diff --git "a/corpus/csb_chevron_final.txt" "b/corpus/csb_chevron_final.txt" new file mode 100644--- /dev/null +++ "b/corpus/csb_chevron_final.txt" @@ -0,0 +1,4470 @@ +U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD  + +FINAL INVESTIGATION REPORT + + +REPORT NO. 2012-03-I-CA +JANUARY 2015 + + +CHEVRON RICHMOND REFINERY +PIPE RUPTURE AND FIRE + + + + + + + + + + + + + + +CHEVRON RICHMOND REFINERY #4 CRUDE UNIT +RICHMOND, CALIFORNIA +KEY ISSUES: AUGUST 6, 2012 + CHEVRON PROCESS SAFETY PROGRAMS + CHEVRON EMERGENCY RESPONSE + MECHANICAL INTEGRITY INDUSTRY STANDARD DEFICIENCIES + LEAK EVALUATION AND RESPONSE INDUSTRY STANDARD DEFICIENCIES + +Chevron Richmond Refinery Investigation Report January 2015 +ii + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + + + + + + +[This page intentionally left blank.] + + + +Chevron Richmond Refinery Investigation Report January 2015 +iii + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Table of Contents +1.0  +Executive Summary ......................................................................................................................... 1  +1.1  +Incident Summary .............................................................................................................................. 1  +1.2 +Chevron Interim Report ..................................................................................................................... 2 +1.3  +Chevron Regulatory Report .............................................................................................................. 4  +1.4  +Chevron Final Investigation Report ................................................................................................... 5  +1.4.1  +Technical Findings .................................................................................................................... 5  +1.4.2  +Organizational Findings ............................................................................................................ 7  +1.4.3  +Emergency Response Findings ............................................................................................... 12  +1.4.4  +Safety Culture Findings .......................................................................................................... 13  +1.4.5  +Industry Codes and Standards Findings .................................................................................. 14  +1.4.6  +Regulatory Findings ................................................................................................................ 14  +1.5  +Recommendations ............................................................................................................................ 15  +2.0  +Richmond Refinery Process Description ..................................................................................... 17  +2.1  +Chevron Background ....................................................................................................................... 17  +2.2  +Richmond Refinery .......................................................................................................................... 17  +2.3  +#4 Crude Unit ................................................................................................................................... 18  +2.4  +4-Sidecut Line .................................................................................................................................. 19  +3.0  +The Incident .................................................................................................................................... 22  +3.1  +Leak Discovery ................................................................................................................................ 22  +3.2  +Leak Response ................................................................................................................................. 25  +3.3  +Consequences ................................................................................................................................... 29  +4.0  +Technical Analysis ......................................................................................................................... 34  +4.1  +Sulfidation Corrosion ....................................................................................................................... 34  +4.2  +Sulfidation Corrosion Inspection Techniques .................................................................................. 41  +4.3  +Silicon Characterization Techniques ............................................................................................... 43  +4.4  +Inherently Safer Design ................................................................................................................... 45  +5.0  +Incident Analysis ............................................................................................................................ 48  +5.1  +Organizational Analysis ................................................................................................................... 50  +5.1.1  +Chevron Energy Technology Company .................................................................................. 53  +5.1.1.1  +ETC Sulfidation Failure Prevention Guidance ................................................................... 54  +5.1.1.2  +Chevron ETC Conclusions .................................................................................................. 56  +5.1.2  +Chevron Turnaround Management ......................................................................................... 58  + +Chevron Richmond Refinery Investigation Report January 2015 +iv + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.2.1  +2007 Crude Unit Turnaround .............................................................................................. 59  +5.1.2.1.1  +Recommendations Regarding 4-Sidecut Line for 2007 Turnaround .......................... 60  +5.1.2.2  +2011 Crude Unit Turnaround .............................................................................................. 64  +5.1.2.2.1  +Recommendations Regarding 4-Sidecut Line for 2011 Turnaround .......................... 65  +5.1.2.3  +Chevron Richmond Refinery Turnaround-Planning Conclusions ...................................... 66  +5.1.3  +Chevron Unit Reliability Improvement Process ..................................................................... 68  +5.1.4  +Chevron Fixed Equipment Reliability Business Improvement Network ................................ 71  +5.1.5  +Chevron Minimum Pipe Thickness Program .......................................................................... 74  +5.1.6  +Chevron Process Safety Indicators Program ........................................................................... 78  +5.1.7  +Stop Work Authority ............................................................................................................... 79  +5.1.8  +Chevron Organizational Conclusions ..................................................................................... 83  +5.2  +Industry Sulfidation Corrosion Guidance ........................................................................................ 85  +5.2.1 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil +Refineries ....................................................................................................................................... 85  +5.2.2 +API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining Industry ..... 89  +5.2.3  +API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of +Piping Systems ........................................................................................................................ 90  +5.2.4  +API RP 578: Material Verification Program for New and Existing Alloy Piping Systems .... 91  +5.2.5  +API RP 574: Inspection Practices for Piping System Components ........................................ 91  +5.3  +Chevron Emergency Response to Process Leaks ............................................................................. 92  +5.3.1  +Area Control and Hazardous Area Assessment ...................................................................... 92  +5.3.2  +Miscommunication regarding 4-sidecut properties ................................................................. 93  +5.3.3  +Leak Mitigation and Discovery Attempts Worsened Leak ..................................................... 94  +5.3.4  +Chevron’s New Leak Response Protocol ................................................................................ 97  +5.3.5  +Leak Response Conclusions .................................................................................................. 100  +5.4  +Industry Leak Response Guidance ................................................................................................. 100  +5.4.1  +API RP 574: Inspection Practices for Piping System Components ...................................... 100  +5.4.2  +API RP 2001: Fire Protection in Refineries ......................................................................... 101  +5.4.3  +API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of +Piping Systems ...................................................................................................................... 103  +5.4.4  +ASME PCC-2-2011: Repair of Pressure Equipment and Piping .......................................... 103  +5.4.5  +Industry Leak Response Guidance Conclusions ................................................................... 104  +5.5  +Chevron Richmond Refinery Safety Culture ................................................................................. 104  +5.5.1  +Normalization of Deviance ................................................................................................... 105  + +Chevron Richmond Refinery Investigation Report January 2015 +v + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.5.2  +Chevron Richmond Refinery Safety Culture Surveys .......................................................... 106  +5.5.2.1  +Stop Work Authority Safety Culture Survey Responses .................................................. 107  +5.5.2.2  +Mechanical Integrity Safety Culture Survey Responses ................................................... 108  +5.5.2.3  +Process Safety Analysis Safety Culture Survey Responses .............................................. 109  +5.5.2.4  +Regulator Overview of Safety Culture Survey Action Items ............................................ 110  +6.0  +Recommendations ........................................................................................................................ 112  +6.1  +American Petroleum Institute ........................................................................................................ 112  +6.2  +American Society of Mechanical Engineers .................................................................................. 115  +6.3  +Chevron USA ................................................................................................................................. 115  +6.4  +Board of Supervisors, Contra Costa County, California ................................................................ 116  +6.5  +Mayor and City Council, City of Richmond, California ................................................................ 116  +Appendix A—Chevron Leak Response Protocol Developed Post-incident ............................................. 117  +Appendix B—Contra Costa County Community Warning System .......................................................... 120  +Appendix C—Usage of Clamps at Chevron Richmond Refinery ............................................................ 121  + + + + + + + + + + + + + + +Chevron Richmond Refinery Investigation Report January 2015 +vi + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +List of Figures +Figure 1. To-scale schematic of 4-sidecut piping identifying key wall thickness values. ....................... 11  +Figure 2. Aerial view of the Chevron Richmond Refinery. ..................................................................... 18  +Figure 3. Schematic of C-1100 Crude Unit atmospheric column and upstream process equipment. ...... 20  +Figure 4. 4-sidecut line configuration and rupture location. .................................................................... 21  +Figure 5. Photo of rupture on 4-sidecut 52-inch component. .................................................................. 22  +Figure 6. CSB animation depicting operator identifying the leaking 4-sidecut pipe. .............................. 23  +Figure 7. Photo taken of the leaking 4-sidecut pipe on August 6, 2012, at the Chevron Richmond +Refinery ..................................................................................................................................... 24  +Figure 8. Timeline of events on August 6, 2012. ...................................................................................... 25  +Figure 9. Example leak repair clamp for piping. It is installed over the leak location to prevent process +fluid leakage to the atmosphere. ............................................................................................... 26  +Figure 10. CSB animation of contractors erecting scaffolding beneath the leak location. ........................ 27  +Figure 11. CSB animation of firefighters attempting to remove the 4-sidecut insulation, the resulting fire, +and fire extinguishing. ............................................................................................................... 28  +Figure 12. CSB animation of firefighters who dropped to their hands and knees to escape the vapor +cloud. ......................................................................................................................................... 29  +Figure 13. CSB animation of firefighter who was inside the fire engine when the light gas oil ignited. .. 31  +Figure 14. Photo of the burned remains of the fire truck that was consumed by the fire. .......................... 32  +Figure 15. Initial vapor cloud formation (white cloud) and subsequent ignition (black smoke) as seen from +a pier in San Francisco, California. ........................................................................................... 33  +Figure 16. Vapor cloud and ignition seen from Marin County. ................................................................. 34  +Figure 17. Graph of sulfidation corrosion rates with respect to silicon content in carbon steel ................. 36  +Figure 18. Locations of metallurgical samples taken from 8-inch 4-sidecut piping post-incident. ........... 38  +Figure 19. Locations of metallurgical samples taken from 12-inch 4-sidecut piping post-incident. ......... 39  +Figure 20. 4-sidecut piping sample (E-017-8) analyzed by Anamet Labs showing the relative thickness of +low silicon piping on the left and the high silicon piping on the right. ..................................... 40  +Figure 21. CML placement on 8-inch 4-sidecut piping. ............................................................................. 42  +Figure 22. Modified McConomy Curves from API RP 939-C. .................................................................. 45  +Figure 23. Hierarchy of controls ................................................................................................................. 46  +Figure 24. Acci-Map of August 6, 2012, Chevron Refinery Fire. .............................................................. 49  +Figure 25. Organizational decision-making schematic showing attempts to have carbon steel 4-sidecut +piping 100 percent component inspected or replaced with a higher chromium steel alloy ...... 50  +Figure 26. Key events at the Richmond refinery between 2002 and 2012. ............................................... 51  +Figure 27. Chevron Corporation refinery process safety programs. .......................................................... 52  +Figure 28. Chevron Energy Technology Company (ETC) organizational roles ........................................ 54  +Figure 29. Presentation slide of ETC training course that guided refinery staff to perform 100 percent +component inspection on high-temperature lines susceptible to sulfidation corrosion. ............ 55  +Figure 30. Presentation slide of ETC training course that guided refinery staff on ways to reduce risk +from sulfidation corrosion. ........................................................................................................ 56  +Figure 31. Work list criteria requirements specified in the Framing Document used during Chevron’s +2007 Crude Unit turnaround. .................................................................................................... 60  + +Chevron Richmond Refinery Investigation Report January 2015 +vii + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Figure 32. Sample Inspection Database report analyzed by unit inspectors when determining piping +remaining life and when making piping replacement recommendations .................................. 61  +Figure 33. Crude column schematic indicating the piping downstream of the 4-sidecut pumps that was +replaced during the 2007 Crude Unit turnaround ..................................................................... 63  +Figure 34. Work list criteria requirements specified in the Framing Document used during Chevron’s +2011 Crude Unit turnaround. .................................................................................................... 64  +Figure 35. Presentation slide showing corporate reliability leader’s findings that the Richmond Refinery +was not complying with the ETC Sulfidation Failure Prevention Initiative. ............................ 73  +Figure 36. To-scale schematic of 4-sidecut piping original wall thickness (0.322-inch), Minimum Alert +Thickness (0.13-inch), and Minimal Required Thickness (0.11-inch) using API RP 574 default +values. ....................................................................................................................................... 75  +Figure 37. Text from Chevron design engineer indicating structural minimum thickness (t(min)) +calculation results for small sections of suction piping upstream of the 4-sidecut pumps. ...... 76  +Figure 38. Photo from API RP 939-C of a low-silicon pup piece that ruptured at a BP refinery .............. 86  +Figure 39. Schematic from API RP 939-C of the piping shown in Figure 38. .......................................... 87  +Figure 40. Photo of undamaged (top) and burned during incident (bottom) pike pole used in early +attempts to remove 4-sidecut insulation. ................................................................................... 95  +Figure 41. Photo showing that tip of fire pole matches apparent puncture location in failed 52-inch +component of 4-sidecut piping. ................................................................................................. 96  +Figure 42. Ideal communication flow to Incident Commander during refinery process fluid leak incident. + .................................................................................................................................................................... 98  +Figure 43. Chevron’s new Leak Response Protocol, developed post-incident .......................................... 99  + + + +Chevron Richmond Refinery Investigation Report January 2015 +viii + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +List of Tables +Table 1. Silicon characterization technique. .............................................................................................. 43  +Table 2. Total number of employees surveyed and job functions of respondents in 2008 and 2010 +Chevron Richmond Refinery staff safety culture surveys. ......................................................... 106  +Table 3. Chevron Richmond Refinery 2008 Safety Culture Survey responses to question of “Stopping +Unsafe Work” by operators and mechanics ............................................................................... 107  +Table 4. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to question of +“Stopping Unsafe Work” by managers and engineers ............................................................... 108  +Table 5. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to question of +“How do people feel about stopping unsafe work?” by operators and mechanics ..................... 108  +Table 6. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to the question: +“How do we take care of equipment?” ....................................................................................... 109  +Table 7. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to the question +“Process Safety Issues Are…” .................................................................................................. 109  + + + + +Chevron Richmond Refinery Investigation Report January 2015 +ix + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Acronyms and Abbreviations +⁰C +degrees Celsius +⁰F +degrees Fahrenheit +ABU +Area Business Unit +A/C +Additional Considerations +AcciMap +Accident Map +API +American Petroleum Institute +API 570 +Piping Inspection Code: In-Service Inspection, Rating, Repair, and +Alteration of Piping Systems +API RP 2001 +Fire Protection in Refineries +API RP 571 +Damage Mechanisms Affecting Fixed Equipment in the Refining +Industry +API RP 574 +Inspection Practices for Piping System Components +API RP 578 +Material Verification Program for New and Existing Alloy Piping +Systems +API RP 754 +Process Safety Performance Indicators for the Refining and +Petrochemical Industries +API RP 939-C +Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in +Oil Refineries +API RP +API Recommended Practice +ASME +American Society of Mechanical Engineers +ASME PCC-2-2011 +Repair of Pressure Equipment and Piping +ASTM +American Society for Testing and Materials +bpd +barrels per day +C/A +Corrective Actions +CML +Condition Monitoring Location +Cr +Chromium +CS +Carbon Steel +CSB +U.S. Chemical Safety and Hazard Investigation Board +CWS +Community Warning System +DRB +Decision Review Board +ETC +Chevron Energy Technology Company +FER BIN +Fixed Equipment Reliability Business Improvement Network +IMPACT +Initiative for Managing Pacesetter Turnarounds +ISO +Industrial Safety Ordinance +Mo +Molybdenum + +Chevron Richmond Refinery Investigation Report January 2015 +x + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +MOC +Management of Change +NFPA +National Fire Protection Association +NFPA 471 +Recommended Practice for Responding to Hazardous Materials +Incidents +OERI +Operational Excellence and Reliability Intelligence +PHA +Process Hazard Analysis +PMI +Positive Material Identification +PPE +Personal Protective Equipment +psig +pounds per square inch gauge +PSM +Process Safety Management +RAGAGEP +Recognized and Generally Accepted Good Engineering Practices +RISO +Richmond Industrial Safety Ordinance +RLOP +Richmond Lube Oil Project +RT +Radiographic Testing +S/D +Shutdown +Si +Silicon +SIP +shelter-in-place +SIS +Safety Instrumented Systems +SME +Subject Matter Expert +STL +Shift Team Leader +SWA +Stop Work Authority +T-min +minimum thickness +TML +Thickness Measurement Location or Thickness Monitoring Location +TOP +Triangle of Prevention +URB +Unit Reliability Brief +URIP +Unit Reliability Improvement Process +UT +Ultrasonic Testing + + + + + +Chevron Richmond Refinery Investigation Report January 2015 +1 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +1.0 Executive Summary +1.1 Incident Summary +On August 6, 2012, the Chevron U.S.A. Inc. Refinery in Richmond, California (“the Chevron Richmond +Refinery”) experienced a catastrophic pipe rupture in the #4 Crude Unit. The incident occurred from +piping referred to as the “4-sidecut” stream, one of several process streams exiting the refinery’s C-1100 +Crude Unit Atmospheric Column.1 The pipe rupture occurred on a 52-inch long component2 of the 4- +sidecut 8-inch line (the 52-inch component). At the time of the incident, light gas oil3 was flowing +through the 8-inch line at a rate of approximately 10,800 barrels per day (bpd).4 +The ruptured pipe released flammable, high temperature light gas oil, which then partially vaporized into +a large, opaque vapor cloud that engulfed 19 Chevron U.S.A. Inc. (Chevron) employees.5 At 6:33 p.m., +approximately two minutes following the release, the released process fluid ignited.6 Eighteen of the +employees safely escaped from the vapor cloud just before ignition; one employee, a Chevron refinery +firefighter, was inside a fire engine that was caught within the fireball when the process fluid ignited. +Because he was wearing full-body fire-fighting protective equipment, he was able to make his way +through the flames to safety. Six Chevron employees suffered minor injuries during the incident and +subsequent emergency response efforts. +The release, ignition, and subsequent burning of the hydrocarbon process fluid resulted in a large plume +of vapor, particulates, and black smoke, which traveled across the surrounding area. This chain of events +resulted in a Community Warning System (CWS) Level 3 alert,7 and a shelter-in-place8 advisory (SIP) +was issued at 6:38 p.m.9 for the cities of Richmond, San Pablo, and North Richmond. It was lifted later + +1 The atmospheric column separates crude oil feed into different streams through distillation. These streams are +further processed in other units in the refinery. The location of the 4-sidecut, light gas oil stream was shown in +Figure 4 (page 12) of the Interim Investigation Report Chevron Richmond Refinery Fire. See +http://www.csb.gov/assets/1/19/Chevron_Interim_Report_Final_2013-04-17.pdf (accessed January 21, 2015). +2 “Component” refers to a portion of piping between welds. It includes straight run piping and pipe fittings. +3 Light gas oil is a component of crude oil with a boiling point range between 401°F and 653°F. +4 This quantity is the equivalent of 315 gallons per minute (gpm). A barrel is equivalent to 42 gallons. +5 This number is based on statements made to the CSB by each of the 19 employees caught in the vapor cloud. +6 Surveillance footage was provided by Chevron. Chevron clarified to the CSB that the video time stamp is +approximately 5 minutes out of sync. The video is available at +http://www.csb.gov/videoroom/detail.aspx?VID=69 (accessed February 8, 2013). +7 A Community Warning System Level 3 alert indicates that a facility within Contra Costa County has had a release +that has offsite impact and is categorized by any of the following conditions: +1. Off-site impact that may cause eye, skin, nose and/or respiratory irritation to the general population. +2. Fire, explosion, heat, or smoke with an off-site impact. Example: On a process unit/storage tank where +mutual aid is requested to mitigate the event and the fire will last longer than 15 minutes. +3. Hazardous material or fire incident where the Incident Commander or unified command, through +consultation with the Contra Costa Health Services Hazardous Material Incident Response Team, requests +that sirens should be sounded. +See http://cchealth.org/hazmat/pdf/incident_notification_policy.pdf (accessed April 9, 2013). +8 Contra Costa County considers a shelter-in-place to include going inside a home or nearest building, closing doors +and windows, and turning off heating, ventilation, and air conditioning. See http://cchealth.org/emergencies/shelter- +in-place.php (accessed February 6, 2013). +9 Chevron U.S.A. Inc. "30 Day Follow-Up Notification Report," September 5, 2012. + +Chevron Richmond Refinery Investigation Report January 2015 +2 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +that night, at 11:12 p.m., after the fire was fully under control. In the weeks following the incident, +approximately 15,000 people from the surrounding communities sought medical treatment at nearby +medical facilities for ailments including breathing problems, chest pain, shortness of breath, sore throat, +and headaches. Approximately 20 of these people were admitted to local hospitals as inpatients for +treatment.10 +1.2 Chevron Interim Report +The U.S. Chemical Safety Board (CSB) released its first report on the Chevron incident in April 2013 +(“the Interim Report”), which highlighted technical findings and safety system deficiencies. The report +issued recommendations to Chevron; the city of Richmond, California; Contra Costa County, California; +the State of California; the California Air Quality Management Divisions; the California Environmental +Protection Agency; and the U.S. Environmental Protection Agency, summarized below. As of January +2015, these groups have made progress in implementing the recommendations, summarized below, to +improve the regulatory requirements for petroleum refineries in California. +________________________________________________________________ +Chevron U.S.A (Urgent) +At all Chevron U.S. refineries and as part of the Process Hazard Analysis cycle, engage a diverse team of +qualified personnel to perform a documented damage mechanism hazard review that identifies potential +process damage mechanisms and consequences of failure and ensures safeguards are in place to control +hazards presented by those damage mechanisms. Include in this review applicable industry best practices, +Chevron Energy Technology Company findings and recommendations, and inherently safer systems to +the greatest extent feasible. Report leading and lagging process safety indicators at all California Chevron +U.S.A. refineries to the applicable regulatory agencies. +________________________________________________________________ +Mayor and City Council, City of Richmond, California; Board of Supervisors, +Contra Costa County, California; California State Legislature, Governor of +California +Require that Process Hazard Analyses include documentation of the recognized methodologies, rationale +and conclusions used to claim that safeguards intended to control hazards will be effective. Require the +documented use of inherently safer systems analysis and the hierarchy of controls to the greatest extent +feasible in establishing safeguards for identified process hazards. The goal shall be to drive the risk of +major accidents to As Low As Reasonably Practicable (ALARP). +________________________________________________________________ +California State Legislature, Governor of California +Require California petroleum refineries to engage a diverse team of qualified personnel to perform a +documented damage mechanism hazard review as part of the Process Hazard Analysis cycle that +identifies potential process damage mechanisms and consequences of failure and ensures safeguards are + +10 Based on information provided to the CSB by local hospitals. + +Chevron Richmond Refinery Investigation Report January 2015 +3 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +in place to control hazards presented by those damage mechanisms. Require the analysis and +incorporation of applicable industry best practices and inherently safety systems to the greatest extent +feasible into this review. +For all California oil refineries, identify and require the reporting of leading and lagging process safety +indicators, such as the action item completion status of recommendations from damage mechanism hazard +reviews, to state and local regulatory agencies that have chemical release prevention authority. +Establish a multi-agency process safety regulatory program for all California oil refineries to improve the +public accountability, transparency, and performance of chemical accident prevention and mechanical +integrity programs. +________________________________________________________________ +The U.S. Environmental Protection Agency +Jointly plan and conduct inspections with Cal/OSHA [California Division of Occupational Safety and +Health], California EPA and other state and local regulatory agencies with chemical accident prevention +responsibilities to monitor the effective implementation of the damage mechanism hazard review process. +________________________________________________________________ +The Board of Supervisors, Contra Costa County, California; The Mayor and City +Council, City of Richmond, California; The California Air Quality Management +Divisions; The U.S. Environmental Protection Agency; and The California +Environmental Protection Agency +Participate in the joint regulatory program to monitor the effective implementation of the damage +mechanism hazard review process with Cal/OSHA and the U.S. Environmental Protection Agency. +________________________________________________________________ + + + +Chevron Richmond Refinery Investigation Report January 2015 +4 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +1.3 Chevron Regulatory Report +The CSB released its second finalized investigation report on the August 6, 2012, Chevron incident in +October 2014 (the “Chevron Regulatory Report”). The report examines California process safety +regulatory gaps and enforcement issues which contributed to the August 6th incident. The Chevron +Regulatory Report also evaluates whether a rigorous goal-setting regulatory approach requiring +employers to demonstrate that they have driven major accident risk to as low as reasonably practicable +(ALARP) could be a more effective, prevention-focused regulatory system to reduce major accidents in +California petroleum refineries. The Chevron Regulatory Report made the following recommendations: +________________________________________________________________ +California State Legislature, Governor of California +Enhance and restructure California’s process safety management (PSM) regulations for petroleum +refineries by including the goal-setting attributes identified in this report for petroleum refineries in the +state of California. +________________________________________________________________ +Mayor and City Council, City of Richmond, California +Implement or cause to be implemented a compensation system to ensure regulator capability in process +safety oversight and policy development in Richmond, California. +________________________________________ +Board of Supervisors Contra Costa County, California +Implement a compensation system to ensure regulator capability in process safety oversight and policy +development in Contra Costa County, California. +________________________________________ + + + +Chevron Richmond Refinery Investigation Report January 2015 +5 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +1.4 Chevron Final Investigation Report +The following Chevron Final Investigation Report addresses additional investigation findings not covered +in the two previous reports, including analysis of (1) the Chevron organization, emergency response, and +safety culture; (2) industry leak response standards; and (3) mechanical integrity industry standards. This +report supplements the information already published in the Interim Report and Regulatory Report. This +is the third and final report the CSB is publishing on this incident. +1.4.1 Technical Findings +This report highlights the following technical findings. (An in-depth discussion appears in the Chevron +Interim Report.) +1. The rupture of the 4-sidecut piping resulted from the 52-inch component being extremely thin due +to a damage mechanism11 known as sulfidation corrosion. Sulfidation corrosion, also known as +sulfidic corrosion,12 is a damage mechanism that causes thinning in iron-containing materials, +such as steel, due to the reaction between sulfur compounds and iron at temperatures ranging +from 450°F to 1,000°F.13 This damage mechanism causes pipe walls to gradually thin over time. +(See Section 4.1.) +2. Sulfidation corrosion is common in crude oil distillation,14 where naturally occurring sulfur and +sulfur compounds found in crude oil feed, such as hydrogen sulfide,15 react with steel piping and +equipment. Process variables that affect corrosion rates include the total sulfur content of the oil, +the sulfur species present, the flow conditions, and the system temperature. Virtually all crude oil +feeds contain sulfur compounds; as a result, sulfidation corrosion is a damage mechanism present +at every refinery that processes crude oil. Sulfidation corrosion can cause thinning to the point of +pipe failure when not properly monitored and controlled. (See Section 4.1.) +3. The Chevron Richmond Refinery 4-sidecut piping was constructed of carbon steel, which +corrodes at a much faster rate from sulfidation than other typical alternative materials of +construction, such as higher chromium-containing steels. In addition to its inherently faster rate +of sulfidation corrosion when compared with higher chromium steels, carbon steel also +experiences significant variation in corrosion rates due to possible variances in silicon content, a +component used in the steel manufacturing process. Carbon steel piping containing silicon + +11 Piping damage mechanisms are any type of deterioration encountered in the refining and chemical process +industry that can result in flaws/defects that can affect the integrity of piping (e.g., corrosion, cracking, erosion, +dents, and other mechanical, physical or chemical impacts). See API 570: Piping Inspection Code: In-Service +Inspection, Rating, Repair, and Alteration of Piping Systems. 3rd ed., Section 3.1.1.5, November 2009. +12 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 3.1.6, May 2009. +13 Ibid., Section 1. +14 Distillation separates mixtures into broad categories of its components by heating the mixture in a distillation +column where different products boil off and are recovered at different temperatures. See +http://www.eia.gov/todayinenergy/detail.cfm?id=6970 (accessed April 4, 2013). +15 Hydrogen sulfide is the most aggressive sulfur compound that causes sulfidation corrosion. + +Chevron Richmond Refinery Investigation Report January 2015 +6 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +content less than 0.10 weight percent can corrode at accelerated rates,16 up to 16 times faster than +carbon steel piping containing higher percentages of silicon. (See Section 4.1.) +4. Carbon steel piping components in refineries throughout the U.S. are susceptible to highly +variable sulfidation corrosion rates. Carbon steel piping is manufactured to meet certain +specifications, including American Society for Testing and Materials (ASTM) A53B,17 ASTM +A106,18 and American Petroleum Institute (API) 5L.19 ASTM A53B and API 5L do not contain +minimum silicon content requirements for carbon steel piping,20 while ASTM A106 requires the +piping to be manufactured with a minimum silicon content of 0.10 weight percent. As a result, +manufacturers have used different levels of silicon in the carbon steel pipe manufacturing +process. Thus, sulfidation corrosion rates could vary depending on the manufacturing +specification for silicon content in the carbon steel installed in refinery processes. In the mid- +1980s, pipe manufacturers began to simultaneously comply with all three specifications, so most +carbon steel piping purchased since then for refinery operations likely has a minimum of 0.10 +weight percent silicon content. However, over 95 percent of the 144 refineries in the U.S., +including the Chevron Richmond Refinery, were built before 1985. Therefore, the original +carbon steel piping components in these refineries likely contain varying percentages of silicon, +so they may experience highly variable sulfidation corrosion rates. (See Section 4.1.) +5. The Chevron Richmond Refinery 4-sidecut piping circuit containing the 52-inch component that +failed was constructed of ASTM A53B carbon steel, which had no minimum specification for +silicon content. Post-incident testing of samples of the 4-sidecut piping from the Chevron +Richmond Refinery identified silicon content ranging from 0.01 weight percent to 0.2 weight +percent. Of 12 samples taken from the 8-inch and the adjacent 12-inch 4-sidecut line, six had a +silicon concentration of less than 0.10 weight percent. The 52-inch pipe component that ruptured +on the day of the incident had a silicon content of only 0.01 weight percent. The elbow +component directly upstream of the 52-inch component that failed had a silicon concentration of +0.16 weight percent, showing considerably less thinning. (See Section 4.1.) +6. Determining silicon content in existing carbon steel piping and equipment in the field is a +difficult undertaking. Every component must be inspected to properly characterize the silicon +content in each component of a piping circuit. This is known as 100 percent component +inspection. Two techniques are used to inspect a component in an existing carbon steel piping +circuit with unknown chemical composition for low silicon content and resulting variable +corrosion rates: (1) performing laboratory-based chemical analysis of the carbon steel (a +“destructive test,” meaning it requires removal of a sample of the steel), or (2) performing pipe + +16 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 6.2.3.2, May 2009. +17 ASTM Standard A53/A53M-12: Standard Specification for Pipe, Steel, Black and Hot-Dipped, Zinc-Coated, +Welded and Seamless. 2012. +18 ASTM Standard A106/A106M–1:Standard Specification for Seamless Carbon Steel Pipe for High-Temperature +Service. 2011. +19 API Specification 5L: Specification for Line Pipe. 45th ed., December 2012. +20. ASTM Standard A53/A53M-12: Standard Specification for Pipe, Steel, Black and Hot-Dipped, Zinc-Coated, +Welded and Seamless. 2012. + +Chevron Richmond Refinery Investigation Report January 2015 +7 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +wall thickness measurements. Measuring pipe wall thickness of every component is useful as a +means to ascertain silicon content only if the piping circuit has been exposed to sulfidation +corrosion for a long enough time period so that variances in corrosion rate caused by differences +in silicon content may be detected. Steel alloys containing at least 9 weight percent chromium +are more resistant to sulfidation corrosion than carbon steel and do not present the hazard of +extreme variations in corrosion rates in components within the same piping circuit due to slight +differences in chemical composition.21 Thus, alloys with higher chromium content are an +inherently safer choice in high-temperature sulfidation corrosion environments.22 (See Section +4.2 and Section 4.4.) +7. Effectively implementing inherently safer design provides an opportunity for preventing major +chemical incidents. The August 6, 2012, incident at Chevron and other incidents23 throughout the +refining industry highlight the difficulty in preventing failure caused by sulfidation corrosion in +low-silicon carbon steel piping solely through inspection, a procedural safeguard that is low on +the hierarchy of controls. Using inherently safer design concepts to eliminate the hazard of +variation in corrosion rate in carbon steel piping due to hard-to-determine silicon content will +prevent future similar failures in refineries. (See Section 4.4.) +1.4.2 Organizational Findings +8. Chevron did not effectively implement internal recommendations to help prevent pipe failures +due to sulfidation corrosion. In the 10 years prior to the incident, a small number of Chevron +personnel with knowledge and understanding of sulfidation corrosion recommended on several +occasions either a one-time inspection of every component within the 4-sidecut piping circuit— +known as 100 percent component inspection—or an upgrade of the material of construction of the +4-sidecut piping. The recommendations were not implemented effectively, and the 52-inch +component remained in service until it failed on August 6, 2012. (See Section 5.1.) +9. Chevron failed to perform internally recommended 100 percent component inspections. An +independent corporate entity within Chevron, the Chevron Energy Technology Company (ETC), +provides technology solutions and technical expertise for Chevron operations worldwide. +Chevron ETC metallurgists released within Chevron a formal report dated September 30, 2009 +(nearly 3 years before the incident), titled Updated Inspection Strategies for Preventing +Sulfidation Corrosion Failures in Chevron Refineries (ETC Sulfidation Failure Prevention +Initiative). The initiative specifically recommends that inspectors perform 100 percent +component inspection on high-temperature carbon steel piping susceptible to sulfidation +corrosion. The initiative defines a priority ranking system to help focus the inspection + +21 The protective scale, FeCr2S4, begins to be the dominant scale formed in steels containing a chromium content of +five weight percent. The 5Cr steel alloy can be manufactured to contain anywhere from 4 percent to 6 percent +chromium. Thus, “the sulfidation corrosion rate can vary dramatically in 5Cr steels even in the same operating +environment.” See Niccolls, E. H., J. M. Stankiewicz, J. E. McLaughlin, and K. Yamamoto. "High Temperature +Sulfidation Corrosion in Refining." 17th International Corrosion Congress. Las Vegas: NACE International, +2008. +22 Steels with higher chromium content are inherently safer than carbon steel with respect to sulfidation corrosion. +However, analysis is still required to ensure that the best material of construction is selected. +23 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., May +2009. + +Chevron Richmond Refinery Investigation Report January 2015 +8 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +implementation efforts. The process conditions of the 4-sidecut stream placed it in the highest +priority for 100 percent component inspection. However, the 4-sidecut piping was not 100 +percent component inspected prior to the August 2012 incident. (See Section 5.1.1.) +10. The CSB found that the Richmond refinery’s turnaround planning group rejected the +recommendations to 100 percent component inspect or replace the portion of the 4-sidecut piping +that ultimately failed24. The turnaround work scope and approval process is guided by +predetermined criteria in what Chevron calls a “Framing Document.” Turnaround work requests +are approved or denied by the turnaround planning group based on the document criteria. The +Framing Document sets the criteria for work items that can be automatically accepted as +turnaround work items during the planned turnaround. Less urgent items and those that may be +performed on the run (while the unit is operating) or during the next turnaround are not included +by default in the turnaround work scope. Inspection data for the 4-sidecut piping, where +measurements were historically taken on high-silicon fittings,25 indicated the 4-sidecut piping +could safely operate through 2016. Therefore, recommendations to replace the 8-inch 4-sidecut +piping during the 2007 and 2011 turnarounds were denied in accordance with the Framing +Document criteria. The Sulfidation Failure Prevention Initiative developed by the ETC +metallurgist experts was not considered a valid mandate for justifying turnaround work which +otherwise fell outside the acceptance criteria of the Framing Document. (See Section 5.1.2.) +11. A Crude Unit metallurgical analysis recommendation to perform 100 percent volumetric +inspection26 of the 4-sidecut line submitted for the 2007 turnaround was approved by the Crude +Unit’s Area Business Unit (ABU) Manager. Chevron installed experimental “Guided Wave +bracelets”27 which were designed to continuously perform 100 percent volumetric inspection. +However, the guided wave bracelets were only installed on a small portion of the 4-sidecut line +which did not include the 52-inch component that ultimately failed. In addition, when the Guided +Wave bracelets were found to be unreliable, manual 100 percent component inspection was not +conducted in its place. (See Section 5.1.2.1.1.) +12. If a submitted turnaround work item recommendation was not accepted under the Chevron +Richmond Refinery turnaround Framing Document—for example, an “Industry Best Practice” +that Chevron may not interpret as being supported by hard data needed to justify the work, or a +profit-improvement project—there was an informal appeal process. A case for approval for the +work had to be made to the ABU Manager for the unit where the turnaround was to occur. +However, this approach was never attempted by Chevron inspection or metallurgical staff who +submitted the recommendations to replace the 4-sidecut piping. In addition, no high-level +manager was assigned responsibility to ensure that the ETC Sulfidation Failure Prevention +Initiative or other ETC sulfidation recommendations were included in the turnaround scope, so all +responsibility to implement the ETC recommendations was placed on lower-level employees, +who did not have decision-making or funding authority. (See Section 5.1.2.3.) + +24 Other portions of the 4-sidecut were replaced in 2007 and 2011. +25 A 2011 effort added an additional 12 CML locations on straight-run piping components. A CML was not placed, +however, on the low-silicon 52-inch component that failed on August 6, 2012. +26 Common volumetric inspection techniques include ultrasonic and radiography testing. +27 Guided Wave bracelets are continuous monitoring probes that can, if proven reliable, remove the need for manual +inspection of piping. + +Chevron Richmond Refinery Investigation Report January 2015 +9 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +13. Chevron relies on its Unit Reliability Improvement Process (URIP) and its associated programs, +including Unit Reliability Briefs (URBs) and Reliability Steering Committee meetings, to steward +mechanical reliability at its various refineries. Employees meeting within the various URIP +programs discussed the ETC Sulfidation Failure Prevention Initiative. However, the +metallurgical and inspection staff assigned by the URB and Reliability Steering Committee to +implement the ETC Sulfidation Failure Prevention Initiative routed all recommendations through +the turnaround planning process. The turnaround planning group denied these recommendations +because they did not meet turnaround Framing Document requirements. In addition, no high- +level refinery managers who attended URBs and Reliability Steering Committee meetings took or +were assigned responsibility for the ETC Sulfidation Failure Prevention Initiative and ETC +sulfidation mitigation recommendations to assure their effective implementation within the +Richmond refinery. (See Section 5.1.3.) +14. Chevron’s Fixed Equipment Reliability Business Improvement Network (FER BIN) program did +not effectively gain the necessary commitment from refinery management to implement the ETC +Sulfidation Failure Prevention Initiative or other ETC recommendations to upgrade susceptible +carbon steel piping to inherently safer, higher chromium steel. The FER BIN is intended to be a +“best practice” network across all Chevron refineries for bringing up-to-date changes in industry +standards and best practices into the organization. The FER BIN is headed by a technically +qualified subject-matter expert, the FER BIN Leader. The individual who was in the FER BIN +Leader role when the ETC Sulfidation Failure Prevention Initiative was issued retired in +September 2010, before the initiative was fully developed and implemented at the Richmond +refinery. A replacement for the FER BIN Leader was not assigned until four months after the +previous FER BIN Leader’s retirement—in January 2011. The onboarding process for the new +FER BIN Leader’s roles and responsibilities took additional time because of the hiring delay. +When the new FER BIN Leader visited the Chevron Richmond Refinery in early 2012, he +identified that the refinery was not successfully implementing the ETC Sulfidation Failure +Prevention Initiative. However, he met only with inspection and reliability personnel—not with +refinery management who had the authority to implement his recommendations to adhere to the +ETC Sulfidation Failure Prevention Initiative guidance. (See Section 5.1.4.) +15. Sulfidation corrosion causes pipe walls to thin, which eventually leads to the need to replace the +thinned piping. Chevron determines the date for replacing thinned piping by using a piping +“Minimum Alert Thickness” and a piping “Minimum Required Thickness” (Figure 1).28 When +piping reaches its Minimum Alert Thickness, an engineering evaluation is triggered to calculate +the piping’s Minimum Required Thickness, or the lowest thickness that can withstand the +pressure and structural stresses of the piping circuit, to determine whether the piping must be +replaced immediately or if replacement can be safely delayed. This evaluation may result in the +lowering of the Minimum Alert Thickness to 0.1-inch. Evaluation of the inspection thickness +data obtained on the 4-sidecut piping during the 2011 turnaround indicated that the 4-sidecut +piping would thin below its 0.14-inch Minimum Alert Thickness before the next turnaround +scheduled for 2016. A minimum structural thickness value of 0.036-inch had been calculated for + +28 Chevron’s term for “Minimum Alert Thickness” is “Flag Thickness,” and its term for “Minimum Required +Thickness” is “T-min.” + +Chevron Richmond Refinery Investigation Report January 2015 +10 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +a small piping component within the 4-sidecut piping earlier during the turnaround. This 0.036- +inch value was applied to the full length of the 8-inch 4-sidecut piping circuit. This calculation +was used as a technical justification to reduce the 8-inch 4-sidecut Minimum Alert Thickness to +0.1-inch, and the piping wall thickness was predicted to stay above this Minimum Alert +Thickness until after the next turnaround. The 4-sidecut line was therefore allowed to continue +operating with replacement scheduled for the next turnaround in 2016. API RP 574: Inspection +Practices for Piping System Components provides users with a default minimum structural +thickness of 0.11-inch for piping with a diameter of 8-inches—which can be used as the +Minimum Required Thickness for piping in lieu of detailed engineering calculations.29 Chevron +performed a detailed calculation to determine the 4-sidecut Minimum Required Thickness and the +API RP 574 default minimum structural thickness was not used. However, had Chevron used the +API RP 574 default minimum structural thickness value of 0.11-inch as the 4-sidecut Minimum +Required Thickness, the remaining life of the piping circuit would have been predicted to be less +than ten years, and a turnaround planning group discussion should have been triggered to discuss +replacement options for the 8-inch 4-sidecut piping. Such a discussion could have resulted in the +decision to replace the 8-inch 4-sidecut piping during the 2011 turnaround, and the August 6, +2012, pipe rupture could have been prevented. In addition, Chevron does not require a formal +multi-person review process to be performed to verify that available inspection data is reliable +considering the relevant piping circuit damage mechanisms prior to changing the minimum +thickness values used to project the remaining life of a piping circuit. (See Section 5.1.5.) + + +29 This minimum thickness is specified for piping between 6 and 18 inches in diameter that operates at temperatures +under 400 ⁰F. The 4-sidecut piping operated at a higher temperature, likely requiring a greater minimum +thickness. + +Chevron Richmond Refinery Investigation Report January 2015 +11 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 1. To-scale schematic of 4-sidecut piping identifying key wall thickness values. These include the +original wall thickness (0.322-inch), “Minimum Alert Thickness” (0.13-inch), and “Minimum Required +Thickness” (0.11-inch) using API RP 574 default values. +16. Inspection data obtained during the 2011 Crude Unit turnaround identified that components of the +12-inch portion of the 4-sidecut piping had become so thin due to sulfidation corrosion that much +of it had to be replaced during the turnaround. Even though the 12-inch 4-sidecut piping was +manufactured from the same specification of carbon steel, contained the same process fluid, and +experienced similar process conditions30 as the 8-inch 4-sidecut piping, Chevron turnaround +management did not consider that components in the 8-inch 4-sidecut piping could also be too +thin to allow the piping to continue in operation. Chevron personnel involved with the decision +to replace portions of the 12-inch 4-sidecut piping concluded, based upon available inspection +data, that all of the 8-inch 4-sidecut piping that had not been inspected, including the 52-inch +component that ultimately failed, was acceptable for continued operation. (See Section 5.1.2.2.1.) +17. Chevron does not effectively use its online dashboard, Operational Excellence and Reliability +Intelligence (OERI), which tracks 26 different process safety indicators, to track the +implementation status of ETC recommendations and new industry guidance. OERI visually +displays the status of many different process safety indicators. Management reviews these + +30 The CSB notes that the process conditions of the 8-inch and 12-inch 4-sidecut piping were not identical. + +Chevron Richmond Refinery Investigation Report January 2015 +12 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +metrics weekly and schedules monthly meetings to discuss the items that need attention. The +Chevron Richmond Refinery leadership team is held accountable for the status of these metrics. +The Refinery manager and the president of global manufacturing meet regularly with members of +the Chevron Richmond Refinery leadership team to discuss status of the metrics they oversee, +and they incorporate into all leadership team members’ performance reviews their effectiveness +in managing these metrics. Chevron does not track in OERI the implementation status of ETC +recommendations or new industry guidance. Such an indicator could have ensured that the status +of the ETC Sulfidation Failure Prevention Initiative at the Chevron Richmond Refinery received +greater management attention. (See Section 5.1.6.) +1.4.3 Emergency Response Findings +18. Chevron did not effectively identify in the Incident Command structure the damage mechanisms +that could have caused the 4-sidecut piping leak on the day of the incident. The OSHA +Hazardous Waste Operations and Emergency Response (HAZWOPER) standard states that the +Incident Commander “shall identify, to the extent possible, all hazardous substances or conditions +present”31 in an emergency response situation. However, the appropriate technical expertise +necessary to identify the potential for low-silicon, more rapidly corroding piping components in +the 4-sidecut piping was not effectively consulted in the Incident Command structure on August +6, 2012. This lack of knowledge of all potential causes of the 4-sidecut piping leak led the +Incident Commander to direct emergency responders to take actions that may have ultimately +exacerbated the leak and put many Chevron personnel in harm’s way. It also led the Incident +Commander to limit the “hot zone” to a small area that did not consider the possibility of pipe +rupture. When the 4-sidecut piping ruptured, personnel and firefighting equipment positioned in +the “cold zone” were engulfed in the large vapor cloud. (See Section 5.3.) +19. Process conditions were not effectively identified and communicated in the Incident Command +structure on the day of the incident. The 4-sidecut leak response and mitigation strategy +developed following an assessment of the leaking pipe by Chevron Fire Department leaders and +other key Chevron operations personnel involved stripping insulation from the hot piping to +identify the leak location. The CSB found that several Chevron Fire Department personnel +responding to the leaking 4-sidecut pipe were not properly informed of the operating temperature +of the line. CSB interviews identified that some firefighters believed the line was operating at a +temperature of about 130°F rather than the actual temperature approaching 640°F. CSB +interviews indicate that, had the responders been aware of the actual operating temperature, some +likely would have raised concerns to their supervisors about the safety of performing aggressive +leak response actions on a hot pipe. (See Section 5.3.2.) +20. Chevron did not recognize or accommodate the shortcomings of reliance on Stop Work Authority +in averting major process hazards. The CSB learned that some personnel participating in the +insulation removal process while the 4-sidecut line was leaking were uncomfortable with the +safety of this activity because of potential exposure to the flammable process fluid. Some +individuals even recommended that the Crude Unit be shut down, but they left the final decision + +31 29 CFR §1910.120(q)(3)(ii) (2012). + +Chevron Richmond Refinery Investigation Report January 2015 +13 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +to the management personnel present. No one formally invoked their Stop Work Authority.32 In +addition, Chevron safety culture surveys indicate that between 2008 and 2010, personnel had +become less willing to use their Stop Work Authority. Regardless of how a Stop Work program +is portrayed, there are a number of reasons why such a program may fail related to the ‘human +factors’ issue of decision-making; these reasons include belief that the Stop Work decision should +be made by someone else higher in the organizational hierarchy, reluctance to speak up and delay +work progress, and fear of reprisal for stopping the job.33 (See Section 5.1.7 and Section 5.5.2.1.) +21. On the day of the incident, Chevron had no leak response guidance or formal protocol for +operations personnel, refinery management, emergency responders, or the Incident Commander +to refer to when determining how to handle a process leak. Without a protocol, Chevron had no +formal system to ensure the right people were gathering all important information before deciding +on leak mitigation strategies. Such an evaluation could have led to the conclusion that the cause +of the leak was general thinning due to sulfidation corrosion, and clamping the pipe—a mitigation +strategy being considered—was not a viable solution because the pipe likely did not have the +structural integrity to support a clamp. This realization likely would have resulted in deciding to +immediately shut down the unit. Following this incident, Chevron improved its internal policies +by developing and implementing a leak response protocol for determining how to assess and +mitigate leaks within the refinery.34 The new leak response protocol would require unit shutdown +if a similar leak were to occur in a Chevron refinery. (See Section 5.3.4.) +1.4.4 Safety Culture Findings +22. The CSB identified several contributing causes of the August 6, 2012, incident relating to the +Chevron Richmond Refinery’s safety culture: +a. Decision making that encourages continued operation of a unit despite hazardous leaks. +Examples include another leak incident in the Chevron Richmond Refinery in 2010, +which was allowed to continue in operation, releasing high-temperature, flammable +process fluid in an active unit, as well as continued efforts on August 6, 2012, to perform +on-stream mitigation attempts despite high-temperature hydrocarbon vapor release and +the occurrence of a flash fire; +b. Reluctance among employees to use their Stop Work Authority. Recent safety culture +surveys performed at the refinery indicate that employees had become less willing to use +their Stop Work Authority between 2008 and 2010; and + +32 Chevron defines “Stop Work Authority” as the “… responsibility and authority of any individual to stop work +when an unsafe condition or act could result in an undesirable terms.” See +http://upstream.chevron.com/contractorgom/forms_policies/stop_work_authority.aspx (accessed November 5, +2014). +33 A 2010 study by The RAD Group of 2,600 workers (primarily oil and gas service employees) found that the +surveyed employees directly intervene in only 39% of the unsafe acts that they observe on the job. The study +concluded people did not stop unsafe work were primarily because (1) they worry the person who is performing +the unsafe work will become angry or defensive, and (2) they do not believe they can effectively stop unsafe +work. See Ragain, R., Ragain, P., Allen, M. & Allen, M. “Study: Employees Intervene in Only 2 of 5 Observed +Unsafe Acts.” Drilling Contractor. January / February 2011. +34 The entire Chevron leak response protocol is presented in Appendix A. + +Chevron Richmond Refinery Investigation Report January 2015 +14 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +c. Substandard equipment maintenance practices. Those same surveys indicate that +Chevron Richmond Refinery employees saw increased problems in how the refinery +maintained its equipment between 2008 and 2010. (See Section 5.5.)  +1.4.5 Industry Codes and Standards Findings +23. Industry falls short of requiring comprehensive inspection or effective facility upgrades. +American Petroleum Institute (API) Recommended Practice (RP) 939-C: Guidelines for Avoiding +Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries is the primary industry guidance +document on ways to monitor and control sulfidation corrosion. It states that carbon steel piping +can contain components with low silicon concentrations, and these components can corrode at a +faster rate than adjacent piping components. However, API RP 939-C does not specifically +require users to perform 100 percent component inspection or recommend that facilities upgrade +high-risk carbon steel piping circuits to steel alloys that are more resistant to sulfidation +corrosion. (See Section 5.2.1.) +24. Industry guidance is inconsistent in the information presented about carbon steel piping +susceptible to sulfidation corrosion. API has published various codes and recommended practices +in addition to API RP 939-C that discuss sulfidation corrosion, including API RP 571: Damage +Mechanisms Affecting Fixed Equipment in the Refining Industry, API 570: Piping Inspection +Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems, API RP 578: +Material Verification Program for New and Existing Alloy Piping Systems, and API RP 574: +Inspection Practices for Piping System Components. While these documents provide some +information on sulfidation corrosion, the information and guidance is varied and inconsistent. +(See Sections 5.2.2, 5.2.3, 5.2.4, and 5.2.5.) +25. Industry guidance for responding to process leak incidents can be improved. API and the +American Society of Mechanical Engineers (ASME) have published several codes, standards, and +recommended practices that provide information on how to safely control, mitigate, or respond to +hazardous process fluid leaks. However, the guidance is inconsistent, and none of the documents +provide overall, comprehensive guidance to emergency responders, operations personnel, and +facility management to respond safely to hazardous process leak incidents. (See Section 5.4.) +1.4.6 Regulatory Findings +26. In the years leading to the August 6, 2012, incident, the Chevron Richmond Refinery identified +weaknesses in its Stop Work Authority program due to employee hesitation to use Stop Work +Authority when witnessing an unsafe act. The Refinery also identified a decline in employee +perception of its mechanical integrity programs. However, the regulator did not require the +Chevron Richmond Refinery to take quality, constructive steps to improve these areas. Had steps +been taken before the incident to encourage employees to use their Stop Work Authority or to +determine why the refinery’s mechanical integrity programs were seen as deficient, the August 6, +2012, pipe rupture might have been prevented. (See Section 5.5.2.4.) + + +Chevron Richmond Refinery Investigation Report January 2015 +15 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +1.5 Recommendations +As a result of the findings and conclusions of this report, the CSB makes recommendations, summarized +below, to the following recipients (see Section 6.0 for full language of the recommendations): +________________________________________________________________ +American Petroleum Institute +Revise API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries +to establish minimum requirements for preventing catastrophic rupture of low-silicon carbon steel piping. +Revise API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining Industry to increase +awareness of sulfidation corrosion characteristics and refer users to specific API standards that provide +important information to prevent catastrophic rupture of low-silicon carbon steel piping. +Revise API 570: Piping Inspection Code: In-service Inspection, Rating, Repair, and Alteration of Piping +Systems to incorporate language consistent with API RP 939-C: Guidelines for Avoiding Sulfidation +(Sulfidic) Corrosion Failures in Oil Refineries, increase awareness of sulfidation corrosion characteristics, +provide additional information to prevent catastrophic rupture of low-silicon carbon steel piping, and +require users to follow the proposed new leak response guidance in API RP 2001: Fire Protection in +Refineries. +Revise API RP 578: Material Verification Program for New and Existing Alloy Piping Systems, to require +users to establish and implement a program to identify carbon steel piping circuits that are susceptible to +sulfidation corrosion and may contain low-silicon components. +Revise API RP 574: Inspection Practices for Piping System Components (3rd edition) to incorporate as a +normative reference API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in +Oil Refineries and to follow the leak response protocol requirements established in API RP 2001: Fire +Protection in Refineries. +Revise API RP 2001: Fire Protection in Refineries to require users to develop a process fluid leak +response protocol specific to their own facility that must be followed when a process fluid leak is +discovered. Recommend users to incorporate key actions into their leak response protocol to effectively +manage response to potential sulfidation corrosion piping failure. +________________________________________________________________ +American Society of Mechanical Engineers +Refer users to follow the leak response guidance developed by the American Petroleum Institute prior to +conducting leak repairs. +________________________________________________________________ +Chevron U.S.A. +Develop an accountability method at Chevron to identify and track effective implementation of Chevron +or industry best practices to ensure process safety or employee personal safety. + +Chevron Richmond Refinery Investigation Report January 2015 +16 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Develop an auditable process for all recommended turnaround work items related to inspection or +mechanical integrity recommendations that are denied or deferred. This process shall provide the +submitter of the denied or deferred recommendation with a mechanism to further elevate and discuss the +recommendation with higher level management. +Develop an approval process that includes a technical review that must be implemented prior to resetting +the minimum alert thickness to a lower value in the inspection database. +________________________________________________________________ +Board of Supervisors, Contra Costa County, California and Mayor and +City Council, City of Richmond, California +Revise the Industrial Safety Ordinance (ISO) regulations for petroleum refineries to require the +development of an oversight committee comprised of the regulator, the company, the workforce and their +representatives, and community representatives. Among the duties of this committee shall be to oversee +the development and implementation of action items created as a result of safety culture assessment +findings. +______________________________________________________________ + + +Chevron Richmond Refinery Investigation Report January 2015 +17 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +2.0 Richmond Refinery Process Description +2.1 Chevron Background +Chevron was originally founded as the Pacific Coast Oil Company in 1879.35 In 1906, Pacific Coast Oil +Company merged with Iowa Standard to form a new company known as Standard Oil Company of +California.36 The company then acquired Gulf Oil Corporation in 1984 and changed its name to +Chevron.37 +Headquartered in San Ramon, California, Chevron Corporation is the third-largest American company by +revenue.38 Globally, Chevron employs over 60,000 people.39 Chevron includes petroleum operations, +chemicals operations, mining operations, power generation, and energy services.40 It operates seven +petroleum refineries, five of which are in the United States. The five U.S. refineries process a combined +crude oil capacity of approximately one million barrels per day (bpd).41 +2.2 Richmond Refinery +Chevron’s Richmond Refinery is located in Richmond, California, approximately 25 miles northeast of +San Francisco in Contra Costa County. The original refinery units were built in 1902 by Pacific Coast Oil +Company. The Richmond refinery covers approximately 2,900 acres of the San Pablo Peninsula (Figure +2) and processes 250,000 barrels of crude oil per day. Approximately 1,200 people are employed at the +refinery. + + +35 http://www.chevron.com/about/history/ Chevron Company History Page (accessed June 5, 2014). +36 http://www.chevron.com/about/history/1876/ (accessed June 30, 2014). + +37 http://www.chevron.com/about/leadership/ (accessed June 30, 2014). +38 http://money.cnn.com/magazines/fortune/fortune500/2012/full_list/. This ranking is by annual revenue (accessed +June 30, 2014). +39 http://www.chevron.com/about/leadership/ (accessed June 30, 2014). +40 http://www.forbes.com/companies/chevron/ (accessed June 30, 2014). +41 See http://www.chevron.com/documents/pdf/UnitedStatesFactSheet.pdf (accessed December 18, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +18 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 2. Aerial view of the Chevron Richmond Refinery. +2.3 #4 Crude Unit +The Richmond, California Chevron Refinery’s #4 Crude Unit (Crude Unit) performs the initial processing +step in the refining process. Raw crude oil stored in storage tanks is pumped to the Crude Unit. After an +initial “cleaning” of the oil through the use of a desalter, which removes corrosive salts, solids, and +water,42 the oil is pre-heated and enters the C-1100 Crude Unit Atmospheric Column (Crude Column) at +approximately 675 degrees Fahrenheit (°F). The Crude Column separates through distillation various +hydrocarbon component mixtures in the crude feed, creating multiple streams coming off the column with +differing boiling points. These streams include an overhead light hydrocarbon stream, jet oil streams, a + +42 Removing chloride salts and water prevents the formation of hydrochloric acid, which can severely corrode +downstream equipment. Other salts and solids are removed to prevent fouling within equipment such as heat +exchangers, which can significantly reduce heat transfer. + +Chevron Richmond Refinery Investigation Report January 2015 +19 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +diesel stream, a light gas oil stream, and a bottoms stream composed of heavy liquid hydrocarbons. Each +stream is further refined and processed in subsequent units within the refinery. +2.4 4-Sidecut Line +The August 6, 2012, incident occurred from the piping referred to as the “4-sidecut” line, one of several +process streams exiting the Crude Column (Figure 3).43 As shown in Figure 4, light gas oil, the Crude +Unit 4-sidecut process fluid, exits the atmospheric column via a 20-inch nozzle and is split into a 12-inch +line and an 8-inch line. The pipe rupture (Figure 5) occurred on a 52-inch long component44 of the 4- +sidecut 8-inch line (the 52-inch component). The line operated at a temperature near 640°F45,46 and had +an operating pressure of approximately 55 pounds per square inch gauge (psig) at the rupture location. At +the time of the incident, light gas oil was flowing through the 8-inch line at a rate of approximately +10,800 bpd.47 + +43 The atmospheric column separates crude oil feed into different streams through distillation. These streams are +further processed in other units in the refinery. +44 The term “component” refers to a portion of piping between welds or flanges. It includes straight run piping and +pipe fittings. +45 The autoignition temperature for this process, the temperature at which a material will combust in the presence of +sufficient oxygen without an ignition source, was 640°F. This number is based on the Chevron Light Gas Oil +Material Safety Data Sheet. Chemical testing of 4-sidecut samples following the incident indicated lower +autoignition temperatures; however, these samples may not have been representative of typical 4-sidecut process +fluid. +46 Chevron instrumentation indicates that the process fluid entered the 4-sidecut piping at a temperature near 640°F +and cooled to 625°F before reaching the piping circuit pumps downstream of the rupture location. +47 This rate is the equivalent of 315 gallons per minute (gpm). A barrel equals 42 gallons. + +Chevron Richmond Refinery Investigation Report January 2015 +20 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 3. Schematic of C-1100 Crude Unit atmospheric column and upstream process equipment. + + +Chevron Richmond Refinery Investigation Report January 2015 +21 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 4. 4-sidecut line configuration and rupture location. + +Chevron Richmond Refinery Investigation Report January 2015 +22 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 5. Photo of rupture on 4-sidecut 52-inch component. +3.0 The Incident +3.1 Leak Discovery +At approximately 3:50 p.m. on August 6, 2012, an outside operator performing routine checks of piping +and equipment found an 18-inch puddle of what appeared to be a diesel-like material on the refinery +concrete pad (Figure 6). Identifying that the leak was occurring from overhead, the operator observed +intermittent drips as they accumulated on the underside of an insulated pipe 14 feet above ground level. +The leaking pipe was identified to be a portion of the 4-sidecut piping that originated on the Crude +Column. Visually analyzing the piping, the operator determined that the line could not be isolated from +the process. + +Chevron Richmond Refinery Investigation Report January 2015 +23 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 6. CSB animation depicting operator identifying the leaking 4-sidecut pipe. +The operator’s supervisor arrived at the leak location, shortly followed by the shift team leader. These +individuals observed that the leak was dripping at a rate of approximately 40 drips per minute. The +piping was insulated, so the individuals gathered near the leak could not identify its precise source (Figure +7). They concluded that the leak was not significant enough to require a shutdown, but was still a serious +situation. Shortly after 4:00 p.m., they called the Chevron Fire Department to the scene, a typical practice +at the refinery when leaks are discovered. Firefighters began to arrive at approximately 4:07 p.m. and +established an Incident Command structure. A hot zone of 20 feet by 20 feet was established and taped +off around the leak location by the Incident Commander. The area outside of the hot zone was considered +the cold zone, or safe zone.48 + +48 A decontamination corridor is often established in the warm zone, an area established between the hot zone and +the cold zone. “Decontamination involves thorough washing to remove contaminants. It should be performed in +an area upwind of the Hot Zone. An area that is uphill, with good drainage, and easily accessible for responders is +preferred.” See http://chemm.nlm.nih.gov/decontamination.htm (accessed January 21, 2015). + + +Chevron Richmond Refinery Investigation Report January 2015 +24 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 7. Photo taken of the leaking 4-sidecut pipe on August 6, 2012, at the Chevron Richmond Refinery.49 +Insulation obscured the actual leak location. Stain signifies where hydrocarbon process fluid was leaking +from the 4-sidecut piping. + + +49 Photo from http://richmond.chevron.com/Files/richmond/Investigation_Report.pdf (accessed June 27, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +25 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Beginning at approximately 4:15 p.m., many additional personnel were called to the scene of the leak to +assist in the leak analysis. Various operations personnel were called to the leak. Two Chevron inspectors +reported to the leak location to provide information on inspection history of the 4-sidecut line. The lead +Crude Unit process engineer also arrived at the leak location to determine an estimate of the hole size and +the quantity of material leaking so that proper environmental release calculations could be performed. +At approximately 5:00 p.m., the shift team leader left the scene of the leak and went to the control room. +He directed the board operator to reduce the feed to the 4-sidecut line by 5,000 bpd. + + +Figure 8. Timeline of events on August 6, 2012. +3.2 Leak Response +Ultimately, a large group of Chevron employees—40 people in total throughout the hours leading to the +incident—accumulated at the leak location. They began discussing their options to mitigate or stop the +leak. The inspectors informed the group that the 4-sidecut pipe walls were thinning due to sulfidation +corrosion, but data collected as recently as two months prior indicated the 4-sidecut line had sufficient +wall thickness to last until the next turnaround in 2016. This assessment led the group to believe that a +localized mechanism, such as abrasion on the line from a pipe support near the dripping location, was the +likely cause of the leak. The group then called the leak repair contractor to the leak location to assess the +possibility of clamping the line in an effort to stop the leak. A photo of a typical leak repair clamp is +shown in Figure 9. + +Chevron Richmond Refinery Investigation Report January 2015 +26 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 9. Example leak repair clamp for piping.50 It is installed over the leak location to prevent process fluid +leakage to the atmosphere. +The group then decided to remove the insulation from the 4-sidecut pipe to determine the cause of the +leak, a practice Chevron personnel call “daylighting the leak.” This procedure, they determined, would +help in the decision either to repair the leak on-line or to shut down the unit. +During preparation for the daylighting activity, the fire engine was repositioned in the cold zone to a +location approximately 65 feet from the leak, fire monitors51 were set up pointing towards the leak +location, and two hose lines were run from the fire engine to a position near the 4-sidecut piping. Two +teams of three firefighters operated the hoses. All hoses and monitors were at-the-ready, able to respond +should any incident occur. +The first attempt to remove insulation was made by pulling on the insulation bands from the ground using +a pike pole.52 This was unsuccessful. Rather, the piping actually moved from the force of the pulling, so +the group determined it was too dangerous trying to remove the insulation in that way. The group then +decided that scaffolding should be built to provide easy access so that firefighters could manually cut +loose the piping insulation. +At this point, shift change was occurring. Some individuals left for the day, and some volunteered to stay +past their shift end time after their relief showed up. This change resulted in an increase of people +standing near the 4-sidecut leak location. + +50 Photo from http://www.huwa.com/en/vervolgpagina/83/37/HUWA_Split_Barrel/ (accessed June 27, 2014). +51 A fire monitor is a piece of firefighting equipment that sprays water and can be manually aimed and operated. +52 A “pike pole” is a long pole with a hooked metal end commonly used by firefighters. + +Chevron Richmond Refinery Investigation Report January 2015 +27 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Contractors arrived at the scene to build the scaffold (Figure 10). During a pre-response safety meeting +consisting of the Incident Commander, a safety officer, firefighters, and key operations personnel, one +group of firefighters operating a hose was directed to spray the scaffold builders should an incident occur, +and the other group of firefighters operating the second hose was directed to spray the pipe if needed. +Three scaffold contractors then built the scaffold beneath the leaking 4-sidecut pipe. + +Figure 10. CSB animation of contractors erecting scaffolding beneath the leak location. +Once the scaffolding was built, two firefighters were directed to climb the scaffold and remove the +aluminum sheathing and insulation (shown in Figure 7) from the 4-sidecut pipe. The battalion chief was +aware that vapors leaking from under the insulation could mix with air and “light off,” or catch on fire, as +the insulation was removed, so the firefighters made preparations for such a possibility by being at-the- +ready with fire hoses. The firefighters on the scaffolding began to remove the aluminum sheathing +surrounding the insulation by using a hook to pull the bands securing the insulation and sheathing +sufficiently away from the insulation to allow for snipping the bands with cutters. Using this technique, +the firefighters were able to remove several three-foot sections of the aluminum sheathing surrounding the +insulation. +As the firefighters were removing the sheathing of the 4-sidecut line (shown in Figure 7), white +hydrocarbon vapor visibly began to emerge from under the now-exposed insulation material. The +firefighters continued to remove the sheathing despite the formation of hydrocarbon vapor. During the +continued sheathing removal, insulation that was soaked with hot 4-sidecut hydrocarbon autoignited once +exposed to oxygen—only feet from the firefighters. The hose teams immediately put out the fire, and +both firefighters quickly came down from the scaffold (Figure 11). + +Chevron Richmond Refinery Investigation Report January 2015 +28 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 11. CSB animation of firefighters attempting to remove the 4-sidecut insulation, the resulting fire, and +fire extinguishing. + + +Chevron Richmond Refinery Investigation Report January 2015 +29 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +The firefighters on the scaffolding successfully removed much of the aluminum sheathing surrounding +the insulation; however, underlying insulation still obscured the location of the leak. Directed by the +operations personnel, the Chevron Fire Department sprayed the insulation with hard, straight streams +using the fire hoses in an attempt to knock the insulation off the pipe. The hose teams knocked off the +insulation up to the location where the aluminum sheathing had been removed. At this point, they +realized that the leak had significantly worsened; hydrocarbon liquid was now spraying from the pipe. +Several operations managers present then decided to shut the unit down, an action that requires hours to +complete. +3.3 Consequences +A vapor cloud quickly began to accumulate. The hose teams attempted to keep the cloud at bay by +spraying it with firefighting water. Suddenly, the vapor cloud worsened, engulfing 19 firefighters and +operators standing in both the hot zone and cold zone in the hot hydrocarbon cloud. The cloud was dense +and very hot, and many of the individuals caught in the cloud were not able to see anything around them. +One person caught in the cloud told the CSB that he could not see his hand if he had held it directly in +front of his face. +Each person engulfed in the cloud began working their way out of the vapor cloud. Several of the +firefighters operating the two hoses dropped to their hands and knees to follow their hose lines to safety, +feeling their way out of the cloud (Figure 12). + +Figure 12. CSB animation of firefighters who dropped to their hands and knees to escape the vapor cloud. +At approximately 6:30 p.m., two minutes after the large vapor cloud formed, the light gas oil ignited.53 +Eighteen employees safely escaped from the cloud just before ignition. One employee, a firefighter, was +inside a fire engine that was engulfed in the fireball when the light gas oil ignited (Figure 13 and Figure + +53 Shown by surveillance video recording. + +Chevron Richmond Refinery Investigation Report January 2015 +30 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +14). He told CSB interviewers, “… All I could see [was] heavy hydrocarbon-type boiling fire in every +window of that truck. I can’t even begin to describe how hot it was. It was very intense.” He called +“MAYDAY” over his radio but received no response. He informed the CSB, “I figured that everybody +else was dead.” Because he was wearing full body firefighting protective equipment, he was able to +escape through the flames surrounding the fire truck and make his way to safety without physical injury. + + +Chevron Richmond Refinery Investigation Report January 2015 +31 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 13. CSB animation of firefighter who was inside the fire engine when the light gas oil ignited. + +Chevron Richmond Refinery Investigation Report January 2015 +32 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 14. Photo of the burned remains of the fire truck that was consumed by the fire. A firefighter was in +the cab when the light gas oil ignited. The fire truck was positioned in the cold zone approximately 65 feet +from the leak location. +The leak resulted in a large plume of vapor which traveled across the surrounding area. The ignition and +subsequent burning of the hydrocarbon process fluid created a large black cloud of smoke, which also +swept across the surrounding area (Figure 15). This situation resulted in a Community Warning System +(CWS) Level 3 alert,54 and a shelter-in-place55 advisory (SIP) was issued at 6:38 p.m.56 for Richmond, +San Pablo, and North Richmond. It was lifted later that night at 11:12 p.m. after the fire was fully under +control. In the weeks following the incident, nearby medical facilities received over 15,000 members of +the public seeking treatment for ailments including breathing problems, chest pain, shortness of breath, + +54 A Community Warning System Level 3 alert indicates that a facility within Contra Costa County has had a release +that has offsite impact and is categorized by any of the following events: +1. Offsite impact that may cause eye, skin, nose and/or respiratory irritation to the general population. +2. Fire, explosion, heat, or smoke with an offsite impact. Example: On a process unit/storage tank where mutual +aid is requested to mitigate the event and the fire will last longer than 15 minutes. +3. Hazardous material or fire incident where the Incident Commander or unified command, through consultation +with the Contra Costa Health Services Hazardous Material Incident Response Team, requests that sirens +should be sounded. +See http://cchealth.org/hazmat/pdf/incident_notification_policy.pdf (accessed April 9, 2013). +55 Contra Costa County considers a shelter-in-place to include going inside a home or nearest building, closing doors +and windows, and turning off heating, ventilation, and air conditioning. See +http://cchealth.org/emergencies/shelter-in-place.php (accessed February 6, 2013). +56 Chevron U.S.A. Inc. "30 Day Follow-Up Notification Report." September 5, 2012. + +Chevron Richmond Refinery Investigation Report January 2015 +33 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +sore throat, and headaches. According to information provided to the CSB by local hospitals, +approximately 20 people were admitted to local hospitals as inpatients for treatment. + +Figure 15. Initial vapor cloud formation (white cloud) and subsequent ignition (black smoke) as seen from a +pier in San Francisco, California.57 + +57 Photos are from Fototaker.net. + +Chevron Richmond Refinery Investigation Report January 2015 +34 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 16. Vapor cloud and ignition seen from Marin County.58 + +4.0 Technical Analysis +The CSB commissioned Anamet, Inc., a materials engineering and laboratory testing company, to +conduct testing of the 4-sidecut pipe, including the failed 52-inch component. The testing concluded that +the rupture was due to pipe wall thinning caused by sulfidation corrosion.59 +4.1 +Sulfidation Corrosion +Sulfidation corrosion, also known as sulfidic corrosion,60 is a damage mechanism61 that causes thinning in +iron-containing materials, such as steel, due to the reaction between sulfur compounds and iron at +temperatures ranging from 450°F to 1000°F.62 For pipe walls, this damage mechanism causes gradual + +58 Photo is a screen capture from KTVU Channel 2 News. +59 Anamet, Inc. "Metallurgical Evaluation of Samples from the Chevron U.S.A. Inc., Richmond #4 Crude Unit 8 +Inch and 12-Inch 4-Sidecut Piping Involved in the August 6, 2012, Hydrocarbon Release and Fire." Prepared for +The Chemical Safety and Hazard Investigation Board (CSB), February 11, 2013. +60 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 3.1.6, May 2009. +61 Piping damage mechanisms are any type of deterioration encountered in the refining and chemical process +industry that can result in flaws/defects, thus affecting the integrity of piping (e.g., corrosion, cracking, erosion, +dents, and other mechanical, physical or chemical impacts). See API 570: Piping Inspection Code: In-Service +Inspection, Rating, Repair, and Alteration of Piping Systems. 3rd ed., Section 3.1.1.5, November 2009. +62 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 1, May 2009. + +Chevron Richmond Refinery Investigation Report January 2015 +35 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +thinning over time. Sulfidation corrosion is common in crude oil distillation,63 where naturally occurring +sulfur and sulfur compounds found in crude oil feed, such as hydrogen sulfide,64 are available to react +with steel piping and equipment. Process variables that affect corrosion rates include the total sulfur +content of the oil, the sulfur species present, flow conditions, and the temperature of the system. Virtually +all crude oil feeds contain sulfur compounds; therefore, sulfidation corrosion is a damage mechanism +present at every refinery that processes crude oil. Sulfidation corrosion can cause thinning to the point of +pipe failure when not properly monitored and controlled. +Sulfidation corrodes carbon steel at a much faster rate in comparison with its effect on other materials of +construction, such as steels with a higher chromium content. This issue is discussed in depth in the +CSB’s Interim Investigation Report of the August 6, 2012, Chevron incident.65 In addition to its naturally +faster rate of sulfidation corrosion when compared with higher chromium steels, carbon steel can also +experience significant variation in corrosion rates due to variances in silicon content, a component used in +the steel manufacturing process. Carbon steel piping containing silicon content less than 0.10 weight +percent can corrode at accelerated rates,66 up to 16 times faster than carbon steel piping containing higher +percentages of silicon. Figure 17 shows how carbon steel corrosion rates can greatly vary depending on +silicon content. + +63 Distillation separates mixtures into broad categories of its components by heating the mixture in a distillation +column where different products boil off and are recovered at different temperatures. See +http://www.eia.gov/todayinenergy/detail.cfm?id=6970 (accessed April 4, 2013). +64 Hydrogen sulfide is the most aggressive sulfur compound that causes sulfidation corrosion. +65 http://www.csb.gov/assets/1/19/Chevron_Interim_Report_Final_2013-04-17.pdf (accessed April 2, 2014). +66API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., Section +6.2.3.2, May 2009. + +Chevron Richmond Refinery Investigation Report January 2015 +36 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 17. Graph of sulfidation corrosion rates with respect to silicon content in carbon steel. This graph +shows how corrosion rates increase in carbon steel containing decreasing percentages of silicon. This +information can be found in Annex C of API RP 939-C.67 +The refining industry has been aware of increased susceptibility to sulfidation corrosion in low-silicon +carbon steel piping since as early as 1974,68 nearly 40 years before the August 6, 2012, incident and two +years before the Chevron Crude Unit was constructed. Before the incident, Chevron documented its +understanding of the potentially catastrophic nature of failures caused by sulfidation corrosion, as +reflected in Chevron’s Corrosion Prevention and Metallurgy Manual: +Sulfidation corrosion has caused severe fires and fatalities in the refining industry, +primarily because it causes corrosion over a relatively large area, so failures tend to +involve ruptures or large leaks rather than pinhole leaks. It can be insidious in that +moderately high corrosion rates can go undetected for years before failure. Finally, +process changes that increase the temperature or sulfur content can creep up over time +and multiply corrosion rates so that what was thought to be a low corrosion rate system +becomes corrosive enough to fail before the increased corrosion rate is recognized. + +67 The y-axis of this figure is in units of mils per year (mpy). A “mil” is 1/1000 inch. +68 API Publication 943: High-Temperature Crude Oil Corrosivity Studies. September 1974. + +Chevron Richmond Refinery Investigation Report January 2015 +37 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Carbon steel piping is manufactured to meet certain specifications. Prior to the mid-1980s, multiple +carbon steel specifications were commonly and independently in use for refinery piping, including +American Society for Testing and Materials (ASTM) A53B,69 ASTM A106,70 and American Petroleum +Institute (API) 5L.71 ASTM A53B and API 5L do not contain minimum silicon content requirements for +carbon steel piping,72 while ASTM A106 requires the piping to be manufactured with a minimum silicon +content of 0.10 weight percent. As a result, manufacturers have used different levels of silicon in the +carbon steel pipe manufacturing process. Thus, depending on the manufacturing specification for carbon +steel, sulfidation corrosion rates could vary depending on the silicon content within the steel. +In the mid-1980s, pipe manufacturers began to comply simultaneously with all three manufacturing +specifications (ASTM A53B, ASTM A106, and API 5L) when manufacturing carbon steel piping, which +resulted in piping being manufactured with at least 0.10 weight percent silicon content due to the ASTM +A106 requirement. As a result, the majority of carbon steel piping purchased following this time period +for refinery operations likely has a minimum of 0.10 weight percent silicon content. However, piping +purchased and installed prior to the mid-1980s could still contain low silicon components susceptible to +high, variable sulfidation corrosion rates. +The timing of this manufacturing change has a profound impact on the susceptibility of refineries to +variable sulfidation corrosion rates today. Over 95 percent of the 144 refineries in operation in the United +States, including the Chevron Richmond Refinery,73 were built before 1985,74 before piping +manufacturers began producing carbon steel in compliance with all three manufacturing specifications. +Therefore, the original carbon steel piping components in these refineries is likely to contain varying +percentages of silicon content and may experience highly variable sulfidation corrosion rates. +The Chevron Richmond Refinery 4-sidecut piping circuit containing the 52-inch component that failed +was constructed of ASTM A53B carbon steel, which had no minimum specification for silicon content. +Post-incident testing of samples of the 4-sidecut piping from the Chevron Richmond Refinery identified +silicon content ranging from 0.01 weight percent to 0.2 weight percent. Of 12 samples taken from the 8- +inch and the adjacent 12-inch 4-sidecut line, six had a silicon concentration of less than 0.10 weight +percent (Figure 18 and Figure 19). The 52-inch pipe component that ruptured on the day of the incident +had a silicon content of only 0.01 weight percent. Illustrating the inherent variability in ASTM A53B +carbon steel sulfidation corrosion rates, the elbow component directly upstream of the 52-inch component +that failed had a silicon concentration of 0.16 weight percent and showed considerably less thinning +(Figure 20). + +69 ASTM Standard A53/A53M-12: Standard Specification for Pipe, Steel, Black and Hot-Dipped, Zinc-Coated, +Welded and Seamless, 2012. +70 ASTM Standard A106/A106M–11: Standard Specification for Seamless Carbon Steel Pipe for High-Temperature +Service, 2011. +71 API Specification 5L: Specification for Line Pipe. 45th ed., December 2012. +72 ASTM Standard A53/A53M-12: Standard Specification for Pipe, Steel, Black and Hot-Dipped, Zinc-Coated, +Welded and Seamless, 2012. +73 The Chevron Richmond Refinery was constructed in 1902. +74 See http://www.eia.gov/tools/faqs/faq.cfm?id=29&t=6 (accessed February 14, 2013). + +Chevron Richmond Refinery Investigation Report January 2015 +38 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 18. Locations of metallurgical samples taken from 8-inch 4-sidecut piping post-incident. The seven +samples taken are boxed in orange. Four of the seven components sampled (shown with an orange fill) were +found to have a silicon content less than 0.10 weight percent. + + +Chevron Richmond Refinery Investigation Report January 2015 +39 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 19. Locations of metallurgical samples taken from 12-inch 4-sidecut piping post-incident. The five +samples taken are boxed in orange. Two of the five components sampled (shown with an orange fill) were +found to have a silicon content less than 0.10 weight percent. + + +Chevron Richmond Refinery Investigation Report January 2015 +40 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 20. 4-sidecut piping sample (E-017-8) analyzed by Anamet Labs showing the relative thickness of low +silicon piping on the left and the high silicon piping on the right. The ruptured 52-inch pipe component (left) +contained 0.01 weight percent silicon, and the upstream elbow component (right) contained 0.16 weight +percent silicon.75 The initial nominal thickness of this piping was 0.322-inch. + + +75 Anamet, Inc. "Metallurgical Evaluation of Samples from the Chevron U.S.A. Inc., Richmond #4 Crude Unit 8- +Inch and 12-Inch 4-Sidecut Piping Involved in the August 6, 2012, Hydrocarbon Release and Fire." Prepared for: +The Chemical Safety and Hazard Investigation Board (CSB), February 11, 2013. + +Chevron Richmond Refinery Investigation Report January 2015 +41 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +4.2 Sulfidation Corrosion Inspection Techniques +As evidenced by the chemical analysis performed on the Chevron 4-sidecut piping post-incident, carbon +steel piping components within a single circuit76 can contain varying percentages of silicon, resulting in a +large variation in sulfidation corrosion rates by component. Current corrosion inspection guidance +documents allow for the measurement of pipe thickness at a minimal number of permanent Condition +Monitoring Locations (CMLs)77 along the piping length.78,79 These CMLs are most frequently placed on +elbows and fittings80 because higher turbulence in these areas usually results in the fastest metal loss.81 +However, due to details of the manufacturing process, carbon steel elbows and pipe fittings, even when +manufactured to the ASTM A53B specification, generally contain relatively high percentages of silicon.82 +When measurements are taken only at high silicon-containing fittings, the measurements can fail to +identify high corrosion rates within a pipe circuit occurring within low-silicon, straight-run piping +components. +API Recommended Practice (RP) 939-C Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion +Failures in Oil Refineries describes the challenges when attempting to inspect carbon steel lines +susceptible to sulfidation corrosion. The recommended practice states that older ASTM A53B piping, such +as the Chevron piping that failed on August 6th, creates a “major inspection challenge”83 and “unless the +refinery is fortunate enough to have located an inspection point on that particular [low silicon] section of +pipe or fitting, it is very difficult to detect the thinning component.”84 It states that in some applications, +carbon steel will appear to be adequate based on measured corrosion rates until failure occurs at some +undocumented or unidentified low-silicon component.85 +At the Chevron Richmond Refinery, the 8-inch 4-sidecut piping had a total of 19 CMLs86 on piping and +fittings. Historically, most of the CMLs measured corrosion rates at high silicon pipe-fitting components, +such as elbow components. An effort in 2011 added additional CMLs on straight-run components within + +76 A piping circuit is a length of pipe and the fittings associated with a particular process service that operate at +similar conditions. A circuit usually begins and ends at either a branch or a piece of process equipment, such as a +vessel or a pump. Reference to piping by circuits allows piping to be grouped conveniently by proximity and +operating service. Piping circuits may also be referred to as piping runs. +77 A condition monitoring location (CML) is a designated area where periodic thickness examinations are +conducted. Each CML represents as many as four inspection locations located circumferentially around the pipe. +CMLs are also referred to as thickness monitoring locations (TMLs). CMLs were historically referred to as +corrosion (rather than condition) monitoring locations, and that terminology is sometimes still used within the +industry. +78 API 570: Piping Inspection Code: In-service Inspection, Rating, Repair, and Alteration of Piping Systems, 3rd +ed., Section 5.6.3, November 2009. +79 For most damage mechanisms that affect an entire piping circuit, the whole circuit loses metal at a similar rate. +Monitoring pipe thickness at a minimal number of CMLs is considered representative of the entire pipe. +80 A fitting is a piping component usually associated with a change in direction or diameter. +81 API 570: Piping Inspection Code: In-service Inspection, Rating, Repair, and Alteration of Piping Systems, 3rd +ed., Section 5.6.2, November 2009. +82 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries, 1st ed., +Section 7.1.5, May 2009. +83 Ibid., Section 9. +84 Ibid., Section 7.1.5. +85 Ibid., Section 6.2.3.2. +86 Many of these CMLs were added during the 2011 turnaround. + +Chevron Richmond Refinery Investigation Report January 2015 +42 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +the 8-inch 4-sidecut piping circuit, although 100 percent component inspection was not performed. The +CSB found that, although a CML was located on the adjacent upstream elbow, no CMLs were placed on +the low silicon piping component that failed (Figure 21). Chevron identified corrosion in the 52-inch +component during a supplemental 2002 inspection,87 but the inspection results were not entered into the +CML-tracking portion of the inspection database, and no new CML was required to be added to ensure +future monitoring. As a result, the 52-inch component was never inspected again (Section 5.1.2.1.1). + +Figure 21. CML placement on 8-inch 4-sidecut piping. Nineteen CMLs were used to monitor corrosion rates +in the 8-inch 4-sidecut piping. A CML was not placed on the 52-inch component that ultimately failed. A +CML was placed, however, on the upstream elbow adjacent to the 52-inch component. +This inspection data gathered using the 19 CMLs did not reflect the corrosion rates of the quickly +corroding, low-silicon components of the 4-sidecut piping. As illustrated by the Chevron incident, +traditional inspection techniques alone—using only a limited amount of CMLs—may not accurately +identify the most aggressive sulfidation corrosion rates throughout an entire circuit of carbon steel piping. + +87 The inspector was supervising contractors who were conducting measurements on a CML located on the adjacent +elbow. Prompted by a recent Chevron corrosion study, he decided to have them also measure the straight pipe +about one foot before and after the elbow, on the other side of the welds, by radiographic thickness techniques +(RT). These were not formal CMLs. In 2002, the 52-inch component had lost roughly a third of its original wall +thickness. The only documentation of this measurement was a note in a history brief in the inspection database. + +Chevron Richmond Refinery Investigation Report January 2015 +43 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Low-silicon components can remain uninspected and unidentified until failures such as the August 6, +2012, Chevron incident occur. +As discussed in the CSB’s Interim Investigation Report of the Chevron incident and again in this report, +upgrading material of construction to steels with higher chromium content is a more effective means of +managing sulfidation corrosion. Indeed, Chevron’s internal “New Construction Guidelines” recommend +that piping installed in high temperature and high sulfur service be constructed with 9-Chrome steel. +4.3 Silicon Characterization Techniques +Determining silicon content in existing carbon steel piping and equipment in the field is a difficult +undertaking. To characterize the silicon content in each component of a piping circuit properly, every +component must be inspected. This is known as 100 percent component inspection. Two techniques are +used to identify low-silicon content and resulting variable corrosion rates in existing carbon steel piping +circuits with unknown chemical composition: (1) performing chemical analysis; or (2) performing pipe +wall thickness measurements of every carbon steel component.88 +Silicon Characterization Technique +Description +Chemical composition analysis of each component + +Shavings of piping are analyzed in a laboratory to +determine silicon content. Requires weld +identification and insulation removal. + +Thickness measurement of each component + + +Identifies gross differences in component +thicknesses due to differing silicon concentrations. +Must be performed on piping that has been in- +service for long enough time to detect corrosion +rate differences. Requires weld identification and +insulation removal. + +Table 1. Silicon characterization technique. +Many field-portable instruments used for Positive Material Identification89 (PMI) cannot adequately +identify silicon content.90,91 If original manufacturing quality assurance data92 are not available, as is + +88 Pipe wall thickness measurement is difficult because of the high operating temperatures of piping subject to +sulfidation corrosion. These measurements are commonly made only when the piping is out of service and cool, +for example, during a unit maintenance activity. +89 Positive Material Identification is the identification and chemical analysis of various metal alloys through +nondestructive methods. +90 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 7.1.5, May 2009. +91 Recent technological advances may allow for nondestructive silicon detection using a hand-held composition +analyzer. See http://www.olympus-ims.com/en/applications/using-handheld-xrf-to-manage-sulfidation-corrosion- +in-carbon-steel/ (accessed October 29, 2014) + +Chevron Richmond Refinery Investigation Report January 2015 +44 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +generally the case with older plants, then chemical verification requires destructive testing. Metal +shavings must be taken from each carbon steel piping component for chemical analysis in a laboratory.93 +Care must be taken not to contaminate the sample with bits of metal from the tools used to gather the +sample. +Carbon steel components containing low concentrations of silicon can also be identified by performing +one-time thickness measurements of every component within a carbon steel circuit.94 This practice is +only useful as a means to ascertain silicon content if the piping circuit has been exposed to sulfidation +corrosion for a long enough time period so that variances in corrosion rate caused by differences in silicon +content may be detected. Chemical analysis is, therefore, the most reliable technique to identify low- +silicon carbon steel components. +Both characterization techniques require identification of each piping component, typically by removing +insulation (so every weld seam can be located), a time consuming and costly undertaking. Weld seams +can be located through insulation using specialized equipment and examination techniques, but this +method can be less accurate than when weld seams are identified manually. Both silicon characterization +techniques can be technically difficult and physically hazardous for inspectors because of the high +operating temperatures of piping subject to sulfidation. It is common to make thickness measurements or +take shaving samples only when the piping is out of service and cool, for example, during a unit +maintenance turnaround.95 +Unlike silicon concentration, the chromium concentration of steel can easily be verified in the field using +portable positive material identification instruments. In addition, steel alloys containing at least 9 weight +percent chromium are more resistant to sulfidation corrosion and do not run the risk of extreme variations +in corrosion rates within components in the same piping circuit.96 This makes alloys with at least 9 +weight percent chromium content an inherently safer choice97 in high temperature sulfidation corrosion +environments because the hazard presented by varying corrosion rates within a single piping circuit is +eliminated. As shown in the Modified McConomy Curves98 from API RP 939-C (Figure 22), higher +chromium steels are also a better safeguard than carbon steel because sulfidation corrosion rates are + +92 Manufacturing quality assurance data, also known as mill data, provides the chemical composition of the steel. +93 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., +Section 7.1.5, May 2009. +94 Ibid., Section 7.1.5. +95 A “turnaround” is a scheduled shutdown of a process unit to perform maintenance, repairs, upgrades, and +inspection of process equipment. +96 The protective scale, FeCr2S4, begins to be the dominant scale formed in steels containing a chromium content of +five weight percent. The 5Cr steel alloy can be manufactured to contain anywhere from 4 percent to 6 percent +chromium. Thus, “the sulfidation corrosion rate can vary dramatically in 5Cr steels even in the same operating +environment.” See Niccolls, E. H., J. M. Stankiewicz, J. E. McLaughlin, and K. Yamamoto. "High Temperature +Sulfidation Corrosion in Refining." 17th International Corrosion Congress. Las Vegas: NACE International, +2008. +97 Steels with higher chromium content are inherently safer than carbon steel with respect to sulfidation corrosion +because they can eliminate the hazard of gross variations in corrosion rates within a single piping circuit. +However, analysis is still required to ensure that the best material of construction is selected. +98 Modified McConomy Curves are the set of curves API RP 939-C uses to predict sulfidation corrosion rates +versus temperature for several steel alloys. + +Chevron Richmond Refinery Investigation Report January 2015 +45 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +greatly reduced. Carbon steel99 corrodes approximately nine times faster than 9-Chrome steel, which +contains 9 percent chromium, and carbon steel corrodes approximately 120 times faster than stainless +steel, which contains 18 percent chromium.100,101 + +Figure 22. Modified McConomy Curves from API RP 939-C. +4.4 Inherently Safer Design +Efforts to improve chemical process safety require the identification of process hazards, followed by the +elimination, mitigation, or control of process hazards to reduce the overall risk of a process. A hazard can +be defined as a “situation with the potential for harm.” Risk is then defined as a function of both the +consequence (hazard) and likelihood (frequency).102 +Risk reduction can be achieved by using, in order of robustness, inherently safer design, passive +safeguards, active safeguards, and procedural safeguards.103 This can be thought of as a tiered or +hierarchical approach to risk management, commonly referred to as a “hierarchy of controls.” The further + +99 ASTM A53B carbon steel contains a maximum of 0.40 weight percent chromium. +100 9-Chrome contains 9 weight percent chromium. +101 These values were calculated using the McConomy Curves at 630°F. +102 Center for Chemical Process Safety (CCPS). Inherently Safer Chemical Processes – A Life Cycle Approach. 2nd + ed., Section 2.1, 2009. +103 Ibid. +100.0 + +Chevron Richmond Refinery Investigation Report January 2015 +46 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +up the hierarchy, the more effective the risk reduction achieved (Figure 23). Inherently safer design104 +reduces risk by permanently reducing or eliminating a defined hazard itself, while safeguards defined as +passive (design features), active (detection and automatic response), and procedural (policies, procedures, +training, inspection, use of personal protective equipment) reduce risk by reducing the ultimate +consequence or likelihood of the hazard.105 + +Figure 23. Hierarchy of controls. The further up the hierarchy (further to the left as shown here), the more +effective the risk reduction achieved. Figure developed from concept presented in Kletz, Trevor; Amyotte, +Paul. Process Plants: A Handbook for Inherently Safer Design, 2nd ed; 2010. +These definitions are published by the Center for Chemical Process Safety (CCPS), a not-for-profit +corporate membership organization within the American Institute of Chemical Engineers that identifies +and addresses process safety needs within the chemical, pharmaceutical, and petroleum industries.106 The +CCPS book Inherently Safer Chemical Processes, 2nd ed. designates a process as inherently safer “if it +reduces or eliminates the hazards associated with materials and operations used in the process and this +reduction or elimination is permanent and inseparable.”107 A facility can approach its inherently safer +design strategy by looking for opportunities to minimize, substitute, moderate, or simplify.108 +The August 6, 2012, sulfidation corrosion pipe rupture at the Chevron Richmond Refinery highlights a +missed opportunity to incorporate inherently safer design strategies through the use of more robust +materials of construction. CCPS states: +The concept of inherent robustness … applies to designing equipment to be impervious to +the corrosion mechanisms that are present given the materials of construction and within +the process, and the operating conditions (i.e., temperature, pH, concentration, viscosity, +etc.). The use of certain alloys will eliminate certain types of corrosive attack. […] +Although robust equipment design may be considered to be a passive safeguard rather +than an inherently safer design, it considerably simplifies the remainder of the process +design. Therefore, it fits within the definition of simplification [an inherently safer design +strategy]. It is also highly effective in eliminating the possibility of an uncontrolled loss + +104 The concept of “inherently safer design” was first established by Trevor Kletz in 1977 in response to the 1974 +Flixborough explosion in England. He presented a lecture titled “What You Don’t Have, Can’t Leak” at the +Jubilee Lecture for the Society of Chemical Industry. +105 Center for Chemical Process Safety (CCPS). Inherently Safer Chemical Processes – A Life Cycle Approach. 2nd +ed., Section 2.1, 2009. +106 www.aiche.org/ccps/about (accessed February 14, 2013). +107 Center for Chemical Process Safety (CCPS). Inherently Safer Chemical Processes – A Life Cycle Approach. 2nd +ed., Section 2.2, 2009. +108 bid., Section 3.5. + +Chevron Richmond Refinery Investigation Report January 2015 +47 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +of containment. In a general sense, the removal of this possibility from a process design +must be considered to be inherently safer.109 +Thus, the use of a higher chromium steel alloy, such as 9-Chrome, is an inherently safer design strategy +that could have prevented the Chevron Richmond Refinery pipe rupture. Installation of 9-Chrome piping +during turnaround opportunities prior to the August 6, 2012, incident to replace the 4-sidecut carbon steel +would have both eliminated the hazard of silicon-based variable corrosion rates within the components of +the 4-sidecut piping circuit and at the same time also greatly reduced the underlying inherent rate of +sulfidation corrosion. The use of 9-Chrome steel also simplifies the procedural inspection safeguards +required, as 100 percent component inspection would not be required to monitor corrosion rates in steels +containing at least 9 percent chromium; the typical inspection strategy of monitoring corrosion rates using +a minimal amount of CMLs would be sufficient. +It is important to remember that inherently safer design strategies are relative; a specific design can only +be described as inherently safer when compared to a different design with regard to a specific hazard or +risk.110 A design may be inherently safer by eliminating one hazard, but can inadvertently introduce or +aggravate another hazard.111 For this reason, performing a comprehensive, documented hazard analysis is +vital for identifying the individual hazards and the likelihood of those hazards occurring, followed by +identifying how they can be effectively controlled to minimize overall risk. The review should include +risks of personal injury, environmental harm, and lost production, as well as evaluating economic +feasibility.112 An inherently safer systems review incorporated as an integral part of this hazard analysis +generates an optimized list of hazard control choices, offering various degrees of inherently safer design +strategies. +Effectively implementing inherently safer design provides an opportunity for preventing major chemical +incidents. The August 6, 2012, incident at Chevron and other incidents113 throughout the refining +industry highlight the difficulty in preventing failure caused by sulfidation corrosion in low silicon carbon +steel piping solely through inspection—a procedural safeguard that is thus among the least effective on +the hierarchy of controls. Implementing inherently safer design concepts to the greatest extent feasible by +Chevron and other refiners and chemical plant operators will avoid hazards such as variation in +sulfidation corrosion rate in carbon steel piping due to hard-to-determine silicon content, and it will +provide a higher degree of protection from incidents similar to the one that occurred on August 6, 2012. + +109 Center for Chemical Process Safety (CCPS). Inherently Safer Chemical Processes – A Life Cycle Approach. 2nd +ed., Section 4.5.1, 2009. +110 Center for Chemical Process Safety (CCPS). Guidelines for Engineering Design for Process Safety. 2nd ed., +Section 5.2, 2012. +111 Center for Chemical Process Safety (CCPS). Inherently Safer Chemical Processes – A Life Cycle Approach. 2nd +ed., Section 3.6, 2009. +112 Ibid., Section 8.6.4. +113 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries. 1st ed., May +2009. + +Chevron Richmond Refinery Investigation Report January 2015 +48 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.0 Incident Analysis +The CSB investigation team developed an accident map (AcciMap)114 for the Chevron investigation +(Figure 24). The AcciMap is a multilayered causal diagram that depicts immediate causes115 as well as +higher level contributing causes at the corporate, governmental, and regulatory levels. This diagram +includes five levels: +1. Outcomes: the impact of the August 6, 2012, event to workers present and the surrounding +community; +2. Physical Events and Conditions: the immediate causes of the incident as displayed in a traditional logic +tree; +3. Chevron: company rules and policies; conduct of turnarounds; risk management; identification of +hazards and evaluation of safeguards; adoption of internal recommendations; safety programs; and +emergency response; +4. Industry Codes and Standards: good practice guidelines that provide safety requirements and +recommendations on topics including mechanical integrity and emergency response; and +5. Government: laws and legislation developed to regulate process safety at refineries. +Some of these contributing factors are discussed in the Chevron Interim Report and the Chevron +Regulatory Report. Refer to the AcciMap in Figure 24116 as a guide to locating information. + +114 The AcciMap was originally developed by Jens Rasmussen in the article Rasmussen, Jens. “Risk Management in +a Dynamic Society: A Modelling Problem.” Safety Science. Vol. 27, No 2/3, 1997; pp. 183-213. The AcciMap +was subsequently used and popularized by Andrew Hopkins, in Hopkins, Andrew. “Lessons From Longford: +The Esso Gas Plant Explosion.” CCH Australia Limited: Sydney, 2000; Chapter 10. +115 Immediate causes are the events or conditions that lead directly to an incident, such as mechanical failure or +human error. +116 A high-resolution graphic of the AcciMap can be found on the CSB website. + + + +Chevron Richmond Refinery Investigation Report January 2015 +49 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 24. Acci-Map of August 6, 2012, Chevron Refinery Fire. + +Chevron Richmond Refinery Investigation Report January 2015 +50 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1 Organizational Analysis +In the ten years prior to the incident, a small number of Chevron personnel with knowledge and +understanding of sulfidation corrosion made recommendations to increase inspections or upgrade the +material of construction in the 4-sidecut piping. Their recommendations were not effectively +implemented. The process to implement important, safety-critical projects within the Chevron Richmond +Refinery was not fully effective. As discussed in the following sections and depicted in Figure 25, a +combination of (1) reliance on a turnaround management program that depended on only a fraction of +necessary data to make important process safety decisions, (2) an unsuccessful bottom-up approach— +with no management oversight or accountability—for implementing a crucial safety program, and (3) no +formal method to track to completion the Chevron expert group’s findings and recommendations +ultimately caused these recommendations to not be implemented. + +Figure 25. Organizational decision-making schematic showing attempts to have carbon steel 4-sidecut piping +100 percent component inspected or replaced with a higher chromium steel alloy. Attempts failed due to lack +of accountability and lack of authority to ensure recommendation implementation, and a rigid turnaround +planning process that could not approve the 4-sidecut piping replacement recommendations. +Figure 26 shows the sequence of sulfidation corrosion-related recommendations and events within +Chevron Corporation and the Chevron Richmond Refinery between 2002 and the day of the incident. +These events are discussed in subsequent sections of this report. + +Chevron Richmond Refinery Investigation Report January 2015 +51 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 26. Key events at the Richmond refinery between 2002 and 2012. +The following sections discuss Chevron programs designed to improve equipment reliability and process +safety. These programs are summarized in Figure 27 and Figure 28. + +Chevron Richmond Refinery Investigation Report January 2015 +52 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 27. Chevron Corporation refinery process safety programs. + + + +Chevron Richmond Refinery Investigation Report January 2015 +53 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.1 Chevron Energy Technology Company +In this section: + + + +Within Chevron is a separate business unit called the Chevron Energy Technology Company (ETC). +This unit provides technology solutions and technical expertise for Chevron operations worldwide (Figure +28). Chevron ETC technical staff has considerable knowledge and expertise regarding sulfidation +corrosion, specifically with respect to corrosion rate variations caused by differing silicon concentration +in carbon steel piping. Chevron ETC employees have authored industry papers on sulfidation corrosion +and had significant influence in the development of the industry sulfidation corrosion recommended +practice, API RP 939-C. This recommended practice, first published in 2009,117 was developed under +Chevron leadership. Metallurgists within ETC had shared their knowledge on sulfidation corrosion via +many outlets over the years to employees at Chevron refineries, as discussed in the following sections. +Despite all of this institutional expertise, the 4-sidecut line ruptured due to sulfidation corrosion at the +Chevron Richmond Refinery. + +117 As of January 2015, the 2009 edition of API RP 939-C is the active edition of this standard. + +Chevron Richmond Refinery Investigation Report January 2015 +54 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 28. Chevron Energy Technology Company (ETC) organizational roles. ETC provides technology +solutions and technical expertise throughout all Chevron (and affiliated companies) operations. + +5.1.1.1 ETC Sulfidation Failure Prevention Guidance +At the approximate time of publication of API RP 939-C, Chevron ETC metallurgists released within +Chevron a formal report dated September 30, 2009, nearly three years before the incident, titled Updated +Inspection Strategies for Preventing Sulfidation Corrosion Failures in Chevron Refineries (hereinafter +referred to as ETC Sulfidation Failure Prevention Initiative). +The ETC Sulfidation Failure Prevention Initiative clearly indicates that Chevron technical experts +understood the high likelihood that the consequence of a sulfidation corrosion failure could be a rupture +or catastrophic failure. It specifically calls out Chevron’s need for action: +Sulfidation corrosion failures are not common in Chevron or in the industry but they +are of great concern because of the comparatively high likelihood of blowout or +catastrophic failure…. This can happen because corrosion occurs at a relatively +uniform rate over a broad area so a pipe can get progressively thinner until it actually +bursts rather than leaking at a pit or local thin area. In addition the process fluid is often +above its autoignition temperature. The combination of these factors means that +sulfidation corrosion failures frequently result in large fires.… [S]everal case histories +of sulfidation corrosion failures … have occurred in Chevron or in the industry, several +of which are blowouts. +The Chevron ETC Sulfidation Failure Prevention Initiative specifically recommends that inspectors +perform 100 percent component inspection on high temperature carbon steel piping susceptible to +sulfidation corrosion. The initiative defines a priority ranking system to help focus the inspection efforts. +The process conditions of the 4-sidecut stream—operating temperatures greater than 600°F— placed it in +the highest priority category for inspection. +In 2010, Chevron ETC technical experts issued a corporate newsletter focusing on materials and +corrosion, again warning of the potential consequence of sulfidation failures. This newsletter reiterated + +Chevron Richmond Refinery Investigation Report January 2015 +55 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +the recommendation from the 2009 ETC Sulfidation Failure Prevention Initiative to conduct 100 percent +component inspection of carbon steel piping systems that operated over 600°F. The Richmond Crude +Unit 4-sidecut piping fell within this high-priority inspection category. This newsletter was accessible to +all employees on Chevron’s company intranet. +Chevron ETC also regularly hosted training sessions for refinery personnel. One of these classes, +designed for refinery inspection staff but also attended by refinery engineers and senior operators, focused +specifically on crude units and on corrosion mechanisms within crude units—including sulfidation +corrosion. One of the main messages from this training was that sulfidation corrosion in piping +containing low-silicon components can result in catastrophic rupture, and that the means to prevent +rupture from occurring is performing 100 percent component inspection or upgrading to a higher +chromium steel (Figure 29 and Figure 30). + +Figure 29. Presentation slide of ETC training course that guided refinery staff to perform 100 percent +component inspection on high-temperature lines susceptible to sulfidation corrosion. + +Chevron Richmond Refinery Investigation Report January 2015 +56 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 30. Presentation slide of ETC training course that guided refinery staff on ways to reduce risk from +sulfidation corrosion. +The 2009 ETC Sulfidation Failure Prevention Initiative report was circulated to reliability managers and +metallurgists at the individual refineries, as well as to the corporate reliability expert. The authors of the +Chevron ETC report chose to send the report to these individuals specifically because they seemed to be +the right people to perform and advocate for the necessary inspection and replacement work. However, +the CSB discovered that the Reliability Department at the Chevron Richmond Refinery believed that the +100 percent component inspection initiative recommended by the ETC sulfidation failure prevention +guidance, while important safety work, was an ambitious, unfunded, and unsupported initiative which +could not be effectively performed solely within the normal budget and headcount resources of the +Reliability group. Also, the Reliability group made attempts to forego the burdensome inspection +initiative and simply implement the ETC alternative sulfidation mitigation strategy to improve sulfidation +corrosion resistance through using higher chromium steel by recommending replacement of the 8-inch 4- +sidecut piping with inherently safer higher chromium steel.118 However, these recommendations were +denied on multiple occasions during the Chevron turnaround planning process (Section 5.1.2). As +discussed in the following sections, the Chevron programs created to ensure that necessary work was +performed to maintain reliable, safe operation and corporate-wide process safety were not successful in +preventing the rupture that had the potential to seriously harm 19 Chevron employees on August 6, 2012. +5.1.1.2 Chevron ETC Conclusions +Despite many attempts by Chevron ETC to warn and educate refinery personnel, neither 100 percent +component inspection of high-risk carbon steel piping nor upgrading susceptible piping to an inherently +safer material of construction was fully performed at the Chevron Richmond Refinery. In practical terms, +because Chevron ETC was a separate business entity within the Chevron corporation and had no direct +authority over the reliability management within the Chevron refineries, Chevron ETC sulfidation +corrosion experts had limited influence on what actually occurred within refineries in their areas of +expertise. These individuals did not participate in refinery-specific processes such as Process Hazard + +118 Shown in Figure 30, ETC’s senior inspector and analysis training taught inspectors that sulfidation corrosion +resistance could be improved by upgrading susceptible piping circuits to steels containing higher percentages of +chromium. + +Chevron Richmond Refinery Investigation Report January 2015 +57 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Analyses (PHAs)119 (see Chevron Interim Investigation Report) and did not affect decisions concerning +monitoring and control of sulfidation corrosion during the Crude Unit turnaround process (Section 5.1.2). +No formal system is in place at the Chevron Richmond Refinery to communicate to the refinery +management and to track to completion ETC findings and recommendations. While Chevron does use an +indicators tracking program (Section 5.1.6), the program does not measure the implementation status of +ETC recommendations and new industry guidance determined by Chevron technical experts to be critical +in ensuring continued safe operations at Chevron refineries. + + + + + + + + + + + + + + + +119 A process hazard analysis is a hazard evaluation to identify, evaluate, and control the hazards involved in a +process. Facilities that process a threshold quantity of hazardous materials, such as the Chevron Richmond +refinery, are required to conduct a process hazard analysis per the California Code of Regulations Title 8 Section +5189. Process Safety Management of Acutely Hazardous Materials (1992). PHAs are also required by the +California Accidental Release Prevention Program and the federal EPA Risk Management Program. See the +CSB’s Chevron Interim Investigation Report for a full analysis on this subject. + +Chevron Richmond Refinery Investigation Report January 2015 +58 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.2 Chevron Turnaround Management +In this section: + + +Chevron uses a turnaround planning and implementation process called Initiative for Managing Pacesetter +Turnarounds, known within Chevron as the IMPACT process. It is organized into several different +phases, each with a specific objective. The IMPACT process begins several years before each scheduled +turnaround. Chevron’s stated goal of the IMPACT process is to perform turnarounds efficiently and +effectively, ensuring that only the necessary work items that must be performed during the turnaround are +performed, and all other items that can be performed on-the-run (while the unit is operating) or during the +next turnaround are not included in the work scope. The turnaround work approval process is guided by +what Chevron calls a “Framing Document.” It is developed over a year before the planned turnaround, by +a group including turnaround planning management, a decision review board, and specific process unit +managers. An IMPACT “Core Team” comprised of an operations representative, inspector, design +engineer, process engineer, capital project representative, and maintenance representative are the main +decision makers regarding what potential work items meet the requirements of the Framing Document +and thus have the potential to be automatically included in the turnaround. The work items that pass the + +Chevron Richmond Refinery Investigation Report January 2015 +59 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Framing Document test are then prioritized.120 A group of refinery managers determine a priority cutoff, +and only items in this high-priority, Framing Document-approved list are automatically included in the +turnaround scope. The items below the priority cutoff are reviewed by refinery managers and can be +included in the turnaround scope on a case-by-case basis. The Core Team follows the strict criteria +described in the turnaround Framing Document when approving or denying requested turnaround work +submitted by refinery employees. If the Core Team determines that a potential work item does not meet +the Framing Document requirements, it never reaches the prioritization step, so it is not part of a review +by the refinery managers. +5.1.2.1 2007 Crude Unit Turnaround +The Framing Document used for the 2007 Crude Unit turnaround specified that the primary requirements +for the turnaround included performing any work needed to assure a minimum of 10 years (two +turnaround cycles) before the next inspection or maintenance was required, performing all required +compliance inspections, and recertifying state operating permits for boilers. The detailed work list criteria +for the turnaround are shown in Figure 31. + +120 The prioritization was based on a combined severity and likelihood of impacts in four individually-weighted +categories: health, safety, environmental, and production. The higher the sum of these four values, the higher the +priority for the potential work item. + +Chevron Richmond Refinery Investigation Report January 2015 +60 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 31. Work list criteria requirements specified in the Framing Document used during Chevron’s 2007 +Crude Unit turnaround. +5.1.2.1.1 +Recommendations Regarding 4-Sidecut Line for 2007 Turnaround +The first recommendations to upgrade or 100 percent component inspect the 4-sidecut line were made for +the 2007 Crude Unit turnaround. One such recommendation was based upon the findings from the 2002 +turnaround. In August 2002, a Chevron Richmond Refinery employee analyzed sulfidation corrosion +rates in the Crude Unit and identified potentially vulnerable areas based on process conditions. The +employee discovered that the 4-sidecut operating temperature had increased, concluding this increase +would cause more hydrogen sulfide to evolve, leading to increased sulfidation corrosion rates. The +employee’s study recommended increased inspection of the 4-sidecut piping and noted that this piping +might need upgrading from carbon steel to 5-Chrome, a steel alloy containing five percent chromium that +is more resistant to sulfidation corrosion. In 2002, proactively following up on this study, the Crude Unit +inspector conducted additional piping inspection and identified corrosion in a 52-inch 4-sidecut +component, which is the component that ultimately failed on August 6, 2012. The corrosion was found +during inspection on a component that was not a typical inspection location (i.e., not an official CML). + +Chevron Richmond Refinery Investigation Report January 2015 +61 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Chevron inspection guidelines require that findings from additional discretionary inspection must be +recorded only as notes in a “history brief” rather than input in the inspection database as is typically done +for CMLs. Therefore, documentation of the corrosion identified in 2002 was recorded only as notes and +was not formally input as data into the inspection database of permanent CML measurements. (An +example of inspection database data appears in Figure 32.) The inspector formally recommended +upgrading this piping during the next shutdown in 2007 to the IMPACT core team. + +API does not require new CMLs to be established on components with non-uniform corrosion in +potentially low silicon carbon steel piping circuits. Such a requirement in API RP 939-C would help to +ensure that components experiencing non-uniform corrosion are effectively monitored, managed, and +replaced to prevent sulfidation corrosion failures. + + +Figure 32. Sample Inspection Database report analyzed by unit inspectors when determining piping +remaining life and when making piping replacement recommendations. This Inspection Database report is +also analyzed by the IMPACT core team to determine Framing Document applicability. + +Adhering to the Chevron turnaround work scope procedures, the IMPACT core team analyzed the +available inspection data of the 4-sidecut line from the inspection database to determine whether +upgrading the 4-sidecut piping met the Framing Document requirements. The recommendation to +upgrade the 4-sidecut piping did not meet the 2007 turnaround Framing Document requirement #10, as all + +Chevron Richmond Refinery Investigation Report January 2015 +62 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +recorded data in the inspection database (from existing CMLs on relatively high-silicon piping fittings) +indicated that the 4-sidecut piping upstream of the pumps had sufficient thickness to continue to safely +operate until the next turnaround before requiring further inspection or maintenance. The +recommendation to upgrade the 4-sidecut piping also did not meet the 2007 turnaround Framing +Document requirement #5, as the IMPACT core team believed it was not a substantiated high-risk +reliability issue because existing inspection data did not indicate an imminent reliability problem. It was +determined, however, that the 4-sidecut piping downstream of the pumps required replacement,121 based +on the same process of analysis, and only that piping was replaced with an upgraded, inherently safer +material of construction, 9-Chrome (Figure 33). + +Also prior to the 2007 turnaround in February 2006, a team consisting of a materials and corrosion +engineer, an inspector, a process engineer, a metallurgist, and a design engineer issued a Corrosion +Mitigation Plan for the Chevron Richmond Refinery Crude Unit. This report was developed specifically +for the Area Business Unit (ABU) Manager, the highest ranking manager for the Crude Unit. The report +specifically identified the 4-sidecut piping to be at risk from high temperature sulfidation corrosion. The +team issued the following recommendation: +Recommendation – Install Guided Wave bracelets [on the 4-sidecut piping] during the +[2007 Crude Unit] Turnaround so that 100% volumetric inspection of the line can be +done to ensure that there are no piping sections in the line that are corroding faster than +the majority of the line…. [The] piping needs to be monitored in anticipation of future +replacement and additionally, industry experience shows that sections of piping with +low silica [sic] content will corrode at higher rates. Monitoring this section of line +using global inspection technique like guided wave is the fastest way to determine if +there are thin piping pups in the system. + +121 The piping downstream of the pumps operates at a higher pressure, and thus the Minimum Required Thickness +calculated was thicker than for the piping on the suction side of the pumps. + +Chevron Richmond Refinery Investigation Report January 2015 +63 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 33. Crude column schematic indicating the piping downstream of the 4-sidecut pumps that was +replaced during the 2007 Crude Unit turnaround. The portion of the carbon steel line containing the +component that failed on August 6, 2012 was not replaced. +The recommendation to perform 100 percent volumetric inspection through the use of experimental +guided wave technology was accepted and implemented by the ABU Manager. 122 The inspection +recommendation met the framing document requirement #5 and requirement #8, as it was specified to be +based upon past industry sulfidation experience and was approved by the ABU Manager. However, this +recommendation was only partially implemented. The guided wave bracelets were installed only on a +small portion of the 4-sidecut line, which did not include the 52-inch component that ultimately failed on + +122 Common volumetric inspection techniques include ultrasonic and radiography testing. + +Chevron Richmond Refinery Investigation Report January 2015 +64 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +August 6, 2012.123 Furthermore, when the experimental guided wave bracelet data proved to be +unreliable, manual 100 percent component inspection was not implemented to address the low-silicon +piping component corrosion concerns raised in the February 2006 recommendation. In addition, because +the Crude Unit inspector’s discretionary inspection observations were not input into the CML-tracking +portion of the inspection database as an official CML (discussed earlier in this section), the 52-inch +component in which the inspector identified corrosion in 2002—the component that failed on August 6, +2012—was never inspected again. +5.1.2.2 2011 Crude Unit Turnaround +The Framing Document for the 2011 Crude Unit Turnaround states that some of the goals of the 2011 +turnaround were to perform work to ensure a five year run; perform compliance inspections and +requirements; and perform safety, environmental, and process improvement work that required a +shutdown. The work list criteria requirements for the turnaround appear in Figure 34. + +Figure 34. Work list criteria requirements specified in the Framing Document used during Chevron’s 2011 +Crude Unit turnaround. + +123 It was later determined that the guided wave data was unreliable, so the Crude Unit inspector resumed +inspections using ultrasonic thickness measurements. + +Chevron Richmond Refinery Investigation Report January 2015 +65 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +A key change from the 2007 Framing Document directly affected whether work items such as 100 percent +component inspection or replacement of the 4-sidecut could obtain Core Team approval. In 2007, the +Core Team could approve a work item as meeting 2007 Framing Document criterion #10 for inspection, +repair or replacement of equipment or piping if it was needed to ensure a minimum of 10 years (two +turnaround cycles) before the next inspection or if maintenance was required. The guided wave +inspection devices on the 4-sidecut met this criteria. In 2011, this Framing Document criterion was +deleted. The Core Team in 2011 should refuse any work item for inspection, repair, or replacement of +equipment or piping unless it required a rate cut or shutdown within only 5 years, before the next +turnaround (items #2 and #6). +5.1.2.2.1 + Recommendations Regarding 4-Sidecut Line for 2011 Turnaround +Following the release of the Chevron ETC Sulfidation Failure Prevention Initiative report, discussed in +Section 5.1.1, the Chevron Richmond Refinery materials group completed the risk-ranking of the carbon +steel piping in the Richmond Lube Oil Project Unit (RLOP) and in the Crude Unit, two units with high +temperature piping known to be susceptible to sulfidation corrosion. The group identified the Crude Unit +4-sidecut line as a high-risk line based on the report ranking guidance. Instead of requesting funding to +perform the 100 percent component inspection, the group recommended the 4-sidecut for replacement +with 9-Chrome. Just as when planning for the 2007 turnaround, the IMPACT team denied the +recommendation because the inspection data available for the 4-sidecut piping did not support a material +upgrade during the 2011 turnaround. The IMPACT team did not consider the lack of data on potentially +more susceptible 4-sidecut straight-run pipe components. +Chevron also conducts “Intensive Process Reviews” prior to turnarounds. This process involves +knowledgeable individuals including Business Improvement Network leaders, process engineers, +metallurgical engineers, design engineers, and turnaround planners. The review aims to identify key unit +issues that should be addressed and repaired during the unit turnaround. Before the 2011 Crude Unit +turnaround, Chevron personnel conducted an Intensive Process Review of the Crude Unit and specifically +recommended that the 4-sidecut carbon steel piping “should be upgraded to [5-Chrome] […] due to +sulfidation.” Although the Intensive Process Review identified sulfidation problems in the 4-sidecut line, +this activity was ineffective. The 4-sidecut piping was not upgraded during the 2011 Crude Unit +turnaround because the IMPACT core team determined that it did not meet the turnaround framing +document requirements. +During the 2011 turnaround, a portion of the 12-inch 4-sidecut piping was identified as unacceptably thin +and was replaced. The 12-inch 4-sidecut piping was the same age, material of construction, and +contained the same process fluid with similar process conditions124 as the 8-inch 4-sidecut piping. +However, because Chevron relies on existing data to make equipment replacement decisions, employees +did not consider that these significant sulfidation corrosion findings could be indicative of similar +thinning that could be occurring in the 8-inch piping. This was another missed opportunity during the +2011 Crude Unit turnaround to identify that the 8-inch 4-sidecut piping needed replacement. + +124 The CSB notes that the process conditions of the 8-inch and 12-inch 4-sidecut piping were not identical. + +Chevron Richmond Refinery Investigation Report January 2015 +66 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.2.3 Chevron Richmond Refinery Turnaround-Planning Conclusions +There were no 4-sidecut line inspection data indicating that the 8-inch 4-sidecut piping had thinned +significantly enough from sulfidation corrosion to require replacement of the piping. Rather, the limited +CML data, extrapolated to apply to the entire pipe, including the portion containing the component that +failed on August 6, 2012, indicated that the pipe could remain in service until the 2016 turnaround. This +oversight occurred for two reasons: (1) 100 percent component inspection was never performed, and (2) +recorded data existed only on high-silicon components that corroded at much slower rates than the low- +silicon 52-inch component. +Chevron’s data-driven turnaround management framework led to unintended negative consequences. The +current Chevron Richmond Refinery turnaround planning framework denies potential, discretionary +turnaround work that does not yet have hard data gathered from refinery equipment to support it, even if +the work request is based upon guidance issued by the industry trade association, American Petroleum +Institute. This rejection is true even if, as in the case of 100 percent component inspection for sulfidation +damage, the purpose of the work request is to actually generate the hard data. The only way a Chevron +employee can have a work request approved based solely on industry guidance is to appeal to the ABU +Manager for the work as an exception to the turnaround framing document criteria. (See 2007 turnaround +Framing Document work criteria #8 and 2011 turnaround Framing Document work criteria #9.) +The CSB cannot conclusively state whether even this method would have resulted in the approval to +replace the 4-sidecut line, but Chevron reliability and metallurgical staff never attempted it. These +individuals had not previously been in the position of having to convince management of the importance +of their turnaround work recommendations, so advocating the Sulfidation Failure Prevention Initiative +and persuading upper management to implement the ETC recommendations would be a foreign work area +for them. In addition, no high-level manager was assigned responsibility to ensure that the ETC +Sulfidation Failure Prevention Initiative or other ETC sulfidation recommendations were included in the +turnaround scope. As a result, lower level employees who did not have decision-making or funding +authority were burdened with convincing Chevron Richmond Refinery management to implement new +industry guidance and the ETC recommendations. +The requirement for hard data to justify turnaround work even affected decisions to mitigate hazards +identified during a turnaround. During the 2011 turnaround, thinning from sulfidation corrosion in the +12-inch portion of the 4-sidecut piping was found to be so severe that the 12-inch piping had to be +replaced immediately. Yet, Chevron replaced only the portions where hard data was available on the +specific pipe to support the replacement. The hazardous condition did not prompt Chevron turnaround +management to inspect all of the 4-sidecut piping or to preemptively implement the longstanding +recommendation to replace all of the carbon steel portions. +Post-incident, Chevron performed 100 percent component inspection of all Chevron Richmond Refinery +Crude Unit piping susceptible to sulfidation corrosion. Four carbon steel piping components out of the +4,600 components inspected were identified to have higher corrosion rates than adjacent piping.125 Each + +125 Steve Wildman (Chevron) letter to Randall Sawyer (Contra Costa Health Services), April 12, 2013. +http://www.ci.richmond.ca.us/DocumentCenter/View/26802 (accessed June 30, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +67 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +of these four piping segments of which Chevron was unaware might have ultimately failed like the 52- +inch 4-sidecut segment. This inspection activity may have prevented additional pipe ruptures in the unit +due to accelerated sulfidation corrosion in low-silicon carbon steel. + + + +Chevron Richmond Refinery Investigation Report January 2015 +68 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.3 Chevron Unit Reliability Improvement Process +In this section: + + +Chevron’s Unit Reliability Improvement Process (URIP) formally integrates a broad range of reliability +activities. This program applies to all of Chevron’s wholly owned refineries. The overall URIP consists +of several sub-processes, including Reliability in Asset Integrity, Resolution of Significant Reliability +Opportunities, Risk Assessment and Asset Strategy, Condition Monitoring and Surveillance, Proactive +Maintenance, and Maintenance and Failure Prevention. +Beginning in 2009 as part of the URIP process, Chevron refineries began holding monthly Unit +Reliability Briefs (URBs) as a forum for discussing short-term and long-term equipment reliability issues. +Participants in these meetings include operators, process engineers, area inspectors, materials engineers, +machinery reliability employees, maintenance employees, and management, such as the ABU Manager. +These meetings proved ineffectual in securing safety critical improvements on several occasions. +Documentation indicates that the high risk of sulfidation corrosion in the 4-sidecut piping was discussed +at least twice in the URB meetings in the years leading to the incident. In 2010, a discussion on the need + +Chevron Richmond Refinery Investigation Report January 2015 +69 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +to replace the 4-sidecut piping to 9-Chrome resulted in a recommendation to replace the piping during the +2011 turnaround. However, the IMPACT Core Team ultimately denied replacement (Section 5.1.2.2.1). +In 2012, URB meetings raised the need for additional inspection of the 4-sidecut line. In early 2012, a +plan was implemented to increase inspection on the 8-inch 4-sidecut line, but the plan did not include the +complete 100 percent component inspection recommended by the ETC Sulfidation Failure Prevention +Initiative. Upper management, with decision-making and funding authority, was not assigned to—nor +took ownership of—assessing implementation of the ETC Sulfidation Failure Prevention Initiative or +similar ETC strategies in the refinery. The net result was that the URB meetings were not successful in +effectively advocating for the ETC Sulfidation Failure Prevention Initiative. The action items developed +in the URB meetings were not implemented, and high-risk piping susceptible to sulfidation corrosion was +not properly inspected or replaced. +Also part of the URIP process is a Reliability Steering Committee (RSC) meeting which occurs twice per +month at the Chevron Richmond Refinery. Participants include the reliability manager, operations +manager, ABU Manager, and materials and design personnel. These meetings aim to help steward +reliability in the Chevron refineries. Specifically, this committee is responsible for “[d]evelop[ing] a +Long-Term Reliability Plan consisting of prioritized, sequenced, and resource scoped recommendations +for achieving long-term reliability objectives.” This committee also monitors long-term reliability +improvement plan work requests to completion. The RSC discussed the ETC Sulfidation Failure +Prevention Initiative and assigned the refinery’s materials engineers to ensure its effective implementation +at the Richmond refinery. The materials engineers risk-ranked piping in the Richmond refinery based on +the ETC report’s guidance. +In late 2010, the refinery materials engineers and the inspectors presented a case to the IMPACT core +team for the Richmond Lube Oil Project (RLOP) turnaround to perform 100 percent component +inspection of various piping segments during its 2011 turnaround. This group also presented the +recommendations of the ETC Sulfidation Failure Prevention Initiative report to the IMPACT core team, +informing the group that this initiative was what prompted the 100 percent component inspection +recommendations. Because the IMPACT team concluded that the inspection could be performed on-the- +run when the unit was operating, they denied the 100 percent component inspection work request for the +2011 RLOP turnaround. Nevertheless, disagreements surfaced among the IMPACT team and the +individuals who submitted the recommendations regarding the feasibility, safety, and accuracy of +measuring thickness on high-temperature piping while the unit was operating. Some employees felt that +performing thickness measurements on-the-run was unsafe. In fact, the inspection database was set up to +define any pipe over 450°F as too hot to safely inspect on-the-run. In addition to safety concerns, +accuracy was also suspect. Thickness testing on hot piping is extremely difficult to do accurately. +However, this objection was also overruled, in this case based on the belief that accuracy was not +important because the on-the-run inspection would only be looking for gross differences in pipe +component thicknesses. +To implement the ETC recommendations in the Chevron Richmond Refinery Crude Unit, the unit’s +materials engineer and inspector submitted a request to replace the 4-sidecut piping with 9-chrome, but as +discussed in Section 5.1.2.2.1, this recommendation was denied because it did not meet the 2011 + +Chevron Richmond Refinery Investigation Report January 2015 +70 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +turnaround Framing Document requirement of performing only work to ensure a five-year run. As yet, no +data supported that the 4-sidecut piping could not remain in operation for at least another five years. +In the case of the 2011 RLOP and Crude Unit turnarounds, Chevron Richmond Refinery staff had +attempted to include work items to implement the ETC Sulfidation Failure Prevention Initiative. They +made recommendations based on their authority in the Chevron Richmond Refinery inspection and +materials engineering departments, and under the auspices of the URIP process and its URB and +Reliability Steering Committee programs. However, they were unsuccessful; they were thwarted by the +IMPACT turnaround planning process and the rigidity of its Framing Documents. No additional +implementation efforts or appeals to refinery management were attempted, and no high-level refinery +managers who attended URBs and Reliability Steering Committee meetings took responsibility of the +ETC Sulfidation Failure Prevention Initiative and ETC sulfidation mitigation recommendations to ensure +their effective implementation in the Richmond refinery. + + + +Chevron Richmond Refinery Investigation Report January 2015 +71 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.4 Chevron Fixed Equipment Reliability Business Improvement Network +In this section: + + +Chevron uses a corporate-wide equipment reliability expert group, called the Fixed Equipment Reliability +Business Improvement Network (FER BIN), to monitor ongoing reliability efforts at each Chevron +refinery as well as to promote new reliability improvement programs and align reliability practices at all +of the Chevron refineries. The FER BIN is intended to be a “best practice” network that brings up to date +changes in industry standards into Chevron. It is headed by a technically qualified subject-matter expert, +the FER BIN Leader, who advocates for the implementation of new industry best practices or new +reliability initiatives, such as the ETC Sulfidation Failure Prevention Initiative. +Significant change occurred in the Chevron Business Improvement Network (BIN),126 and the FER BIN +in particular, right after the ETC Sulfidation Failure Prevention Initiative was issued. The individual who + +126 The Chevron Business Improvement Network (BIN) incorporates all areas of expertise needed to ensure process +safety and mechanical integrity in Chevron refineries (e.g., expertise in specific chemical processes, rotating +equipment expertise, and fixed equipment reliability expertise). + +Chevron Richmond Refinery Investigation Report January 2015 +72 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +was in the FER BIN Leader role when the ETC Sulfidation Failure Prevention Initiative was issued +retired in September of 2010, before the initiative was fully developed and implemented. This previous +FER BIN Leader had been in the role for many years, and had a close working relationship with ETC, +including those responsible for developing the ETC Sulfidation Failure Prevention Initiative. Also in +2010, the organization to which the various BIN Leaders reported was restructured. They now reported to +a new organization encompassing process safety, reliability, and energy management for the entire +manufacturing organization. The structure of the new organization was changed as a result, and every +position was filled “from a clean sheet of paper.” A replacement for the FER BIN Leader position was +not assigned until four months after the previous FER BIN Leader’s retirement, in January 2011, and the +onboarding process for the new FER BIN Leader’s roles and responsibilities took additional time because +of the hiring delay. +The FER BIN meets periodically to discuss status of ongoing reliability improvement strategies. One of +the strategies that the FER BIN was tasked to focus on was implementing the ETC Sulfidation Failure +Prevention Initiative at all of the Chevron refineries. The FER BIN 2012 business plan included a task +item to develop and implement “shaping plans” at Chevron refineries to inspect for high temperature +sulfidation. The FER BIN Leader was charged with tracking progress of the shaping plans at each +refinery. However, the CSB found that the new FER BIN Leader had minimal authority to enforce +implementation of the ETC Sulfidation Failure Prevention Initiative at the Chevron Richmond Refinery. +No employees within the refinery directly reported to the FER BIN Leader. In addition, despite the +IMPACT core team’s decision power, it did not analyze the fixed equipment reliability shaping plans +when making turnaround work item decisions. No requirements existed in the IMPACT Framing +Documents to comply with shaping document directives. +In March 2012, five months prior to the incident, the FER BIN Leader visited the Chevron Richmond +refinery and identified that inspection of all carbon steel components susceptible to sulfidation corrosion +was not being performed as recommended by the ETC Sulfidation Failure Prevention Initiative (Figure +35). In addition to identifying that CML placement for piping may need to be reassigned, this review +found that the IMPACT team was denying critical inspection recommendations during the turnaround +planning process. The FER BIN Leader identified that Richmond refinery leadership needed to review +and implement the 2009 Chevron ETC Sulfidation Failure Prevention Initiative report and +recommendations. + +Chevron Richmond Refinery Investigation Report January 2015 +73 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 35. Presentation slide showing corporate reliability leader’s findings that the Richmond Refinery was +not complying with the ETC Sulfidation Failure Prevention Initiative. +When the FER BIN Leader visited refineries, he met solely with inspection managers and inspectors to +track progress. He did not meet with higher management within the Richmond refinery, such as the ABU +Manager, to give updates on whether the inspection group was meeting corporate expectations. His +assumption was that the individual refinery lead inspectors would use the knowledge he provided to +shepherd new safety programs outlined in the refinery FER shaping plan. However, that implementation +strategy did not work at the Chevron Richmond Refinery. (See Section 5.1.2.) +Despite the existence of the FER BIN, it was not successful in ensuring important fixed equipment +reliability work was being performed at the Chevron Richmond Refinery. The FER BIN program did not +effectively gain commitment from refinery management—the individuals capable of ensuring that the +necessary reliability work was being performed—to implement the ETC Sulfidation Failure Prevention +Initiative or other ETC recommendations to upgrade susceptible carbon steel piping to inherently safer, +higher chromium steel. + + + +Chevron Richmond Refinery Investigation Report January 2015 +74 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.5 Chevron Minimum Pipe Thickness Program +In this section: + + +Inspection staff at the Chevron Richmond Refinery determines the necessary inspection practices, such as +condition monitoring location (CML) placement, inspection time intervals, and minimum allowable pipe +thicknesses, by following the Richmond Refinery Piping Inspection Guideline. The inspection staff +followed this guideline when monitoring the 4-sidecut piping corrosion rates and “remaining life,” the +amount of time before the piping would become unacceptably thin and require replacement. +The Richmond Refinery Piping Inspection Guideline refers to two pipe thicknesses that must be known to +properly determine remaining life of a pipe: + +“Minimum Alert Thickness” (Chevron calls this Flag Thickness) – The “wall thickness +value used for triggering the need for quantitative [“Minimum Required Thickness”] and + +Chevron Richmond Refinery Investigation Report January 2015 +75 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +half-life assessments.”127 The Richmond Refinery Piping Inspection Guideline assigns the +4-sidecut piping a “Minimum Alert Thickness” of 0.14-inch. This value may be reduced +to 0.10-inch based upon a “thorough technical review.” + +“Minimum Required Thickness” – The minimum thickness of piping that can withstand +the existing pressure and structural stresses. Piping must be replaced before it reaches its +“Minimum Required Thickness.” +A visual depiction of the 4-sidecut original wall thickness, its Minimum Alert Thickness, and its +Minimum Required Thickness appears in Figure 36. +Chevron uses a database to store inspection findings, notes, and piping wall thickness values, and to track +corrosion rates. This database allows the inspector to input a pipe’s “Min Value,” which can be either the +piping’s Minimum Alert Thickness or Minimum Required Thickness—to help to determine a piping +circuit’s remaining life. + +Figure 36. To-scale schematic of 4-sidecut piping original wall thickness (0.322-inch), Minimum Alert +Thickness (0.13-inch), and Minimal Required Thickness (0.11-inch) using API RP 574 default values. + +127 API 570 Section 6.3.3 states, “thickness measurements should be scheduled at intervals that do not exceed the +lesser of one half the remaining life determined from the corrosion rates…or the maximum intervals +recommended” by API 570 Table 2. + +Chevron Richmond Refinery Investigation Report January 2015 +76 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Inspection thickness data obtained during the 2011 turnaround indicated that the piping corrosion rates +would result in the 4-sidecut piping wall thickness being reduced to below its 0.14-inch Minimum Alert +Thickness before the next shutdown scheduled for 2016. According to the Richmond Refinery Piping +Inspection Guideline, the next step would be to calculate a Minimum Required Thickness for the 8-inch +4-sidecut line and then determine if it needed to be replaced immediately, or if replacement could be +safely delayed. This evaluation can also be used to lower the Minimum Alert Thickness to 0.1-inch +following a thorough technical review. A structural minimum thickness value of 0.036-inch had been +calculated for a small piping component on the suction of a 4-sidecut pump earlier during the turnaround. +The inspector received this calculation in writing (Figure 37). + +Figure 37. Text from Chevron design engineer indicating structural minimum thickness (t(min)) calculation +results for small sections of suction piping upstream of the 4-sidecut pumps. +A communication breakdown occurred in reviewing these results. The design engineer understood +“suction piping” to refer to only a small section of piping upstream of the 4-sidecut pump P-1149, while +the inspector understood “suction piping” to refer to the entire 4-sidecut circuit upstream of pump P-1149. +They never met to clarify the calculation results. The calculated minimum structural thickness value of +0.036-inch for a small portion of the P-1149 suction piping was applied to the full length of the 8-inch 4- +sidecut piping circuit. This calculation was used as a technical justification to reduce the 8-inch 4-sidecut +Minimum Alert Thickness to 0.1-inch, and the piping wall thickness was predicted to stay above this +Minimum Alert Thickness for at least six years. The 4-sidecut line was therefore allowed to continue +operating with replacement scheduled for the next turnaround in 2016. +API RP 574: Inspection Practices for Piping System Components gives specific guidance to users on +Minimum Alert Thickness and Minimum Required Thickness. API RP 574 provides guidance on +minimum thickness values only for piping that operates under 400°F. Piping that operates above this +reference temperature of 400°F, such as Chevron’s 4-sidecut piping circuit, could be expected to require +even greater minimum thickness values. API RP 574 provides an example of a minimum alert thickness + +Chevron Richmond Refinery Investigation Report January 2015 +77 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +of 0.13-inch for piping 6 to 18 inches in diameter. Chevron’s 0.14-inch Minimum Alert Thickness is a +conservative value based on API RP 574’s guidance. API RP 574 also provides users with a default +minimum structural thickness of 0.11-inch for piping with a diameter of 8-inches—which can be used as +the Minimum Required Thickness for piping in lieu of detailed engineering calculations.128 Chevron +performed a detailed calculation to determine the 4-sidecut Minimum Required Thickness and the API RP +574 default minimum structural thickness was not used. However, had Chevron used the API RP 574 +default minimum structural thickness value of 0.11-inch as the 4-sidecut Minimum Required Thickness, +the remaining life of the piping circuit would have been predicted to be less than ten years, and a +turnaround planning group discussion should have been triggered to discuss replacement options for the +8-inch 4-sidecut piping. Such a discussion could have resulted in the decision to replace the 8-inch 4- +sidecut piping during the 2011 turnaround, and the August 6, 2012, pipe rupture could have been +prevented. +Chevron allowed adjusting the minimum thickness value in the inspection database based upon an +evaluation of existing inspection thickness data and minimum structural thickness calculations. However, +its inspection procedures caution the inspector to validate the quality of the data. The Piping Inspection +Guideline poses the questions “Were enough measurement points taken; [a]re measurements being taken +at the right locations?” Had these questions been effectively considered, evaluation of the Chevron +Sulfidation Failure Prevention Initiative could have aided in the determination that there was not +sufficient thickness data gathered on the 4-sidecut piping to justify the minimum thickness value change. +Chevron does not require a formal multi-person review process to be performed prior to changing +minimum thickness values and remaining life predictions. Such a process may have identified that the +inspection data was unreliable and insufficient for carbon steel piping susceptible to sulfidation corrosion, +and the piping could have been replaced per Chevron’s policies during the 2011 turnaround. +Chevron’s minimum pipe thickness program is intended to obtain the maximum life out of piping, yet +replace piping before it becomes dangerously thin. However, the program allowed changes to minimum +thickness values without a formal multi-person review process and lacked sufficient oversight to ensure +the safety questions were adequately considered before minimum thickness values were altered. As a +result, the 4-sidecut piping that ultimately failed in 2012 was allowed to continue in operation following +the 2011 turnaround inspection findings. The Chevron Richmond Refinery should strengthen its +minimum pipe thickness program when determining a piping circuit’s remaining life. + + +128 This minimum thickness is specified for piping between 6 and 18 inches in diameter that operates at +temperatures under 400 ⁰F. The 4-sidecut piping operated at a higher temperature, likely requiring a greater +minimum thickness. + +Chevron Richmond Refinery Investigation Report January 2015 +78 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.6 Chevron Process Safety Indicators Program +API RP 754: Process Safety Performance Indicators for the Refining and Petrochemical Industries states, +“A comprehensive leading and lagging indicators program provides useful information for driving +improvement and when acted upon contributes to reducing risks of major hazards….”129 Indicators can +reveal safety gaps before an incident occurs. One goal of the use of indicators is to drive continuous +safety improvement. Lagging indicators are facts about previous events, such as process safety incidents, +that meet a certain severity threshold. Leading indicators are measurements that predict future +performance. They help facilities maintain safety protection layers and operating discipline by +monitoring items such as equipment selection, engineering design, and specification, technique, and +frequency of inspection.130 The CSB’s 2007 BP Texas City investigation report describes the importance +of analyzing leading and lagging indicators: +Process safety [indicators] provide important information on the effectiveness of safety +systems, and an early warning of impending catastrophic failure. The sole use of lagging +safety indicators, such as injury rates or numbers of incidents, has been described as +trying to drive down the road looking only in the rear view mirror—it tells you where you +have been but not where you are headed. Process safety good practice guidelines +recommend using both leading and lagging indicators for process safety. Leading +indicators provide a check of system functioning—whether needed actions have been +taken, such as equipment inspections completed by the target date or PSM action item +closure. Lagging indicators, such as near-misses, provide evidence that a key outcome +has failed or not met its objective. “Active monitoring” of both leading and lagging +indicators is important to the health of process safety systems.131, 132 +Chevron uses an online dashboard which was developed in 2009 called Operational Excellence and +Reliability Intelligence (OERI) to track 26 different process safety indicators. OERI visually displays the +status, represented in red, yellow, or green, of many different process safety indicators: green represents a +good indicator status, yellow identifies a couple of action items are necessary, and red represents the need +to complete many action items. Management reviews these metrics weekly and schedules monthly +meetings to discuss the yellow or red items. OERI also has the ability to project 30 days into the future to +show the status of metrics at that future date should no action be taken on these items. The Chevron +Richmond Refinery leadership team is held accountable for the status of metrics that they oversee. The +refinery manager and the president of global manufacturing meet regularly with the Chevron Richmond +Refinery leadership team to discuss status of the metrics they oversee, and this is incorporated into each +leadership team member’s performance review. +Chevron tracks the following 26 process safety indicators in the OERI database: + +129 ANSI/API RP 754: Process Safety Performance Indicators for the Refining and Petrochemical Industries. 1st ed., +Foreword, April 2010. +130 Center for Chemical Process Safety (CCPS). Guidelines for Process Safety Metrics. Section 3.1, 2010. +131 U.K. Health and Safety Executive (HSE). Developing Process Safety Indicators: A Step-By-Step Guide For +Chemical And Major Hazard Industries, 2006. +132 U.S. Chemical Safety Board (CSB). Investigation Report: Refinery Explosion and Fire (15 Killed, 180 Injured), +BP Texas City, Texas, page 185, March 2007. + +Chevron Richmond Refinery Investigation Report January 2015 +79 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +PHA Recommendation Implementation Overdue + +Safety Instrumented Systems (SIS) Functions Disabled + +SIS Functional Test Overdue + +Open Safety Work Requests + +Overdue Preventative Maintenance + +Inspections Overdue + +Overdue training + +Training due in 30 days + +Permanent MOCs Overdue + +Temporary MOCs Overdue + +Mechanical Availability + +Incident Solutions Overdue + +Investigations + +Audit Action Items + +Pre-Startup Safety Review Exceptions + +Overdue Testing of Over Speed Trips + +Overdue PRDs (Pressure relief valves) Testing + +Days Exceeding Alarm Limit + +Critical Process Variable Deviations + +Routine Duties not Completed + +Work Order Schedule Adherence + +Open Temporary Leak Repairs + +Utilization (Mechanical Utilization) + +Reliability Clock (Mechanical Reliability) + +Industrial Safety Ordinance Recommendation Implementation Overdue + +Overdue Compliance Assurance Program tasks +While Chevron’s OERI database is an excellent framework for tracking leading and lagging indicators to +continuously monitor and improve process safety, it does not track the implementation of ETC process +safety recommendations or new industry guidance as determined, for example, by Chevron technical +experts to be critical to ensuring process safety in Chevron refineries. Such an indicator could have +ensured that the ETC Sulfidation Failure Prevention Initiative and its status at the Chevron Richmond +Refinery were at the forefront of management’s attention. Including an indicator into the OERI system +for tracking the implementation of key ETC process safety recommendations or new industry guidance +will aid in preventing future incidents at Chevron refineries. +5.1.7 Stop Work Authority +Chevron’s corporate-wide Stop Work Authority policy applies to upstream drilling operations and +downstream refining and manufacturing processes. It states: + +Chevron Richmond Refinery Investigation Report January 2015 +80 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Stop Work Authority (SWA) establishes the responsibility and authority of any +individual to stop work when an unsafe condition or act could result in an undesirable +event. In general terms, the SWA process involves a stop, notify, correct, and resume +approach for the resolution.133 +In theory, Stop Work Authority is a safety critical power that workers can use to halt operations if they +see an unsafe condition or act occurring. On August 6, 2012, Stop Work Authority was not used to +require immediate and safe shutdown of the Crude Unit.134 Instead, the unit continued to operate for an +extended time, during which the potentially risky removal of insulation from the 4-sidecut pipe took +place. The CSB learned in interviews that some personnel participating in the insulation removal process +while the 4-sidecut piping was leaking were uncomfortable with the operation and the possible exposure +to flammable process fluid. Some individuals recommended that the Crude Unit be shut down, but they +did not formally invoke their Stop Work Authority. They left the final decision to the management +personnel present. One employee stated to CSB investigators: +If we can’t isolate [the 4-sidecut piping] then we’re going to, you know, we should shut +down…. At that time, once I gave my opinion, I walked away because I let the head +operator handle the decisions, right? +Stop Work Authority has been used successfully at the Chevron Richmond Refinery in unsafe work +situations (e.g., skipping a step in a procedure, working in unsafe weather conditions, wearing improper +personal protective equipment (PPE), employing improper safety precautions when working at heights). +The difficulty arises when faced with a process safety situation—a leak, vibration, process upset— +especially where shutdowns are being considered. Under these circumstances, there are significant +limitations to a Stop Work Authority initiative, the most familiar being the reliance on the individual +employee to assert a dissenting viewpoint in an atmosphere where a group of individuals may not agree. +Groups of employees working together to solve a problem can be hindered by the “group think” mindset: +Without conflict, or without enough conflict, a phenomenon called group think can result. +This occurs when group members do not express their personal opinions but rather +willingly submit to what the group as a whole thinks. Group think can lead to bad +decisions and inappropriate actions.135 + +133 See http://upstream.chevron.com/contractorgom/forms_policies/stop_work_authority.aspx (accessed July 14, +2014). +134 In its Chevron Regulator Report, the CSB recommended that California enhance and restructure its process safety +management (PSM) regulations for petroleum refineries by including specific goal-setting attributes. The +recommendation included language to strengthen stop work authority, “The regulation should provide workers +and their representatives with the authority to stop work that is perceived to be unsafe until the employer resolves +the matter or the regulator intervenes.” +http://www.csb.gov/assets/1/19/Chevron_Regulatory_Report_11102014_FINAL_-_post.pdf (accessed +December 18, 2014). +135 Society of Manufacturing Engineers. “Personal Effectiveness,” Fundamentals of Manufacturing. 3rd ed. Philip +D. Rufe, editor, 2013, page 596. + +Chevron Richmond Refinery Investigation Report January 2015 +81 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Regardless of how a Stop Work program is portrayed, there are a number of reasons why such a program +may fail related to the ‘human factors’ issue of decision-making; these reasons include belief that the Stop +Work decision should be made by someone else higher in the organizational hierarchy, reluctance to +speak up and delay work progress, and fear of reprisal for stopping the job.136 Another significant +limitation is that, by design, Stop Work Authority is a decision process embedded into the chaos of the +event itself. It becomes an option only when all other barriers have failed—often during a stressful +atmosphere such as an emergency situation. Another employee stated to CSB investigators: +We asked them, you know, shouldn’t we reevaluate this job, you know, stop it and try to +figure something else out, because I’m thinking the leak is not where you say it is. It +could be leaking up higher, and it would be a safer thing to shut this line down…. [The +operations management present] said ‘This is an emergency. We need it done right now.’ +… Everybody seemed to be in agreement that it needed to get done, and I didn’t want to +argue anymore, because I don’t want to take any flack for stopping the job myself. +These significant shortcomings of Stop Work Authority have been identified in previous CSB +investigations. The CSB’s Investigation Report analyzing the refinery fire that occurred at the Tosco +Avon Refinery in Martinez, California, on February 23, 1999 states: +Tosco management stated that workers had the authority to stop unsafe work activity +and should have stopped the line replacement job. However, stop work authority— +though a desirable safety policy if properly encouraged—is a less effective measure for +incident prevention than good job preplanning for the following reasons: + +It is exercised during the execution of work, when pressures to get the job done +are generally greater.137 + +It relies on the assertiveness of individual workers. To attempt to stop a job, a +worker may need to assert a position that runs contrary to direct instructions from +a supervisor. + +Once the job has begun the idling of contractors and equipment can result in +significant financial cost to the facility, which can add to the pressure to get the +job done without delay.138 + +136 A 2010 study by The RAD Group of 2,600 workers, primarily oil and gas service employees, found that the +surveyed employees directly intervene in only 39% of the unsafe acts that they observe on the job. The study +concluded that they did not stop unsafe work because (1) they worry the person who is performing the unsafe +work will become angry or defensive and (2) they do not believe they can effectively stop unsafe work. See +Ragain, R., Ragain, P., Allen, M. & Allen, M. “Study: Employees Intervene in Only 2 of 5 Observed Unsafe +Acts,” Drilling Contractor, January / February 2011. +137 In discussing the management dilemma of production versus process safety, CCPS guidelines state: “The +continuity of operations can be best addressed at the planning stage.” See American Institute of Chemical +Engineers (AIChE), Center for Chemical Process Safety (CCPS). Plant Guidelines for Technical Management of +Chemical Process Safety. 1995c. page 17. +138 U.S. Chemical Safety Board. Investigation Report: Refinery Fire Incident (4 Dead, 1 Critically Injured), Tosco +Avon Refinery, March 2001, page 43. See http://www.csb.gov/assets/1/19/Tosco_Final_Report.pdf (accessed +November 14, 2013). + +Chevron Richmond Refinery Investigation Report January 2015 +82 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Rather than relying on Stop Work Authority after an emergency process safety situation is identified, a +more effective process is to rely upon formal procedures that reduce reliance on the individual,139 for +example, having an established predetermined leak response plan. One should not rely on Stop Work +Authority as a safeguard because it is not a formal procedure. Rather, it is a “residual reduction” +technique, falling below “procedural safeguards” on the hierarchy of controls (Figure 23). With specific +decision-making criteria in place, those responding to an emergency process safety event should not have +to evaluate risk in the heat of an event, but only determine whether the event meets the predetermined +criteria to stop operations and shut down a unit. +At the time of the incident, Chevron did not recognize or accommodate the shortcomings of Stop Work +Authority in averting major process hazards. The Chevron Stop Work Authority program was not +designed to assist operations and emergency response personnel in determining whether taking aggressive +emergency response actions to remove insulation from a leaking pipe was a wise decision. Since the +incident, Chevron has created a Leak Response Protocol (Section 5.3.4) to lead emergency responders, +operators, and other plant personnel in deciding how to handle a leaking pipe. Used effectively, this +protocol could alleviate pressure from individuals to rely on their Stop Work Authority during potentially +hazardous process operations. Other refiners and petrochemical producers should also take such action to +develop process Leak Response Protocols for their facilities to help prevent incidents like the August 6, +2012, Chevron Richmond Refinery pipe rupture and fire. + + + +139 “Experience indicates that effective systems require quite a high degree of formality. The purpose of these +systems of work is to ensure a personal and collective discipline, to exploit the experience gained by the +organization, and to provide checks to minimize problems and errors. The framework of such systems is +typically a set of standing orders or instructions which lay down requirements for the conduct of particular +activities.” Mannan, Sam. “Management and Management Procedures.” Lees’ Loss Prevention in the Process +Industries: Hazard Identification, Assessment and Control. Volume 1, 4th ed., 2005, page 6/5. + +Chevron Richmond Refinery Investigation Report January 2015 +83 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.1.8 Chevron Organizational Conclusions +The CSB found that Chevron management, engineers, inspectors, and operators all see the importance of +having good process safety systems and the value of ensuring that work processes are safe and equipment +is reliable. (See Section 5.1.) Despite this mindset and the existing programs, the Chevron Richmond +Refinery was unsuccessful in preventing the 4-sidecut pipe from rupturing. A desire to be safe is not +enough; to ensure process safety, organizations must have a well-designed, integrated system, rigorous +programs, and strong leadership for these programs. +The critical flaw in Chevron’s safety programs is their reliance largely on individual personnel assertions +and initiatives to implement new important safety programs—a bottom-up approach. While this can +occasionally be a successful method, it is not a reliable way to implement safety-critical programs. Lees’ +Loss Prevention in the Process Industries states: +It is fundamental that responsibility for [safety and loss prevention] should be shared by +all concerned in the project…. This does not mean, however, that reliance should be +placed simply on individual competence and conscientiousness. It is essential to support +the competent people with appropriate systems of work. Experience indicates that +effective systems require quite a high degree of formality.140 +To get the necessary work implemented, the ETC Sulfidation Failure Prevention Initiative and other ETC +training course recommendations relied on the persuasive abilities of individual inspectors and +metallurgists—who did not have final decision-making and funding ability. The Chevron turnaround +IMPACT process relied solely on a data-driven decision process that did not account for all information, +such as ETC publications and industry best practices. Inclusion of any yet-to-be justified work into +Chevron Richmond Refinery turnarounds, even under the auspices of industry guidance and company +experts, required the willingness of individual engineers or inspectors to step forward and advocate for the +effort in the face of an already official IMPACT core team rejection. +The Unit Reliability Improvement Process and the Business Improvement Network, among other +programs, were ineffective in encouraging implementation of initiatives, such as the ETC Sulfidation +Failure Prevention Initiative. Even analytical programs based on hard data and analysis, such as the +Richmond Refinery Piping Inspection Guideline and the Operational Excellence and Reliability +Intelligence program were not successful. And when these failures resulted in a dangerous leak in the 4- +sidecut line, the Stop Work Authority program was ineffective because it relied on individuals to step out +of the group-think mindset to persuade others that insulation removal might be dangerous. +The failure to prevent this incident is indicative of a fragmented process safety management approach that +placed responsibility to implement key process safety recommendations on lower-level employees +without sufficient recommendation-approval and funding authority. These systems might have been +successful in other incidents before August 6, 2012. However, depending on non-formalized individual + +140 Mannan, Sam. “Management and Management Procedures.” Lees’ Loss Prevention in the Process Industries: +Hazard Identification, Assessment and Control. Volume 1, 4th ed. 2005, page 6/5. + +Chevron Richmond Refinery Investigation Report January 2015 +84 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +employee performance to eliminate low frequency, high consequence events like the August 6, 2012, pipe +rupture is often ineffective when systemic failures are present. +Chevron can ensure the effectiveness of implementing new safety-critical programs at the refinery level, +such as the ETC Sulfidation Failure Prevention Initiative, by developing a formalized system that +identifies one individual or group with decision-making authority within each refinery to be responsible +and accountable for program implementation. The implementation efforts can then be tracked as a +leading indicator, such as in Chevron’s OERI system. CCPS’s Plant Guidelines for Technical +Management of Chemical Process Safety states: +Each technical element in a process safety management program needs to have a specific +person or organizational unit clearly designated as responsible for its design, +implementation, and maintenance as well as for proper review. Having this designated +“champion” for the activity helps assure that it receives adequate management attention +and support.141 +At the Chevron Richmond Refinery, many individuals attempted to implement the ETC Sulfidation +Failure Prevention Initiative and ETC training course recommendations either to 100 percent component +inspect sulfidation-susceptible carbon steel piping or to replace sulfidation-susceptible carbon steel piping +with an inherently safer, higher chromium material of construction. However, none of these individuals +were held accountable for the implementation status of the ETC Sulfidation Failure Prevention Initiative +or other ETC sulfidation prevention recommendations, nor did they have the authority to ensure the +initiative and recommendations were implemented. As a result, the 4-sidecut piping was never 100 +percent component inspected, nor was it ever upgraded to higher chromium steel before the incident. + + + + +141 Center for Chemical Process Safety of the American Institute of Chemical Engineers. Plant Guidelines for +Technical Management of Chemical Process Safety. Revised Edition, Appendix 2A, 1995, page 10. + +Chevron Richmond Refinery Investigation Report January 2015 +85 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.2 Industry Sulfidation Corrosion Guidance +Industry organizations and trade associations, such as the American Petroleum Institute (API), the +American Society of Mechanical Engineers (ASME), and the National Fire Protection Association +(NFPA), develop codes, standards, and recommended practices which define requirements and +recommendations to conduct operations safely. Codes, standards, and recommended practices are +developed by a committee of experts on the basis of consensus and are often updated on fixed-year +intervals. Codes can be adopted as requirements by regulatory agencies or authorities having +jurisdiction.142 In addition, since these requirements are often considered Recognized and Generally +Accepted Good Engineering Practices (RAGAGEP), regulators can cite industrial facilities for not +following them. +Codes and standards developed by API give specific information and guidance to industry on the +technical details of sulfidation corrosion and ways to inspect piping and equipment susceptible to it. +However, the CSB identified significant gaps in these standards. There is varying, sometimes conflicting +information in many of API’s standards and recommended practices that describe sulfidation corrosion. +All of these publications should align to deliver a constant message to users on inspecting for sulfidation +corrosion and preventing sulfidation failures in low-silicon carbon steel. +5.2.1 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion +Failures in Oil Refineries +API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries is the +primary industry guidance document on ways to monitor and control sulfidation corrosion. It aims “to +provide practical guidance to inspectors, maintenance, reliability, project, operations and corrosion +personnel on how to address sulfidation corrosion in petroleum refining operations.”143 It was published +in 2009 following a string of sulfidation corrosion-related incidents in the early 2000s. +The recommended practice cautions that low-silicon carbon steel piping can corrode at an accelerated +rate.144 It states that carbon steel will appear to be of sufficient thickness based upon measured corrosion +rates, typically at CMLs placed on elbows and fittings with higher silicon content,145 until failure occurs +at an unmonitored or unidentified low-silicon piping component.146 +API RP 939-C specifically discusses risks associated with sulfidation corrosion in low-silicon carbon steel +piping. It acknowledges that older carbon steel piping can have low silicon content, creating: + +142 American Petroleum Institute: Procedures for Standards Development. 4th ed., 2009, Section 5.4. See +http://www.api.org/publications-standards-and-statistics/~/media/Files/Publications/FAQ/2011-Procedures- +Final.ashx (accessed September 15, 2014). +143 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries, Introduction, +2009. +144 Ibid., Section 6.2.3.2. +145 Ibid., Section 7.1.5. +146 Ibid., Section 6.2.3.2. + +Chevron Richmond Refinery Investigation Report January 2015 +86 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +a major inspection challenge, because small piping sections (pups) or fittings with low +[silicon] may corrode at rates 2 to 10 times faster than surrounding higher [silicon] +piping. Unless the refinery is fortunate enough to have located an inspection point on +that particular section of pipe or fitting, it is very difficult to detect the thinning +component.147 +The document also communicates the risk of sulfidation corrosion failures, stating “ruptures are possible +leading to the potential release of large quantities of hydrocarbon streams,”148 and sulfidation corrosion +“continues to be a significant cause of leaks leading to equipment replacements, unplanned outages, and +incidents associated with large property losses and injuries.”149 It shows an example of a rupture that +occurred due to unmonitored low-silicon carbon steel components (Figure 38 and Figure 39). + +Figure 38. Photo from API RP 939-C of a low-silicon pup piece that ruptured at a BP refinery. The +surrounding piping had higher silicon content, and the pup piece’s accelerated corrosion rate was +unmonitored. + +147 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries, Section 9, +2009. +148 Ibid., Section 4. +149 Ibid., Introduction. + +Chevron Richmond Refinery Investigation Report January 2015 +87 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 39. Schematic from API RP 939-C of the piping shown in Figure 38. The piping component that +ruptured was significantly thinner than the surrounding piping. +Despite the known risks of unmonitored sulfidation corrosion rates in potentially low-silicon carbon steel +piping components, the CSB found that API RP 939-C specifically refrained from requiring companies to +search for low-silicon piping components in their facilities. All guidance given on methods for +identifying these components is written in a permissive way that does not require action by the operating +companies. Specifically, API RP 939-C requires no action by the operating companies, as it states: + +“Some refiners have instituted an approach similar to [positive material identification (PMI)] for +identification of these materials. These approaches may involve an initial risk assessment to +focus inspections on the circuits representing the highest risk. When mill certificates are + +Chevron Richmond Refinery Investigation Report January 2015 +88 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +available, some operators have used them to determine whether low-Si steels were procured and +will try to locate the low-Si spools.”150 + +“Many field portable instruments used for PMI cannot identify silicon to the level needed to +distinguish between high and low-Si-containing steel. Chemical verification requires that metal +shavings of all components be taken and analyzed in a lab.”151 + +“As an alternative, insulation can be stripped and each piping segment can be exposed for UT +inspection…. […] If a low-Si content material is identified, a risk assessment should be +performed to determine if and when it should be replaced. Unless all components in a carbon +steel system have been checked for either silicon content or thickness, the inspector should +assume that low-Si steel may be present in the system and may corrode at much higher than +nominal rates under some conditions.”152 +While API RP 939-C informs the user that it “is preferable to specify higher alloy for better corrosion +resistance to minimize the reliance on inspection,”153 it does not recommend that the user take such an +approach. Susceptible piping contains sulfur species and operates between 450°F and 1000°F.154 +Upgrading to a steel alloy that contains at least 9 percent chromium is an inherently safer choice in high- +temperature sulfidation environments, and it is higher in the hierarchy of controls155 than inspection. +High-chromium steels corrode due to sulfidation at a much slower rate than carbon steel and do not run +the risk of extreme variations in corrosion rates within components of the same piping circuit.156 +To ensure that a low-silicon carbon steel rupture does not again occur at a U.S. refinery, API RP 939-C +should establish minimum requirements to prevent another catastrophic incident as a result of pipe rupture +in low-silicon carbon steel piping. It should require 100 percent component inspection of existing in- +service carbon steel piping susceptible to sulfidation corrosion that could contain low-silicon +components,157 and should recommend users to replace carbon steel piping susceptible to sulfidation +corrosion that could contain low-silicon components with a steel alloy that is more resistant to sulfidation +corrosion to avoid the necessity to perform the 100 percent component inspection. Had API RP 939-C +phrased these enhanced inspection strategies as requirements before the Chevron August 6, 2012, pipe +rupture, the 100 percent component inspection and material of construction upgrade would not have been + +150 API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries, Section +7.1.5, 2009. +151 Ibid. +152 Ibid. +153 Ibid., Section 7.1.9. +154 Ibid., Section 1. +155 An effectiveness ranking of techniques used to control hazards and the risk they represent as a hierarchy of +controls. See U.S. Chemical Safety Board. Interim Investigation Report: Chevron Richmond Refinery Fire. April +2013. http://www.csb.gov/assets/1/19/Chevron_Interim_Report_Final_2013-04-17.pdf (accessed November 15, +2013). +156 U.S. Chemical Safety Board. Interim Investigation Report: Chevron Richmond Refinery Fire. April 2013, page +22, paragraph 20. http://www.csb.gov/assets/1/19/Chevron_Interim_Report_Final_2013-04-17.pdf (accessed +November 15, 2013). +157 Two techniques are used to inspect a component in an existing carbon steel piping circuit with unknown +chemical composition for low silicon content and resulting variable corrosion rates: (1) performing laboratory- +based chemical analysis of the carbon steel (a “destructive test,” meaning it requires removal of a sample of the +steel), or (2) performing pipe wall thickness measurements. + +Chevron Richmond Refinery Investigation Report January 2015 +89 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +considered “discretionary work items,” and Chevron management likely would have ensured that its +refineries complied with the API requirements. +5.2.2 API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining +Industry +API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining Industry is a summary +guidance document on the “most likely damage mechanisms affecting common alloys used in the refining +and petrochemical industry and is intended to introduce the concepts of service-induced deterioration and +failure modes.”158 Sulfidation corrosion is one of the 66 damage mechanisms159 summarized in the +recommended practice. +API RP 571 includes API RP 939-C as a reference document. However, the body of API RP 571 poses +several clarity problems in its discussion of sulfidation corrosion: + +Section 4.4.2, Sulfidation, summarizes API RP 939-C in a condensed format. While it does +inform the reader that sulfidation corrosion can occur as localized corrosion160 and resistance is +achieved by upgrading to a higher chromium alloy,161 its description of high-corrosion rate +problems in low-silicon carbon steel is sparse. In fact, silicon is mentioned only once in this +section, and this mention is in a figure caption,162 not in the text body. + +Section 4.4.2.3, Critical Factors, does not explain that low-silicon carbon steel piping corrodes at +a much faster rate than higher silicon carbon steel. + + Section 4.4.2.5, Appearance or Morphology of Damage, does not specify that sulfidation +corrosion rates can be significantly faster in just a few, individual piping components. + +Section 4.4.2.7, Inspection and Monitoring, does not specify that 100 percent component +inspection is necessary to identify any low-silicon components in a carbon steel piping circuit. +API RP 571 should become more aligned with the content in API RP 939-C so that the information is +presented to users in a more consistent way. + + +158 API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining Industry, 2nd ed., Section 1.2, +2011. +159 Ibid., Table of Contents. +160 Ibid., Section 4.4.2.5. +161 Ibid., Section 4.4.2.6. +162 Ibid., Figure 4-117. + +Chevron Richmond Refinery Investigation Report January 2015 +90 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.2.3 API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and +Alteration of Piping Systems +API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems +specifies “the in-service inspection and condition-monitoring program that is needed to determine the +integrity of piping.”163 It discusses different inspection strategies that can be employed based on the type +of damage mechanism the piping is susceptible to. The latest version of this document was released six +months following the release of API RP 939-C, yet it fails to mention the localized corrosion possibilities +in carbon steel susceptible to sulfidation corrosion, nor does it include API RP 939-C as a normative +reference.164 +Section 5.6.3, CML Selection, provides general guidance to inspectors for determining where to place +piping CMLs for various corrosion mechanisms. It states: +A number of corrosion processes common to refining and petrochemical units are +relatively uniform in nature, resulting in a fairly constant rate of pipe wall reduction…. +Examples of such corrosion phenomena include high-temperature sulfur corrosion …. In +these situations, the number of CMLs required to monitor a circuit will be fewer than +those required to monitor circuits subject to more localized metal loss.165 +These statements directly oppose the inspection techniques required to identify low-silicon components in +sulfidation-susceptible carbon steel piping circuits. Sulfidation corrosion in carbon steel piping can be +localized to only a few components; therefore, 100 percent component inspection is required to identify +low-silicon components. In addition, the use of “high-temperature sulfur corrosion” nomenclature rather +than “sulfidation corrosion” (as is used in other API publications) can lead to confusion among users. +To provide a consistent description of sulfidation corrosion throughout all pertinent API documents, +content and nomenclature should be aligned. Sulfidation corrosion should be referenced using the same +terminology in all API publications, and the potential for localized corrosion must also be emphasized. In +addition, because API 570 is a piping inspection code, this document should also establish the 100 percent +component inspection requirements necessary to identify low-silicon components in carbon steel piping +circuits susceptible to sulfidation corrosion. + +163 API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems, +Section 1.1.2, 2009. +164 A “normative reference” is a reference to another code, standard, recommended practice, or regulation that +provides additional useful information. +165 API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems, +Section 5.6.3, 2009. + +Chevron Richmond Refinery Investigation Report January 2015 +91 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.2.4 API RP 578: Material Verification Program for New and Existing Alloy +Piping Systems +API RP 578: Material Verification Program for New and Existing Alloy Piping Systems “provides the +guidelines for material control and material verification programs….”166 This document does include API +RP 939-C as a normative reference167 and discusses specifically the increased susceptibility of low-silicon +carbon steel to sulfidation corrosion. Section 4.3.3.4, Process Units Susceptible to Sulfidation, states: +Carbon steels with low silicon (<0.10%) content can corrode at an accelerated rate when +exposed to hydrogen-free sulfidation conditions. These phenomena are discussed more +extensively in API 571 and API 939-C. Owner/users with assets at risk from this type of +degradation should consider the risks and the requirements to apply [positive material +identification] control in order to determine silicon levels and the extent to which the +material may corrode.168 +While this guidance document does describe the accelerated corrosion rate that occurs in low-silicon +carbon steel piping, it does not require facilities to establish and implement a program to identify low- +silicon components that may not have been manufactured to ASTM A106 requirements in carbon steel +piping circuits susceptible to sulfidation corrosion. Such a requirement is essential for facilities to +successfully identify low-silicon components susceptible to sulfidation corrosion to prevent future pipe +rupture incidents similar to the August 6, 2012, Chevron incident. +5.2.5 API RP 574: Inspection Practices for Piping System Components +API RP 574: Inspection Practices for Piping System Components “supplements API 570 by providing +piping inspectors with information that can improve skill and increase basic knowledge and practices.”169 +The recommended practice lists 50 reference documents, including other API standards, that are +“indispensible for the application”170 of the recommended practice. Several of these documents +specifically discuss certain damage mechanisms. However, API RP 939-C is not listed among the +reference documents. In addition, API RP 574 does specifically inform the reader that corrosion rates can +be localized in carbon steel piping: “Nonsilicon-killed steel pipe (e.g. ASTM A53 and API 5L) can +corrode at higher rates than silicon-killed steel pipe (e.g. ASTM A106) in high-temperature sulfidation +environments.”171 However, it does not specifically point the reader to API RP 939-C to learn more +information. To align the messages presented in all piping inspection-related API guidance documents, +API RP 574 should refer the reader to the more enhanced information in API RP 939-C to increase +understanding of important sulfidation corrosion characteristics and failure prevention strategies. + +166 API RP 578: Material Verification Program for New and Existing Alloy Piping Systems, 2nd ed., Section 1, +2010. +167 API RP 578: Material Verification Program for New and Existing Alloy Piping Systems. 2nd ed., Section 2, +2010. +168 Ibid., Section 4.3.3.4. +169 API RP 574: Inspection Practices for Piping System Components, 3rd ed., Section 1, 2009. +170 Ibid., Section 2. +171 Ibid., Section 7.4.6.2. + +Chevron Richmond Refinery Investigation Report January 2015 +92 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.3 Chevron Emergency Response to Process Leaks +Following the identification of the leak in the 8-inch 4-sidecut piping circuit on August 6, 2012, a series +of decisions ultimately put many people in harm’s way. Chevron has since developed a leak response +protocol to be used when determining how to respond to future leaks in the refinery. +5.3.1 Area Control and Hazardous Area Assessment +While the 4-sidecut line was leaking high temperature flammable process fluid, 40 individuals entered the +Crude Unit. Many of these individuals entered the unit to assist in determining how to handle the leak. +Individuals who entered the Crude Unit included the hydroprocessing refinery business manager, the +Crude Unit section head, 14 operations personnel, the process engineering team lead, a field safety +coordinator, two inspectors, the pipe clamp contractor, three scaffold builders, a pipe fitter, and 15 +firefighters. API RP 574: Inspection Practices for Piping System Components warns: “Those who +investigate [on-stream piping leaks] may be particularly at risk to the consequence associated with release +of the process fluid.”172 A safer practice is to establish a safe location away from the active process leak +to perform an analysis of the situation and to determine a path forward. Occupational Safety and Health +Administration (OSHA) regulations for emergency response require limiting the number of personnel in +the immediate vicinity of the incident “to those who are actively performing emergency operations.”173 +Typically, management coordination is provided from an emergency operations center located remotely +from the hazards of the emergency situation.174 +When Chevron fire department personnel took control of the leak response, the fire fighters created and +taped off a 20 foot by 20 foot “hot zone” around the leak location. Chevron defines a hot zone as “the +immediate release area [emphasis in the original] of the incident where there is risk of exposure or +injuries due to flame contact, radiant heat, or hazardous materials.” The size and location of the hot zone +is determined by the Incident Commander.175 For this incident, Chevron had also established a “cold +zone” immediately outside of the hot zone perimeter. In this area, key operations staff, additional fire +department staff, and the Incident Commander were positioned to provide expert support during the leak +response decision-making. The Chevron fire truck that was ultimately destroyed in the fire was also +located in the designated “cold zone.”176 +When the 4-sidecut pipe ruptured, a very large vapor cloud formed which engulfed all personnel both +within the hot zone and standing in what was considered the “cold zone,” or the area where personnel +should be safe from the adverse affects of a fire.177 The hot zone designated prior to the pipe rupture was +not of sufficient size to ensure that individuals outside of the hot zone were safe from the high- +temperature, flammable 4-sidecut process fluid when the sulfidation failure of the pipe occurred. As + +172 API RP 574: Inspection Practices for Piping System Components, 3rd ed., Section 9.3, 2009. +173 29 CFR §1910.120(q)(3)(v) (2012). +174 API RP 2001: Fire Protection in Refineries, 9th ed., Section 9.2, 2012. +175 Chevron defines the Incident Commander as the “senior emergency response official” at an incident site who is +responsible for overall incident objectives and controlling emergency operations at the site. Additional +responsibilities include site hazard assessment to the extent possible all hazardous substances or conditions +present are identified, establish hot zone, and address exposure control and PPE selection.” +176 The fire truck was positioned approximately 65 feet from the leak location. +177 National Fire Protection (NFPA) 600: Standard on Industrial Fire Brigades, Section 3.3.30, 2010. + +Chevron Richmond Refinery Investigation Report January 2015 +93 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +discussed in subsequent sections of this report, it is beneficial for various personnel with differing areas of +process expertise to report their knowledge of the leak properties and potential causes of the leak to the +Incident Commander. Had this been effectively communicated on the day of the incident, the Incident +Commander might have been informed that the 4-sidecut piping had the potential to catastrophically +rupture. This information could have led the incident command team to establish a much larger hot zone +area. +5.3.2 Miscommunication regarding 4-sidecut properties +The CSB found that many personnel responding to the leaking 4-sidecut pipe were not properly informed +through information disseminated in the Incident Command structure of the operating temperature of the +line. Interviews show that some firefighters believed the line was operating at a temperature of about +130°F rather than the actual temperature which approached 640°F. The CSB identified that this +misunderstanding might have occurred because, during the initial incident response, much of the focus +was on determining the flash point of the 4-sidecut fluid. Little to no discussion occurred about the actual +operating temperature of the 4-sidecut line, which could have provided the most insight into the hazards +of the situation. Furthermore, in a “Scene Safety and Action Plan” that was developed immediately +before the leak response, the “Hazard Evaluation” section only identified as a hazard the pressure of the +4-sidecut line, recorded as 25 psi. The temperature of the piping circuit was not recorded. This +inattention to the temperature hazard likely resulted in the miscommunication and misunderstanding of +the actual operating temperature of the piping. +Following the incident, Chevron Fire Department personnel developed an “Event Critique,” which, in +part, was used to document areas that did not go well during the response activities. The Event Critique +states, “Somewhere in the process, impression was given to [the Chevron Fire Department] that [the 4- +sidecut temperature] was only 130 degrees F. [The Chevron Fire Department] [b]elieved [the +temperature] to be far below the autoignition temperature and below [the flash point].” Regarding the +white vapor formation that occurred during insulation removal (referred to as “smoke”), the Event +Critique also states that the firefighters felt they “[n]ever put two and two together that additional smoke +[formation was] because product [was] much hotter than 130 degrees.” CSB interviews indicate that had +the responders been aware of the actual operating temperature, some likely would have raised concerns +about the safety of removing insulation from the hot, leaking piping and concerns regarding the +responders’ close proximity to the leak to their supervisors. +In addition, no individuals determining how to handle the leak were aware that the leak was coming from +a pipe component that had thinned so severely so that it could no longer contain the process fluid. All +emergency response activities were conducted under the assumption that the 4-sidecut pipe was of +acceptable thickness beneath the insulation. This incorrect conclusion was based on measurements of an +adjacent CML—a high silicon-containing elbow—which had been found to be of acceptable thickness +during the 2011 turnaround. +The OSHA Hazardous Waste Operations and Emergency Response (HAZWOPER) standard states that +the Incident Commander “shall identify, to the extent possible, all hazardous substances or conditions + +Chevron Richmond Refinery Investigation Report January 2015 +94 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +present.”178 The Incident Commander, as well as the individuals providing technical input to the Incident +Commander, did not realize that the leak could be due to a highly thinned, low-silicon carbon steel piping +component that could exist within the 4-sidecut piping circuit. Had the potential for a pipe thinned to the +point of leak and loss of containment been recognized and communicated, the emergency response would +likely have been handled very differently. The group might have decided to immediately shut down the +Crude Unit. +The firefighter post-incident Event Critique document highlights the need for all individuals to be made +aware of operating conditions and potential failure modes, stating that hazard assessments performed +prior to response activities must ask “all the proper questions for the hazard being addressed.” This +comment indicates that the Incident Command structure did not have sufficient technical expertise +reporting to it to provide the necessary information to determine the safest response to the leaking pipe. +In this case, that response would have been to shut down the Crude Unit. +5.3.3 Leak Mitigation and Discovery Attempts Worsened Leak +Operations and fire department personnel discussed their options to stop the leak. Since the 4-sidecut line +could not be isolated from the process, the team decided their options were to 1) install a clamp on the +leak while the 4-sidecut line was operating, or 2) shut down the unit to stop the leak and perform +maintenance while the line was not in operation. The decision-makers tried to visually confirm the leak +location by removing insulation covering the piping before determining whether to clamp the line or shut +down the unit. However, attempts to remove the insulation actually worsened the leak, resulting in the +ultimate pipe rupture and endangerment to the lives of everyone responding. Post-incident metallurgical +analysis indicates that the firefighter pike pole used in an attempt to remove insulation may have stabbed +through the highly thinned pipe, worsening the leak (Figure 40 and Figure 41). + + +178 29 CFR §1910.120(q)(3)(ii) (2012). + +Chevron Richmond Refinery Investigation Report January 2015 +95 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 40. Photo of undamaged179 (top) and burned during incident (bottom) pike pole used in early attempts +to remove 4-sidecut insulation. + +179http://www.safetyfirstweb.com/firefighting/accessories.html?page=shop.product_details&flypage=flypage.tpl&pr +oduct_id=2411&category_id=96 (accessed July 14, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +96 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 41. Photo showing that tip of fire pole matches apparent puncture location in failed 52-inch +component of 4-sidecut piping. +To note, because the 4-sidecut leaking component was so thin, clamping the line would not have been a +viable option because the pipe likely did not have the structural integrity to support a clamp. However, +the potential for a thin pipe was not known by or communicated to the responding operators and Chevron +fire fighters. In addition, the high-temperature of the piping and the process fluid introduced a significant +hazard to both the individuals who would install the clamp and to individuals who would work near the +piping before the clamp installation. If the decision-makers had been aware that the leak might have +resulted from pipe thinning to the point of loss of containment, they would have been more likely to shut +down the unit without removing the insulation, and this incident could have been prevented. In addition, +had it been unacceptable within the Chevron organization to allow high-temperature, flammable process +lines to continue leaking until a clamp could be installed, the unit would have been shut down, effectively +preventing this incident. + +Chevron Richmond Refinery Investigation Report January 2015 +97 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.3.4 Chevron’s New Leak Response Protocol +OSHA requires that all individuals and organizations performing emergency response operations to +follow the HAZWOPER standard.180 In part, HAZWOPER details the organizational structure and +response elements to be performed when planning for and responding to an emergency. It requires +facilities to develop an Emergency Response Plan that “shall be developed and implemented to handle +anticipated emergencies prior to the commencement of emergency response operations.”181 Pursuant to +this requirement, Chevron developed an emergency response plan comprised of a number of Refinery +Instructions182 outlining the required response activities for specific types of emergencies. For example, +the Chevron Richmond Refinery developed Refinery Instructions to be followed in the case of a fire in +the refinery, oil spills to the adjacent bay, and releases of hydrogen sulfide, among other emergency +situations. However, at the time of the incident, the refinery did not have a specific Refinery Instruction +on how to assess and respond to hazardous process fluid leaks in the refinery. Chevron did not recognize +this gap in the Chevron Richmond Refinery emergency response plan before the incident. In addition, +current industry guidance on developing response plans to hazardous process fluid leaks is limited +(Section 5.4). +During emergency response operations, such as to a process leak, it is difficult to recall all of the +necessary safety precautions and to ensure they are performed. It is therefore essential that process safety +emergency response procedures are pre-established and followed to ensure that all safety critical steps are +taken before performing any mitigation attempts. If Chevron had used a pre-established response +procedure that required consultation with various subject-matter experts (Figure 42), the Incident +Commander could have identified that sulfidation-induced failures have historically resulted in large +blowouts and catastrophic failures.183 A clamp to mitigate a sulfidation leak would thus have to +encompass the entire affected piping segment. The time required to engineer and build a clamp for this +type of specialized application can be several days. In addition, the 4-sidecut process fluid was at a +temperature near 640°F, which made it very hazardous to work on while in operation. It also meant that +if the anticipated catastrophic-type sulfidation failure did occur, the resulting large release would +potentially auto ignite. The CSB analysis suggests that had an effective leak response protocol been in +place during the August 6, 2012, incident, it likely would have been clear there was little to no chance that +the 4-sidecut leak could be stopped and that the Crude Unit should not continue to operate for any +significant period. The analysis of the end result for every potential leak mitigation action likely would +have resulted in the same decision: to shut down the Crude Unit immediately. + +180 29 CFR §1910.120. +181 29 CFR §1910.120(q)(1) (2012). +182 Chevron calls its important internal, refinery-wide policies and procedures “Refinery Instructions.” +183 Chevron ETC Sulfidation Failure Prevention Initiative. See the block quote in Section 5.1.1.1 of this report. + +Chevron Richmond Refinery Investigation Report January 2015 +98 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Figure 42. Ideal communication flow to Incident Commander during refinery process fluid leak incident. In +industrial process fluid leak emergency situations, it is essential that various personnel with different areas of +expertise communicate their relevant knowledge to the Incident Commander. +Since the incident, Chevron has developed a leak response protocol184 to assist operators and fire +department personnel when they are deciding how to handle a process leak. The protocol will assist +Incident Command in identifying and gathering the pertinent process information prior to the +performance of any aggressive action. This new protocol has been incorporated into the Chevron +Richmond Refinery emergency response plan as a Refinery Instruction that must be followed when a +potentially hazardous process leak is identified in the refinery. +Under the new protocol, when a process fluid leak is identified in a Chevron refinery, several steps must +be taken immediately. The Chevron emergency response team must deploy to the leak site, the area must +be cleared of non-essential personnel, and a group of individuals with various areas of expertise (e.g., +operators, managers, unit inspectors, materials engineers, and chemical engineers) must gather in a safe +location to discuss the likely cause of the leak and mitigation options. + +184 The entire Chevron leak response protocol is presented in Appendix A. + +Chevron Richmond Refinery Investigation Report January 2015 +99 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Chevron’s new leak response protocol also includes a checklist to be completed during the pre-response +meeting. It requires the group to answer questions including: + +What is the likely cause [of the leak]?;  + +What is the current operating pressure and temperature?;  + +What is the pertinent pipe, vessel, structural integrity, or corrosion history of this leak?;  + +Is the product at or above its auto ignition temperature?;   + +Is the leak toxic?   +This checklist facilitates evaluating all potential hazards when determining how to handle leaks in the +refinery. This identification of hazards should allow the Incident Commander to more effectively +determine an appropriately sized “hot zone” and safe distances to stage key equipment, such as +responding fire engines. +The new leak response protocol also provides examples of scenarios when it is required to shut down the +unit. (See Figure 43) Effectively using it greatly reduces some of the inherent human factors concerns at +play when invoking Stop Work Authority, as discussed in Section 5.1.7. + +Figure 43. Chevron’s new Leak Response Protocol, developed post-incident. The new protocol advises plant +personnel on when to shut down a unit due to a piping or equipment leak. + +Chevron Richmond Refinery Investigation Report January 2015 +100 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.3.5 Leak Response Conclusions +The piping rupture and subsequent hydrocarbon release occurred two hours after the original leak was +identified, which would have been enough time to execute an emergency shutdown of the Crude Unit, or +at a minimum, to initiate the shutdown and bring the unit to a much safer condition with no liquid in the +4-sidecut line. Early in this incident, Chevron personnel determined that the 4-sidecut pipe could not be +isolated from the crude column. This resulted in the decision to reduce the feed rate to the Crude +Column. However, had Chevron decided to shut down the unit once staff knew the line could not be +isolated, the pipe rupture and the endangerment of the community and Chevron personnel could have +been avoided. At the time of the incident, Chevron did not have procedures to direct when a unit should +be shut down. Since the incident, Chevron has developed a leak response protocol that should be used to +guide decisions in future leak incidents. If a similar leak were to occur in a Chevron refinery, the new +leak response protocol would require unit shutdown. +5.4 Industry Leak Response Guidance +API and ASME have issued several documents, discussed here, that provide guidance on leak response in +refineries and chemical plants. +5.4.1 API RP 574: Inspection Practices for Piping System Components +As specified in Section 5.2.5 of this report, API RP 574 is a guidance document for piping inspectors to +improve their skills and practices. The CSB found that this document gives the most specific guidance on +how to safely respond to leaks in refineries and chemical plants. Section 9.3, Investigation of Leaks, +provides the following information: +On-stream piping leaks in process units can occur for various reasons. Those who +investigate the leak may be particularly at risk to the consequence associated with release +of the process fluid. A site may want to create a general safety procedure to be followed +during a piping leak investigation. A further precaution is to hold a safety review before +any leak investigation. The review would consider the state of a piping system in terms of +pressure, temperature, remaining inventory of process fluids, potential damage +mechanisms and similar factors. +The safety review team should define: +a) a “hot zone” around the leak site, and establish PPE and additional firefighting +equipment requirements to perform work inside this zone; +b) decontamination requirements upon exit from the hot zone and other requirements +necessary to protect personnel and the environment 185 +API RP 574 even cautions the reader about potential consequence escalation: “The safety review team +must be careful making assumptions about the leak’s cause. Incidents have occurred where investigative + +185 API RP 574: Inspection Practices for Piping System Components, 3rd ed., Section 9.3, 2009. + +Chevron Richmond Refinery Investigation Report January 2015 +101 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +personnel assume they knew the cause of a small leak on an operating line and were caught unprepared +when the leak suddenly became quite large.”186 +Despite its positive aspects, API RP 574 should be improved to require facilities to develop a site-specific +leak response protocol to be followed when a process fluid leak is discovered to help prevent and control +future pipe leak incidents. For instance, users, such as Chevron, are not required to follow any of the +guidance issued in Section 9.3, Investigation of Leaks. Permissive language informing users that they +“may want to create” a safety procedure, or that a safety review “would consider” certain damage +mechanisms requires no action by the operating companies. +In addition, API RP 574: + +Does not recommend limitation of site access around the leak to essential personnel only; + +Does not specify employee job functions or leak analysis roles that should be established prior to +performing the safety review of the leak (e.g., inspection staff, process engineers, metallurgical or +mechanical engineers, operators, emergency responders); + +Does not recommend evaluation of whether the leaking process fluid is near its autoignition +temperature; + +Does not recommend the determination of whether the leak is toxic; and + +Does not recommend the user to evaluate the worst-case leak scenario. +In addition, API RP 574 does not recommend the leaking piping circuit to be isolated—or recommend +unit shutdown if the piping cannot be isolated—if leak response personnel cannot prove it is safe to +continue operating the leaking line. Specifically, API RP 574 does not recommend piping isolation or +unit shutdown if: + +The leak poses immediate danger to safety, health, or the environment; or + +The leak cannot be safely isolated or mitigated while the piping circuit is in operation. +5.4.2 API RP 2001: Fire Protection in Refineries +API RP 2001: Fire Protection in Refineries “provide[s] a better understanding of refinery fire protection +and the steps needed to promote safe storage, handling, and processing of petroleum and petroleum +products in refineries. A basic principle of this standard is that fire prevention provides the fundamental +foundation for fire protection.”187 This document also “examines fire protection concepts that should be +covered in operating and maintenance practices and procedures ….”188 + +186 API RP 574: Inspection Practices for Piping System Components, 3rd ed., Section 9.3, 2009. +187 API RP 2001: Fire Protection in Refineries, 9th edition, Section 1.1, 2012. +188 Ibid., Section 1.2. + +Chevron Richmond Refinery Investigation Report January 2015 +102 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Like API RP 574, API RP 2001 gives guidance to users on techniques for responding to process fluid +leaks. Section 7.4, Loss of Containment, gives users guidance for both liquid leaks (Section 7.4.2) and +gas leaks (Section 7.4.3). Interestingly, in areas where API RP 574 is lacking, API RP 2001 fills in some +of the gaps. API RP 2001 suggests that response to control a leak should consider: +a) Protection of personnel against exposure, +b) Utilization of emergency response personnel and resources, +c) Isolation of the fuel release or leak at the upstream source, +d) Isolation of transfer medium, +e) Isolation of ignition sources, +f) Containment of product, +g) Downwind and off-site impact, +h) Displacement and/or removal of liquids still at risk, +i) Reduction of hazard zone via application of firefighting foam for vapor suppression, and +j) Development of mitigation cleanup strategies.189 +Like API RP 574, API RP 2001 does not require users to follow its good practice guidance. Rather, it +uses language like “Considerations… should include,” requiring no effective action by the operating +companies to develop their own site-specific leak response protocol. +Furthermore, API RP 2001 does not provide sufficient guidance on other safety critical leak response +actions. It does not recommend: + +Conducting a pre-response meeting with knowledgeable personnel to analyze the pressure, +temperature, remaining inventory of process fluids, or potential damage mechanisms in the piping +or equipment; + +Limiting site access around the leak to essential personnel only; + +Evaluating whether the leaking process fluid is near its autoignition temperature; + +Determining whether the leak is toxic; and + +Determining the worst-case leak scenario. + +189 API RP 2001: Fire Protection in Refineries, 9th ed., Section 7.4.1, 2012. + +Chevron Richmond Refinery Investigation Report January 2015 +103 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +5.4.3 API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and +Alteration of Piping Systems +API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems +specifies practices that must be performed regarding inspection, rating, repair, and alteration of metallic- +and fiberglass-reinforced plastic piping systems.190 In Section 8, Repairs, Alterations, and Rerating of +Piping Systems, the code specifies requirements when performing on-stream piping repairs, such as weld +repairs, installing a clamp, or wrapping the piping. However, API 570 does not require safety evaluation +of the leak—nor does it refer to any document that outlines the necessary safety precautions and +evaluations—before attempting on-stream repairs. To better align the API standards that address leak +repair and leak response, API 570 should require users to follow the process leak response safety +requirements established in other standards. +5.4.4 ASME PCC-2-2011: Repair of Pressure Equipment and Piping +ASME PCC-2-2011: Repair of Pressure Equipment and Piping is a standard that “provides methods for +repair of equipment and piping within the scope of ASME Pressure Technology Codes and Standards +after they have been placed in service.”191 It gives requirements for installing leak mitigation devices, +such as clamps, which Chevron personnel considered installing on the 4-sidecut leak location before the +pipe rupture. While the standard does discuss safety requirements before installing a clamp, they are +vague and are lacking needed safety preventative measures. +The safety requirements discussed in Article 3.6, Mechanical Clamp Repair state: +Personnel shall be aware of hazards in installing clamps on degraded components, and +shall take the necessary precautions to avoid unacceptable risks. A risk review shall be +conducted before a clamp is installed. Personnel shall take any necessary precautions to +avoid unacceptable risks. […] If the component is leaking or has the potential to leak +during installation, and if the contents are hazardous, additional precautions should be +taken and those precautions should be addressed during the pre-job hazard review +meeting (e.g., need for fresh air suit, etc.).192 +Article 2.4, Welded Leak Box Repair, gives similar safety guidance.193 This article goes on to state, “If +the component is leaking prior to repair, consideration should be given to stopping the leak prior to +welding the leak box.”194 +Article 4.1, Nonmetallic Composite Repair Systems: High-Risk Applications, states the requirements for +repairing leaks, or repairing piping and vessels that have defects from internal corrosion, among other + +190 API 570: Piping Inspection Code: In-Service Inspection, Rating, Repair, and Alteration of Piping Systems, 3rd +ed., Section 1.1.1, 2009. +191 ASME PCC-2-2011. Repair of Pressure Equipment and Piping, Section 1. 2011. +192 Ibid., Article 3.6. Sections 2.4 and 2.5. +193 Ibid., Article 2.4, Section 2.4. +194 Ibid., Article 2.4, Section 4.5. + +Chevron Richmond Refinery Investigation Report January 2015 +104 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +applications.195 It requires users to perform “an assessment of the risks associated with the defect and +repair method.” It requires users before installation to consider: +(1) Assessment of the nature and location of the defects +(2) Design and operating conditions for the component and contents (including pressure, +temperature, sizes, and combinations thereof) +(5) Hazards associated with system service +(9) Failure modes 196 +While ASME PCC-2-2011 does describe general safety precaution requirements, they are not detailed +enough to provide much value. Referencing other standards and recommended practices that give +guidance on leak mitigation and response would add significant value to the user. +5.4.5 Industry Leak Response Guidance Conclusions +Many industry standards, recommended practices, and guidance documents exist to aid refining and +petrochemical personnel and facility management in industrial leak response. However, the documents +are inconsistent, and none of them provide overall, comprehensive guidance or requirements for +operations personnel and facility management to safely respond to hazardous process fluid leaks. The +CSB found that existing API guidance language could be strengthened to control and prevent major +process fluid releases and to ensure the safety of facility personnel. +5.5 Chevron Richmond Refinery Safety Culture +The CSB found that weaknesses in the Chevron Richmond Refinery safety culture contributed to the +August 6, 2012, pipe rupture. The CSB’s investigation report on the March 23, 2005, BP Texas City +refinery incident presents the following definitions of the concept of safety culture: +The U.K. Health and Safety Executive describes safety culture as “the product of +individual and group values, attitudes, competencies and patterns of behaviour that +determine the commitment to, and the style and proficiency of, an organization’s health +and safety programs” (HSE, 2002). The CCPS cites a similar definition of process safety +culture as the “combination of group values and behaviors that determines the manner in +which process safety is managed” (CCPS, 2007, citing Jones, 2001). Well-known safety +culture authors James Reason and Andrew Hopkins suggest that safety culture is defined +by collective practices, arguing that this is a more useful definition because it suggests a +practical way to create cultural change. More succinctly, safely culture can be defined as +“the way we do things around here” (CCPS, 2007; Hopkins, 2005). An organization’s +safety culture can be influenced by management changes, historical events, and economic +pressures. + +195 ASME PCC-2-2011. Repair of Pressure Equipment and Piping, Article 4.1, Section 1.2, 2011. +196 Ibid., Article 4.1, Section 1.3. + +Chevron Richmond Refinery Investigation Report January 2015 +105 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Expanding on the above definitions, which are often applied to the petrochemical industry, the +nuclear industry has developed definitions of safety culture that can be applied to all industrial +sectors. The U.S. Nuclear Regulatory Commission (NRC) defines safety culture as “the core +values and behaviors resulting from a collective commitment by leaders and individuals to +emphasize safety over competing goals to ensure protection of people and the environment.”197 +The Institute of Nuclear Power Operations (INPO), a nuclear power industry group, further +expands upon this definition: + +Nuclear safety is a collective responsibility. The concept of nuclear safety culture +applies to every employee in the nuclear organization, from the board of directors +to the individual contributor. No one in the organization is exempt from the +obligation to ensure safety first.198 + +Discussed in the following sections are several aspects of safety culture at the Chevron Richmond +Refinery that the CSB found contributed to the occurrence of the August 6, 2012, incident. +5.5.1 Normalization of Deviance +“Normalization of deviance” is the acceptance of events that are not supposed to happen.199 Objective, +outside observers see a situation as deviant, while people inside the situation see it as normal and +acceptable.200 The August 6, 2012, pipe leak and subsequent response, as well as a previous incident in +2010, demonstrate efforts by Chevron employees to try to keep a unit operating during a hazardous leak, +suggesting a culture at the Chevron Richmond Refinery that normalized this behavior. +During response activities on August 6, 2012, Chevron firefighters performed physical actions that placed +them in hazardous conditions by removing insulation on the high-temperature 4-sidecut piping while it +was leaking flammable hydrocarbon process fluid. Even when hydrocarbon vapor visibly emerged from +the pipe and a flash fire occurred during insulation removal attempts, the group decided to continue +efforts to remove insulation from the on-stream pipe. This activity was acceptable to the individuals +making the leak response decisions on the evening of the incident. +A similar incident occurred before the 2012 incident. In April 2010, a pipe was found to be leaking on a +high-temperature jet fuel pipe in the hydroprocessing unit at the Chevron Richmond Refinery. The +operations staff reported the leak to management. However, no timely action was taken to repair the leak +or shut down the unit. Unit operators expressed serious safety concerns with keeping the unit online with +an active hazardous process leak. Nevertheless, the pipe remained in operation, still leaking, until the + +197 Nuclear Regulatory Commission Final Safety Culture Policy Statement, 76 Fed. Reg. 34773 (June 14, 2011). +198 Institute of Nuclear Power Operations. Traits of a Healthy Nuclear Safety Culture. INPO 12-012. December +2012. +199 Vaughan, Diane, The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA, +University of Chicago Press, 1996. +200 Interview: Diane Vaughan, Consulting Newsline, May, 2008. See http://www.consultingnewsline.com/Info/Vie +du Conseil/Le Consultant du mois/Diane Vaughan (English).html (accessed July 20, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +106 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +leak significantly worsened two days later. The unit was then shut down, and the leak was repaired. Both +the April 2010 and August 2012 incidents are examples of decision making that encouraged and tolerated +continued operation of a unit despite the presence of hazardous leaks in the Chevron Richmond Refinery. +5.5.2 Chevron Richmond Refinery Safety Culture Surveys +The Chevron Richmond Refinery facilitated safety culture surveys of its staff.201 The surveys were +designed as tools “by which the [Chevron Richmond Refinery] personnel’s perceptions about safety are +revealed, explored, and developed.” One company conducted the safety culture surveys of Chevron +Richmond staff in 2008 and 2010,202 providing the opportunity to identify any areas in which workers’ +perceptions changed significantly and areas the Chevron Richmond Refinery may need to improve. The +employees’ responses were divided into five groups based upon job categories: Operators and Mechanics, +First Line Supervisors, Second Line Supervisors, Managers and Engineers, and “Other.” The number of +employees surveyed and the job functions of the respondents are shown in Table 2. +Refinery Job Position +2008 Survey +Number of Respondents +2010 Survey +Number of Respondents +Operator +208 +163 +Mechanic +181 +202 +First Line Supervisor +53 +103 +Second Line Supervisor +47 +66 +Manager +18 +46 +Engineer +29 +93 +Other +125 +263 +Total Respondents +661 +936 + Table 2. Total number of employees surveyed and job functions of respondents in 2008 and 2010 Chevron +Richmond Refinery staff safety culture surveys. +Two types of comparison data were collected during these surveys. First, for each topic in the survey, the +employees selected a statement supplied in one of four categories, in order of improving safety culture, to +indicate their perception of the culture within the refinery. In the tables in the following sections, these +selections are labeled “Current Conditions.” The change in employee perception of safety culture can be +determined by examining the change in these answers from 2008 to 2010. The second type of data +compared the employees’ view of the current environment at the refinery, or “Current Conditions,” with +how they think things should be at the refinery, or “Hoped-for Conditions.” Comparing answers in the +“Current Conditions” with the “Hoped-for Conditions” provides insight into the gaps in safety culture as +seen by the employees. + +201 Safety culture assessments are required by the City of Richmond RISO (See Section 5.5.2.4). However, the +Chevron Richmond Refinery performed these surveys before they were a regulatory requirement. Safety Culture +Assessments were not required by the RISO on August 6, 2012. The City of Richmond adopted this requirement +in February 2013. See http://www.ci.richmond.ca.us/ArchiveCenter/ViewFile/Item/4988 (accessed December +21, 2014). +202 A safety culture survey was also performed in 2009, showing similar results to the 2008 and 2010 surveys. +However, the company conducting the safety culture survey did not include the 2009 results in its 2008 and 2010 +comparisons. + +Chevron Richmond Refinery Investigation Report January 2015 +107 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Below is an analysis of three survey topics from the 2008 and 2010 safety culture surveys at the Chevron +Richmond Refinery which evaluate two key safety culture characteristics having direct impact on an +incident like the August 6, 2012, 4-sidecut piping failure: stop work authority and equipment +maintenance. The CSB performed a statistical analysis of the results using a Chi-Square test.203 The +information presented in the following section shows statistically significant changes in responses. +5.5.2.1 Stop Work Authority Safety Culture Survey Responses +Both the 2008 and 2010 surveys polled workers on their perception of their own Stop Work Authority. +The 2008 survey found that while 95 percent of operators and mechanics indicated that they desired to +use their Stop Work Authority at any time they witnessed unsafe activity, only 68 percent said they would +do so (Table 3). The analysis of this discrepancy concludes this “may imply a perceived barrier” to using +one’s Stop Work Authority. + +Question: +Stopping Unsafe +Work +I would rarely, if +ever do this. +I would ask the +safety person to do +it. +I do this with my +own team. +I do this with +anyone and +anytime there is +unsafe activity. +Operators & +Mechanics +2008 +Current Conditions +7% +(27 responses) +6% +(23 responses) +19% +(74 responses) +68% +(265 responses) +Operators & +Mechanics +2008 +Hoped-for +Conditions +0% +3% +(12 responses) +2% +(8 responses) +95% +(369 responses) +Table 3. Chevron Richmond Refinery 2008 Safety Culture Survey responses to question of “Stopping Unsafe +Work” by operators and mechanics. Chevron Richmond Refinery 2008 process safety culture survey +identified discrepancy between operators’ and mechanics’ desire and personnel willingness to use Stop Work +Authority. A total of 389 operators and mechanics were polled for this survey. + +Between 2008 and 2010, there also was a decrease in Chevron Richmond Refinery employees’ +willingness to use their Stop Work Authority beyond their own work group. In 2010, a statistically +significant portion of managers and engineers reported that they were less willing to use their Stop Work +Authority at any time they witnessed unsafe activity (Table 4). + +203 The Chi-Square test permits the determination of whether a significant difference exists between two sets of +categorical data, an "observed" set and an "expected" set. It permits an answer to the question, "How well does +our observed distribution fit the hypothetical distribution?" + +Chevron Richmond Refinery Investigation Report January 2015 +108 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Question: +Stopping Unsafe +Work +I would rarely, if +ever do this. +I would ask the +safety person to do +it. +I do this with my +own team. +I do this with +anyone and +anytime there is +unsafe activity. +Managers & +Engineers +2008 +Current Conditions +6% +(3 responses) +2% +(1 response) +15% +(7 responses) +77% +(36 responses) +Managers & +Engineers +2010 +Current Conditions +6% +(8 responses) +4% +(6 responses) +26% +(36 responses) +64% +(89 responses) +Table 4. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to question of +“Stopping Unsafe Work” by managers and engineers. Between 2008 and 2010, a significant portion of +managers and engineers became less willing to use their Stop Work Authority at any time. In 2008, the +survey polled 47 managers and engineers. In 2010, the survey polled 139 managers and engineers. +A similar question was then asked about how the refinery as a whole perceives Stop Work Authority. +Between 2008 and 2010, a statistically significant, increased portion of operators and mechanics began to +feel that they could get in trouble when using their Stop Work Authority (Table 5). These trends could +explain why no individuals used their Stop Work Authority on the day of the incident despite some +participants reporting in interviews with the CSB that they were not comfortable with the hazardous work +activity taking place. + +Question: How +do people feel +about stopping +unsafe work? +It could get you in +trouble. +It’s probably best +to point it out to a +supervisor first. +They do it and +know it might +slow down the job. +They will do it and +know they will be +backed up. +Operators & +Mechanics +2008 +Current Conditions +7% +(27 responses) +26% +(101 responses) +25% +(97 responses) +42% +(164 responses) +Operators & +Mechanics +2010 +Current Conditions +12% +(44 responses) +24% +(88 responses) +25% +(91 responses) +39% +(142 responses) +Table 5. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to question of “How +do people feel about stopping unsafe work?” by operators and mechanics. Between 2008 and 2010, a +significant portion of operators and mechanics began to feel they could get in trouble when using their Stop +Work Authority. In 2008, the survey polled 389 operators and mechanics. In 2010, the survey polled 365 +operators and mechanics. +5.5.2.2 Mechanical Integrity Safety Culture Survey Responses +The 2008 and 2010 surveys polled Chevron Richmond Refinery employees on their perception of how +equipment is maintained at the refinery. Table 6 summarizes the responses to this survey question. + +Chevron Richmond Refinery Investigation Report January 2015 +109 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +Question: How +do we take care +of equipment? +Equipment is not +cared for, and we +often have +breakdowns and +near misses. +We have +procedures for +updating and +maintaining our +equipment but +they are not +always followed. +We use procedures +for updating and +maintaining our +equipment but +they are not +always up to date. +We work hard to +think about what +can go wrong, and +fix the equipment +before it causes +harm. +All Employees +2008 +Current Conditions +8% +(53 responses) +26% +(172 responses) +35% +(231 responses) +31% +(205 responses) +All Employees +2008 +Hoped-for +Conditions +1% +(7 responses) +5% +(33 responses) +8% +(53 responses) +86% +(568 responses) +All Employees +2010 +Current Conditions +12% +(112 responses) +27% +(253 responses) +30% +(281 responses) +31% +(290 responses) +All Employees +2010 +Hoped-for +Conditions +2% +(19 responses) +4% +(37 responses) +9% +(84 responses) +85% +(796 responses) +Table 6. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to the question: “How +do we take care of equipment?” +Both surveys revealed that most employees felt procedures to maintain equipment were not always up to +date, were not always followed, or that equipment was not properly maintained despite most individuals’ +desire that the refinery fix equipment before it causes harm. Between the 2008 and 2010 surveys, a +statistically significant portion of operators and mechanics began to feel that equipment was not cared for +at the refinery. In addition, a statistically significant number of managers and engineers expressed that +procedures for updating and maintaining equipment were not always followed. These results indicate that +Chevron Richmond Refinery employees identified increased weakness in their mechanical integrity +programs, which could result in equipment failures, such as the incident on August 6, 2012. +5.5.2.3 Process Safety Analysis Safety Culture Survey Responses +The 2008 and 2010 surveys tasked Chevron Richmond Refinery employees to provide feedback on how +process safety failures are investigated at the refinery. The survey results are shown in Table 7. + +Question: +Process Safety +Issues Are… +…usually not +investigated. +…investigated but +not always +resolved. +…investigated and +resolved if it’s in +the budget. +…routinely +investigated, +resolved and the +lessons learned are +shared with others. +All Employees +2008 +Current Conditions +2% +(13 responses) +24% +(159 responses) +21% +(139 responses) +53% +(350 responses) +All Employees +2008 +Hoped-for +Conditions +1% +(6 responses) +1% +(7 responses) +3% +(20 responses) +95% +(628 responses) +All Employees +2010 +Current Conditions +3% +(28 responses) +26% +(243 responses) +17% +(159 responses) +54% +(506 responses) +All Employees +2010 +Hoped-for +Conditions +2% +(19 responses) +1% +(9 responses) +4% +(37 responses) +93% +(871 responses) +Table 7. Chevron Richmond Refinery 2008 and 2010 Safety Culture Survey responses to the question +“Process Safety Issues Are…” Responses indicate employees desire process safety issues are investigated +more thoroughly. + +Chevron Richmond Refinery Investigation Report January 2015 +110 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +The survey responses show a discrepancy between the “Current Conditions” and “Hoped-for Conditions.” +While nearly all of the employees polled desired that process safety issues were “routinely investigated, +resolved and the lessons learned [were] shared with others”, only about half of the employees reported +that they believed the refinery achieved that goal. The safety culture survey analysis of this difference +concluded, “There continues to be a very strong desire to investigate more thoroughly to ensure there is +adequate learning.” +5.5.2.4 Regulator Overview of Safety Culture Survey Action Items +Contra Costa County issues regulations to covered facilities within the county through its Industrial +Safety Ordinance (ISO).204 Although the City of Richmond is located in Contra Costa County, the county +does not have jurisdiction over industrial facilities located within city limits. Thus, the ISO is not +enforceable within the City of Richmond. On December 18, 2001, the City of Richmond adopted its own +industrial safety ordinance (RISO), based on the ISO, to extend jurisdiction of a similar sort over facilities +located in the city. The RISO covers two facilities—one of which is the Chevron Richmond Refinery. +Pursuant to an agreement between the two parties, Contra Costa County inspects these two facilities and +implements the RISO for the City of Richmond. Both the Contra Costa County ISO and the City of +Richmond RISO require covered facilities to perform safety culture assessments at least once every five +years.205 In addition, Contra Costa County published an “Industrial Safety Ordinance Guidance +Document” that establishes additional requirements covered facilities must perform. Section F: Safety +Culture Assessments requires facilities to maintain the following records, which can be audited by the +regulator: +a. Safety Culture Assessment reports; +b. Stated facility goals and objectives regarding safety culture and related topics; +c. Documentation of the appropriateness of the participation level targeted and achieved; +d. Assessment methodologies used for each work group and criteria for successful implementation; +e. Criteria used for rejection of any results or findings; +f. Criteria used for determining if no action(s) will be taken on assessment results or +recommendations; +g. Summary of the assessment components with key findings; +h. Improvement plan with clear list of action items and identifiable milestones; +i. +Rationale for prioritizing action items and justification for the action items; + +204 For more information on the Contra Costa County Industrial Safety Ordinance, and the City of Richmond +Industrial Safety Ordinance, see the CSB’s draft Regulatory Report, http://www.csb.gov/chevron-refinery-fire/ +(accessed September 9, 2014). +205 The Contra Costa County ISO has required covered facilities to perform Safety Culture Assessments at least +every five years since 2006, but Richmond did not adopt these requirements until February 2013. See +http://www.ci.richmond.ca.us/documentcenter/view/26375, page 22 (accessed July 11, 2014) and +http://www.ci.richmond.ca.us/ArchiveCenter/ViewFile/Item/4988 (accessed December 21, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +111 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +j. +Documentation of communications to workforce; and +k. Qualitative and quantitative comparisons in subsequent assessments of whether improvement +plans affected observable safety behavior, or culture.206 +As items (h) and (k) show, Contra Costa County and the City of Richmond require facilities to develop an +improvement plan based on the safety culture assessment findings. The facilities must also monitor +whether the improvement plan results in observable changes to safety behavior and culture. These items +set a requirement for facilities to strive for continuous improvement in process safety culture. However, +no means are in place for the regulator to ensure that the action items are of sufficient quality to promote +cultural change. As is currently written in the Contra Costa County guidance document, the regulator +only requires facilities to develop action items following the conduct of safety culture assessments. The +quality of these action items can be subpar, lacking in capability to significantly change culture, yet the +regulator must still approve them in its document-verification audits. +In the years leading to the August 6, 2012 incident, the Chevron Richmond Refinery identified +weaknesses in its Stop Work Authority program due to employees’ hesitation to use their Stop Work +Authority when they witnessed an unsafe act occurring. The refinery also identified a deteriorating +employee perception of the mechanical integrity programs used at the refinery. However, the Chevron +Richmond Refinery was not required to take quality, constructive steps to improve these areas.207 Had +steps been taken prior to the incident to find ways to encourage employees to use their Stop Work +Authority or to determine why the refinery’s mechanical integrity programs were seen as deficient, the +August 6, 2012, pipe rupture might have been prevented. +Contra Costa County and the City of Richmond should enhance the ISO and RISO, respectively, to +require the development of an oversight committee to monitor the development and implementation of +action items created as a result of safety culture assessment findings. This committee should also assess +whether the action items that result from the safety culture assessments have the potential to effectively +lead to improved process safety culture in the facility. This oversight committee should be comprised of +regulator representatives, company representatives, and members of the workforce and their +representatives. Many citizens of Contra Costa County and Richmond, California, are concerned about +the environmental impacts of industrial process incidents on their community, so they passionately +advocate for improved industrial process safety at the petrochemical facilities in the county. For this +reason, it is important that Contra Costa County community members play an active role in overseeing +and providing input into process safety culture improvement efforts at petrochemical facilities in Contra +Costa and Richmond. Ideally, one to two community members—who are selected by their peers based +upon their ability to effectively (1) communicate the concerns of community members and (2) provide +valuable input into the process safety culture improvement plans—will also be member(s) of this +oversight committee. + + + +206 http://cchealth.org/hazmat/pdf/iso/section_f.pdf (accessed July 14, 2014). +207 Under the RISO, Chevron was not required to conduct safety culture assessments until February 2013. + +Chevron Richmond Refinery Investigation Report January 2015 +112 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +6.0 Recommendations +Under the authority of 42 U.S.C. §7412(r)(6)(C)(i) and (ii), and in the interest of promoting safer +operations at petroleum refineries and protecting workers and communities from future accidents +nationwide, the CSB makes the following safety recommendations: +6.1 American Petroleum Institute +2012-03-I-CA-R26 +Revise API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil +Refineries to establish minimum requirements for preventing catastrophic rupture of low-silicon +carbon steel piping. At a minimum: +a. Require users to identify carbon steel piping circuits susceptible to sulfidation corrosion that +may contain low-silicon components. These circuits have the potential to contain carbon steel +components that were not manufactured to the American Society for Testing and Materials +(ASTM) A106 specification and may contain less than 0.10 weight percent silicon content. +b. For piping circuits identified to meet the specifications detailed in 2012-03-I-CA-R26(a), +require users to either (1) enact a program to inspect every component within the piping circuit +once, known as 100 percent component inspection (per the requirements established pursuant to +recommendation 2012-03-I-CA-28(c)), or (2) replace the identified at-risk carbon steel piping +with a steel alloy that is more resistant to sulfidation corrosion. +c. If low-silicon components or components with accelerated corrosion are identified in a carbon +steel piping circuit meeting the specifications detailed in 2012-03-I-CA-R26(a), require +designation of these components as permanent Condition Monitoring Locations (CMLs) until +the piping components are replaced.  +2012-03-I-CA-R27 +Revise API RP 571: Damage Mechanisms Affecting Fixed Equipment in the Refining Industry to: +a. Describe the potential for increased rates of sulfidation corrosion occurring in low-silicon +carbon steel in Section 4.4.2.3 Critical Factors;   +b. Specify that sulfidation corrosion rates in carbon steel piping can be significantly faster in a +few, individual piping components in section 4.4.2.5 Appearance or Morphology of Damage; +and +c. Refer the reader to the 100 percent component inspection or pipe replacement requirements +detailed in API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in +Oil Refineries (pursuant to recommendation 2012-03-I-CA-26) and API 570: Piping Inspection +Code: In-service Inspection, Rating, Repair, and Alteration of Piping Systems (pursuant to + +Chevron Richmond Refinery Investigation Report January 2015 +113 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +2012-03-I-CA-28(c)) for carbon steel piping circuits susceptible to sulfidation corrosion that +may contain low-silicon components. +2012-03-I-CA-R28 + +Revise API 570: Piping Inspection Code: In-service Inspection, Rating, Repair, and Alteration of +Piping Systems to: + +a. Use terminology consistent with API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) +Corrosion Failures in Oil Refineries and other API standards and recommended practices +discussed in this report. Replace the terminology “high-temperature sulfur corrosion” with +“sulfidation corrosion”; +b. Specify that sulfidation corrosion rates in carbon steel piping can be significantly faster in some +individual piping components than in others; +c. Establish a new section that details inspection requirements to identify low-silicon piping +components in carbon steel circuits susceptible to sulfidation corrosion. This section shall +require users to identify carbon steel piping circuits at risk to contain low-silicon components +by following the requirements detailed in API RP 939-C: Guidelines for Avoiding Sulfidation +(Sulfidic) Corrosion Failures in Oil Refineries (pursuant to 2012-03-I-CA-26(a)) and API RP +578: Material Verification Program for New and Existing Alloy Piping Systems (pursuant to +2012-03-I-CA-29). At a minimum, require users to either: +i. Inspect every component within all carbon steel piping circuits susceptible to +sulfidation corrosion that may contain low-silicon components once. The +purpose of this practice is to identify any low-silicon components that are +corroding at accelerated rates. Inspection may be performed through ultrasonic +thickness measurements to establish corrosion rates for each component, +destructive laboratory analysis, or other methods. Following the inspection, +require users to follow the low-silicon corrosion rate monitoring requirements +established in 2012-03-I-CA-R26(c); or +ii. Replace the identified at-risk carbon steel piping with a steel alloy that is more +resistant to sulfidation corrosion. + +d. Incorporate as a “normative reference” API RP 939-C: Guidelines for Avoiding Sulfidation +(Sulfidic) Corrosion Failures in Oil Refineries; and +e. Require users to follow the minimum leak response guidance established in API RP 2001: Fire +Protection in Refineries, developed in response to recommendation 2012-03-I-CA-R31. + + + +Chevron Richmond Refinery Investigation Report January 2015 +114 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + +2012-03-I-CA-R29 +Revise API RP 578: Material Verification Program for New and Existing Alloy Piping Systems, to +require users to establish and implement a program to identify carbon steel piping circuits that are +susceptible to sulfidation corrosion and may contain low-silicon components. These circuits have +the potential to contain carbon steel components that were not manufactured to the American +Society for Testing and Materials (ASTM) A106 specification and may contain less than 0.10 +weight percent silicon content. Refer the reader to the 100 percent component inspection or pipe +replacement requirements detailed in API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) +Corrosion Failures in Oil Refineries (pursuant to recommendation 2012-03-I-CA-26(b)) and API +570: Piping Inspection Code: In-service Inspection, Rating, Repair, and Alteration of Piping +Systems (pursuant to 2012-03-I-CA-28(c)) for carbon steel piping circuits susceptible to sulfidation +corrosion that may contain low-silicon components. +2012-03-I-CA-R30 + +Revise API RP 574: Inspection Practices for Piping System Components (3rd edition) to: +a. Incorporate as a normative reference API RP 939-C: Guidelines for Avoiding Sulfidation +(Sulfidic) Corrosion Failures in Oil Refineries; +b. Reference API RP 939-C: Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in +Oil Refineries when discussing that nonsilicon-killed carbon steel is susceptible to sulfidation +corrosion; and +c. In Section 9.3 Investigation of Leaks, require users to follow the leak response protocol +requirements established in API RP 2001: Fire Protection in Refineries (pursuant to 2012-03-I- +CA-R31). +2012-03-I-CA-R31 +Revise API RP 2001: Fire Protection in Refineries to require users to develop a process fluid leak +response protocol specific to their own facility that must be followed when a process fluid leak is +discovered. Recommend users to incorporate the following actions into their leak response +protocol: +a. Establish an Incident Command structure upon identification of a process fluid leak; +b. Conduct a pre-response meeting with personnel with specific technical expertise (e.g., +inspectors, operators, metallurgists, engineers, and management) and the Incident Commander +to determine pressure, temperature, remaining inventory of process fluids, potential damage +mechanisms that caused the leak, and worst-case leak scenario; + +Chevron Richmond Refinery Investigation Report January 2015 +115 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +c. Establish a hot zone that identifies the area of risk of exposure or injuries due to flame contact, +radiant heat, or contact to hazardous materials, taking into consideration the worst-case leak +scenario; +d. Limit site access around leak location to essential personnel only; +e. Isolate the leaking piping or vessel, or if isolation is not possible, shutdown of the unit when the +leaking process fluid poses immediate danger to safety, health, or the environment—such as +piping fluid that is toxic or near the autoignition temperature. +6.2 American Society of Mechanical Engineers +2012-03-I-CA-R32 + +Revise ASME PCC-2-2011: Repair of Pressure Equipment and Piping to require users to follow the +minimum process fluid leak response requirements established in API RP 2001: Fire Protection in +Refineries, developed in response to recommendation 2012-03-I-CA-R31, before conducting +process fluid leak repair. +6.3 Chevron USA +2012-03-I-CA-R33 +Develop a method to assign accountability at Chevron to determine whether any new Energy +Technology Company (ETC) recommended program or industry best practice, such as API +guidance must be followed to ensure process safety or employee personal safety. This method shall +include monitoring of these practices and guidance at a refining system level and at the refinery +level. Develop a tracking system to monitor the progress of implementing these selected practices +and guidance to completion. +2012-03-I-CA-R34 +Develop an auditable process to be available for all recommended turnaround work items necessary +to address mechanical integrity deficiencies or inspection recommendations that are denied or +deferred. This process shall provide the submitter of the denied or deferred recommendation with +the option to seek further review by his or her manager, who can further elevate and discuss the +recommendation with higher level management, such as the Area Business Unit Manager. +Maintain an auditable log of each of these potential turnaround work items, including the ultimate +determination of approval, deferral, or rejection, justification determination, and the person or team +responsible for that decision. +2012-03-I-CA-R35 +Develop an approval process that includes a technical review that must be implemented prior to +resetting the minimum alert thickness to a lower value in the inspection database. + +Chevron Richmond Refinery Investigation Report January 2015 +116 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +6.4 Board of Supervisors, Contra Costa County, California +2012-03-I-CA-R36 +Revise the Industrial Safety Ordinance (ISO) regulations for petroleum refineries to require a +process safety culture continuous improvement program including a written procedure for periodic +process safety culture surveys across the work force. Require an oversight committee comprised of +the regulator, the company, the company’s workforce and their representatives, and community +representatives. This oversight committee shall: +a. Select an expert third party that will administer a periodic process safety culture survey; +b. Review and comment on the third party expert report developed from the survey; +c. Oversee the development and effective implementation of action items to effectively +address identified process safety culture issues; and +d. Develop process safety culture indicators to measure major accident prevention +performance. +The periodic process safety culture report shall be made available to the plant workforce. +6.5 Mayor and City Council, City of Richmond, California +2012-03-I-CA-R37 +Revise the Richmond Industrial Safety Ordinance (RISO) regulations for petroleum refineries to +require a process safety culture continuous improvement program including a written procedure for +periodic process safety culture surveys across the work force. Require an oversight committee +comprised of the regulator, the company, the company’s workforce and their representatives, and +community representatives. This oversight committee shall: +a. Select an expert third party that will administer a periodic process safety culture survey; +b. Review and comment on the third party expert report developed from the survey; +c. Oversee the development and effective implementation of action items to effectively +address identified process safety culture issues; and +d. Develop process safety culture indicators to measure major accident prevention +performance. +The periodic process safety culture report shall be made available to the plant workforce. + +Appendix A + +117 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Appendix A—Chevron Leak Response Protocol Developed Post-incident + + + +Appendix A + +118 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + + + +Appendix A + +119 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + + + + + +Appendix B + +120 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Appendix B—Contra Costa County Community Warning System +Following the incident, Contra Costa County’s Community Warning System was used to notify the +surrounding community of the hazardous material incident and order a shelter-in-place. But for many +people, the warning came hours after the Chevron fire began. The Community Warning System uses +sirens, the news media, and phone calls to residents to initiate the shelter-in-place. Contra Costa County +issued the shelter-in-place advisory on August 6, 2012, at 6:38 p.m. for the cities of Richmond, San +Pablo, and North Richmond, and lifted the shelter-in-place later that evening at 11:12 p.m. However, +some phone calls notifying residents of the shelter-in-place advisory did not occur until over four hours +after the release. This delay could have resulted in nearby residents unnecessarily and unknowingly being +exposed to materials released to the atmosphere during the Chevron process leak and fire. +Since the incident, Contra Costa County has made efforts to improve the Community Warning System. It +has contracted with a new vendor that will automatically call Contra Costa County residents in an +emergency.208 + +208 See http://concord-ca.patch.com/groups/politics-and-elections/p/contra-costa-county-testing-new-community- +alert-system and http://www.contracosta.ca.gov/documentcenter/view/8161. + +Appendix C + +121 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD + +Appendix C—Usage of Clamps at Chevron Richmond Refinery +The CSB committed to analyzing Chevron’s culture of using clamps to temporarily stop a process fluid +leak in “Additional Issues Currently Under Investigation” in its Interim Investigation Report on the +August 6, 2012, Chevron Richmond Refinery incident. Following the August 6, 2012 incident, +Cal/OSHA issued a citation to Chevron for nine temporary nonwelding repairs that had not been removed +at the most recent turnaround.209 The CSB analyzed this citation and all available evidence on clamp +usage at the Chevron Richmond Refinery. The CSB could not take a conclusive stance on whether the +refinery over-relied on temporary leak repair clamps in its mechanical integrity program based upon +available evidence. In addition, the CSB did not find any direct or relevant linkage between the specific +clamps for which the citations were issued and the incident. As a result, this report does not analyze the +use of leak repair clamps at the Chevron Richmond Refinery. + +209 Division of Occupational Safety and Health, Cal/OSHA Process Safety Management District Office. Citation +and Notification of Penalty. Inspection Number 314332370; Citations Issued to Chevron U.S.A. Inc.; Issuance +Date 01/30/2013. Citation 8 Item 1. See +http://www.dir.ca.gov/dosh/citations/Chevron_314332370_cites_issued_1-30-13.pdf#zoom=100 (accessed +November 5, 2014). + +Chevron Richmond Refinery Investigation Report January 2015 +122 + U.S. CHEMICAL SAFETY AND HAZARD INVESTIGATION BOARD \ No newline at end of file