Prediction of rolling bearing performance degradation based on whale optimization algorithm and backpropagation model

Abstract

The dissertation proposes a prediction model that enhances the BP (Backpropagation) neural network using the WOA (Whale Optimization Algorithm) to address the issue of local convergence during prediction. The model optimizes the initial weights and thresholds of the network using the whale optimization algorithm, resolving the problem of BP’s local convergence. Feature data is selected using the correlation coefficient theory to obtain the model inputs. Bearing performance degradation assessment is conducted using the predicted remaining life index. The superiority of the model is validated by comparing its predictions with those of BP and LSTM prediction models.