.

If these conditions are input into the Baker-Strehlow model, the peak overpressure achieved by an open-areanatural gas vapor cloud is 0.16 psig. This is considerably below the threshold for injury to humans (andbelow the threshold of structural damage to any building). Based on these observations, vapor cloud explo-sions were not considered further in this analysis.

3.5 Summary of Consequence Analysis Results

Table 3-10 presents a summary of the releases evaluated for this project. The maximum ground level distanceto the specified hazard endpoints for each of the three hole sizes evaluated is listed. As seen in the table, thehazard that consistently dominates this analysis is the thermal radiation from torch fires.

4.0 ACCIDENT FREQUENCY

The likelihood of a particular accident occurring within some specific time period can be expressed in dif-ferent ways. One way is to state the statistical probability that the accident will occur during a one-yearperiod. This annual probability of occurrence can be derived from failure frequency data bases of similaraccidents that have occurred with similar systems or components in the past.

Most data bases (e.g., CCPS [1989], OREDA [1984]) that are used in this type of analysis contain failurefrequency data (e.g., on the average, there has been one failure of this type of equipment for 347,000 hoursof service). By using the following equation, the annual probability of occurrence of an event can be calcu-lated if the frequency of occurrence of the event is known.

= p ( )1 te λ−−

where: = annual probability of occurrence (dimensionless)p= annual failure frequency (failures per year)λ

= time period (one year)t

If an event has occurred once in 347,000 hours of use, its annual failure frequency is computed as follows.

= = λ 1 8,760347,000

event hourshours year

i 0.0252 /events year

QUEST-19-

Tab

le 3

-10

Lar

gest

Haz

ard

Dis

tanc

es fo

r R

elea

ses f

rom

Pip

elin

es

Rel

ease

Fro

mR

elea

se H

ole

Size

(inch

es)

Max

imum

Dow

nwin

d D

ista

nce

(fee

t) fr

om R

elea

se to

Fat

ality

Lev

el

Flas

h Fi

re(L

FL)

Tor

ch F

ire

or P

ool F

ire

The

rmal

Rad

iatio

n

Imm

edia

te Ig

nitio

nD

elay

ed Ig

nitio

n

100%

1%50

%99

%1%

50%

99%

Line

401

361,

198

1,49

61,

333

1,31

891

285

877

7

2

57

17

0

150

14

817

014

914

8

¼

7

3

2

32

3

1 3

2 3

2 3

1

Line

002

26

835

90

6

895

88

458

853

048

0

2

52

15

9

142

14

015

914

214

0

¼

6

3

1

30

3

0 3

1 3

0 3

0

Line

162

A

10

422

48

7

430

40

948

743

040

9

2

31

10

7

96

9

010

7 9

6 9

0

¼

3

2

5

25

2

4 2

5 2

5 2

4

Line

176

6

97

30

4

277

27

430

427

727

4

2

31

10

7

96

9

010

7 9

6 9

0

¼

3

2

5

25

2

4 2

5 2

5 2

4

Line

049

9 (c

rude

)

18

‡

129

9

4

84

217

163

155

2‡

9

4

69

5

715

711

310

6

¼‡

3

3

29

2

5 5

0 4

2 3

2

†

Flam

mab

le g

as c

once

ntra

tion

does

not

exi

st b

eyon

d th

e liq

uid

pool

.

QUEST-20-

The annual probability of occurrence of the event is then calculated as follows.

= = p ( )0.0252 11 e −− i0.0249

Note that the frequency of occurrence and the probability of occurrence are nearly identical. (This is alwaystrue when the frequency is low.) An annual probability of occurrence of 0.0249 is approximately the sameas saying there will probably be one event per forty years of use.

Due to the scarcity of accident frequency data bases, it is not always possible to derive an exact probabilityof occurrence for a particular accident. Also, variations from one system to another (e.g., differences indesign, operation, maintenance, or mitigation measures) can alter the probability of occurrence for a specificsystem. Therefore, variations in accident probabilities are usually not significant unless the variationapproaches one order of magnitude (i.e., the two values differ by a factor of ten).

The following subsections describe the basis and origin of failure frequency rates used in this analysis.

4.1 Gas Transmission Pipeline Failure Rates

Data compiled from DOT statistics on failures of gas pipelines show a failure rate of 1.21 x 10-3

failures/mile/year for steel pipelines in the United States [Jones, et al., 1986]. In addition to failures of buriedpipe, these data include failures of buried pipeline components, such as block valves and check valves, whenthe failure resulted in a release of fluid from the pipeline.

Data gathered by operators of gas transmission pipelines in Europe indicate a failure rate of 1.13 x 10-3

failures/mile/year [EGPIDG, 1988].

These data sets are not sufficiently detailed to allow a determination of the failure frequency as a function ofthe size of the release (i.e., the size of hole in the pipeline). However, British Gas has gathered such data ontheir gas pipelines [Fearnehough, 1985]. These data indicate that well over 90% of all failures are less thana one-inch diameter hole, and only 3% are greater than a three-inch diameter hole.

Data compiled from DOT data on gas pipelines in the United States show a trend toward higher failure ratesas pipe diameter decreases [Jones, et al., 1986]. (Smaller diameter pipes have thinner walls; thus, they aremore prone to failure by corrosion and by mechanical damage from outside forces.)

Based on the data sets described above, the expected failure rates for steel gas transmission pipelines areassumed to be as follows.

For pipelines from six to twelve inches in diameter:

Hole size <1/4 inch 1/4 to 2 inch 2 inch to full ruptureExpected failure rate 1.140 x 10-3/mile/year 0.289 x 10-3/mile/year 0.091 x 10-3/mile/year

For pipelines from fourteen to twenty-two inches in diameter:

Hole size <1/4 inch 1/4 to 2 inch 2 inch to full ruptureExpected failure rate 0.975 x 10-3/mile/year 0.247 x 10-3/mile/year 0.078 x 10-3/mile/year

QUEST-21-

For pipelines from twenty-four to twenty-eight inches in diameter:

Hole size <1/4 inch 1/4 to 2 inch 2 inch to full ruptureExpected failure rate 0.420 x 10-3/mile/year 0.106 x 10-3/mile/year 0.034 x 10-3/mile/year

For pipelines from thirty to thirty-six inches in diameter:

Hole size <1/4 inch 1/4 to 2 inch 2 inch to full ruptureExpected failure rate 0.150 x 10-3/mile/year 0.038 x 10-3/mile/year 0.012 x 10-3/mile/year

4.2 Liquids Pipeline Failure Rates

Data from the Department of Transportation’s (DOT) Office of Pipeline Safety (OPS) include accident ratesfor hazardous liquids pipelines. Based on a data analysis similar to that presented for the gas transmissionlines, the failure rates for the 18-inch crude oil line were calculated.

Hole size <1/4 inch 1/4 to 2 inch 2 inch to full ruptureExpected failure rate 0.900 x 10-3/mile/year 0.228 x 10-3/mile/year 0.072 x 10-3/mile/year

4.3 Hazardous Events Following Flammable Gas Releases

A release of flammable gas to the atmosphere may create one or more hazardous conditions. Depending onevents that occur subsequent to the release, the possibilities are:

(a) No ignition. If a flammable vapor cloud forms but never ignites, there is no hazard.(b) Immediate ignition. If ignition occurs nearly simultaneously with the beginning of the release, the

hazard is heat radiation from a torch fire or a pool fire.(c) Delayed ignition. If there is a time delay between the start of the release and ignition of the release,

a flammable vapor cloud may form. Upon ignition, there may be a vapor cloud fire (flash fire), andpossible vapor cloud explosion, followed by a torch fire or pool fire.

Each of these three possibilities has some probability of occurring, once a release has occurred. The sum ofthese three probabilities must equal one. The ignition/explosion probabilities employed in this study are takenfrom an Institution of Chemical Engineers report [Cox, Lees, and Ang,, 1990]. Estimated values are a func-tion of the “size” of the release.

Consequences of the hazardous events that may occur subsequent to a release of flammable gas are also pro-portional to the “size” of the release. Therefore, when calculating the accident probability, it is necessary toestimate the distribution of releases of various sizes. This is typically done by applying a hole size distri-bution, such as the one presented in Section 4.1 for piping.

The estimates used for hole size and ignition probability are best illustrated by event trees, with a release ofgas as the initial event. One of the event trees prepared for this study is presented in Figure 4-1. It beginswith a release from Line 401, a 36-inch natural gas pipeline. Moving from left to right, the tree first branchesinto three release sizes, each being defined by the diameter (d) of the hole through which the flammable gasis being released. Each of these three branches divides into three branches based on ignition timing andprobability. At the far right of the event tree are the nine “outcomes” that have some probability of occurring

QUEST-22-

Prob

abili

ty(A

ssum

ing

Initi

atin

g Ev

ent

Occ

urs)

Ann

ual

Prob

abili

type

r Foo

tof

Pip

eO

utco

me

Imm

edia

te Ig

nitio

n (1

0%)

0.00

62.

27 x

10-1

0To

rch

Fire

Rup

ture

(6%

)D

elay

ed Ig

nitio

n (2

0%)

0.01

24.

55 x

10-1

0Fl

ash

Fire

/VC

E/To

rch

Fire

1 in

< d

# 3

6 in

No

Igni

tion

(70%

)0.

042

1.59

x 1

0-9

Dis

sipa

tion

Imm

edia

te Ig

nitio

n (2

%)

0.00

41.

44 x

10-1

0To

rch

Fire

Rel

ease

of G

as fr

omLi

ne 4

01M

ajor

Lea

k (1

9%)

Del

ayed

Igni

tion

(5%

)0.

009

3.60

x 1

0-10

Flas

h Fi

re/V

CE/

Torc

h Fi

re

Pipe

Dia

met

er =

36

in

(Ann

ual P

roba

bilit

y3.

788

x 10

-8 ft

/yr)

0.25

in <

d #

1 in

No

Igni

tion

(93%

)0.

177

6.69

x 1

0-9

Dis

sipa

tion

Imm

edia

te Ig

nitio

n (2

%)

0.01

55.

68 x

10-1

0To

rch

Fire

Min

or L

eak

(75%

)D

elay

ed Ig

nitio

n (5

%)

0.03

81.

42 x

10-9

Fl

ash

Fire

/VC

E/To

rch

Fire

d #

0.2

5 in

No

Igni

tion

(93%

)0.

698

2.64

x 1

0-8

Dis

sipa

tion

Figure 4-1Example Event Tree for a Flammable Gas Release

QUEST-23-

if the initiating release occurs. To arrive at the probability of a specific outcome, the overall failure rate ismodified by the probability at each applicable branching of the event tree. The estimated annual probabilityof occurrence of each possible outcome, per foot of pipe, is listed on the event tree.

In general, small releases are the most likely to occur, and the least likely to be ignited (small probability ofreaching an ignition source. The largest releases are the least likely to occur, the most likely to be ignited(highest probability of reaching an ignition source), and the most likely to be ignited immediately (the forceneeded to cause a large release may also be capable of igniting the release).

Similar event trees were constructed for releases of gas over the range of pipe sizes, and for releases from thecrude oil line. The outcome probabilities from the event trees are combined with consequence outcomes inthe risk mapping analysis described in Section 5.

5.0 RISK ANALYSIS METHODOLOGY

The pipelines evaluated in this analysis pose no health hazards to the public as long as they do not releaseflammable gas or liquids into the environment. In the event of an accident that results in a release, personsnear the release point may be at risk due to a fire created by the release.

The risk associated with the use of hazardous materials is often expressed as the product of the probabilityof occurrence of a hazardous event and the consequences of that event. Therefore, in order to quantify therisk associated with the transmission of gas and oil, it is necessary to quantify both the probability of releasingmaterial into the environment, and the consequences of such releases.

The analysis follows four major steps.

1) Failure scenario definition (described in Sections 1 and 4)2) Hazard zone analysis (described in Section 3)3) Failure frequency definition (described in Section 4)4) Risk quantification

The risk quantification methodology used in this study has been successfully employed in several QRAstudies that have undergone regulatory review in several countries worldwide. The following is a briefdescription of the steps involved in quantifying the risk from the pipelines evaluated in this work.

Conceptually, performing a risk analysis for the pipeline is straightforward. For example, the analysis canbe divided into the following steps.

Step 1. For each pipeline of the project, determine the potential credible releases that would create a flam-mable gas cloud or fire.

Step 2. Determine the frequency of occurrence of each of these releases.

Step 3. Calculate the size of each potentially fatal hazard zone created by each of the releases identified inStep 1.

i. The hazards of interest are:a. Fire radiation from flash fires and torch fires.b. Overpressures from vapor cloud explosions.

QUEST-24-

ii. The size of each hazard zone is a function of one or more of the following factors.a. Orientation of the release.b. Wind speed.c. Atmospheric stability.d. Local terrain (including diking and drainage).e. Composition, pressure, and temperature of gas being released.f. Hole size.g. Pipeline inventories.

Step 4. Determine the risk to the persons in the vicinity of the pipeline.i. For a specific wind direction, the potential exposure of any individual to a specific

hazard zone depends on the following factors.a. Size (area) of the hazard zone.b. Location of the individual relative to the release location.c. Wind direction.

ii. Determine the exposure of the persons to each potential hazard zone.a. Perform vapor dispersion (flash fire) hazard zone calculations for all wind

directions, wind speeds, atmospheric stabilities, and release orientations.b. Perform torch fire and pool fire hazard zone calculations for all wind

speeds, wind directions, and release orientations.iii. Modify each of the above exposures by its probability of occurrence. Probabilities

are divided into the following groups.a. P(wd,ws,stab) = probability that the wind blows from a specified direction

(wd), with a certain wind speed (ws), and a given atmospheric stabilityclass, A through F (stab). Meteorological data are generally divided intosixteen wind directions, six wind speed classes, and six Pasquill-Giffordatmospheric stability categories. Although all 576 combinations of theseconditions do not exist, a significant number will exist for each releasestudied. Figure 5-1 represents the wind speed versus stability distributionfor the meteorological data.

b. P(acc) = probability of occurrence of each release identified in Step 1 (seeSection 4).

c. P(ii) = probability of immediate ignition (i.e., probability that ignitionoccurs nearly simultaneously with the release) (see Section 4).

d. P(di) = probability of delayed ignition (i.e., probability that ignition occursafter a vapor cloud has formed) (see Section 4).

e. P(orientation) = probability that hazardous fluid is released into the atmo-sphere in a particular orientation.

iv. Sum the potential exposures from each of the hazards for all releases identified inStep 1. This summation requires modifying each potential hazard zone by itsprobability of occurrence (i.e., the probability of a specific delayed torch fire isP(acc) P(orientation) P(ws,wd,stab) P(di)).i i i

The result of the analysis is a prediction of the risk posed by the pipeline. Risk may be expressed in severalforms (e.g., individual risk contours, average individual risk, societal risk, etc.). For this analysis, the focuswas on the prediction of risk transects and individual risk contours.

QUEST-25-

Figure 5-1Wind Speed/Atmospheric Stability Categories

6.0 RISK ANALYSIS RESULTS

This section presents the results of the risk analysis. The consequences were calculated using the informationpresented in Sections 2 and 3, probabilities computed with information from Section 4, and the risk contoursgenerated using the methods presented in Section 5.

6.1 Hazard Footprints and Vulnerability Zones

For each release from the pipeline system, one particular combination of conditions will create the largestpotentially lethal hazard zone for any one type of hazard. For a release from Line 401 (as discussed in Sec-tion 3), the largest flash fire hazard zone is associated with a rupture of the pipe, with a horizontal orientation.The maximum extent of a flash fire hazard extends 1,198 feet from the point of release. The hazard zoneassociated with this unique event is illustrated in two ways in Figure 6-1. One method presents it as a corridorvulnerability zone (that parallels the pipeline), with an offset distance of 1,198 feet (the lightly hatched areain Figure 6-1). This presentation is misleading since all locations within this corridor cannot besimultaneously exposed to the potentially lethal hazard from any single accident. A more realisticpresentation of the maximum potential hazard zone associated with a release from Line 401 is the heaviercross-hatched area in Figure 6-1. This is the hazard footprint that would be expected if the pipe were torupture, and the wind is blowing perpendicular to the pipeline (from the north toward an east-west pipelinein this example), and the wind speed is moderate (4.63 m/s), and the atmosphere is relatively stable (class E),and the release is oriented horizontally, and the gas does not ignite until it disperses to its maximum extent.The probability of the simultaneous occurrence of these conditions is 2.046 x 10-8 per year, per mile of pipe.Thus, the probability of generating the hazard zone (footprint) seen in Figure 6-1 becomes one chance inabout 49,000,000 per year, per mile, that a flash fire will present a hazard exactly as shown.

QUEST-26-

Figure 6-1Maximum Flash Fire Hazard Corridor and Hazard Footprint from Line 401

When the hazard vulnerability zone (the corridor) on Figure 6-1 is presented, there is no associated probabil-ity since the hazard zone cannot cover the entire area at one time. In addition, there are many other hazardzone possibilities from the same accident scenario. This risk analysis considered 21 wind speed/atmosphericstability combinations and 16 wind directions. These conditions are combined with three release hole sizes,three release orientations, and several event outcomes (flash fire, torch fire, etc.). The scenario presented inFigure 6-1 is just one of these many possible outcomes. Thus, vulnerability zones are not a meaningfulmeasure of risk. Vulnerability zones simply provide information about which areas could potentially beexposed to one unique accident, but provide no information about the probability of exposure.

6.2 Individual Risk Results

6.2.1 Risk Transects

Risk transects (measurements of individual risk as a function of distance from a pipeline) were constructedfor the following areas:

• Along the pipeline corridor that includes Lines 401, 002, and 0499• Along Line 162A• Along Line 176

The data that are generated when calculating the risk transects provide a picture of how the risk decreases asthe distance from the pipeline or pipeline corridor increases. Table 6-1 provides a summary of this data,presenting the distances to several individual risk levels.

QUEST-27-

Figure 6-2Risk Transect for Pipeline Corridor Containing

Lines 401, 002, and 0499

Table 6-1Approximate Distances from Pipeline or Pipeline corridor

to Individual Risk Levels

DesignationDistance (ft) to Individual Risk Level

10-5 10-6 10-7 10-8

401 † † 765 1,270

002 † 70 555 840

162A † 135 370 475

176 † 30 200 270

0499 (crude) † 2 110 155

401/002/0499 corridor † 250 805 1,270

† - Individual risk level does not exist.

Each risk transect provides a detailed graphical description of data in Table 6-1 (individual risk decreasingas the distance from the pipeline or pipeline corridor increases). Figure 6-2 presents the risk transect for the401/002/0499 corridor. The risk transect for Line 162A is presented in Figure 6-3, while the transect for Line176 is shown in Figure 6-4.

QUEST-28-

Figure 6-4Risk Transect for Line 176

Figure 6-3Risk Transect for Line 162A

QUEST-29-

6.2.2 Individual Risk Contours

Individual risk contours were constructed for the system of pipelines around the Mountain House Develop-ment in order to describe the combination of risk due to multiple pipelines that are not in the same corridor.Risk contours are constructed in the same manner as risk transects, but offer a two-dimensional representationof the risk around the pipelines. Figure 6-5 shows the individual risk contours in the area of the MountainHouse Development.

6.2.3 Individual Risk Summary

Any level of risk shown by a risk transect or risk contour is the risk of lethal exposure to any potential hazardassociated with the possible releases from one of the pipelines. For example, the individual risk contourlabeled 10-6 in Figure 6-5 represents one chance in one million per year (1.0 x 10-6/year) of being exposed toa fatal hazard from any of the possible releases of natural gas or crude oil from Line 401, Line 002, Line162A, Line 176, or Line 0499. Because the risk contours are based on annual data, this level of risk is depen-dent on an individual being in the location where the 10-6 contour is shown for 24 hours a day, 365 days peryear. For this reason, individual risk contours do not describe the risk to populations that are inherentlymobile, such as traffic on roadways or persons at a retail store for a small portion of a day. Their risk levelis lower, proportional to the amount of time they are present at a specific location.

It is important to note that for the pipeline releases studied, the number of people along the pipeline route doesnot affect the calculation of the individual risk contours. Thus, whether there are 1 or 100 people continuous-ly standing at a location near the pipeline, each person’s “risk” of exposure to a fatal release of gas from thepipeline would be the same.

From Table 6-1 and Figures 6-2 through 6-5, several things can be observed:

(a) There is no location where an annual individual risk level of 1.0 x 10-5 (per year) is reached.(b) All of the pipelines (except Line 401) produce risk levels greater than 1.0 x 10-6 at or near the

pipeline.(c) The individual risk level drops to a negligible value at approximately 1,500 feet from the combined

corridor, 500 feet from Line 162A, and 300 feet from Line 176. These distances roughly correspondto the maximum downwind hazards from Lines 401, 162A, and 176, respectively. Thus, beyondthese distances, the risk of being fatally affected by a release from a pipeline is zero.

6.3 Study Conclusions

Quest Consultants Inc. performed a quantitative risk analysis for five pipelines in the area of the proposedMountain House Specific Plan III Development. The study was composed of four distinct tasks.

Task 1. Determine potential releases that could result in significant hazardous conditions to the publicin the vicinity of the pipelines.

Task 2. For each potential release identified in Task 1, calculate the potentially lethal hazard zones.Task 3. For each potential release identified in Task 1, derive the annual probability of the release.Task 4. Using a consistent, accepted methodology, combine the potential release consequences from

Task 2 with the release frequencies from Task 3 to arrive at a measure of the “risk” the pipelinespose to the public.

QUEST-30-

Figure 6-5Individual Risk Contours in the area of the Mountain House Development

QUEST-31-

When reviewing the results presented in this report, it should be kept in mind that the risk levels presentedare higher than those that would actually exist. This is due to two primary factors.

(1) The calculations assume the public is present 24 hours a day, 365 days a year, at all locations alongthe pipeline route. The actual risk would be lower if the actual residence time of the population wasused in the study.

(2) In this analysis, several simplifying assumptions were made to reduce the computation requirementsand streamline the overall study. In each case, the simplifying assumption led to an overpredictionof the potential risk to persons in the area surrounding the pipeline. Several of the assumptions are:

(A) Local terrain. Although the terrain along the pipeline route is generally uniform, obstructionsto vapor travel within the area are potentially significant. In this analysis, no additional dilutiondue to obstructions being in the travel path of the vapor cloud was taken into account. Thisassumption is applicable to all releases that disperse near grade level and results in anoverprediction of the potential hazards.

(B) Release orientation. All horizontal and 45° gas releases were assumed to be oriented such thatthey are pointing in the direction the wind is blowing. This orientation allows the gas to travelthe maximum distance before diluting below the lower flammable limit. Any other releasedirection (upwind, crosswind, etc.) would result in smaller hazard zones.

(C) Flammable Vapor Cloud Ignition. If the pipeline gas does not ignite immediately uponrelease, the flammable vapor cloud is assumed to reach its maximum extent before any potentialignition source is found. This will overpredict the true hazard since the gas may find an ignitionsource before reaching its maximum extent.

In summary, this analysis provided a comprehensive estimation of the risk associated with the five pipelinesin the area of the proposed Mountain House Development. The results of the analysis show that the riskposed to the area surrounding the pipelines is similar to those surrounding natural gas and crude oil pipelinesystems present in many parts of the United States.

7.0 REFERENCES

Baker, Q. A., M. J. Tang, E. Scheier, and G. J. Silva (1994), “Vapor Cloud Explosion Analysis.” 28th LossPrevention Symposium, American Institute of Chemical Engineers (AIChE), 1994.

Baker, Q. A., C. M. Doolittle, G. A. Fitzgerald, and M. J. Tang (1998), “Recent Developments in the Baker-Strehlow VCE Analysis Methodology.” Process Safety Progress, 1998: p. 297.

CCPS (1989), Guidelines for Process Equipment Reliability Data, with Data Tables. Center for ChemicalProcess Safety of the American Institute of Chemical Engineers, 345 East 47th Street, New York, NewYork 19917, 1989.

Chang, Joseph C., Mark E. Fernau, Joseph S. Scire, and David G. Strimaitis (1998), A Critical Review ofFour Types of Air Quality Models Pertinent to MMS Regulatory and Environmental Assessment Missions.Mineral Management Service, Gulf of Mexico OCS Region, U.S. Department of the Interior, NewOrleans, Louisiana, November, 1998.

QUEST-32-

Cox, A. W., F. P. Lees, and M. L. Ang (1990), Classification of Hazardous Locations. The Institution ofChemical Engineers (IChemE), 1990.

DOT (1988), Hazardous Materials Information System. U.S. Department of Transportation, Materials Trans-portation Bureau, Washington, D.C., 1988.

EGPIDG (1988), Gas Pipeline Incidents. A report of the European Gas Pipeline Incident Data Group, April,1988.

Fearnehough, G. D. (1985), “The Control of Risk in Gas Transmission Pipelines.” Presented at the Institutionof Chemical Engineers Symposium on Assessment and Control of Major Hazards, Manchester, UnitedKingdom, April, 1985.

Hanna, S. R., D. G. Strimaitis, and J. C. Chang (1991), “Uncertainties in Hazardous Gas Model Predictions.”International Conference and Workshop on Modeling and Mitigating the Consequences of AccidentalReleases of Hazardous Materials, Center for Chemical Process Safety of the American Institute ofChemical Engineers, New Orleans, Louisiana, May 20-24, 1991.

Harris, R. J., and M. J. Wickens (1989), “Understanding Vapour Cloud Explosions—An ExperimentalStudy.” The Institution of Gas Engineers, Communication No. 1408, 1989.

Hirst, W. J. S., and J. A. Eyre (1982), “Maplin Sands Experiments 1980: Combustion of Large LNG andRefrigerated Liquid Propane Spills on the Sea.” Proceedings of the Second Symposium on Heavy Gasesand Risk Assessment, Frankfurt am Main, May 25-26, 1982: pp. 211-224.

HSE (1991), Major Hazard Aspects of the Transport of Dangerous Substances. Health and Safety Executive,Advisory Committee on Dangerous Substances, London, United Kingdom, 1991.

Johnson, D. M., P. Sutton, and M. J. Wickens (1991), “Scaled Experiments to Study Vapour Cloud Explo-sions.” Institution of Chemical Engineers (IChE) Symposium Series No. 124, Hazards XI, New Direc-tions in Process Safety, Manchester, United Kingdom, April, 1991: pp. 67-85.

Jones, D. J., G. S. Kramer, D. N. Gideon, and R. J. Eiber (1986), An Analysis of Reportable Incidents forNatural Gas Transmission and Gathering Lines, 1970 through June 1984. Pipeline Research Committee,American Gas Association, NG-18, Report No. 158, March, 1986.

OREDA (1984), OREDA, Offshore Reliability Data Handbook (First Edition). OREDA, Post Office Box370, N-1322 Hovik, Norway, 1984.

TRC (1991), Evaluation of Dense Gas Simulation Models. Prepared for the U.S. Environmental ProtectionAgency by TRC Environmental Consultants, Inc., East Hartford, Connecticut 06108, EPA Contract No.68-02-4399, May, 1991.

Tsao, C. K., and W. W. Perry (1979), Modifications to the Vulnerability Model: A Simulation System forAssessing Damage Resulting from Marine Spills. U.S. Coast Guard Report CG-D-38-79, Washington,D.C., March, 1979.

van Wingerden, C. J. M., and J. P. Zeeuwen (1983), “Flame Propagation in the Presence of RepeatedObstacles: Influence of Gas Reactivity and Degree of Confinement.” Journal of Hazardous Materials,Vol. 8, 1983: pp. 139-156.

QUEST-33-

Wiekema, B. J. (1984), “Vapour Cloud Explosions—An Analysis Based on Accidents, Part I.” Journal ofHazardous Materials, Vol. 8, 1984: pp. 285-311.

QUESTA-1

APPENDIX ACANARY BY QUEST® MODEL DESCRIPTIONS

The following model descriptions are taken from the CANARY by Quest User Manual.

Section A Engineering PropertiesSection B Pool Fire Radiation ModelSection C Torch Fire and Flare Radiation ModelSection E Fluid Release ModelSection F Momentum Jet Dispersion ModelSection I Vapor Cloud Explosion Model

CANARY by Quest User’s Manual Section A. Engineering Properties

July, 2003 Section A - Page 1

Engineering Properties

Purpose

The purpose of this model is to provide an accurate means of computing physical and thermodynamic prop-erties of a wide range of chemical mixtures and pure components using a minimum of initial information.

Required Data

(a) Fluid composition(b) Temperature and pressure of the fluid prior to release

Methodology

Basic thermodynamic properties are computed using the Peng-Robinson equation of state [Peng and Robin-son, 1976]. The necessary physical and thermodynamic properties are calculated in the following manner.

Step 1: The temperature and pressure of the fluid at storage conditions and the identity and mole fraction ofeach component of the fluid are obtained. Mixture parameters are determined using data from theextensive properties data base within CANARY.

Step 2: Each calculation begins with the computation of the vapor and liquid fluid composition. For caseswhere the temperature and pressure result in only one phase being present, the vapor or liquid com-position will be the same as the initial feed composition. The composition calculation is an iterativeprocedure using a modification of the techniques described by Starling [1973].

Step 3: Once the vapor and liquid compositions are known, the vapor and liquid densities, enthalpies,entropies, and heat capacities can be computed directly. Other physical properties (viscosity, thermalconductivity, surface tension, etc.) are computed using correlations developed in Reid, Prausnitz, andPoling [1987].

Step 4: A matrix of properties is computed over a range of temperatures and pressures. Physical and thermo-dynamics properties required by other models within CANARY are then interpolated from this table.

Basic Thermodynamic Equations

= 0 (1)( ) ( ) ( )3 2 2 2 31 3 2Z B Z A B B Z A B B B− − + − − − − −i i i i i

where: = fluid compressibility factor, , dimensionlessZ P VR Ti

i

= system pressure, kPaP= fluid specific volume, m3/kmolV

CANARY by Quest User’s Manual Section A. Engineering Properties

July, 2003 Section A - Page 2

= gas constant, 8.314 m3 kPa/(kmol K)R i i

= absolute temperature, KT

= A 2 2

a PR T

i

i

= a2 2

0.45724c

R TP

αi

i i

= α ( )20.51 1 rm T + − i

= m 20.37464 1.54226 0.26992ω ω+ −i i

= acentric factorω

=rTc

TT

= pseudo-critical temperature, KcT= pseudo-critical pressure, kPacP

= B b PR Ti

i

= b 0.0778 c

c

TRP

i i

= (2)H 20

o P P dH R T P TT

ρ

ρ

ρρ ρ

∂ + − + − ∂

⌠⌡

i i i

where: = enthalpy of fluid at system conditions, kJ/kgH= enthalpy of ideal gas at system temperature, kJ/kgoH

= (3)S ( ) 20

lno P dS R R T RT

ρ

ρ

ρρ ρρ

∂ − + − ∂

⌠⌡

i i i i i

where: = entropy of fluid at system conditions, kJ/(kg K)S i

= entropy of ideal gas at system temperature, kJ/(kg K)oS i

= (4)ln io

i

fR Tf

i i ( ) ( )o oi i i iH H T S S − − − i

where: = fugacity of component kPaif ,i

= standard state reference fugacity, kPaoif

CANARY by Quest User’s Manual Section A. Engineering Properties

July, 2003 Section A - Page 3

References

Peng, D., and D. B. Robinson, “New Two-Constant Equation of State.” Industrial Engineering ChemistryFundamentals, Vol. 15, No. 59, 1976.

Reid, R. C., J. M. Prausnitz, and B. E. Poling, The Properties of Gases and Liquids (Fourth Edition).McGraw-Hill Book Company, New York, New York, 1987.

Starling, K. E., Fluid Thermodynamic Properties for Light Petroleum Systems. Gulf Publishing Company,Houston, Texas, 1973.

CANARY by Quest User’s Manual Section B. Pool Fire Radiation Model

July, 2003 Section B - Page 1

Pool Fire Radiation Model

Purpose

The purpose of this model is to predict the impact of fire radiation emitted by flames that are fueled by vaporsemanating from liquid pools. Specifically, the model predicts the maximum radiant heat flux incident upona target as a function of distance between the target and the flame.

Required Data

(a) Composition of the liquid in the pool(b) Temperature of the liquid in the pool(c) Wind speed(d) Air temperature(e) Relative humidity(f) Elevation of the target (relative to grade)(g) Elevation of the pool (relative to grade)(h) Dimensions of the free surface of the pool(i) Orientation of the pool (relative to the wind direction)(j) Spill surface (land or water)

Methodology

Step 1: The geometric shape of the flame is defined. The flame column above a circular pool, square pool,or rectangular pool is modeled as an elliptical cylinder.

Step 2: The dimensions of the flame column are determined. The dimensions of the base of the flame aredefined by the pool dimensions. An empirical correlation developed by Thomas [1965] is used tocalculate the length (height) of the flame.

=L( )

0.61

0.542 ha h

mDg Dρ

i ii i

where: = length (height) of the flame, mL= hydraulic diameter of the liquid pool, mhD= mass burning flux, kg/(m2 s)m i

= density of air, kg/m3aρ

= gravitational acceleration, 9.8 m/s2g

Notes: Mass burning fluxes used in the Thomas equation are the steady-state rates for pools on land(soil, concrete, etc.) or water, whichever is specified by the user.

CANARY by Quest User’s Manual Section B. Pool Fire Radiation Model

July, 2003 Section B - Page 2

For pool fires with hydraulic diameters greater than 100 m, the flame length, is set equal,Lto the length calculated for 100 m.hD =

Step 3: The angle to which the flame is bent from vertical by the wind is calculated using an empirical( )Φcorrelation developed by Welker and Sliepcevich [1970].

= tan( )cos ( )

ΦΦ

0.70.07 0.62

3.2 h a v

a h a

D u ug D

ρ ρµ ρ

−

i ii i i

i

where: = angle the flame tilts from vertical, degreesΦ= wind speed, m/su = viscosity of air, kg/(m s)aµ i

= density of fuel vapor, kg/m3vρ

Step 4: The increase in the downwind dimension of the base of the flame (flame drag) is calculated using ageneralized form of the empirical correlation Moorhouse [1982] developed for large circular poolfires.

= wD0.0692

1.5 xx

uDg D

i ii

where: = downwind dimension of base of tilted flame, mwD= downwind dimension of the pool, mxD

Step 5: The flame is divided into two zones: a clear zone in which the flame is not obscured by smoke; anda smoky zone in which a fraction of the flame surface is obscured by smoke. The length of the clearzone is calculated by the following equation, which is based on an empirical correlation developedby Pritchard and Binding [1992].

=cL ( )1.13 2.49

0.1790.655.05 1ha

m CD uHρ

−− +

i i i i

where: = length of the clear zone, mcL

= carbon/hydrogen ratio of fuel, dimensionlessCH

Step 6: The surface flux of the clear zone is calculated using the following equation.

= c zq ( )1 hb Ds mq e−− ii

where: = surface flux of the clear zone, kW/m2c zq

= maximum surface flux, kW/m2s mq

= extinction coefficient, m-1b

CANARY by Quest User’s Manual Section B. Pool Fire Radiation Model

July, 2003 Section B - Page 3

Average surface flux of the smoky zone, is then calculated, based on the following assumptions.,s zq

• The smoky zone consists of clean-burning areas and areas in which the flame is obscured bysmoke.

• Within the smoky zone, the fraction of the flame surface that is obscured by smoke is afunction of the fuel properties and pool diameter.

• Smoky areas within the smoky zone have a surface flux of 20 kW/m2 [Hagglund and Pers-son,1976].

• Clean-burning areas of the smoky zone have the same surface flux as the clean-burning zone.• The average surface flux of the smoky zone is the area-weighted average of the surface

fluxes for the smoky areas and the clean-burning areas within the smoky zone.

(This two-zone concept is based on the Health and Safety Executive POOLFIRE6 model, as describ-ed by Rew and Hulbert [1996].)

Step 7: The surface of the flame is divided into numerous differential areas. The following equation is thenused to calculate the view factor from a differential target, at a specific location outside the flame,to each differential area on the surface of the flame.

= for [ ] and [ ] < 90N

t fdA dAF →

( ) ( )2

cos cost ffdA

rβ β

πi

ii

tβ fβ

where: = view factor from a differential area on the target to a differential area on thet fdA dAF →

surface of the flame, dimensionless= differential area on the flame surface, m2

fdA= differential area on the target surface, m2

tdA= distance between differential areas and mr tdA ,fdA= angle between normal to and the line from to degreestβ tdA tdA ,fdA= angle between normal to and the line from to degreesfβ fdA tdA ,fdA

Step 8: The radiant heat flux incident upon the target is computed by multiplying the view factor for eachdifferential area on the flame by the appropriate surface flux or and by the appropriate( c zq )s zqatmospheric transmittance, then summing these values over the surface of the flame.

= a iqt f

f

sf dA dAA

q F τ→∑ i i

where: = attenuated radiant heat flux incident upon the target due to radiant heat emitted by thea iqflame, kW/m2

= area of the surface of the flamefA= radiant heat flux emitted by the surface of the flame, kW/m2 equals either ors fq ( s fq c zq

as appropriate),s zq= atmospheric transmittance, dimensionlessτ

Atmospheric transmittance, is a function of absolute humidity and the path length between dif-,τ ,rferential areas on the flame and target [Wayne, 1991].

Step 9: Steps 7 and 8 are repeated for numerous target locations.

CANARY by Quest User’s Manual Section B. Pool Fire Radiation Model

July, 2003 Section B - Page 4

0 2000 4000 6000 8000 10000

Pool Fire Test Data, Btu/hr-ft2

0

2000

4000

6000

8000

10000

CA

NAR

Y Pr

edic

tions

, Btu

/hr-f

t2

40 x 40 ft20 x 20 ft10 x 10 ft5 x 5 ft

Figure B-1

Validation

Several of the equations used in the Pool Fire Radiation Model are empirical relationships based on data frommedium- to large-scale experiments, which ensures reasonably good agreement between model predictionsand experimental data for variables such as flame length and tilt angle. Comparisons of experimental dataand model predictions for incident heat flux at specific locations are more meaningful and of greater interest.Unfortunately, few reports on medium- or large-scale experiments contain the level of detail required to makesuch comparisons.

One source of detailed test data is a report by Welker and Cavin [1982]. It contains data from sixty-one poolfire tests involving commercial propane. Variables that were examined during these tests include pool size(2.7 to 152 m2) and wind speed. Figure B-1 compares the predicted values of incident heat flux withexperimental data from the sixty-one pool fire tests.

References

Hagglund B., and L. Persson, The Heat Radiation from Petroleum Fires. FOA Rapport, Forsvarets For-skningsanstalt, Stockholm, Sweden, 1976.

Moorhouse, J., “Scaling Criteria for Pool Fires Derived from Large-Scale Experiments.” The Assessment ofMajor Hazards, Symposium Series No. 71, The Institution of Chemical Engineers, Pergamon Press Ltd.,Oxford, United Kingdom, 1982: pp. 165-179.

Pritchard, M. J., and T. M. Binding, “FIRE2: A New Approach for Predicting Thermal Radiation Levels fromHydrocarbon Pool Fires.” IChemE Symposium Series, No. 130, 1992: pp. 491-505.

CANARY by Quest User’s Manual Section B. Pool Fire Radiation Model

July, 2003 Section B - Page 5

Rew, P. J., and W. G. Hulbert, Development of Pool Fire Thermal Radiation Model. HSE Contract ResearchReport No. 96/1996.

Thomas, P. H., F.R. Note 600, Fire Research Station, Borehamwood, England, 1965.

Wayne, F. D., “An Economical Formula for Calculating Atmospheric Infrared Transmissivities.” Journalof Loss Prevention in the Process Industries, Vol. 4, January, 1991: pp. 86-92.

Welker, J. R., and W. D. Cavin, Vaporization, Dispersion, and Radiant Fluxes from LPG Spills. Final ReportNo. DOE-EP-0042, Department of Energy Contract No. DOE-AC05-78EV-06020-1, May, 1982 (NTISNo. DOE-EV-06020-1).

Welker, J. R., and C. M. Sliepcevich, Susceptibility of Potential Target Components to Defeat by ThermalAction. University of Oklahoma Research Institute, Report No. OURI-1578-FR, Norman, Oklahoma,1970.

CANARY by Quest User’s Manual Section C. Torch Fire and Flare Radiation Model

July, 2003 Section C - Page 1

Torch Fire and Flare Radiation Model

Purpose

The purpose of this model is to predict the impact of fire radiation emitted by burning jets of vapor. Specific-ally, the model predicts the maximum radiant heat flux incident upon a target as a function of distancebetween the target and the point of release.

Required Data

(a) Composition of the released material (b) Temperature and pressure of the material before release(c) Mass flow rate of the material being released(d) Diameter of the exit hole(e) Wind speed(f) Air temperature(g) Relative humidity(h) Elevation of the target (relative to grade)(i) Elevation of the point of release (relative to grade)(j) Angle of the release (relative to horizontal)

Methodology

Step 1: A correlation based on a Momentum Jet Model is used to determine the length of the flame. Thiscorrelation accounts for the effects of:

• composition of the released material,• diameter of the exit hole,• release rate,• release velocity, and• wind speed.

Step 2: To determine the behavior of the flame, the model uses a momentum-based approach that considersincreasing plume buoyancy along the flame and the bending force of the wind. The followingequations are used to determine the path of the centerline of the flame [Cook, et al., 1987].

= (downwind)XΦ ( ) ( ) ( ) ( )0.50.5 sin cosja u uρ θ ϕ ρ∞ ∞+i i i i

= (crosswind)YΦ ( ) ( ) ( )0.5 sin sinja uρ θ ϕi i i

= (vertical)ZΦ ( ) ( ) ( ) ( )0.50.5 1cosja biu u

nρ θ ρ∞

++i i i i

CANARY by Quest User’s Manual Section C. Torch Fire and Flare Radiation Model

July, 2003 Section C - Page 2

where: = momentum flux in directionX Y ZΦ , ,X Y Z = density of the jet fluid at ambient conditions, kg/m3

jaρ = average axial velocity of the flame, m/su = release angle in plane (relative to horizontal), degreesθ X Z− = release angle in plane (relative to downwind), degreesϕ X Y−

= density of air, kg/m3ρ∞

= wind speed, m/su∞

= density of combustion products, kg/m3bρ = buoyancy velocity, m/sbu

= number of points taken along the flame lengthn

These correlations were developed to predict the path of a torch flame when released at variousorientations. The model currently does not allow a release angle in a crosswind direction; the releaseangle is confined to the downwind/vertical plane (i.e., = 0).ϕ

Step 3: The angle of flame tilt is defined as the inclination of a straight line between the point of release andthe end point of the flame centerline path (as determined in Step 2).

Step 4: The geometric shape of the flame is defined as a frustum of a cone (as suggested by several flare/fireresearchers [e.g., Kalghatgi, 1983, Chamberlain, 1987]), but modified by adding a hemisphere to thelarge end of the frustum. The small end of the frustum is positioned at the point of release, and thecenterline of the frustum is inclined at the angle determined in Step 3.

Step 5: The surface emissive power is determined from the molecular weight and heat of combustion of theburning material, the release rate and velocity, and the surface area of the flame.

Step 6: The surface of the flame is divided into numerous differential areas. The following equation is thenused to calculate the view factor from a differential target, at a specific location outside the flame,to each differential area on the surface of the flame.

= for [ ] and [ ] < 90°t fdA dAF →

( ) ( )2

cos cost ffdA

rβ β

πi

ii

tβ fβ

where: = view factor from a differential area on the target to a differential area on thet fdA dAF →

surface of the flame, dimensionless= differential area on the flame surface, m2

fdA= differential area on the target surface, m2

tdA= distance between differential areas and mr tdA ,fdA= angle between normal to and the line from to degreestβ tdA tdA ,fdA= angle between normal to and the line from to degreesfβ fdA tdA ,fdA

Step 7: The radiant heat flux incident upon the target is computed by multiplying the view factor for eachdifferential area on the flame by the surface missive power and by the appropriate atmospheric trans-mittance, then summing these values over the surface of the flame.

CANARY by Quest User’s Manual Section C. Torch Fire and Flare Radiation Model

July, 2003 Section C - Page 3

=a iqt f

f

sf dA dAA

q F τ→∑ i i

where: = attenuated radiant heat flux incident upon the target due to radiant heat emitted by thea iqflame, kW/m2

= area of the surface of the flamefA= radiant heat flux emitted by the surface of the flame, kW/m2

s fq= atmospheric transmittance, dimensionlessτ

Atmospheric transmittance, is a function of absolute humidity and the path length between,τ ,rdifferential areas on the flame and target [Wayne, 1991].

Step 8: Steps 6 and 7 are repeated for numerous target locations.

Validation

Several of the equations used in the Torch Fire and Flare Radiation Model are empirical relationships basedon data from medium- to large-scale experiments, which ensures reasonably good agreement between modelpredictions and experimental data for variables such as flame tilt angle. Comparisons of experimental dataand model predictions for incident heat flux at specific locations are more meaningful and of greater interest.Unfortunately, few reports on medium- or large-scale experiments contain the level of detail required to makesuch comparisons.

One reasonable source of test data is a report by Chamberlain [1987]. It contains data from seven flare testsinvolving natural gas releases from industrial flares, with several data points being reported for each test.Variables that were examined during these tests include release diameter (0.203 and 1.07 m), release rate andvelocity, and wind speed. Figure C-1 compares the predicted values of incident heat flux with experimentaldata from the seven flare tests.

References

Chamberlain, G. A., “Developments in Design Methods for Predicting Thermal Radiation from Flares.”Chemical Engineering Research and Design, Vol. 65, July, 1987.

Cook, D. K., M. Fairweather, G. Hankinson, and K. O’Brien, “Flaring of Natural Gas from Inclined VentStacks.” IChemE Symposium Series #102, Pergamon Press, 1987.

Kalghatgi, G. T., “The Visible Shape and Size of a Turbulent Hydrocarbon Jet Diffusion Flame in a CrossWind.” Combustion and Flame, Vol. 52, 1983: pp. 91-106.

Wayne, F. D., “An Economical Formula for Calculating Atmospheric Infrared Transmissivities.” Journalof Loss Prevention in the Process Industries, Vol. 4, January, 1991: pp. 86-92.

CANARY by Quest User’s Manual Section C. Torch Fire and Flare Radiation Model

July, 2003 Section C - Page 4

0 2 4 6 8 10

Flare Test Data, kW/m2

0

2

4

6

8

10

CA

NA

RY

Pre

dict

ions

, kW

/m2

Test Series 3Test Series 4

Figure C-1

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 1

Fluid Release Model

Purpose

The purpose of the Fluid Release Model is to predict the rate of mass release from a breach of containment.Specifically, the model predicts the rate of flow and the physical state (liquid, two-phase, or gas) of therelease of a fluid stream as it enters the atmosphere from a circular breach in a pipe or vessel wall. The modelalso computes the amount of vapor and aerosol produced and the rate at which liquid reaches the ground.

Required Data

(a) Composition of the fluid(b) Temperature and pressure of the fluid just prior to the time of the breach(c) Normal flow rate of fluid into the vessel or in the pipe(d) Size of the pipe and/or vessel(e) Length of pipe(f) Area of the breach(g) Angle of release relative to horizontal(h) Elevation of release point above grade

Methodology

Step 1: Calculation of Initial Flow Conditions

The initial conditions (before the breach occurs) in the piping and/or vessel are determined from theinput data, coupled with a calculation to determine the initial pressure profile in the piping. Thepressure profile is computed by dividing the pipe into small incremental lengths and computing theflow conditions stepwise from the vessel to the breach point. As the flow conditions are computed,the time required for a sonic wave to traverse each section is also computed. The flow in any lengthincrement can be all vapor, all liquid, or two-phase (this implies that the sonic velocity within eachsection may vary). As flow conditions are computed in each length increment, checks are made todetermine if the fluid velocity has exceeded the sonic velocity or if the pressure in the flow incrementhas reached atmospheric. If either condition has been reached, an error code is generated andcomputations are stopped.

Step 2: Initial Unsteady State Flow Calculations

When a breach occurs in a system with piping, a disturbance in flow and pressure propagates fromthe breach point at the local sonic velocity of the fluid. During the time required for the disturbanceto reach the upstream end of the piping, a period of highly unsteady flow occurs. The portion of thepiping that has experienced the passage of the pressure disturbance is in accelerated flow, while theportion upstream of the disturbance is in the same flow regime as before the breach occurred.

To compute the flow rate from the breach during the initial unsteady flow period, a small timeincrement is selected and the distance that the pressure disturbance has moved in that time incrementis computed using the sonic velocity profile found in the initial pressure profile calculation. The

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 2

disturbed length is subdivided into small increments for use in an iterative pressure balancecalculation. A pressure balance is achieved when a breach pressure is found that balances the flowfrom the breach and the flow in the disturbed section of piping. Another time increment is added,and the iterative procedure continues. The unsteady period continues until the pressure disturbancereaches the upstream end of the pipe.

Step 3: Long-Term Unsteady State Flow Calculations

The long-term unsteady state flow calculations are characterized by flow in the piping system thatis changing more slowly than during the initial unsteady state calculations. The length of acceleratedflow in the piping is constant, set by the user input pipe length. The vessel contents are being deplet-ed, resulting in a potential lowering of pressure in the vessel. As with the other flow calculations,the time is incremented and the vessel conditions are computed. The new vessel conditions serve asinput for the pressure drop calculations in the pipe. When a breach pressure is computed thatbalances the breach flow with the flow in the piping, a solution for that time is achieved. The solu-tion continues until the ending time or other ending conditions are reached.

The frictional losses in the piping system are computed using the equation:

= (1)h24

2ls

c e

f L Ug D

i i i

i i

where: = head (pressure) loss, ft of fluidh = friction factorf = length of system, ftL = average flowing velocity, ft/secU= gravitational constant, 32.2 lbm • ft/(lbf • sec2)cg= equivalent diameter of duct, fteD

The friction factor is computed using the following equation:

= (2)1f

102 18.71.74 2.0 log

eD Re fε − +

ii

i

where: = pipe roughness, ftε= Reynolds number, , dimensionlessRe /eD U ρ µi i

= fluid density, lb/ft3ρ= fluid viscosity, lb/(ft • sec)µ

Equations (1) and (2) are used for liquid, vapor, and two-phase flow regimes. Since the piping issubdivided into small lengths, changes in velocity and physical properties across each segment areassumed to be negligible. At each step in the calculation, a check is made to determine if the fluidvelocity has reached or exceeded the computed critical (sonic) velocity for the fluid. If the criticalvelocity has been exceeded, the velocity is constrained to the critical velocity and the maximum massflow rate in the piping has been set.

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 3

If the fluid in the piping is in two-phase flow, the Lockhart and Martinelli [1949] modification toEquation (1) is used. The Lockhart and Martinelli equation for head loss is shown below:

= (3)TPh2

2 42

ls

c e

f L Ug D

Φ

i i ii

i i

where: = head loss for two-phase flow, ft of fluidTPh= empirical parameter correlating single- and two-phase flow, dimensionlessΦ= superficial liquid velocity (velocity of liquid if liquid filled the pipe), ft/seclsU

This equation is valid over short distances where the flowing velocity does not change appreciably.

Validation

Validation of fluid flow models is difficult since little data are available for comparison. Fletcher [1983]presented a set of data for flashing CFC-11 flowing through orifices and piping. Figures E-1 through E-4compare calculations made using the Fluid Release Model with the data presented by Fletcher. Figure E-1compares fluid fluxes for orifice type releases. These releases had length-to-diameter (L/D) ratios less than0.88. Figure E-2 compares computed and experimental release fluxes for an L/D ratio of 120 at several levelsof storage pressure. Figure E-3 compares similar releases for an L/D of 37.5. Figure E-4 shows predictedand experimental release fluxes at a given pressure for L/D ratios from 1 to 200.

Figures E-5 and E-6 compare computed and experimental gas discharge rates for the complete breach of twopipes. One pipe had an internal diameter of 6.2 inches (0.157 m); the other had a diameter of 12 inches (0.305m). These pipes were initially pressurized to 1,000 psia with air and then explosively ruptured. Theexperimental values were reported in a research paper for Alberta Environment, authored by Wilson [1981].

Aerosols and Liquid Droplet Evaporation

Liquids stored at temperatures above their atmospheric pressure boiling point (superheated liquids) will giveoff vapor when released from storage. If the temperature of storage is sufficiently above the normal boilingpoint, the energy of the released vapor will break the liquid stream into small droplets. If these droplets aresmall enough, they will not settle, but remain in the vapor stream as aerosol droplets. The presence of aerosoldroplets in the vapor stream changes its apparent density and provides an additional source of vapor. Dropletslarge enough to fall to the ground will lose mass due to evaporation during their fall.

The prediction of aerosol formation and amount of aerosol formed is based on the theoretical work performedfor the Center for Chemical Process Safety (CCPS) by CREARE. CREARE’s work has been extended andcorrected by Quest. The extension to the model computes the non-aerosol drop evaporation. In Figure E-7,the four experimental data sets available for comparison (chlorine (Cl2), methylamine (MMA), CFC-11, andcyclohexane) are compared to the values computed by the CANARY Aerosol Model.

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 4

0 1 2 3 4 5System Pressure, barg

0

5000

10000

15000

20000

Mas

s Fl

ux, k

g/(m

2 -s)

ExperimentalComputed

Figure E-1Comparison of CFC-11 Orifice Releases as a Function of System Pressure

0 1 2 3 4System Pressure, barg

0

1000

2000

3000

4000

5000

6000

7000

Mas

s Fl

ux, k

g/(m

2 -s)

ExperimentalComputed

Figure E-2CFC-11 Release Rate Comparison with L/D of 120

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 5

0 1 2 3 4System Pressure, barg

0

1000

2000

3000

4000

5000

6000

7000

Mas

s Fl

ux, k

g/(m

2 -s)

ExperimentalComputed

Figure E-3CFC-11 Release Rate Comparison with L/D of 37.5

0 20 40 60 80 100 120 140 160 180 200

Length/Diameter Ratio

0

4000

8000

12000

16000

Mas

s Fl

ux, k

g/(m

2 -s)

ExperimentalComputed

Figure E-4CFC-11 Release Rate Comparison at Varying L/D Ratios

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 6

0 10 20 30 40 50 60 70Time, sec

10

100

2

3

4

5

6

789

2

3

4

Mas

s R

ate,

kg/

sec

ExperimentalComputed

Figure E-5Air Discharge Rates for 0.157 m Diameter Piping

0 20 40 60 80 100Time, sec

100

1000

8

9

2

3

4

5

6

7

8

9

Mas

s R

ate,

kg/

sec

ExperimentalComputed

Figure E-6Air Discharge Rates for 0.305 m Diameter Piping

CANARY by Quest User’s Manual Section E. Fluid Release Model

July, 2003 Section E - Page 7

230 250 270 290 310 330 350 370 390 410

Temperature, K

0

10

20

30

40

50

60

70

80

90

100

110

Perc

ent o

f Liq

uid

to G

roun

d (N

orm

aliz

ed to

100

% a

t NB

P)

Cl2 MMA CFC-11 Cyclohexane

Figure E-7Aerosol Formation as a Function of Storage Temperature

References

Fletcher, B., “Flashing Flow Through Orifices and Pipes.” Paper presented at the AIChE Loss PreventionSymposium, Denver, Colorado, 1983.

Lockhart, R. W., and R. C. Martinelli, “Proposed Correlation of Data for Isothermal Two-Phase, Two-Component Flow in Pipes.” Chemical Engineering Progress, Vol. 45, 1949: p. 39.

Wilson, D. J., “Expansion and Plume Rise of Gas Jets from High Pressure Pipeline Ruptures.” ResearchPaper, Pollution Control Division, Alberta Environment, April, 1981.

CANARY by Quest User’s Manual Section F. Momentum Jet Dispersion Model

July, 2003 Section F - Page 1

Momentum Jet Dispersion Model

Purpose

The purpose of this model is to predict the dispersion of a jet release into ambient air. It is used to predictthe downwind travel of a flammable or toxic gas or aerosol momentum jet release.

Required Data

(a) Composition and properties of the released material(b) Temperature of released material(c) Release rate of material(d) Vertical release angle relative to wind direction(e) Height of release(f) Release area(g) Ambient wind speed(h) Ambient Pasquill-Gifford stability class(i) Ambient temperature(j) Relative humidity(k) Surface roughness scale

Methodology

Step 1: An assumption is made that flow perpendicular to the main flow in the plume is negligible, that thevelocity and concentration profiles in the jet are similar at all sections of the jet, that molecular trans-port in the jet is negligible, and that longitudinal turbulent transport is negligible when compared tolongitudinal convective transport. The coordinate system is then defined in and where is thes ,r spath length of the plume and is the radial distance from the plume centerline. The angle betweenrthe plume axis and horizontal is referred to as Relationships between the downwind coordinate, .θ ,xvertical coordinate, and plume axis are given simply by:,y

= (1)dxds

( )cos θ

and

= (2)d yd s

( )sin θ

Step 2: Velocity, concentration, and density profiles are assumed to be cylindrically symmetric about theplume axis and are assumed to be Gaussian in shape. The three profiles are taken as:

= (3)( ), ,u s r θ ( ) ( ) ( )

2

2*cosr

b saU u s eθ

−

+i i

CANARY by Quest User’s Manual Section F. Momentum Jet Dispersion Model

July, 2003 Section F - Page 2

where: = plume velocity, m/su= ambient wind speed, m/saU= plume velocity relative to the wind in the downwind direction at the plume axis, m/s*u= characteristic width of the plume at distance from the release, m( )b s s

= (4)( ), ,s rρ θ ( ) ( )

2

2 2*rb s

a s e λρ ρ−

+ ii

where: = plume density, kg/m3ρ = density of ambient air, kg/m3 aρ

= density difference between plume axis and ambient air, kg/m3( )* sρ = turbulent Schmidt number, 1.352λ

= (5)( ), ,c s r θ ( ) ( )

2

2 2*r

b sc s e λ−ii

where: = pollutant concentration in the plume, kg/m3c= pollutant concentration at plume centerline, kg/m3( )*c s

Step 3: The equation for air entrainment into the plume and the conservation equations can then be solved.The equation for air entrainment is:

(6)( )2

02bd u dr

dsρ π∫ i i i i

= ( ) ( ) ( ){ }*1 2 32 sin cosa ab u s U uπ ρ α α θ θ α ′+ +i i i i i i i i

where: = entrainment coefficient for a free jet, 0.0571α = entrainment coefficient for a line thermal, 0.52α = entrainment coefficient due to turbulence, 1.03α = turbulent entrainment velocity (root mean square of the wind velocity fluctuation isu′ used for this number), m/s

Step 4: The equations of conservation of mass, momentum, and energy are given as:

= 0 (7)( )2

02bd c u dr

dsπ∫ i i i i

(8)( )( )( )2 2

0cos 2bd u dr

dsρ θ π∫ i i i i i

= ( ) ( ) ( ){ }*1 2 32 sin cosa ab u s U uπ ρ α α θ θ α ′+ +i i i i i i i i i

+ ( )2 sind a aC b Uπ ρ θi i i i

CANARY by Quest User’s Manual Section F. Momentum Jet Dispersion Model

July, 2003 Section F - Page 3

(9)( )( )2 2

0cos 2

bd u drds

ρ θ π∫ i i i i i

= ( ) ( ) ( )2 2

0sin cos

b

a d a ag r dr C b Uρ ρ π π ρ θ θ− ±∫ i i i i i i i i i

(10)2

00

1 1 2b

a

d u r drd s

ρ πρ ρ

−

∫ i i i i i

= ( ) ( ) ( ){ }*1 2 3

0

1 12 | | sin | cosa aa a

b u s U úρ π α α θ θ αρ ρ

− + +

i i i i i i i i

The subscript refers to conditions at the point of release. These equations are integrated along the0path of the plume to yield the concentration profiles as a function of elevation and distance down-wind of the release.

Step 5: After the steady-state equations are solved, an along-wind dispersion correction is applied to accountfor short-duration releases. This is accomplished using the method outlined by Palazzi, et al. [1982].

Step 6: If the plume reaches the ground, it is coupled to the Heavy Gas Dispersion Model (described inSection G) and the dispersion calculations continue.

Validation

The Momentum Jet Dispersion Model used in CANARY was validated by comparing results obtained fromthe model with experimental data from field tests. Data used for this comparison and the conditions used inthe model were taken from an American Petroleum Institute (API) study [Hanna, Strimaitis, and Chang,1991]. For this model, comparisons were made with the Desert Tortoise, Goldfish, and Prairie Grass seriesof dispersion tests. Results of these comparisons are shown in Figure F-1.

References

Astleford, W. J., T. B. Morrow, and J. C. Buckingham, Hazardous Chemical Vapor Handbook for MarineTank Vessels (Final Report – Phase II). U.S. Coast Guard Report No. CG-D-12-83, April, 1983.

Hanna, S. R., D. G. Strimaitis, and J. C. Chang, Hazard Response Modeling Uncertainty (A QuantitativeMethod), Evaluation of Commonly Used Hazardous Gas Dispersion Models, Volume II. Study co-sponsored by the Air Force Engineering and Services Center, Tyndall Air Force Base, Florida, and theAmerican Petroleum Institute; performed by Sigma Research Corporation, Westford, Massachusetts,September, 1991.

Havens, J., and T. Spicer, LNG Vapor Dispersion Prediction with the DEGADIS Dense Gas DispersionModel. Gas Research Institute Contract No. 5086-252-1287 with the University of Arkansas, September,1990: pp. 37-48.

Ooms, G., “A New Method for the Calculation of the Plume Path of Gases Emitted by a Stack.” AtmosphericEnvironment, Vol. 6, 1972: pp. 889-909.

CANARY by Quest User’s Manual Section F. Momentum Jet Dispersion Model

July, 2003 Section F - Page 4

10-2 10-1 100 101 102 103 104 105 10610-2

10-1

100

101

102

103

104

105

106

Desert Tortoise (NH3)Goldfish (HF)Prarie Grass (SO2)

OVERPREDICTION

UNDERPREDICTION

Flammable RangeTrace Gas Range

Field Data Concentration (ppm)

CAN

ARY

Pre

dict

ed C

once

ntra

tion

(ppm

)

Figure F-1

Ooms, G., A. P. Mahieu, and F. Zelis, “The Plume Path of Vent Gases Heavier than Air.” First InternationalSymposium on Loss Prevention and Safety Promotion in the Process Industries, C. H. Buschman, Editor,Elsevier Press, 1974.

Palazzi, E., M. De Faveri, G. Fumarola, and G. Ferraiolo, “Diffusion from a Steady Source of Short Dura-tion.” Atmospheric Environment, Vol. 16, No. 12, 1982: pp. 2785-2790.

CANARY by Quest User’s Manual Section I. Vapor Cloud Explosion Model

July, 2003 Section I - Page 1

Vapor Cloud Explosion Model

Purpose

The purpose of this model is to predict the overpressure field that would be produced by the explosion of apartially confined and/or obstructed fuel-air cloud, based on the Baker-Strehlow methodology. Specifically,the model predicts the magnitude of the peak side-on overpressure and specific impulse as a function ofdistance from the source of the explosion.

Required Data

(a) Composition of the fuel (flammable fluid) involved in the explosion(b) Total mass of fuel in the flammable cloud at the time of ignition or the volume of the partially-confined/

obstructed area(c) Fuel reactivity (high, medium, or low)(d) Obstacle density (high, medium, or low)(e) Flame expansion (1-D, 2-D, 2½-D, or 3-D)(f) Reflection factor

Methodology

Step 1: The combustion energy of the cloud is estimated by multiplying its mass by the heat of combustion.If the volume of the flammable cloud is input, the mass is estimated by assuming that a stoichiometricmixture of gas and air exists within that volume.

Step 2: The combustion energy is multiplied by the reflection factor to account for blast reflection from theground or surrounding objects.

Step 3: Flame speed is determined from the fuel reactivity, obstacle density, and flame expansion parameters,as presented in Baker, et al. [1994, 1998].

Fuel reactivity and obstacle density each have low, medium, and high choices. The flame expansionparameter allows choices of 1-D, 2-D, 2.5-D, and 3-D. The choices for these three parameters createa matrix of 36 possibilities, thus allowing locations that have differing levels of congestion or con-finement to produce different overpressures. Each matrix possibility corresponds to a flame speed,and thus a peak (source) overpressure. The meanings of the three parameters and their options are:

Fuel Reactivity (High, Medium, or Low). The fuels considered to have high reactivity areacetylene, ethylene oxide, propylene oxide, and hydrogen. Low reactivity fuels are (pure)methane and carbon monoxide. All other fuels are medium reactivity. If fuels from differentreactivity categories are mixed, the model recommends using the higher category unless theamount of higher reactivity fuel is less than 2% of the mixture.

CANARY by Quest User’s Manual Section I. Vapor Cloud Explosion Model

July, 2003 Section I - Page 2

Obstacle Density (High, Medium, or Low). High obstacle density is encountered whenobjects in the flame’s path are closely spaced. This is defined as multiple layers of obstruc-tion resulting in at least a 40% blockage ratio (i.e., 40% of the volume is occupied byobstacles). Low density areas are defined as having a blockage ratio of less than 10%. Allother blockage ratios fall into the medium category.

Flame Expansion (1-D, 2-D, 2.5-D, or 3-D). The expansion of the flame front must be char-acterized with one of these four descriptors. 1-D expansion is likened to an explosion in apipe or hallway. 2-D expansion can be described as what occurs between flat, parallel sur-faces. An unconfined (hemispherical expansion) case is described as 3-D. The additionaldescriptor of 2.5-D is used for situations that begin as 2-D and quickly transition to 3-D.

Step 4: Based on the calculated flame speed, appropriate blast curves are selected from the figures in Baker,et al., 1994. For flame speeds not shown on the graph, appropriate curves are prepared by interpola-tion between existing curves.

Step 5: The Sachs scaled distance, is calculated for several distances using the equation:,R

= R 1/ 3

0

R

EP

where: = distance from the center of the explosionR= total energy calculated in step 2, aboveE= atmospheric pressure0P

Step 6: The peak side-on overpressure and specific impulse at each scaled distance are determined from theblast curves in Baker, et al., 1994.

References

Baker, Q. A., M. J. Tang, E. Scheier, and G. J. Silva, “Vapor Cloud Explosion Analysis.” 28th Loss Preven-tion Symposium, AIChE, 1994.

Baker, Q. A., C. M. Doolittle, G. A. Fitzgerald, and M. J. Tang, “Recent Developments in the Baker-StrehlowVCE Analysis Methodology.” Process Safety Progress, 1998: p. 297.