Skip to content

Enterprise Pascagoula Gas Plant Explosion and Fire

Overview

On June 27, 2016, explosions and fire occurred at the Enterprise Products Pascagoula Gas Plant (PGP) in Pascagoula/Moss Point, Mississippi. The CSB final report concluded the probable cause was failure of a brazed aluminum heat exchanger (BAHX) due to thermal fatigue, with the absence of a reliable process to ensure mechanical integrity contributing to the catastrophic failure. The incident caused extensive damage in the A-Train process area and shut the site down for almost six months.

Incident Snapshot

Field Value
Facility / Company Enterprise Products
Location Moss Point, MS
Incident Date 06/27/2016
Investigation Status The CSB's final report was released on February 13, 2019.
Accident Type Chemical Distribution - Fire and Explosion
Final Report Release Date 02/13/2019

What Happened

  • Sometime before 11:22 p.m. on June 27, 2016, a major loss of containment (LOC) resulted in the release of methane, ethane, propane, and several other hydrocarbons at PGP.
  • The hydrocarbons ignited, initiating a series of fires and explosions.
  • At 11:22 p.m., a sudden explosion and fire occurred.
  • Within a minute of the initial explosion, the operators activated the emergency shut down systems at the plant and sheltered in the control room.
  • The board operator initiated ESD 1000 about a minute after the initial explosion, and then hit the ESD 9900 button 23 seconds later.
  • The plant operations supervisor immediately called 9-1-1.
  • Emergency responders from the Pascagoula Fire Department were the first to arrive on scene at 11:30 p.m.
  • Around 12:00 a.m., PGP management met up with emergency responders at a command post on a nearby highway overpass.
  • Unified Command decided to continue to cut the feed into the plant and allow the contained fire to burn itself out, while cooling equipment where feasible.
  • The fire began to diminish around midnight.
  • At approximately 1:00 a.m., emergency responders approached the process area to conduct reconnaissance.
  • At approximately 2:30 a.m., emergency responders returned to set up unmanned fire monitors to cool vessels and equipment near the fire.
  • The fire finally burned out around 6:00 p.m. the following day.
  • Emergency responders discontinued safety and fire watch operations on the morning of June 29.

Facility and Process Context

  • The Pascagoula Gas Plant (PGP) receives raw natural gas via a pipeline from Gulf of Mexico deepwater oil wells and processes it into natural gas liquids and a natural gas fuel stream.
  • PGP has a design capacity of 1.5 billion cubic feet per day of raw gas and up to 50,000 barrels per day of NGLs.
  • At the time of the incident, throughput was approximately 500 million standard cubic feet per day of raw gas and 16,000–18,000 bpd of NGLs.
  • On the night of June 27, 2016, PGP was operating A-Train, one of its cryogenic process lines.
  • A- and B-Trains were part of the original design of the plant and were commissioned in 1999 after the plant was built.
  • C-Train was added in 2003 and does not operate at cryogenic temperatures.
  • PGP has NGL storage bullets that allow operators some flexibility to accumulate product until full flow to the Tri-States pipeline is reestablished, but storage capacity is limited.
  • The plant has two emergency shut-down (ESD) systems, ESD 1000 and ESD 9900.
  • PGP’s emergency response plan was to vent fuel to the flare or vent to atmosphere and to let any existing fire burn until it self-extinguished.
  • The board operator was in the control room, a blast-resistant module without windows.

Consequences

  • Fatalities: 0
  • Injuries: Two workers were on the night shift when the incident occurred and were uninjured.
  • Environmental release: Enterprise reported that 104,000 pounds of total emissions were released on June 27, 2016; most of the released hydrocarbon was burned off by the ensuing fires and flares. Onsite chemical monitoring detected no hydrocarbon traveling beyond the plant’s property.
  • Facility damage: Equipment, piping, and vessels in the A-Train process area were extensively damaged. After the incident, 12 different breaches were identified in the A-Train equipment, piping, and vessels.
  • Operational impact: The site was shut down for almost six months. Remaining fuel burned out around 6:00 p.m. on June 28, 2016, and units from the Forts Lake Fire Department remained onsite to provide safety and fire watch until released from the incident scene at 8:47 a.m. on June 29, 2016.

Key Findings

Immediate Causes

  • The probable cause of this incident was the failure of a brazed aluminum heat exchanger (BAHX) due to thermal fatigue.
  • The CSB concludes that the first loss of containment most likely originated at a BAHX when it lost core integrity due to accumulated thermal fatigue.
  • The ACSR most likely failed as a result of process fluids leaking into outer layers of the exchanger due to thermal fatigue damage.
  • These outer layers were blocked, with no relief venting, after a previous repair for thermal fatigue–induced cracking.
  • They became over-pressurized and catastrophically ruptured.
  • DCS data from temperature and differential pressure sensors suggest that the first loss of containment most likely originated at the cold side reboiler shortly before 11:21 p.m.

Contributing Factors

  • The absence of a reliable process to ensure the mechanical integrity of the heat exchanger contributed to the catastrophic failure of the equipment.
  • The BAHX had blocked layers within the exchanger.
  • Thermal fatigue cracks developed in the parting sheet adjacent to the blocked layers.
  • At least one layer separated from the core of the exchanger.
  • The accumulated fluid in the blocked layer expanded due to heating.
  • There was an absence of venting or draining to allow the accumulated fluid to readily escape from the blocked layer.
  • PGP’s BAHXs were repeatedly subjected to temperature changes that exceeded industry-recommended practices.
  • External plant conditions and internal operational practices caused these temperature fluctuations.
  • Upstream and downstream activities impacted the site, often in unpredictable ways, requiring PGP to reduce or completely stop production.
  • The BAHXs were subjected to conditions conducive to thermal fatigue damage, where significant and sometimes rapid temperature fluctuations of the hydrocarbons circulating through the process often exceeded the recommended operational guidance of the BAHXs at the site.
  • The CSB found that none of the industry guidelines give criteria for using current and historical process data to assess the potential for thermal fatigue risks associated with BAHXs.
  • PGP’s stream temperature data indicates that the ACSR operated above the ROC recommended by industry guidelines for more than 10 percent of its operational time over several years.
  • PGP alarmed the metal temperature sensors, but not the stream temperature sensors, to alert operators to temperature problems in the exchangers.
  • Several gaps in the safety management of the BAHXs at PGP reveal deficiencies in hazard assessment, MOC, and mechanical integrity programs at the site.
  • The lack of a company representative at the command post delayed response decision making, contrary to the requirements of a unified command structure.
  • Unified Command did not initiate the reverse 9-1-1 system called CodeRED®.
  • An official evacuation was not ordered, nor was a shelter-in-place.

Organizational and Systemic Factors

  • BP had operational control of PGP for almost its entire existence up to the time of the June 2016 event.
  • The transition from BP to Enterprise operatorship resulted in a change in corporate engineering support teams, but not in onsite plant operations.
  • For the most part, the personnel and operational procedures were the same on June 27 as they had been before Enterprise’s full acquisition of the site.
  • Eventually, site management and personnel focused their assessment of the BAHX hazards on relatively minor safety/health concerns and financial costs for repairing thermal fatigue cracks.
  • PGP chose to modify its operating procedures after leaks, but it is unclear why the procedural change exceeded the recommended operating practices.
  • BP superseded the recommendation to replace the exchanger with a subsequent recommendation to insert temperature sensors for improved monitoring.
  • PGP’s approach to the mechanical integrity of its BAHXs was leak detection and repair.
  • Throughout its operational ownership of PGP, BP used administrative controls, modifying its operating procedures and installing temperature ROC alarms, to manage the integrity of its BAHXs.
  • The company did not replace the leaking BAHXs with new ones or switch to another type of exchanger.
  • Not until 2015, after more than nine leaks from the BAHXs at PGP, did BP make plans to replace the first BAHX by 2018.
  • Enterprise used a variety of approaches for the BAHXs at its other facilities, sometimes replacing its exchangers after a single leak, sometimes preemptively before any leak occurred.
  • PGP established this “let it burn” philosophy when it designed and built the plant, and its staff and responders generally followed the plan on June 27, 2016.
  • PGP did not assign someone to stay at the command post where the unified command was established.
  • The planned emergency response activities stipulate that the plant will participate in the unified command structure.
  • The CSB noted that the PGP incident demonstrates an opportunity to address the need for a more robust and engaged LEPC/community alert network—one that includes social media and the ability to dialogue with the community throughout the incident.
  • Enterprise and other companies had confidentiality or proprietary reasons for not sharing data.
  • Many midstream gas companies opted not to provide data to the GPA Midstream research effort because of various proprietary and confidentiality concerns.

Failed Safeguards or Barrier Breakdowns

  • No abnormal alarms or other indicators warned the two PGP personnel of any problems.
  • No fixed fire monitors were in the plant to provide water for cooling.
  • The vent holes for the blocked layers were welded closed.
  • The blocked-off layers had no relief protection.
  • PGP did not conduct any calculations to determine the critical size of a crack required to result in a catastrophic failure of the exchanger.
  • PGP did not provide the CSB with MOC documentation detailing the origin and justification of the metal temperature sensor alarm set point.
  • PGP did not provide the “what-if” hazard review with the MOC documentation.
  • PGP did not initiate a reverse 9-1-1 call, relying instead upon the local media outlets and social media to keep the public informed.
  • The company did not replace the leaking BAHXs with new ones or switch to another type of exchanger.
  • On the night of the incident, PGP did not assign someone to stay at the command post where the unified command was established.
  • The CSB is unaware of any chemical monitoring at the PGP boundary conducted between the initial explosion, shortly after 11:20 p.m. on June 27, 2016, and 10:30 a.m. on June 28, 2016.

Recommendations

  1. 2016-02-I-MS-R1 — Recipient: American Petroleum Institute (API) — Status: Closed – Acceptable Action — Summary: Develop a new informational product or incorporate into the next revision of Brazed Aluminum Plate-Fin Heat Exchangers for General Refinery Services 1st ed.; ANSI/API Standard 668 (formerly Standard 662, Part 2), guidance focused on the safe operation, maintenance, and repair of brazed aluminum heat exchangers (BAHX) to advance understanding of thermal fatigue hazards and how to mitigate them.
  2. 2016-02-I-MS-R2 — Recipient: GPA Midstream Association — Status: Closed – Acceptable Action — Summary: Revise GPA Technical Bulletin: Brazed Aluminum Heat Exchangers, or develop a new bulletin, to incorporate the significant lessons learned from this incident, including but not limited to: information on the potential of both minor leaks and catastrophic failure as a result of thermal fatigue; clarification on the optimal placement of BAHX temperature and pressure sensors to better monitor operating conditions, including temperature rates of change; and clarification on the need to safely vent layers that have been blocked off after interpass leak repairs, in all BAHX configurations.
  3. 2016-02-I-MS-R3 — Recipient: GPA Midstream Association — Status: Closed – Reconsidered/Superseded — Summary: Develop a database for operators to submit BAHX operational data for collaborative industry learning and analysis.
  4. 2016-02-I-MS-R4 — Recipient: GPA Midstream Association — Status: Closed – No Longer Applicable — Summary: Using available operational process data of BAHXs in midstream gas plant operation collected in fulfillment of 2016-02-I-MS-R3, continue data analysis efforts to determine what, if any, correlation exists between operational process data and the frequency or timing of thermal fatigue-generated cracking to more accurately predict the service life of a BAHX.
  5. 2016-02-I-MS-R5 — Recipient: Jackson County Local Emergency Planning Committee — Status: Closed – Acceptable Action — Summary: Work with members (industry, emergency response, community) to explicitly define the communication methods for community notification and incident updates and the expectations for their use, so that members of the public can efficiently and effectively obtain current safety information.

Key Engineering Lessons

  • Thermal fatigue in brazed aluminum heat exchangers can accumulate over time and can lead to both minor leaks and catastrophic failure.
  • Blocked-off layers in a BAHX must be safely vented; otherwise trapped fluid can become over-pressurized during heating.
  • Reliance on temperature sensors, procedures, and operator intervention alone does not address the accumulative nature of thermal fatigue damage.
  • Industry guidance cited in the report did not provide criteria for using current and historical process data to assess thermal fatigue risk in BAHXs.
  • Mechanical integrity programs for BAHXs need a reliable process that goes beyond leak detection and repair.
  • Temperature and pressure sensor placement matters for monitoring operating conditions and temperature rates of change in BAHXs.
  • Community notification systems should be explicitly defined and exercised so the public can receive timely incident information.

Source Notes

  • Priority 1 final report used to resolve incident cause, consequences, and recommendations.
  • Priority 3 recommendation status summaries used to update recommendation statuses for R1 through R5.
  • Priority 4 supporting documents used only for technical detail on the ACSR metallurgical examination and thermal expansion calculations where directly relevant.
  • Location metadata listed Moss Point, MS; final report text identifies Pascagoula, Mississippi. Both refer to the same incident location context in the provided sources.

Similar Incidents

Incidents sharing the same equipment, root causes, or hazard types.

Same Equipment

Same Root Cause

Same Hazard

← View in Knowledge Graph