Operating state prediction for wind turbine generator bearing based on ULSSVM and QPSO

Abstract

Aiming at the problem of operating state prediction of generator bearing, a prediction method based on quantum particle swarm optimization (QPSO) and united least squares support vector machine (ULSSVM) is proposed. Firstly, the time least squares support vector machine (TLSSVM) model is established in accordance with the change law of characteristic parameters over time. Space least squares support vector machine (SLSSVM) model is established in accordance with the law of mutual influence between characteristic parameters. Secondly, the QPSO algorithm is used to optimize the parameters of each least squares support vector machine (LSSVM) model. When the difference between the predicted value and the measured value reaches the minimum, the optimal LSSVM parameter set is output. Then the improved Dempster-Shafer (D-S) theory is used to determine the weights of TLSSVM and SLSSVM. A united model of time LSSVM and space LSSVM is established. The characteristic parameters are predicted. The prediction results and the reference matrix are fused and reduced in dimension. Finally, the generator bearing operating status is predicted based on the location of the prediction results. The results show that the proposed method is helpful to realize the operating state prediction of the wind turbine bearing.