Regression Trend Prediction of Rolling Bearing Performance based on Integrated Soft Competition ART

In order to improve the accuracy and stability of rolling bearing performance prediction,a prediction method combining soft predictive ART-RBF integrated forecasting model and confidence CV value is proposed. The soft ART is introduced into the RBF neural network to establish the soft ART-RBF neural network prediction model. Combining with weighted average technology,the establishment of integrated soft ART-RBF neural network prediction model is carried out. And the confidence degree（CV） value with rich fault information is obtained through the self-organizing map（SOM） network as a comprehensive index to characterize the degradation of rolling bearing performance. Finally,the above method is verified by the acceleration signal obtained by the accelerated fatigue test of the rolling bearing. The results show that the method can effectively improve the accuracy and stability of the prediction of the degradation trend of rolling bearings.