Conventional failure of a hydrogen compressor
Discharge pressure of a hydrogen compressor on a chloride plant fell to alarm level without any prior warning and simultaneously three bumps were heard. Smoke was seen coming from the hot motor bearing and the machine had to be stopped manually as it had not tripped. No external fire occurred and the hydrogen product was out to stack, in which the chlorine plant continued to run as required. Examination of the compressor showed that the studs of the driven end first stage cylinder head had failed and subsequent examination of the piston rod showed that this had failed at the threaded attachment to the cross head, the motor bearing showed severe heating by oil smoke and paint discoloration. The following conclusions were made: 1. The piston rod failed by fatigue initiated at the cross head end threads. 2. No evidence was found of any individual error in assembly or operation which contributed directly to the initiation of this incident. 3. Indications are that the vibration trip system was not in function at the time of this piston rod failure, increasing the secondary damage at the motor and resulting in the provision of a possible ignition source for the escaping hydrogen. 4. The damage to the compressor indicates that both the compressor and the driver motor, acting as a flywheel, may have been subjected to extreme impulsive forces, with possible damage being caused to frames or shafts. However, subsequent inspections did show distortions to be negligible and no cracks have been detected. 5. There is some evidence that liquid passes into this compressor in sufficient quantities under some conditions to cause mechanical distress.
Event Date
October 3, 1979
Record Quality Indicator
Region / Country
Event Initiating System
Classification of the Physical Effects
Nature of the Consequences
Causes
Cause Comments
The INITIATING CAUSE was the failure of the compressor piston due to fatigue, under conitions which worsened the consequence. This is an incident where a hydroge nsystem was involved, but failed by an pparently conventional intiating cause (unless the pisotn failure was caused/affected by hydrogen attack). The ROOT CAUSE is related to material defects.
