Pump Reliability Issue

G,day all, here is another interesting job I got called to

 

Background:

This pump and motor had a history of reliability failures from bearings, shaft shearing and pipework flanges leaking. This was a pair of pumps on separate base frames but secured to the same concrete floor with a pipework common outlet.

I performed vibration analysis with phase analysis and diagnosed a foundation and structural problems as the root cause.

 

Vibration Data:

Pump Vibration Data:

Figure 1 shows the overall Velocity vibration trend from our first visit and second visit. This is gathered at the motor DE.

From this trend you can see a marked increase in the velocity vibration levels from 6.304 mm/s RMS and 8.388 mm/s RMS.

Fig 1:

 

Fig 2 compares the acceleration time waveforms from the motor drive end bearings

From this comparison you can see the lower levels of pump A (Blue Plot) and the very high impacting from pump B (Green Plot)

Fig 2:

 

Fig 3 is the vibration data from the motor drive end bearing

There is high impacting form the motor bearing and damage to the inner and outer raceway

Fig 3:

 

Pump B – Motor Bearing Inspection

Failure Mode:

From inspection the failure mode as per ISO 15243:2004 is 5.3.3.3 False Brinelling.

False Brinelling occurs in the contact area due to micromovements and/or resilience of the elastic contact under cyclic vibrations. Depending on the intensity of the vibrations, lubrication conditions and load, a combination of corrosion and wear can occur, forming shallow depressions in the raceway. In the case of a stationary bearing, the depressions appear at rolling element pitch.

In many cases, it is possible to discern rust at the bottom of the depressions. This is caused by oxidation of the detached particles, which have a large area in relation to their volume, as a result of their exposure to air.

Key Points are:

  • rolling element / raceway contact areas
  • micromovements / elastic deformation
  • vibrations
  • corrosion/wear & shiny or reddish depressions
  • when stationary: at rolling element pitch
  • when rotating: parallel “flutes”

 

Findings:

  1. Depressions appearing at rolling element pitch indicating damage while the pump was in standby stationary bearing (Image 1)
  2. Indications of oxidation of the detached particles, which have a large area in relation to their volume, as a result of their exposure to air.

 

Bearing Inspection: Motor Drive End Bearing – FAG X-lite NU319E.TVP2

Image 1 is the outer raceway, and displays depressions appearing at rolling element pitch which indicates damage to the bearing when the motor was stationary

Image 1:

 

Image 2 is a close up of the depression at rolling element pitch on the outer raceway, this is from the load side of the bearings and also shows the roll over.

Image 2:

 

Image 3 is a microscopic image of a depression on the outer raceway.

Image 3:

 

Image 4 is an image from the inner raceway, this also displays the depressions at the pitch of the rolling elements.

Image 4:

 

Image 5 is a microscopic image of a rolling element.

Image 5:

 

 

Motion Amplification

Even with this vibration data the client was not convinced so I had to use another technology to show the client how the structural and base was causing them their reliability headache.

 

This first video shows how the pipe work was moving, this was the cause of the stress and strain to the flange joints that led to the leaks

 

This second video is of the base plate, this showed the true motion of the pump and how these failures were being induced.

9 thoughts on “Pump Reliability Issue

  1. Weak bases and pipe stress are common causes to premature bearing, seal and even shaft failures that many Engineers, Operators and even Managers don’t seem to understand. Most of the time when they see signs of bearing defects, bearings are replaced without realizing that there is something else in the system that’s causing the bearing/seals to fail prematurely until it becomes a nuisance. Depending on the severity of weak bases, the failures can be quite abrupt with process/load changes ie. Vibration data may have been collected a few days or weeks ago when it seem normal. Seen this happen on Offshore platforms, Cement, Petrochemical and even in light industries.

  2. What type of equipment did you use for your motion amplification video’s? Great tool for people that are need something visual.

  3. Dear Gents,
    It is a good case, visuals also, and very frequent in industry in similar multi-machines running in close proximity, & where we generally don’t see or avoid looking into the details. Great.

    However, I was trying to relate Ultimate Faults (on bearings), Enforcing Root Problem Creator (i.e. structure/frame) and associated Vibration Signature exclusively for this application. But wish to know more and still hungry to know:

    * The fundamental frequency selected to analyse structure/ frame response and the related Spectra. Also, the dominant Phase readings, if you can share.
    * What corrective actions taken to restrict the movements and if latest data is available after corrections.

    • Hi Tarapada

      Apologies for the late response.

      The fundamental frequency to analyse the structure and frame response was: The motion amplification was filtered to running speed of the asset. When we measured the structure with vibration analysis we measured overall velocity to 1000Hz.

      The motor to pump was laser aligned, however vertical phase readings at running speed across the coupling in the vertical direction was 160 degrees, indicting misalignment/bed flex.

      I am still waiting to follow up the corrective actions, however the client was going to fill the gap in the metal frame to floor with concrete. Once I get post repair data I will share.

      Regards.

  4. Hello all, new to the site. Great write up. The motion amplification videos are great for getting buy in on the suggested repair. We bought a camera from RDI at the end of 2017 and it has already proven its worth. Similar situation, VFD driven pump train, poor installation on a poor base that was complicated by a resonant frequency near maximum operating speed. Using the motion amplification camera we were able to pinpoint the issue in 10 minutes. Using other techniques such as Phase and ODS it would have taken hours to get to the root cause.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.