System Description
Two pipeline reciprocating compressors driven by synchronous motors. There is a large flywheel between the motor and the compressor. One VFD shared between the two vibration motors, used as a soft start for the first motor which is switched to across the line”, and then the second unit is started and run on the VFD.
Symptoms
Failures:
Several motor shaft failures in the stub shaft area have occurred, caused by torsional vibrations and improper shaft materials.
Vibrations:
Bearing housing vibrations are not high. Torsional vibrations (TVs) are an issue on the vibration motor side of the flywheel. TVs are lower than predicted on the outboard end of the compressor, but higher than predicted on the outboard end of the motor.
The original symptoms were exhibited in the Unit 1 motor. As the speed was increased and the first torsional critical speed was excited, the TVs peaked and then stayed high at the TNF even though the motor speed was still increasing.
Unit 2 did not exhibit the“1ock-in” TV at a fixed frequency. The amplitude of the TVs on the motor outboard was higher than predicted.
Mechanical Natural Frequencies: not an issue in this case.
Torsional Natural Frequencies (TNFs): The TNF measured in the field is as predicted using computer models. The damping used in the predictions is consistent with the damping measured in the field.
Current Pulsations: Not measured correctly.
Mechanical Changes Attempted
The vibration motor shaft material problems have been corrected. This has had no effect on the vibrations, but will give better motor life.
Electrical or Software Changes
In a recent test, the VFD manufacturer made software changes to the drive. We are not privy to the changes made.
The result of the changes was that the two motors now behave the same when on VFD control. The lock-in constant frequency of TV no longer occurs on Unit 1.
The amplitude of TV at resonance is higher than predicted on both units at the front of the vibration motor.
Conclusions
We believe that the compressor torque fluctuations cause speed fluctuations in the motor, which lead to current fluctuations being generated by the motor and sent to the VFD. The VFD then amplifies the current fluctuations which cause larger speed fluctuations (another way of saying torsional vibrations) in the motor.
The flywheel apparently isolates the compressor from the torsional vibrations (torque fluctuations) coming from the motor. It was originally installed to isolate the motor from torque fluctuations coming from the compressor.
Designers of torsional systems can only assume that Variable Frequency Drives will not amplify the torsional vibrations in a system at this time. In fact, we have seen that amplification does occur.
Therefore, the VFD manufacturer must be made responsible for tuning the drive to match the electrical and mechanical system characteristics, eliminating the amplification.