A Novel Combinatorial Network for Anomaly Monitoring of Wind Turbines

Effective anomaly monitoring is crucial for ensuring the safe operation of wind turbines, necessitating advanced monitoring and data collection techniques. The supervisory control and data acquisition (SCADA) technology, recording a range of parameters relevant to wind turbine operation, is fundamental in this regard. In this paper, a combinatorial network named temporal convolutional network-transformer (TCN- Transformer) is proposed to extract multidirectional features of SCADA data for wind turbine condition monitoring. Firstly, the data is cleaned and parameters with greater relevance to the prediction target are filtered out as input parameters of the model. Then, TCN model is used to extract the temporal features of the input data, and the Transformer model is used to extract the correlation features between the input parameters. Finally, by analyzing the SCADA data of the actual wind farm, it is proved that this proposed method can be accurately and reliably applied to the anomaly detection of wind turbines.