Lesson Learned on Electric Motor Shaft

A 750-horsepower motor for a power station air compressor was sent to the motor shop for a winding failure. It was refurbished and returned to the station. Plant personnel performed a precision alignment of the motor to the compressor, yet high vibration was found on the initial startup.

Alignment tolerances were reviewed, soft foot conditions were minimized, and motor mounting and support adjustments were made to ensure no resonant conditions existed. However, high vibration levels were still found on the motor. Further investigation of the uncoupled motor vibration and phase identified a bent motor shaft.

Alignment readings were all within tolerance for the 3,600 RPM machine. Soft foot problems were minimal and were removed with step shimming. Due to the extremely high vibration readings in the order of 0.8 inch/second velocity, which was predominantly in the horizontal direction, resonance was suspected.

Chart 1. Comparison of motor points with bent shaft.

An investigation of the motor mounting structure revealed a stiff, good supporting structure. Bump tests on the motor found the natural frequency well above and out of range of influence by the motor operating speed. Resonance was ruled out, so the unit was started and phase data was taken.

Horizontal readings found similar amplitudes of 0.7 to 0.8 inch/second at 180 degrees out-of-phase between motor bearings. Much lower vertical readings were found on the outboard (opposite drive end) motor bearing of 0.02 inch/second and a high of 0.26 inch/second at the inboard (drive end) bearing at 180 degrees out-of-phase.

Axial readings on the motor from one end to the other showed acceptable vibration levels, but at 180 degrees out-of-phase. Phase readings across the coupling showed mostly in-phase readings with low, acceptable vibration readings on the compressor.

Misalignment was suspected, but to rule out motor rotor imbalance and a possible bent motor rotor, an uncoupled motor run was requested. Vibration levels were low vertically, but still somewhat elevated horizontally for an uncoupled motor - in the order of 0.1 inch/second at running speed. Phase readings were mostly all in-phase except for axial end-to-end comparison.

The difference in horizontal to vertical readings along with the axial out-of-phase suggested a possible bent rotor. Dial indicator readings did find the motor stub shaft bent by 0.009 inch at the coupling hub mounting location, tapering back to less than 0.002 inch run-out near the shoulder before the inboard bearing.

The effects of the bent shaft were greatly exaggerated in the coupled condition, resulting in the coupled amplitude and phase readings mentioned previously.

Chart 2. Readings before and after the bent shaft.

The motor repair shop stated that they had balanced the rotor to "as low as we could get it"; however, we had not provided any balance criteria, nor had we requested any vibration testing at the motor shop prior to shipment.

With the location of the bent shaft, it is likely that the motor shop would have never seen the effects of the bent shaft with the rotor on a balance machine. Shaft run-out tolerances were never included as part of the repair specifications by our company, and none of the repair documentation was provided by the motor shop.

Had we requested shaft run-out checks during repair and provided acceptance criteria that included a final vibration reading and analysis before the motor was shipped, it is likely we may have caught this problem at the motor shop.

Now, we have no idea if the shaft was bent during initial disassembly, during shipment to the motor shop, while at the motor shop or damaged upon return shipment. We had no case for liability against the motor shop, so the cost of the repairs fell on us.

In the future, we plan to specify balance acceptance criteria, disassembly "As-Found" and "As-Left" criteria, and request review of vibration signatures on repaired motors prior to shipment. We are also developing job plans for alignment work to include checking shaft and coupling hub run-outs as part of the initial steps to performing alignment.