vibration/alignment/balance

Natural frequency? Did you made any bump test for vibrations?

Hi, @Michal Janicek Thank you for the fast response. No, I haven't done the bump test yet. Do you think this frequency is related to natural frequencies? But my question now is: what is the possible excitation for this frequency?

Hard to say when I don't know the device. But I once dealt with something similar. It was the pump too. And a frequency was detected that was not related to the revolutions at all. In the end, it was the natural frequency of the pipeline that was transmitted to the pump.
It was excitated by the flow of water.

Vertical pumps are notorious for what are called Reed Critical Frequency (RCF) problems.  Essentially the first cantilever bending mode of the discharge head and motor as an assembly.  If the amplitude is higher at the top of the motor than the bottom, it is likely the RCF.  The only way to tell for sure is to do a modal test, putting sensors at the top of the motor, bottom of the motor, and at the bottom of the discharge head (baseplate), impacting at the top of motor.  There will typically be different frequencies in the directions parallel and perpendicular to discharge.  If you're looking at the vibration in real time, the excitation of a natural frequency will look very unsteady, I call it "breathing", with the amplitude constantly changing.  Other mechanical excitations, like 1X, will not exhibit much amplitude change.

A has issue first, now B is also showing same vibration and frequency, correct?

Fluctuating vibration levels, are you collecting walk around vibration data or is sensor permanently mounted?

A and B motors are the same?

Do you have any process data, motors amps, temperature, suction/discharge pressure?

Was there any maintenance activities done when the vibration increased?

Dave

ahmad alfakeer,

My reply was deleted.

I can not seem to get back in to replace it.

Good morning, @Michal Jelinek

thanks for your reply , I had been checking the flow and the amper of the motor with vibration before and after the vibration growing up and as We say from the  attached trend,  ther

e are some correlation between vibration and flow of pump ?! we also plan to do bump test at the first opportunity  

Good Morning@Lyn Greenhill , Thank you for the explanations. I agree with you. Do you mean the bump test will support us in finding the natural frequency because I don't have the calibrated hammer on site now? And another question to you: how can we deal with this issue (RCF)? Do you think the best option is to control the excitation or the increase in stiffness in the base plate?

Good Morning@Dave Reynolds , thank you for your inputs : 

A has issue first, now B is also showing same vibration and frequency, correct? yes that is right 

Fluctuating vibration levels, are you collecting walk around vibration data or is sensor permanently mounted? yes there are permanent  sensor in the machines as one reading only ( protection )   and also we collecting the vibration the portable analyzer 

A and B motors are the same? yes the same spectrum behaver 

Do you have any process data, motors amps, temperature, suction/discharge pressure? yes i collect it with operation as attached picture 

Was there any maintenance activities done when the vibration increased? no 

Hi all , I execrated from the DCS the only trend of the NDE vibration with flow as you showed in the attached  picture , do you mind the flow is the excitation of this Strang frequency? 

Thank you for the answers you have provided.

It appears the motor current (purple) and discharge pressure (orange) follow each other quite well. 

The chart shows NDE vibe (red x/blue y) and DE vibe (dark blue x/green y). NDE has 7 mm/s scale and DE has 4 mm/s scale. DE, assume this is lower motor bearing, has half the vibration as the NDE which we find typical with a vertical motor. All four vibration readings react the same with the pressure/amps. When pressure is at its highest vibration is at lowest levels, when pressure decreases vibration increases. Assume the pressure fluctuations is normal for this pump?

How is suction side designed? How deep, approximate pit dimensions, fluid temperature, is the surface of the water smooth or disturbed and swirling.

Example of vortexing: https://www.youtube.com/watch?v=ltWFzjRNBaE

Do you hear any noise change when the pressure drops on the pump?

Assume the low frequency vibration 7.5 hz is always present no matter the discharge pressure?

Dave

Good Morning @Dave Reynolds , thank you so much for your inputs <

first of all , the orange trend for the flow of the pump is not for the pressure , and these pump delivered water to the LP drum. The flow of the pump depend on the capacity of the steam turbine; in other words, when the 1 HRSG is in service, the flow of the pump around 82 kg/s, and when the 2 HRSG is in service, the flow rate becomes ( 140kg/s These depend on the grid.

and the low frequency vibration 7.5 hz is always present  in the spectrum but different in the level and this frequency fluctuated and was not stable for the level ?

Do you think the change in flow is the excitation force for the sub-structure response? or not, as the trend shows?

Greeting all , 

In the updated case , I had been trying  to test the behavior of the vibration when I changed the flow of the pump because I suspected the flow might be the excitation for this mystery frequency , and all the trend results are shown in the attached pictures. I got the trend of the vibration at the NDE ( motor ) with the flow for these cases : 

• before the issue start .
• trend vibration with the flow when two HRSGs  ( high flow ) are in service after the issue happened .
• trend vibration with the flow when one HRSGs ( low flow ) is in service after the issue happened.

Please review the attached trend and your inputs is appreciated.

With the motor having been rebuilt, the pump and process have yet to be ruled out. It appears the vibration is related to process.

Has this issue been present for years or just showed up, your charts are showing a short history of the machine.

How is suction side designed? How deep, approximate pit dimensions, fluid temperature, is the surface of the water smooth or disturbed and swirling, has temperature of liquid being pumped change?

Do you hear any noise change when the pressure drops on the pump?

Have seen suction vortexing, worn "bell' on bottom of pump causing flow changes, pump not submerged enough in liquid, poor mechanical fits between pump base/adapters/pump head, pipe strain, lack of piping support and nearby machinery operating all create low frequency vibration. Have you/can you vary the liquid level on suction side of pump?

If we understand correctly the NDE is the top of the motor which has the 7.5 hz vibration, assume the vibration at 7.5 hz has lower amplitudes on rest of pump? Can you provide FFT/spectrums of NDE x,y,z DE x,y.z and any pump vibration readings.

As others have stated, a series of bump tests should be done on motor and pump to verify a natural frequency is not close to 7.5 hz.

Have you done any running soft foot checks on the motor to pump interface, pump baseplate, adapters?

Dave

Ahmad,

The modal test will determine if you are exciting a resonance at 7.5 Hz.  While having subsync RCF resonances in a vertical pump is common, excessive vibration at the resonant frequency is not.  There are some ways to deal with this kind resonance, but first confirm you indeed have the mode with the modal test.

Hi @Dave Reynolds and @Lynn Greenhill , Sorry for the delay replay. I am doing bump test at the different locations as shown in the pictures attached ( motor case body at the NDE side , motor flange for three directions , pump flange and pump casing), and I also shared the picture of the suction and discharge pipe of the pump and the screen shot of the DCS process parameter. Please show it and give me your input.

Thank you for your support.

I looked at your "modal" data.  It does seem to indicate that the top of the motor has a resonance around your problem frequency.  However, the modal data is nowhere near suitable enough to make a confident conclusion.  You need to do a proper modal test, with multiple simultaneous measurements and impacts in orthogonal horizontal directions at the top of the motor.  Results must be presented in Bode plots (sensitivity and phase).  Your FFT resolution is too coarse, increase the number of lines such that the FFT bin width is no greater than 0.5 Hz, preferably less than that.

Looking at the general arrangement of this motor on the discharge head, I would say the chances of a low frequency RCF is very high.  As I said before, this is not normally a problem.  Something in your process or operation is creating enough broadband excitation to drive the response of the RCF.  I would recommend a very careful evaluation of how these pumps are operating to identify what is driving the RCF and then eliminate/modify the source.

@Lyn Greenhill

 also unfortunately I don't have a tools to doing modal test ( calibrated hammer and multichannel data collector ) at this time , only my capability to doing the bump test 

Greeting @Lyn Greenhill ,

thank you for your reply , I really appreciated  your input in my case , and I agree with you , but unfortunately I don't have the right tools to doing the impact test , and my thinking now is: what are the excitation source for this low frequency( 7.5 hz Because for the motor A ,when the vibration level started to increase in the NDE side m i doing FFT analysis and  We showed in the FFT the bearing frequency and the low frequency, my subsect the losses mounting in the bearing excitation the common RCF , and I recommended to maintenance team to check bearing , and the conclusion of finding when the motor A open , huge corrosion in the bearing place and shaft  and this subsect because the water entering from the leakage water from the above platform of the motor and the Cooling fan in the NDE was loose and ovality , but unfortunately after all maintenance  work had been done the level of the vibration still high and fluctuated as you shown in the previous FFT, but to be honest  I don't trust the work of the maintenance , in the new FFT the bearing frequency and the level of Gs valve decreased but the low frequency ( 7.5 hz) still dominants in the FFT , I don't know , but my thinking now these N.f exited by some thing ??? we doubt for the process and changing flow of the pump the cause of excitation because the grid decided last year to star the steam turbine many times in the day at one HRSG in service and this action will change the flow of the pump from ( 82kg/s ) for one HRSG in service  to ( 140kg/s ) for two HRSG Inservice , we monitor the vibration behavior when the flow changes from 82 to 140 As you can see in the attached trend, there is  no relationship between the vibration and flow changing in the pump; I don't see clear correlation between them. what is your thought ???? 

The last "trend" plot you posted, it appears that when flow goes down vibes go up? Yes, not everytime, but in general the vibration is reacting to the flow. When was the last overhaul of the pumps? The green flow line is constantly changing at higher flow rates...and under lower flow rate it seems more stable. Do you hear any audible noise coming from the unit at either low or high flow?

Obviously these units have been in service for awhile. The natural frequency of the units has always been there from original design. Why get excited now? If you have vortexing in the suction side, this can create low frequency vibration. Is the scaffolding that is erected around the units touching either unit? Is the scaffolding vibrating?

Dave

Greeting @Dave Reynolds 

Sorry for delay replied , 

 I agree with there is a relationship between the flow and the vibration but as you see, the level of vibration is still high when we compare it with the history ,I don't hear any audible noise coming from the unit at either low or high flow, and the area is noisy. Exactly this is my equation now. The natural frequency of the units has always been there from the original design. Why get excited now? and now any type of maintenance has happened at the pump !

How can we evaluate if there is  vortexing in the suction side ? Can you assist me in determining the method to investigate this phenomenon?

No, scaffolding was erected at the pump after the issue started to collect vibration data because there was no access to collect data ? and no vibration observed at the scaffolding at this moment ?

and please show the attached PID Drawing  for the system  

One more thought on the appearance of this strange frequency after the pumps have been running for a while.  Have you checked the accelerometer(s) and complete system?  Obviously the resonance has been with the machines from the start, and unless the process has suddenly changed, there is really no good reason for this subsync peak to appear.

For both pumps to be exhibiting the same symptoms, this would lead me to believe it is a process issue, in my opinion. You changed bearings in at least one motor and nothing changed.

If you have the ability to look at the surface of the water on the suction side, you are looking for a swirl in the water or large drop in pit level from low to high flow.

Is there a suction strainer for the pit or screen mounted on the suction of the pump? Your pumps require X amount of liquid above the suction bell at all times per design specs.

Here is a link showing vortexing https://www.youtube.com/watch?v=lIQQHsHoQYo

If the water level falls too low in the pit, you risk having vortexing. If the liquid entering the pit is full of air, hot or turbulant, this too can cause vortexing. If the "suction" bell shown at the bottom of the pumps in the video is damaged, this can cause vortexing. Worn impeller, worn pump bowl will cause performance change, not able to deliver the same pressure and flow as in the past.

Do you have any trend data on the suction side of the pump? Pressure/flow/temp/level?

Have heard of silt building up on suction side of pump changing the dynamics of the pumping system. If the liquid you are pumping has a lot of fine silt/sand this can build up in the suction side. Do not have any details on your suction side yet.

Dave

Greeting  @Lyn Greenhill , thank you for your response , I am still Search  for the source of excitations , the I&C team had been checking for the sensors and they tell me the senser working very well, but we overserved something strange ,when I collect the vibration data by normal resolution 1600 lines , the vibration was normal , but when we collect high resolution of 6400 lines, the vibration goes up and was clearly high.

Greeting @Dave Reynolds , thank you for your response , 

As you can see in the attached picture, I tested today by raising and lowering the condensate tank's suction side level while keeping an eye on the vibration. When I increased the level, the vibration decreased, but it eventually increased again without the level changing, The fluid is demi-water and the temperature in suction is around 50 c, The last thing I will want to check tomorrow is the strainer on the suction side. 

So you change the resolution on your data acquisition equipment and the peak appears?  Sounds like aliasing.  What is you Fmax?  Shouldn't be more than 1000 Hz.  1600 lines of resolution is a bin width of 0.625 Hz, 6400 lines is 0.156 Hz.  With either resolution, you should still see a peak around 7.5 Hz, the amplitude may not be correct with the lower resolution, but it still should be there.  Can you check your data acquisition system against a known source, like a portable shaker?  What kind of windowing are you using?  For this type of steady vibration, a Hanning window is fine.  I would use 1000 Hz bandwidth, 5000 lines resolution, bin width 0.2 Hz.

Note that fault frequencies of your bearing at 1491 rpm are:

6230:  FTF = 10.546 Hz, BSR = 80.303 Hz, BPFO = 116.011 Hz, and BPFI = 157.339 Hz

7228:  FTF = 10.798 Hz, BSR = 71.430 Hz, BPFO = 172.764 Hz, and BPFI = 224.836 Hz

Nothing above 250 Hz, so with a Fmax of 1000 Hz, you will pick up the first 4 harmonics of any bearing fault frequency.

It would appear pit level does/did change the vibration. If I read your chart correctly, the vibration went from 2 mm/sec up to 6 mm/sec in less than 20 seconds? The pit level was constant at the time of vibration increase, was there any other changes occurring when the vibration went up, noises?

Did the motor amps also go up when the vibration increased while adjusting the pit level? In my opinion, if the amps do not change then the sensor/circuitry is in question. If the amps go up along with vibration, then process or mechanical drag maybe causing the vibration

Odd the vibration changed so rapidly, do both pumps exhibit the same changes in vibration with varying pit level?

Is it possible a bushing on the pump shaft is spinning in the housing?

Dave

Greetings @Lyn Greenhill 

I doing  the test again for the low and high resolutions, and for both cases, the 7.5 hz appears in the spectrum Sorry for this mistake from my side.

Greetings, @Dave Reynolds , no  obvious changes in the apers of the motor , I doing test of level of CE tank for one pump only , but the behavior of the two pumps was the same for everything.

Yesterday , I took the off-route spectrum and live spectrum for the following positions: : suction and discharge pipe of the pump , the flange between pump and motor and motor NDE side, as shown in the attached pictures, and I also opened the suction strainer at the pump to check. We found the strainer partially clogged. I also noticed that the mish of the strainer is very small , the maintenance team cleaned it as much as possible and reinstalled it , and the pump was starting to check , but unfortunately, the issue is still as shown in the attached picture My doubt now is that the NDE colling fan is not attached correctly at the motor shaft ???

@Dave Reynolds & @Lyn Greenhill

Greeting my friends , I am waiting for your inputs about the last data I shared about our case , I really appreciated your effort for this case 

Rebuild the pump.

7.5hz is close to 1/3 operating speed.

Portable vibration and installed sensors both show same vibration levels and frequency?

Flow/pressure/amps goes down, vibration increases. Flow/pressure/amps goes up, vibration decreases

Motor amps increase with flow and pressure, this is normal. IF the vibration changes but motor amps do no increase suspect sensors are not reading correctly

Fix the motor bearings did nothing to change vibration.

Changes in pump discharge pressure will put different forces on the pump shaft. If there is enough wear in pump bearings, the pressure can hold or let the shaft move as the pressure varies. There is also a shaft torsional vibration (twisting mode) that can influence your vibration.

Motor NDE and mounting flange to pump both show 4 mm/s but suction pipe is only .4 mm/s, this points toward the pump as the source.

Not sure why your pressure and amps are constantly changing? 

Have had check valves, air leaks and control valves fail which can cause low frequency vibration.

Had a 3600 rpm vertical pump vibrating at 1200, no vibration at 1x. Found a vacuum pump quite a distance away that was vibrating, the piping from the vacuum pump was touching the piping to the 3600 rpm vertical pump and excited the natural frequency causing it to vibrate at 1200 which was the speed of the vacuum pump.

Dave