AFTER OUR RECENT SHUTDOWN I STARTED DETECTING PROBLEMS ON A 6.6KV SIEMENS MOTOR. ALL THE DATA POINTED TO MOTOR DE BEARING PROBLEM. PRIOR TO THE SHUTDOWN THE STANDBY PUMP WAS SENT TO THE O.M. FOR OVERHAUL. (SO, IT WAS CRITICAL THAT THE OPERATING PUMP IS KEPT IN SERVICE.) FIRST HIGH gSE’s WERE DETECTED. WE LUBRICATED BUT SUSPECTED BLOCKED PORTS. THE SIGNATURE MOVED DOWN IN THE SPECTRUM AND BECAME VISIBLE AS LOW FREQUENCY BPFO,s AND FTF PRESENT IN THE SPECTRUM -LESS THAN 1000Hz. (CONFIRMED WITH HIGH RESOLUTION SPECTRUM). AS SUB SYNC STARTED TO PICK UP WE MADE THE CALL TO HAVE THE MOTOR CHANGED OUT. A DAY BEFORE THE PLANNED CHANGE OUT THE TRAIN TRIPPED. WE STRIPPED THE MOTOR AND FOUND NOTHING WRONG WITH THE BEARINGS!!!. THE PUMP. HOWEVER HAS SEIZED, SHEARING THE SHAFT.

HAVE ANY OF YOU COME ACROSS SOMETHING SIMILAR?
YOUR FEED BACK WILL BE GREATLY APPRECIATED.

6.6kV_siemens_Motor_De_Spectrum.doc (66 Kb, 79 downloads) word doc

I suspect more information revealed with data above the 25 orders displayed.

Do you have waveform Acceleration pk-pk values (g's), up to 3000hz preferred.

Mike

More data of course is needed, but SpikeEnergy clearly points to, at least, BPFO harmonics. Demodulated signal does not always point to a spall. Harmonics there could be due to friction modulation at BPFO due to a slight egg shaped OR as it has been discussed before by others. From the data it is not clear if BPFO is present in the normal spectrum up to 2-3 kHZ which would've helped in making a more reliable call.

There is still a possibility that if the race was examined under a 25-50x microscope, small pits/spalls could be detected.

Was the cause of the pump bearing seizer ever determined?

David

With this being an NU217 bearing, is it correct to assume this is a belt driven pump? If so, the belt tension on the motor bearing might have contributed to the defect signal being more than it actually was.
Is the pump speed 598 rpm?
But if this is not a belt drive, oh well.
What is the 5x in the spectrum of the motor? Vanepass if pump is direct coupled?

So, we have a motor which shows some symtpoms but has no bearing damage attached to a pump that doesn’t show symptoms (is that correct?) and then later the pump bearings seized which led to shearing of the pump shaft (is that correct?)

It’s an interesting scenario. Just brainstorming some possibilities:
1 – The motor bearing vibration had nothing to do with the pump problem.... just a coincidence that you saw some false motor bearing vibration at this point in time (doens’t seem very likely).
2 – The motor bearing vibration was actually coming from a pump bearing fault frequencies transmitted through shaft or foundation and seen on motor bearing housing?
3 – There were some very abnormal stresses in the system (severe misalignment?). These led to both the pump bearing problem as well as creation of the frequencies on the motor bearing.

That’s all I can think of (someone else has an idea?). In the choice between 3 unlikely scenario’s, I would vote for #3. I have seen high bearing temperature give high BPFO fundamental apparently due to reduced internal clearnace in the bearing (even though there was no race damage). What we have here looks like high 1*BPFO and 2*BPFO although we don’t know what lies above that. It doesn’t sound out of the question to me that severe bending stress on the shaft might create something like this. (Although I’ve never seen it... again just trying to pick among unlikely scenarios).

At any rate I agree with David, it would be interesting to know more about the pump bearing and shaft failure. What type of bearing on the pump (same defect frequencies? Could shed light on scenario 2 above). Along with the other questions above: was there any higher frequency stuff present and what type of connection or coupling between motor and load.

Hi all,
Failure of pump found to be blocked suction strainer.(Ralph picked it up - 4x RPM in spectrum) The electronic high DP alarm failed in operation and Instruments never got around to replacing the reported faulty PG gauge, normally used as back up(. I have since reset the BPF parameter lower. Alarm at 1.25mm/sec Critical at 1.75mm/sec.
The second pump also failed with clogged suction strainer. The root cause found: Due to expansion and contraction of vessel after plant restart (shutdown of unit) resulted in corrision peeling off and blocking filter.
We have since put in place:
1) PM04 - Filter clean once per month
2) PM04 - filter clean every 4 days for 3 months after unit Startup.

With regards to the bearing on the motor. Data is taken very close to the bearing. I.E. on the bearing housing. Due to the accel proximity to the fault,the change in amplitudes was marked and I over reacted , by not including this in my analysis. Since,I have made a point to take a bearing housing temp reading to correlate date.

Question; The pump is a 12 stage multistage- sulzer pump.About 2.5m long. With throttle bush in center.I am considering taking an additional point in the middle of the pump(near throttle bush) to detect vane pass.Good or bad idea?