Fusion Chaotic Prediction Model for Bearing Performance by Computer Technique

Abstract

The delay time (DT) and embedding dimension (EM)of the rolling bearing vibration time series are different because of different methods. The improved weighted first-order local prediction model (IWFLPM) based on fusion technology is established. The delay DT obtained by the mutual information method and the ED obtained by the Cao method are used to form the parameter pair, and then the parameter pair sequence is constructed. The IWFLPM is used for one-step prediction. Finally, the bootstrap maximum entropy method is used to fuse the prediction result and MATLAB is used to perform all mathematical operations. The experimental results show that the accuracy of the fusion prediction results is significantly better than the IWFLPM, and the optimal delay time and optimal embedding dimension are obtained.