SEAI has supported EnergyPro over the past 3 years to develop Machine Learning tools. To augment this project EnergyPro have joined forces with the LERO research centre based in the University of Limerick, and the CeADAR research centre based in University College Dublin, to develop a Machine Learning solution to identify common causes of wind turbine underperformance.
Postgraduate students from the two institutes are working closely with EnergyPro’s analysis team to determine which indicators of underperformance causes can be found in turbine SCADA data, and develop an automated solution via Machine Learning that can determine which of these causes is most likely when a turbine is not performing as well as it should.
Turbine underperformance is a very common issue that only becomes more prevalent as turbines age, but it receives much less focus in the industry than faults which cause downtime, and existing solutions often focus on detecting and quantifying underperformance.
By focusing on the root causes EnergyPro will be able to go straight to Operations and Maintenance providers with a suggested solution, saving time on troubleshooting so that O&M teams can restore turbines to full efficiency faster, and the remote SCADA-based approach means the solution could be used on any windfarm in the world without the need for site visits or extra costs to install any specialised equipment.
The initial project will run until October, by which point EnergyPro hope to have a full solution rolled out for our managed sites, and further development plans would allow us to offer it as a separate service to any site with a remote database connection.
We look forward to providing updates on the project over the coming months.