Skip to main content

Steam reformer explosion at a oil production plant (18)

This data is from the European Hydrogen Incidents and Accidents database HIAD 2.1, European Commission, Joint Research Centre.

Steam reformer explosion at a oil production plant

The incident occurred at the steam methane reformer unit producing hydrogen. The event occurred just after that start-up steam was introduced into the reformer furnace inlet. At this moment a very high pressure and a single loud bang was reported by the operations personnel on the furnace structure. The investigation concluded that the mechanical failure of the reformer was due to flash evaporation of water still in liquid form, which caused an abnormal pressure peak. This is therefore an accident involving a hydrogen system, but where hydrogen did not have a role in the accidental sequence. Since the incident occurred at the re-start of the reformer, probably the hydrogen quantities were negligible An emergency shutdown of the furnace was initiated and the start up aborted. At the time of the incident there were 7 people on the furnace. Four were operations staff plus three maintenance personnel.
Event Date
January 30, 2005
Record Quality Indicator
Region / Country
Event Initiating System
Classification of the Physical Effects
Nature of the Consequences
Cause Comments
The INITIATING CAUSE was accidental injection of remnant water into the furnace. The ROOT CAUSE was related to organisation aspects such as inappropriate safety design risk, (see lesson learned)
Facility Information
Application Type
Application
Steam methane reforming
Specific Application Supply Chain Stage
Components Involved
steam reformer, furnace inlet
Storage/Process Medium
Location Type
Operational Condition
Pre-event Summary
The event occurred just after that start-up steam was introduced into the reformer furnace inlet. The plant had been shutdown and the reformer furnace isolated, to perform maintenance work. After the replacement of 6 damaged tubes, the start-up procedure was initiated: it consists in heating furnace up to 350oC (662oF), then introducing 4136 kPa (600 psig) of steam into the radiant tubes, rising the furnace temperature is increased until 500oC (932oF), feeding the methane and finally bringing the unit on line for the production of hydrogen.
Consequences
Number of Injured Persons
1
Number of Fatalities
0
Post-event Summary
A total of 46 reformer tubes were destroyed in the incident. The inlet portions of 5 tube assemblies were ejected from the furnace with sufficient force to severely injure one worker and launch debris beyond the furnace structure. Under normal operation the loss of the hydrogen from this reformer is estimated to restrict production by 4. 5 million barrels of Syncrude Sweet Blend over a 71-day outage. The cost of re-tubing the furnace was approximately $20 million, which does not include other associated shutdown costs. The cost of a new catalyst charge was approximately $1million. The investigation team condemned the entire reformer radiant tube section as there was no method to prove the integrity of the surviving tubes and there was too much damage to simply conduct repairs.
Lessons Learned
Inadequate Safeguards: The system was designed with low point drains to facilitate water removal. However, these were found to be inadequate in both location and size. The fact that the mixed feed pre-heat coil was not self-draining was unknown prior to the incident. A HAZOP (Hazard and Operability Study) revalidation had not been performed on this plant but had been scheduled for 2006. Inadequate Procedure: the startup procedure did not account for a startup of a cold furnace. Therefore, the time to reach the critical “steam in” temperature of 350 C (662 F) was short as compared to previous startups. Moreover, the procedure did not allow for checking that the reformer furnace feed system was dry. Lack of Management of Change: several steps I the procedure were not executed, because thought not required or not applying to this specific start-up. Consequently, the heat-up cycle was artificially shortened. This alteration to the startup sequence was not viewed as a change of operations. However, shutdown and startup procedures are designed to take a unit from safe operation to a zero-energy state and then return it to safe operation. Changing these sequences by an intentional omission is a change and must be properly assessed for risk. Moreover, the decision to leave some steam flow in the steam generating system for this winter shutdown was made to keep the system warm and prevent freezing. This decision was seen as a safeguard from the risk of freezing. But no formal risk assessment was done, to identify possible negative effects of this decision, which provided an opportunity for water to accumulate upstream of the reformer furnace. Non-essential Personnel: at the time of the incident there were 7 people on the furnace structure. Only the operations personnel were essential.
Event Nature
Release Type
Release Substance
H2 100%
Hydrogen Release Concentration (%)
0.00
Ignition Source
Detonation
No
Deflagration
No
High Pressure Explosion
No
High Voltage Explosion
No
References
References

Mike Rogers, Lessons Learned From an Unusual Hydrogen Reformer Furnace Failure Canadian Society of Chemical Enigneers 2005 https://www. cheminst. ca/wp-content/uploads/2019/04/Rogers20-20CSChE202005. pdf (accessed August 2020)

CBC News http://www. cbc. ca/cp/business/050201/b0201158. html (accessed 2016)

Also uptaken by H2TOOLS https://h2tools. org/lessons/hydrogen-reformer-tubes-ruptured-during-startup (accessed decemebr 2025)