Fault early warning of wind turbine gearbox based on multi‐input support vector regression and improved ant lion optimization

Gearbox oil temperature is one of the important indicators for gearbox condition monitoring and faults early warning. Accurately predicting change trend can maintain in advance ensure safety reliability wind turbine gearbox. The purpose this article to analyze supervisory control data acquisition (SCADA) turbines. A method based on multi-input improved ant lion optimization support vector regression (M-IALO-SVR) proposed, which accurately predict temperature. prediction compared with back propagation neural network (BPNN) ALO-SVR methods verify effectiveness M-IALO-SVR method. To further results, 95% confidence interval processing performed residuals model, then trends mean standard deviation moving window are calculated. Testing SCADA from a farm northeast China, test results show that when operating normally, predicted value follows measured very well. When operates abnormally, its deviates normal range, statistical characteristics also change. According characteristics, abnormal state be found time.