Hydrocracking Unit Fire During Pressure Testing
Following a catalyst regeneration a fire occurred on a hydrocracking unit during unit pressure testing prior to start-up. The fire resulted from ignition of high - pressure hydrogen-rich gas that was escaping from two shell-to-shell nozzle flanges on the hydrocracker reactor charge - effluent exchangers. Four pipefitters who were in the process of tightening the leaking flanges when the gas ignited, received first and second degree burns. The unit was depressurised to the relief system and the fire was extinguished after burning for approximately 45 minutes. Unit damage was confirmed to the insulation.The hydrocracking unit was shutdown for a scheduled regeneration of the preheater and hydrocracker reactors and minor maintenance. Following this preparations were made to bring the unit on stream. The unit was nitrogen purged and evacuated several times and nitrogen was then fed into the pre-treating an hydrocracking sections checked for leaks at 100 PSIG. The nitrogen was then vented to the relief system and the system evacuated Hydrogen-rich gas was introduced and again the preheater and hydrocracker were checked for leaks up to 450 PSIG. The make-up hydrogen compressor was placed in service and the pre-treater system was checked for leaks at 1250 PSIG. A moderate leak was successfully stopped. The hydrocracker section was being raised to a final test pressure of 1600 PSIG when two shell-to-shell nozzle flange leaks developed in the inner connections of the hydrocracker reactor charge-effluent exchangers. These severe leaks developed from flanges that had not been disturbed during shutdown regeneration and start-up.The make-up hydrogen compressor was shut down at about the same time due to vibration. The loss of this and the leaks caused the system pressure to fail. Four pipe-fitters were instructed to tighten the exchanger flanges using a steel maul and a steel hammer wrench. By the time the pipefitters had assumed position pressure had dropped to 1050 PSIG. At the instant the hammer wrench was struck for the second time ignition occurred. The gas in the area of both leaks flashed and continuous burning followed at both flanges. The fire extinguished itself after 45 minutes.
Event Date
May 5, 1973
Record Quality Indicator
Region / Country
Event Initiating System
Classification of the Physical Effects
Nature of the Consequences
Causes
Cause Comments
Probable causes of ignition were.1. Spark caused by striking the hammer wrench2. Static build-up from the escaping hydrogen3. Autoignition of hydrogen due to expansion heating4. Autoignition of iron sulfide scale.
Facility Information
Application Type
Application
Specific Application Supply Chain Stage
Components Involved
hydrocracker, flange
Location Type
Location description
Industrial Area
Pre-event Summary
The hydrocracking unit was shutdown for a scheduled regeneration of the preheater and hydrocracker reactors and minor maintenance. Following this preparations were made to bring the unit on stream. The unit was nitrogen purged and evacuated several times and nitrogen was then fed into the pre-treating an hydrocracking sections checked for leaks at 100 PSIG. The nitrogen was then vented to the relief system and the system evacuated Hydrogen-rich gas was introduced and again the preheater and hydrocracker were checked for leaks up to 450 PSIG. The make-up hydrogen compressor was placed in service and the pre-treatment unit system was checked for leaks at 1250 PSIG. A moderate leak was successfully stopped. The hydrocracker section was being raised to a final test pressure of 1600 PSIG when two shell-to-shell nozzle flange leaks developed in the inner connections of the hydrocracker reactor charge-effluent exchangers. These severe leaks developed from flanges that had not been disturbed during shutdown regeneration and start-up.The make-up hydrogen compressor was shut down at about the same time due to vibration. The loss of this and the leaks caused the system pressure to fail. Four pipe-fitters were instructed to tighten the exchanger flanges using a steel maul and a steel hammer wrench. By the time the pipefitters had assumed position pressure had dropped to 1050 PSIG. At the instant the hammer wrench was struck for the second time ignition occurred. The gas in the area of both leaks flashed and continuous burning followed at both flanges. The fire extinguished itself after 45 minutes.
Lessons Learned
Lessons Learned
Flange leaks are not unusual where hydrogen is circulated during start-up. It is common practice to fit known leak prone flanges with stream rings to prevent flash-fires and to heat flanges that are leaking as quickly as possible to normal process temperatures. Leaking flanges not equipped with permanent steam rings to protected with steam lances until the leak stops.Unit operating and current maintenance practices include corrective measures:-1. Continued use of nitrogen for evacuating and testing equipment prior to start-up; nitrogen to be used through the full range of test pressures.2. Should major leaks occur during or after start-up with hydrogen in the system, the system pressure will be reduced to minimise leak. If a leak should continue at low pressure nitrogen will be readmitted before work is initiated.3. Employees will wear protective clothing and equipment as required while stopping leaks occuring during and after start-up.4. Exchanger flanges that have been opened and have leaked in the past will be tightened using hydraulic torque wrench equipment prior to pressure testing. This equipment will also be used in stopping process leaks that are found during unit start-up.
Event Nature
Emergency Action
Unknown
Detonation
No
Deflagration
No
High Pressure Explosion
No
High Voltage Explosion
No
Source Category
References
References
Event description extracted from the UK database ICHEME in PDF
ICHEME database is no longer available for purchase, but data can be download as PDF for free.
https://www.icheme.org/knowledge/safety-centre/resources/accident-data/
(accessed October 2020)