Senior ScientistContact Madhav
The installed wind power capacity is on a strong growth path in Europe and around the world. At the same time, the profit margins for wind energy have been decreasing for many operators. Thus, it is desired to further reduce the operation and maintenance cost of their turbines.
Most commercial wind turbines are remotely monitored around the clock to enable an early detection of operation faults and potential developing damage. Modern wind turbines can be equipped with several hundreds of sensors that continuously monitor the health of the turbine subsystems. Those sensors can acquire hundreds of gigabytes of condition data every day. Thermal and electrical state variables, oil quality, vibration responses, and environmental conditions are typically monitored 24/7. In addition, comprehensive control system data are usually available to complement the monitoring and are even used to act as inexpensive proxies for dedicated sensing systems.
Data-driven machine learning models are in use in monitoring centers to provide decision support and detect faults and developing damages at an early stage. I will provide an overview of opportunities and challenges in Prognostics and Health Management of wind farms. I will also discuss recent progress made in data-driven condition monitoring and diagnostics, including the potential of transfer learning strategies and convolutional autoencoders for vibration-based fault diagnostics tasks based on normal behavior modelling.
Dr. Angela Meyer has been a Professor of Applied Machine Learning and a research group leader at the Bern University of Applied Sciences (BFH) since 2021. Before joined BFH, she was a Senior Lecturer in Smart Maintenance at Zurich University of Applied Sciences. Dr. Meyer's research focuses on increasing the reliability and performance of industrial systems by means of artificial intelligence. Her research is driven by the vision to facilitate the energy transition by means of AI/ML and digitali