Severity
Incident
Leak
Yes
Ignition
Yes

A brazing retort in a shop malfunctioned and resulted in an explosion that propelled the retort shell to the roof of the brazing area and then back to the floor. There were no injuries but damage was sustained by the furnace housing and the retort shell.

Administrative personnel were soon on the scene to make a preliminary assessment of the situation. An expert safety team was retained to assist in the investigation of the explosion. The safety team conducted their initial field investigation on the afternoon of the explosion and again on the following day.

Once it was determined that the explosion was the result of an ignition of a flammable mixture of hydrogen and air, the next step was to determine how air ended up in the retort, given that the retort is nominally pumped down to 28 inches Hg vacuum, and back filled with a positive pressure of nitrogen before the hydrogen is introduced.

A number of potential mechanical/physical possibilities for the presence of air in the retort were considered and eventually rejected. A careful study of the metallic structures involved showed no leaks or cracks that could account for air in the system. Detailed examination by safety specialists ruled out both o-ring and ceramic vacuum seal failure. Tests on the hydrogen and nitrogen storage tanks and their respective piping network to the shop proper showed no flaws within these systems, thereby eliminating the possibility that a leak could have introduced air into the retort system. When the vacuum gage at the inlet to the retort was found to be operating properly, the safety team had eliminated the last of the mechanical/physical possibilities for the air in the system that led to the explosion.

At this juncture the safety team turned its attention to two non- mechanical/physical possibilities for the explosion, i.e., either the written procedure to be used with hydrogen retorts is badly flawed, or the operator failed to follow the procedure. A series of tests showed the procedure itself to be sound (actually allowing for a "robust" level of safety, according to the FaAA findings), so operator failure to follow procedures became the focal point of investigation. Further analysis clearly indicated that the cause of the explosion was the failure of the operator to effectively evacuate the retort and backfill with nitrogen, combined with a hydrogen purge far short of the specified 20 minutes at 50 SCFH.

Incident Date
Mar 06, 1996
Setting
Equipment
  • Heating Equipment
  • Retort
Damage and Injuries
Contributing Factors
When Incident Discovered
Lessons Learned

The primary lesson learned was that the active hydrogen facility and existing operating procedure, at the time of the accident, were sound. While this now has been determined, the previous form(s) of this system, associated documentation, and accident history did not provide sufficient basis to assume continued safe operation. The review of furnace operation subsequent to this accident now provides a basis to evaluate the safe operation of furnace #4 as well as other similar facilities.

The failure to apply the appropriate procedure suggests the need for either additional administrative controls and/or periodic refresher training.