Skip to main content

Fire from a hydro-desulphurisation unit of a refinery (1032)

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

Fire from a hydro-desulphurisation unit of a refinery

In a refinery, a fire occurred at a new reactor for the vacuum distillate hydro-desulphurisation unit. Process gas (hydrogen, hydrocarbon, etc. ) blew from a flange of upper piping in the reactor. It ignited producing 0. 5-2 meters high flames. The reason for the leak was a lack of tightness of the flange. Emergency stop of th plant was executed, thereactor was cooled and purged with nitrogen. The fire department was alarmed.
Event Date
September 27, 1989
Record Quality Indicator
Region / Country
Classification of the Physical Effects
Nature of the Consequences
Cause Comments
The INITIATING CAUSE was considered a loose high-pressure flange due to uneven tightening force, on uneven thermal expansion of materials at a temperature of 390 °C. The tightening operation of hot bolting at start-up three months before was inadequate. The ROOT CAUSE was a shortcoming in the procedure for hot-bolting, tohether with a lack of effective regular inspection. There was also a alck of management actions, becasue the difficulty of hot-bolting was known and several otehr plants were able to prevent occurrence of these events.
Facility Information
Application Type
Application
Hydrodesulphurisation process
Specific Application Supply Chain Stage
Components Involved
highpresure, high tmeprature flange
Storage/Process Medium
Actual Pressure
10
Design Pressure
10
Location Type
Operational Condition
Pre-event Summary
The reactor was new, but the operation was a standard operation.
Consequences
Number of Injured Persons
0
Number of Fatalities
0
Environmental Damage
0
Property Loss (onsite)

Y

Property Loss (offsite)

N

Post-event Summary

Heat damage and some of vacuum distillate oil was lost.

Lessons Learned
It is well-known that to manage the tightening of flanges in high-temperature and high-pressure sections, a quantitative control of the tightening force is required. Critical as well is to determine the number of times this control should be performed, and at which moments of the pressure-temperature history of the flange. Since not all flanges in complex high-pressure high-temperature systems are experiencing the same conditions, the hot bolting procedure should be based on the most severe conditions. The investigation found out that: 1. There was not a sufficient margin for tightening force on hot-bolting. 2. The upper part and the lower part of the flange were different (respectively SUS321 steel and 2-1/4Cr-1Mo steel with an overlay). Therefore, their coefficients of thermal expansion differed. 3. The thermal expansion of the piping acted in the direction of opening the flange. Findings 2 and 3 were specific to this flange only. It is possible that the special combinations of different materials, geometric configuration and harsh working conditions may have been overlooked during the design phase and/or during operation.
Event Nature
Emergency Action
Event sequence: 19:55. In a reactor with a height of 13 m, vacuum distillate containing hydrogen blew from a flange of the upper piping. It ignited and produced a 0. 5-2 m high flame. 20:00. An emergency call to the fire department was made and the emergency shutdown was started. Pressure was reduced from operating pressure of 10 MPa(g). 20:08. Public fire brigades arrived. For fear of a secondary disaster with hydrogen sulphide, they stood by without extinguishing the fire with water. 20:44. The fire extinguishment was confirmed. 20:45. The pressure in the reactor lowered to 0. 6 MPa(g). Nitrogen gas was introduced into the system. 21:20. Pressure lowered to 0. 3 MPa(g) or less. Cooling by feeding nitrogen gas was continued.
Release Type
Release Substance
H2, hydrocarbons
Hydrogen Release Concentration (%)
0.00
Ignition Source
Detonation
No
Deflagration
No
High Pressure Explosion
No
High Voltage Explosion
No
References
References

Orignal source lost. RISCAD is closed, now in database shippai: https://www. shippai. org/fkd/include/fkd_showCase. php?id=CC0000200&text1=CC0000200 (accessed April 2025)

Event no. 1989-059 of the KHK (High Pressure Gas Safety Association): https://www. khk. or. jp/public_information/incident_investigation/hpg_incident/incident_db. html (accessed May 2025)