Wind turbine anomaly detection using normal behavior models based on SCADA data

This paper presents an approach for anomaly detection in wind turbines (WTs) using normal behavior models (NBMs) based on supervisory control and data acquisition (SCADA) data. A genetic algorithm combined with partial least squares regression (GAPLS) is used for input parameter selection to reduce the redundant parameters for anomaly detection in WTs. The NBMs for 14 temperature parameters of SCADA system are developed by using back propagation neural networks (BPNNs). The proposed method is verified by a case of a 1.5MW WT fault. Results show that the NBM has a low prediction error under normally operation condition and a high prediction error prior to the fault. The prediction error can be used as an effective indicator for anomaly detection in WTs.