In this case study an extensive root cause analysis was performed on a compressor test rig following the observation of unexpected high amplitude vibrations during the rig commissioning. Typically, the location and the frequency of an observed vibration is not enough to pinpoint the cause of the behavior in a system with multiple components. The test rig comprising of an electric motor, gearbox and compressor rotor may be subject to rotordynamics, torsional dynamics, mechanical eigenfrequencies and fluid dynamics in the compressor. In order to understand the source of the vibrations observed at the gearbox, lateral vibrations of the casing and torsional vibrations of the rotor were measured simultaneously. In parallel, the system was modeled in MADYN 2000 in order to match the observations with theoretical torsional eigenmodes of the drivetrain. The modeling of the system resulted in a torsional eigenfrequency lying within 10% of the observed gearbox vibration, making it a viable candidate to excite gearbox vibrations through a varying torque. To be able to measure this mode shape, a temporal measurement setup for torsional vibrations had to be installed. The measurements of the test campaign revealed that a torsional resonance is indeed present at the frequency of the observed vibrations. However, a lack of coherence between the torsional and lateral vibrations suggests that the torsional resonance is not the root cause of the gearbox vibrations. Two separate resonances were observed as the gear mesh frequency crossed the torsional eigenfrequency and the observed high amplitude gearbox vibration. Evidence of a structural eigenfrequency of the gearbox was found through an impact test. Although the torsional resonance is ruled out as the root cause, it is presumed that the two resonances do experience a positive coupling, causing the particularly high gearbox vibrations.
