Posters
Come meet the poster presenters to ask them questions and discuss their work
We would like to invite you to come and see the posters at our upcoming conference. The posters will showcase a diverse range of research topics, and will give delegates an opportunity to engage with the authors and learn more about their work. Whether you are a seasoned researcher or simply curious about the latest developments in your field, we believe that the posters will offer something of interest to everyone. So please join us at the conference and take advantage of this opportunity to learn and engage with your peers in industry and the academic community.
PO517: Wind turbine fault detection through a virtual sensor using exogenous variables
Shun Wang, Mr, Universitat Politècnica de Catalunya (UPC)
Abstract
High operations and maintenance (O&M) costs, driven by unexpected drivetrain failures, remain a significant challenge to the economic viability of wind energy. While data-driven normal behavior models are widely used for condition monitoring, they suffer from a fundamental limitation. These models typically rely on a turbine's internal SCADA variables, which are often highly correlated. This creates a critical "fault masking" effect, where the model fails to detect developing faults because the physically coupled components degrade in tandem, maintaining the expected correlations and suppressing prediction errors. To solve this, we introduce a novel monitoring paradigm centered on an exogenous virtual sensor. The method is a data-driven model that generates a real-time "healthy" temperature reading for a component. Crucially, it is calibrated using only external meteorological data, such as ambient temperature and wind speed, making it completely independent of the turbine's operational state or physical condition. In practice, the method runs in parallel with the turbine's physical sensor. A fault is detected when the physical sensor's reading consistently deviates from the stable, healthy benchmark provided by the method. We validated this approach on a commercial turbine with a known gearbox fault. The growing divergence between the physical sensor and our method provided a clear warning three months before failure, a fault entirely missed by traditional models. This framework offers a more robust, reliable, and practical solution for predictive maintenance in the wind industry.
No recording available for this poster.
