In one of the facilities where I’m involved, I found some very curios bearing damage and would like to seek your opinion of possible causes. This is the first time I have had a chance to see this strange wear pattern.
This is 7318 angular contact bearing serving as a thrust for vertical pump. According to gathered information the pump had been in service for several years and higher vibration levels had been observed about a year ago while increasing the flow. The pump was inspected and normal oil level was found but the thrust bearing had all the balls with identical wear (about 0.030” removed from OD)(no damage on radial bearing). Several people had a look at it but so far no good reason for this damage was found. I heard that sister pump had identical wear some time ago but to less degree.
I’m open for discussion on what could cause this kind of wear.

This message has been edited. Last edited by: Alex J,

Bearing.pdf (1,333 Kb, 38 downloads)

Please resize the picture so we can see what it is.

Apparently this wear pattern is due to excessive static axial load keeping ball's axis of rotation in a certain position in space as oppose to normally random rotation within the cage sockets.

Therefore, instead of even around-the-ball-surface wear just certain area on a ball was suffering wear, thus being more intensive. Note that this wear is a snowball process. But the root cause IMO is excessive static thrust loading.

Just a theory...

David,
It was exactly my thought and I asked the pump manufacturer to re-calculate the hydraulic load at most adverse condition. They came back with the values ranging from 3600 to 3800 lbs. This bearing has rated static load of 32,800 lbs.
I still can not find the answer and I'm open for any suggestions.

This message has been edited. Last edited by: Alex J,

The balls look too good to be deteriorated. The edge between the cylindrical and the spherical area is too “clean”. You would think when this bearing was install this manufacturing defect would have been noticed. Maybe it was, but they install it anyway.

Did the vibration data match the ball spin frequency? Do you have any trend data from when the pump was installed? If the ball “tracks” in the cylindrical area, ball spin vibration might not be measured. I have seen manufacturing defect on rollers before, but not near this bad. This is just my opinion.

Thank you all for your comments. The pump bearing failure damage happened several years ago and a couple of years ago they converted to CSI and all the past vibration data is lost.