Intermittent service can be a challenge. The majority of our compressors run constantly. But, for the ones in intermittent service, we have seen a number of problems that have required us to make changes.
We have a spare coker wet gas compressor in one of our cokers. They used to leave that machine in a ready-to-run condition, blocked in from the process and under nitrogen purge. But, that compressor has cone seals that cannot function correctly when they are not spinning. Even with the N2 purge, they would get slight leakage through the block valves which accumulated liquid in the compressor case. The liquid corroded and fouled the rotor resulting in high vibration on start-up. Now, we keep that compressor fully blinded from the process. If we need to run it, it takes 4 to 8 hours to get the blinds pulled before it is ready to start-up.
We have a spare induced draft fan on one of our hydrogen plants. That big blower sits idle most of the time. There are no block valves, only inlet louvers. The louvers leak by and water condenses out of the hot flue gas. This corroded the wheel over time. As the stiffener ribs on the back of the wheel got weaker, the natural frequency of the rotor dropped down into run speed range. We experienced extremely high vibration or start-up. We had to change our operating practices and now we run both fans in parallel all the time to prevent this. We are considering an upgrade to a wheel material that won’t corrode as readily.
We have a large barrel compressor with a steam turbine driver that sits idle during winter. The compressor has not had problems associated with this. But the turbine has had issues. The turbine and associated support systems were not designed to sit idle in extreme cold. They had freeze up problems on the surface condenser, gland condenser and seal steam systems. The operators responded by introducing steam into the condenser. But this is a bottom exhaust turbine. Heat came up from the condenser and we found that the turbine was sitting there idle and quite hot. We were very lucky that they didn’t severely bow the turbine rotor. We offered them two options to address this. They could fully winterize the systems with thermostatically controlled electric tracing and insulation or they could run the turbine year-round. They opted to run the turbine all the time. They idle the turbine down to minimum speed and recirculate the gas on the compressor to keep the entire system hot.
Synchronous motors can suffer hard starts depending on the driven machine. If I was buying a machine for intermittent service, I would probably specify an induction motor rather than a synchronous motor. There would be an efficiency penalty, but the motor could be designed to take the additional starts. We have a coker jet pump that is driven by a 3500 HP induction motor. Because of the nature of decoking, they motor sees as many as 1000 starts per year. The motor is an enclosed motor with air-to-air exchanger. It has a heavy duty rotor with copper bars. The rotor was designed as a stiff rotor design so that it does not have to pass through critical speed on each start-up and shut-down. The motor has been very reliable.
For intermittent service, I would increase the service factor requirement for the coupling from 1.5 up to 2.0 or more. I would insist on a lube oil system with two pumps (no shaft driven oil pump). I would size the lube oil heater to be sufficient to maintain full normal oil temperature (100 °F minimum) when the machine was not running. I would keep the oil in circulation when the compressor was down. This would help reduce the chance of water condensing in the oil system when the compressor was not running. Depending on size, I would insist on a continuous monitoring system armed to trip on radial vibration or thrust position. I would verify that the gearbox design has a substantial overdesign. I would insist on a full torsional and rotor dynamics analysis of each major component and the full train. I would consider increasing the minimum separation margin from any critical speeds beyond the minimum API requirements.
I could be more specific if you provided additional details about the compressor, driver, service, expected number of starts per year, etc.
Johnny Pellin
