Wind Turbine Gear Box Monitoring with Machine Learning

Posted to EPRI in the Digital Utility Group
image credit: Image Credit: EPRI
Jeremy Renshaw's picture
Senior Program Manager, Electric Power Research Institute (EPRI)

Dr. Jeremy Renshaw is the Sr. Program Manager for Artificial Intelligence at the Electric Power Research Institute (EPRI) and has been with EPRI since 2012. Dr. Renshaw manages the AI.EPRI...

  • Member since 2021
  • 13 items added with 12,084 views
  • Aug 2, 2021

A wind turbine gearbox converts low-speed rotations received from blades into high-speed rotations required by generators for electricity production. When a major wind turbine component - like the gearbox - prematurely fails, O&M costs increase and production revenue is lost, both of which increase the cost of energy.


Engineers and operators are constantly searching for ways to avoid failures and the high costs associated with them, as a full gearbox replacement can cost as much as $350,000.

Operators have traditionally relied on physics-based models to detect failures.  These models use equations to analyze gearbox material microstructure and properties for damage but are only about 60% accurate and cannot model and simulate a full gearbox system, its components, and their associated failure modes.

Combining machine learning with physics-based models for the wind turbines adds historic failure information and macro factors into the equation, such as wind speed, temperature, pressure, real-time power generation, vibration, etc. This brings a higher level of confidence in predicting the probability of gearbox damage and improves overall model accuracy.

EPRI has successfully developed a hybrid solution that has worked on gearboxes, main bearings, and even generators. The hybrid model is already 80% accurate and has supported operators in optimizing O&M actions, reducing high-cost repairs by identifying damage in its early stages, and providing 2-7 months lead time ahead of failures.

The model will improve budgeting, forecasting decisions, inventory management, resource planning, and component supplier selection. Operators can recognize damage early and extend the overall lifespan of a gearbox. If a damaged bearing within a gearbox is identified early, the repair may only cost $15,000-70,000 vs. a full replacement at $350,000. This hybrid AI model has the potential to provide significant benefits and efficiencies to wind turbine operators across the entire wind energy industry.  Similar model methodologies could be applied to multiple electric power components across the global industry.

Learn more about the game-changing solutions generated by the convergence of AI and the electric power industry during EPRI’s AI and Electric Power Summit on September 28-29. Register here.

Founded in 1972, EPRI is the world's preeminent independent, non-profit energy research and development organization, with offices around the world.
Matt Chester's picture
Matt Chester on Aug 2, 2021

Why is modeling the predominant method for this-- are there no reasonable sensors that can be installed to track certain characteristics to feed direct information rather than speculated? 

Jeremy Renshaw's picture
Jeremy Renshaw on Aug 2, 2021

Matt, great question. The traditional physics-based models were used due to simplicity and the fact that there was just too much data for people to crunch through. With the use of machine learning models, number crunching is no longer an issue, so the data from gearbox sensors (temperature, pressure, etc.) is utilized in the models mentioned in this work in addition to additional data, such as windspeed, power generated, etc. As these models continue to ingest more data over time, the accuracy and predictions will improve.

Jeremy Renshaw's picture
Thank Jeremy for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »