Hi friends,
we have an 'DURCO' Pump which is having high vibration on Pump inboard vertical (5.5 mm/s rms pk) as per CSI data collector it's dominant peak is 5x.pump operating speed (2970 rpm) and also Skf 6310E brg BPFI defect freq (pl note 6310E is at pump out board side).both the defect frequencies are same.we had carried out bump test under shut down condition which is also 5x .we operated the pump at best eff point 65cu.mt/hr.the impeller to casing clearance is .0018"(.45mm)as per manual We didn't change the 6310 brg because is in out board side where as high vibration is at pump inboard side . we suspect vane passing resonance.Is there any other opinion on this issue? pl give us your valuable guidence

Pal,
If you acquire a spectrum with an fmax of 300 Hz and 6400 lines of resolution, you should be able to separate .0468 Hz or 2.8 CPM. 5X (Vane Pass) is 14790 CPM and 4.98X (BPFI for an SKF 6310E) is 14850 CPM which is 1 Hz difference. 5X is 247.5 Hz so you will be able to see your first order of vane pass and first order of BPFI at 300 Hz fmax. This should tell you if it is vane pass or BPFI.

Ronnie

Of course you might try higher resolution or time sync averaging to see if the frequency is exactly 5x or slighly off.

You might look at the time waveform or modulated time waveform. A distinctive impact/ringdown recurring at a time interval corresponding to bpfi would argue toward bearing defect. This might also show in an enveloped waveform.

You might look at the pattern of harmonics. As far as I have seen, there might be up to 3 or 4 harmonics of blade pass. Bearing defects often have more, typically peaking up in the neighborhood 50kcpm - 150kcpm (when viewed in velocity). Also you might see sidebands 1x around these.

You might look at the sensitivity to fluid conditions. Blade pass frequency often increases as you move further from BEP. Change in magnitude with pumping conditions tends to point more toward blade pass.

Also, vibration on the pump discharge piping at the same frequency tends to point toward blade pass.

Another method:

Connect tachometer/strobe to analyzer and use Time Synchronous Averaging that will remove non-synchronous bearing vibrations from vane pass frequency which is an exact shaft order.

Walt

Could you post a spectrum. It has been my experience that the fundamental frequency of vane pass is normally the highest and the harmonics get less the higher order they are. Bearing faults as a rule are opposite with the higher orders having the most amplitude and amplitudes will get lower as the near the fundamental.
You could also compare waveforms as Pete suggested.

Thank you for all for your valuable suggection. Mr.Ronnie Lynn
We had taken 6400 line spectrum it has dominant 5x vane passing but with out side band. No 2vpf or 3vpf. We shall send you the spectrum in due course !
Mr.Electricpete
we are unable to go beyond BEP as there no sufficient process flow suction side pipe bump test also reveals 5x.time wave form we will try with strobe lite as suggested by Mr.Walt strong and post the spectrum as suggested by Mr.Vibe guy 2004 .as on today we want to try installing vibration absorber at pump in board brg is it a bad idea? But before I want your concurrence pl!

[QUOTE]Originally posted by pal:
Mr.Ronnie Lynn,we had attached 6400 line spectrum it has dominant 5x vane passing but with out side band. No 2vpf or 3vpf.
you can also verify as suggested by Mr.Vibeguy that normally the highest and the harmonics get less the higher order they are. Bearing faults as a rule are opposite with the higher orders having the most amplitude and amplitudes will get lower as the near the fundamental.

Durco_pump_spectra.doc (37 Kb, 56 downloads)

Pal,
On the 6400 LOR spectrum you set a cursor on what appears to be vane pass, 5X. It actually shows 4.997 orders. Did you mark your speed first and if so, what orders do you get on the 12800 LOR spectrum when you set the cursor on 5X? It would be more definitive if you lowered your Fmax also. You are still going out to 120 Kcpm or 2000 Hz. It appears you have the ability to take 12800 LOR and I believe if you lower your Fmax to 500 Hz or 30 Kcpm and mark your speed, the 5X will either be 5X or 4.98 orders. I would think one of your fault frequency lines will be slightly off of the peak.

Ronnie

Ronnie,why the CSI doesn't check 5x or 4.98 orders even with 12800 LOR the attached spectra has vane & 6310E bpfi.CSI claims that it is safer to predict all the possible fault frequencies i will again send you spectra with 500Hz or 30 kcpm with set speed.even if we assume that it is vane passing as per MrVibeguy,( that normally the highest and the harmonics get less the higher order they are. Bearing faults as a rule are opposite with the higher orders having the most amplitude and amplitudes will get lower as the near the fundamental.)we had adjusted imp clearence to .0018"(.45mm)as per manual why the peak is not reducing?

Pal,
My recommendation to take higher resolution is to confirm either vane pass or BPFI. I suspect that it is vane pass but more resolution would be more conclusive. As to why you are unable to reduce the vibration by changing your clearance, I am not sure.

Ronnie

Ronnie,thanks for ur message.pl see the attached spectra where in B-vane passing K-6310E BPFI are distinctively clear.as per ur recommendation LOR 6400 confirms that it is not 6310E BPFI but it is not vane passing also as 2vpf & 3vpf with side band not present.we feel side band alone confirms the severity hence we feel it could be vane passing resonance or still the imp to casing clearance needs correction.i want your and others valuable opinion!

Durco_pump.doc (38 Kb, 33 downloads) Durco pump

I Would look at the cut water clearance. Sometimes you will have a 7" impeller with 10" backplate. This leaves about 3" cut water clearance. The plant I used to work for had 95% durco pumps. Most of all the ones operating at 3600 rpm had vane pass frequencies in the spectrums. Some as high as .6 ips but never gave us any problems. I just added a band to cover the blade pass frequency and increase the alarm level.

JFRENCH,
presently we are also doing the same thing! Since u r using 'durco' pump you can clarify this point.ie all these pumps have impeller, which can be adjusted by rotating the bearing carrier which is not allowing us to maintain exactly .0018"(.45mm) clearance between imp to casing.is it so? This being an non API pump and an ANSI pump does n't have hyd balancing holes.this is also strange. Is it so? Why no body recommend vane passing resonance for the above problem?

Pal

The durco impellers we had did had hyd. balancing holes in them. Some of the Group 1 did not but the Group 2 &3 did. We set the impeller clearance per process operating temperature. I think you will not be able to set the impeller clearance too exactly to .018” by using the marks on the face of the bearing carrier. Each mark is .002” but is close guess because they are not machined marks. Using them will get you close to the impeller clearance you want.

The throat or cut water clearance is all together different. It is located right at the neck of the discharge nozzle inside the casing. Wish I had picture to show you. The smaller the impeller the larger this gap is. I am no expert in the hyd. field but this is what I noticed in the field and taking vibration data.