I’m having some vibration issues with GE 1000 HP frame 8311S motor driving our compressors. These motors are exhibiting vibration levels at ~0.3 inch/sec in vertical direction on both bearings. As you can see on the attached charts the predominant frequency is at 1xrpm but in horizontal direction I have a strong beat between 2xrpm and 2xLF.
The motor was inspected and bearings and bearing fits were found to be OK. After the rotor was re-checked for unbalance it was found to be slightly off (?) and was re-balanced. When put on dyno without the load the readings were low but increasing the load increased the level to 0.3 inch/sec. It seems to be related to thermal condition of the motor.
The air gap variation was less than 8%.
I did some bump tests on the motor bearings but no resonance was found around 1x, 2x or 3x rpm. The motor doesn’t have any rotor bar issues revealed through the motor current analysis. The rotor was also checked for hot spots in the electrical shop and found to be OK.
I wonder if any of you had similar experiences with exactly this frame? Could we have some rotor thermal instability or it is design issue.
What puzzles me is that I have two identical motors and both behave the same way.
Your feedback would be appreciated.

Alex

GE_1000_hp.doc (68 Kb, 39 downloads)

Alex,

Your data does not illustrate a Beat frequency very well. A high resolution spectrum will separate 2xSS from 2xLF and will probably indicate they are nearly the same amplitude, hence beating. This is a good indication of shaft misalignment.

Impacting bearing housings on a large motor can be ineffective (useless). Your impact testing (or ODS test while running) should include the motor casing including upper air box, since this is often where a resonant structure occurs.

Walt

quote:

Originally posted by Alex J:
I’m having some vibration issues with GE 1000 HP frame 8311S motor driving our compressors.

Is this condition 'new'? Are these motors mounted on the same kind of baseplate or machine base? I'm just curious if there is something other than just the motor at play, here. If they are vibe-tested when disconnected from their load they seem okay, but when the load is applied they go a bit squirrely... perhaps there is something about the support for the motors that they are mounted on that doesn't handle the torque as well as it should? I'd be giving some thought to a structural issue... this could be the tail wagging the dog.

Thank you Walt and Mike,
Here are some additional info which can answer your questions:
1. I did high a resolution test and this is why I know I have a beat at 2xrpm and 2xLF, however this is not predominant peak. The major issue is change in 1xrpm from no load to load conditions.
2. Two discussed motors have the same support structure driving 3 stage integrally geared Joy air compressors.
3. I was testing the motor at the electrical shop under load from dymo and the same was happening. It was the same vibration pattern and the same changes from no load to loaded conditions. This excludes structural issues and shaft misalignment.
4. It takes some time for vibration to change after motor is loaded and this is what leads me to shaft thermal instability suspicion. The shop did some rotor testing for hot spots but nothing was detected.

I still have feeling that it could be some design issue and was hoping some other people had similar issues and maybe solutions.
I contacted GE about this issue but so far no feedback.
Thank you,

Alex

Alex,

You are on the right track looking for rotor thermal change. Some large motors that run at 3600 rpm nominal speed operate above their 1st critical or balance resonance. You can verify this by rotor impact test or by a coast-down test measuring 1xSS amplitude and phase. I suggest the coast-down test both with cold and hot conditions.

If the 1st critical is below normal operating speed, then it is a Flexible rotor rather than a Rigid rotor when considering balancing procedure. The common practice of placing balance weights at the ends of the rotor (2-planes) assumes that rotor is rigid, and it does not compensate for mid-span unbalance. I worked on a 4500-hp 3575-rpm nuclear plant feed pump motor that required mid-span balance correction.

Walt

Have you checked the position of the motor from off-line to running condition temp? We may have a similar situation.

Mick,
Could you please elaborate on this one? I'm not sure what position you are describing.

Alex, we have motor similar and the problem ended up being a soft foot after warm-up. 2x line frequency (7200)due to internal alignment and 2x (7170) (due to coupling alignment.

What you can do is use a set of Dodd bars or Essinger bars on your motor compressor set. You set them up when the unit is cold and measure/record the position. After unit is up to operating temperature measure and record it's position again and you will identify whether the machine is moving up, left, right. Does this make sense?

Elevatorguy,
I think we concentrate too much on 2xLF and 2xRPM. The feet were machined and this did not change anything. It was tested on flat stend in electrical shop and later in place and the same results.
I'm more concerned by changes at 1xrpm as this is predominant and generates a lot of energy.

Alex

What type of bearings does the motor have? Ball or Sleeve?

Also, I notice that the trend line on the horizontal increases rapidly about day 38 then decreases until about day 47 and then both horizontal and vertical increase rapidly. What changed?

Dwight,
This is 1000 hp motor with sleeve bearings. The increase around 38 was when the motor was tested at the electric shop without the load. At that time it was strong beating around 2xLF. but 1xrpm in vertical direction was not too bad. What you see later it was when the motor was started with the compressor. The readings were low but after temperature increased 1xrpm increased a lot. This is why I suspect thermal instability.

Alex,

It sounds like a thermal growth issue as others have commented. Have you done any phase analysis on startup and after thermal stabilization? Looking at 1x phase before/after may reveal something across the coupling or across the motor.
In the end, a thermal growth study is most likely needed.

One method that has been used to detect significant hot spots, those that bend the shaft has been to setup dial indicators at shutdown and measure runout. Having installed proximity probes makes the job automatic by taking proper coastdown data.

This is a case where temporary prox probes away from the bearing can possibly help - having the probes at the bearings is good to help define the rotor shape.

For those who like to put probes away from the bearing (not the proper place for monitoring), one may need to use longer range probes (e.g. 11 mm tip vs 8 mm tip). On many large steam turbines one may find such probes used to detect rotor bow or rotor eccentricity. The same principle can be applied here.

Otherwise, try to use proper safety proceedures, stop the machine, and use dial indicators. The prox probes do a much better and safer job.

Of course one needs to compare pre-startup to post-shutdown, before heating to after heating.

Yes, runout may be expected since the shaft has not been prepared. You have to suck it up and deal with it.