All Member Open Forum

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent. 
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected. 

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics. 

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Brayden,

I am not able to see your attachments for some reason.
Cage damage can be caused by poor lubrication, sudden speed changes, moment loads, misalignment, and others.
If possible, I suggest removing the bearing non-destructively so that you can inspect the cage, rolling elements, and rings intact.
If you have contacts at the bearing manufacturer, you can try reaching out to them for support of the bearing failure analysis.
Instead of trying to pick through the vibration data, it may be more useful to determine why the bearing failed, and correlate the failure cause and mode to what you see in the vibration data to truly understand if vibration would have alerted you to bearing failure.

For those that can see the attached vibration data, please reply with the following information:
-bearing part number, including BPFO, BPFI, FTF, and BSF
-operating speed
I

------------------------------
Mark Wolka
Sr. Application Engineer
NSK Americas
Ann Arbor, MI USA
------------------------------

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Hi Mark,

First off, thanks for your feedback. I'm not sure why the photos don't seem to be displaying - they aren't on my end either. I will try attaching them in another format to my original post. 

A few additional details that have come up since I posted yesterday. We did open the bearing up to find indications of high thermal exposure on both raceways and the balls themselves. Almost all of the balls from both raceways are a very deep blue in color. The pattern on both raceways is slightly off-center which may corroborate the misalignment theory. In two or three locations on the outer race there are visible indications of spalling. Three of the balls from one raceway exhibited some amount of material loss, flattening some of the outer surfaces - there is enough material loss to indicate that this did not develop recently. 

Additionally, it's believed that this pump has been in service for nearly 10 years in an almost non-stop application, so fatigue is almost certainly a factor in some capacity, although it's still unclear what the initiating root cause might've been for this bearing to fail - the NDE single ball bearing is in very good condition. 

As much as I think our root cause will be found through the investigation of the damaged components, I think understanding the vibration data in the lead up to the failure is important as well so that we can recognize similar conditions in the future. 

For additional reference information, the double ball bearing is a 5306 C3.

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Original Message:
Sent: 11-17-2021 10:42 AM
From: Mark Wolka
Subject: Small Pump Bearing Cage Fracture

Brayden,

I am not able to see your attachments for some reason.
Cage damage can be caused by poor lubrication, sudden speed changes, moment loads, misalignment, and others.
If possible, I suggest removing the bearing non-destructively so that you can inspect the cage, rolling elements, and rings intact.
If you have contacts at the bearing manufacturer, you can try reaching out to them for support of the bearing failure analysis.
Instead of trying to pick through the vibration data, it may be more useful to determine why the bearing failed, and correlate the failure cause and mode to what you see in the vibration data to truly understand if vibration would have alerted you to bearing failure.

For those that can see the attached vibration data, please reply with the following information:
-bearing part number, including BPFO, BPFI, FTF, and BSF
-operating speed
I

------------------------------
Mark Wolka
Sr. Application Engineer
NSK Americas
Ann Arbor, MI USA

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Brayden,

Yes, knowing the vibration pattern to detect it next time is critical.
You are fortunate to be able to confirm the damage present in the bearing, and THEN go look at the vibration data to confirm bearing fault frequencies.

What is the condition of the lubrication?  Grease or oil?  Discoloration is an indication of poor lubrication or overloading (see page 38 of NSK Bearing Doctor).
The slightly offset race paths are normal since this is a double row angular contact ball bearing (see page 14 of NSK Bearing Doctor).
You can also see how misalignment will show in the wear path.
If the outer ring race path is 360 degrees around the raceway, then you had loss of clearance from the heat generation.
I would dig deeper into your lubrication situation.

As for the vibration analysis, check for misalignment at 1x and 2x if you see in on the bearing raceways.
If there are clear spalls on the outer race, then BPFO should be present in your overall acceleration spectrum.
Do you do high frequency detection, such as Peakvue, Spike Energy, Shock pulse, etc?
That will catch early bearing faults and possibly your lubrication condition.

Look forward to hearing a conclusion on your findings!
Good luck!

------------------------------
Mark Wolka
Sr. Application Engineer
NSK Americas
Ann Arbor, MI USA
------------------------------

Original Message:
Sent: 11-17-2021 02:04 PM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Mark,

First off, thanks for your feedback. I'm not sure why the photos don't seem to be displaying - they aren't on my end either. I will try attaching them in another format to my original post.

A few additional details that have come up since I posted yesterday. We did open the bearing up to find indications of high thermal exposure on both raceways and the balls themselves. Almost all of the balls from both raceways are a very deep blue in color. The pattern on both raceways is slightly off-center which may corroborate the misalignment theory. In two or three locations on the outer race there are visible indications of spalling. Three of the balls from one raceway exhibited some amount of material loss, flattening some of the outer surfaces - there is enough material loss to indicate that this did not develop recently.

Additionally, it's believed that this pump has been in service for nearly 10 years in an almost non-stop application, so fatigue is almost certainly a factor in some capacity, although it's still unclear what the initiating root cause might've been for this bearing to fail - the NDE single ball bearing is in very good condition.

As much as I think our root cause will be found through the investigation of the damaged components, I think understanding the vibration data in the lead up to the failure is important as well so that we can recognize similar conditions in the future.

For additional reference information, the double ball bearing is a 5306 C3.

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com

Original Message:
Sent: 11-17-2021 10:42 AM
From: Mark Wolka
Subject: Small Pump Bearing Cage Fracture

Brayden,

I am not able to see your attachments for some reason.
Cage damage can be caused by poor lubrication, sudden speed changes, moment loads, misalignment, and others.
If possible, I suggest removing the bearing non-destructively so that you can inspect the cage, rolling elements, and rings intact.
If you have contacts at the bearing manufacturer, you can try reaching out to them for support of the bearing failure analysis.
Instead of trying to pick through the vibration data, it may be more useful to determine why the bearing failed, and correlate the failure cause and mode to what you see in the vibration data to truly understand if vibration would have alerted you to bearing failure.

For those that can see the attached vibration data, please reply with the following information:
-bearing part number, including BPFO, BPFI, FTF, and BSF
-operating speed
I

------------------------------
Mark Wolka
Sr. Application Engineer
NSK Americas
Ann Arbor, MI USA

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Hi Brayden,

Due to some reason, I am not able to download the vibration signatures. Did you get a chance to measure 'g' (acceleration) when the pump was running? I suspect this would have been high since the bearing has failed.

Regards
Saigopal Visvanathan

------------------------------
Saigopal Visvanathan
Lead, Asset Strategy
Suncor Energy Canada
Calgary AB
------------------------------

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Hi Saigopal,

The pk-pk measurement of our waveform in the 12 months prior to the failure did not exceed 5.5 Gs. Over the last 6 months it did not exceed 2.5 Gs. For a failure like this I might've expected those values to be higher.

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Original Message:
Sent: 11-17-2021 01:14 PM
From: Saigopal Visvanathan
Subject: Small Pump Bearing Cage Fracture

Hi Brayden,

Due to some reason, I am not able to download the vibration signatures. Did you get a chance to measure 'g' (acceleration) when the pump was running? I suspect this would have been high since the bearing has failed.

Regards
Saigopal Visvanathan

------------------------------
Saigopal Visvanathan
Lead, Asset Strategy
Suncor Energy Canada
Calgary AB

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

5.5 G's sounds high to me. I have an electric motor now running 4 G's and have put in on the pull list just to avoid an unscheduled failure. I may have limits set more conservative than others, but at 5.5 I would have put that on the swap list. Being an electric utility I would rather change something out a little earlier than necessary and avoid potential of having multiple items go down at once.

------------------------------
Christopher Mertz
Sr.Production Tech
Luminant Power
Terrell TX
------------------------------

Original Message:
Sent: 11-18-2021 07:16 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Saigopal,

The pk-pk measurement of our waveform in the 12 months prior to the failure did not exceed 5.5 Gs. Over the last 6 months it did not exceed 2.5 Gs. For a failure like this I might've expected those values to be higher.

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com

Original Message:
Sent: 11-17-2021 01:14 PM
From: Saigopal Visvanathan
Subject: Small Pump Bearing Cage Fracture

Hi Brayden,

Due to some reason, I am not able to download the vibration signatures. Did you get a chance to measure 'g' (acceleration) when the pump was running? I suspect this would have been high since the bearing has failed.

Regards
Saigopal Visvanathan

------------------------------
Saigopal Visvanathan
Lead, Asset Strategy
Suncor Energy Canada
Calgary AB

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

10y continuous operation would expect wearing of races, ball elements including cage. Vibration data analysis should go beyond 12 months. It is possible that Vibration could have trended up before coming down. Cage damage could have been the last thing to occur before catastrophic and may be initiated by rolling elements wear on the cage. Examination of the rolling elements for wear, but this may be mixed w seizure.

---------------------------------
Tee Yeow Hum
Reliability Manager
Shell
Singapore
---------------------------------



Original Message:
Sent: 11-18-2021
From: Brayden Olafson
Subject: RE: Small Pump Bearing Cage Fracture

Hi Saigopal,

The pk-pk measurement of our waveform in the 12 months prior to the failure did not exceed 5.5 Gs. Over the last 6 months it did not exceed 2.5 Gs. For a failure like this I might've expected those values to be higher.

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

Original Message:
Sent: 11-17-2021 01:14 PM
From: Saigopal Visvanathan
Subject: Small Pump Bearing Cage Fracture

Hi Brayden,

Due to some reason, I am not able to download the vibration signatures. Did you get a chance to measure 'g' (acceleration) when the pump was running? I suspect this would have been high since the bearing has failed.

Regards
Saigopal Visvanathan

------------------------------
Saigopal Visvanathan
Lead, Asset Strategy
Suncor Energy Canada
Calgary AB

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------

If I read the failure mode correctly, it is a cage failure?  If so the rest of my comments will be around cage failure.  

When cage failed the rollers would bunch together and remove all the RIC which would lock up the bearing as you describe.  That would also cause motor trips you saw.

If you had known bearing defects, it is really hard to know exactly what type defect it is and the risk level and remaining life.  Cage failure is especially hard to predict. I've seen cage failures where it was worn this from long term lube deficiencies and never picked up by vibration.  I've also seen (several recently) where cage abruptly failed also with no warning.  Sudden events are almost impossible to manage with vibration analysis unless continuous monitoring and even then still don't necessarily prevent failure just reduce collateral damage. 

Severity of defects in vibration analysis can be very tricky and is not as much an exact science.  There are some rules vibration uses but I've seen too many go against these rules.  Bottom line is if there is a defect, the longer you run the more risk is being taken.  Had a recent failure where vibration analysis and other condition monitoring showed nothing hours before catastrophic failure.  A bearing that fractures or spins on shaft can occur very rapidly in between condition analysis with little warning.

Was the cage worn from lube issue or was it an overload?  This will lead the investigation into two different directions.  Worn cage and look at lubrication.  Cage overload with little wear then look at bearing ftis, installation, loading, etc.

------------------------------
Randy Riddell, CMRP, PSAP, CLS
Reliability Manager
Essity
Cherokee AL
------------------------------

Original Message:
Sent: 11-16-2021 09:51 AM
From: Brayden Olafson
Subject: Small Pump Bearing Cage Fracture

Hi Folks,

Looking for some thoughts on a recent undetected failure at our facility on a small (20hp, 3600rpm) water recirculation pump. The following is a timeline of the failure..

1 - Operations attempts to start the pump. It immediately trips on overcurrent.
2 - Electricians measure motor resistance, which is fine - >2Gohm
3 - Discharge pressure trend is reviewed - no anomalies, steep increases or decreases over a 365 day period - pressure is stable at around 70psi.
4 - Vibration data is reviewed for possible fault indication - broadband amplitudes are relatively consistent, sidebands of 0.189x Exist around BPFI, which seems to have developed from more of a "haystacking" pattern around BPFI 4-6 months ago. Subtle increase in 5x peak (which is also a sideband of BPFI - there are 4 impeller vanes on this pump).
5 - Pump is removed and found to be completely seized. Shaft is removed. Drive end double roller bearing cage is found to have a fracture, and appears to be piled up in the cage. Bearing races have not yet been destructively inspected.

Admittedly, my vibration analysis aptitude is a bit rusty, and because of the size of our facility, I don't have much opportunity to focus on it, or have dialogue with colleagues about it - this forum is really my go-to spot for dialogue on reliability topics.

What I'm wondering is - should we have expected this failure based on the vibration readings shown? Despite the sideband frequency being significantly lower than the FTF, and the general amplitude trend being steady, should it have been obvious that a fault existed in this bearing and that a failure was imminent?

------------------------------
Brayden Olafson, CMRP
Reliability Specialist, Canaport LNG, Saint John, NB
brayden.olafson@canaportlng.com
------------------------------