The Problem

VROC was asked to analyse data from an onshore wind farm installation by the asset owners with a view to diagnosing performance and efficiency of each turbine and the entire asset.
Some of the turbines are not able to reach their name plate power output and this study was conducted to provide critical insights to this problem.

The Solution

Advanced AI analytics combined with traditional physics-based modelling are possible on the OASIS platform allowing failure diagnosis. In this case, a healthy bearing was compared with Inner race (IRD) and Outer race (ORD) failures. High resolution vibration data using a sampling frequency of 20khz over a 10 second period from a healthy bearing and defective bearings was ingested for advanced analytics on the VROC platform.

Spectral analysis was done using traditional physics-based models and overlaid on multiple thresholds generated by the VROC AI models in addition to the set-points specified by the OEM.

The Outcome

VROC was able to pinpoint issues with the variation in performance versus wind speed due to switching issues from STAR to DELTA below 8m/s. We were able to visualise the divergence in power production related to wind speed when switching between the two modes.

Another important insight was the deterioration of performance of the turbines under similar operating conditions and over a period of time. Mapping and analytics on the platform was able to confirm that WT05 has been producing a lower range of power at a similar wind speed in comparison to WT07 and 09.

AI modelling on the VROC platform using standard monitoring was able to identify the bearing defects ahead of time for confirmation using high resolution techniques.

Get started with VROC today

Ready to embark on a pilot project or roll-out the innovation enterprise wide? Perhaps you need assistance integrating your systems or accessing your data? We have a solution to help you as you progress through your digital transformation.

Get Started Today