Vibration Based Life - Reducing the Cost of Failures

Your condition monitoring team has let you know that an asset has a likely axial crack on an intermediate stage bearing and the vibration trend is increasing. You inspect the bearing and there is indeed damage…what next? How long do you have before you need to take action? Is the bearing likely to make it to the next low wind season? These are difficult questions that wind farm owners/operators will often find themselves asking. At Romax InSight we have developed Vibration Based Life to help answer these questions and empower the user to make more informed maintenance decisions.

Vibration Based Life is implemented in 3 simple steps:

Step 1 – Early-stage damage is detected as a vibration alarm in our powerful Fleet Monitor software. Lead-time is maximised by utilising advanced processing and detection features.

Figure 1 – Early-stage fault detection example using our Fleet Monitor software. A vibration trend exceeds the yellow alarm threshold (top right) and is diagnosed by reviewing fault frequency harmonics in the associated vibration spectra (bottom right).

Step 2 - The vibration-based life is modelled by taking advantage of our extensive failure database incorporating data from 5GW of assets monitored for more than 5 years.

Figure 2 – Vibration Based Life modelling example. An active fault (vertical yellow line) is tracked relative to the estimated life model (average line). The important P50, P75, and P90 dates, which indicate the advanced stages of failure, are also labelled.

Step 3 - Customers use our easily interpretable Vibration Based Life estimates to balance the risk of failure with the cost of repairs for all active issues across the farm:

  • P50 Date – “Warning” – Maintenance should have been scheduled.
  • P75 Date – “Progressed” – Regularly inspect damage progression.
  • P90 Date – “Critical” – Consider shutting down the turbine.

Maintenance is planned well in advance to coincide with low wind periods and repairs are combined where possible to reduce the cost of failures.

Figure 3 – Multiple Vibration Based Life models from active faults across a farm are combined in a single plot. When managing a site with several failing components the Vibration Based Life forecast is very handy for planning multiple repairs.


Watch the webinar recording below and learn how to save money on your wind fleet O&M by detecting failures well in advance and forecasting the remaining useful life by optimising repairs.