Intelligent prediction of surface roughness of PSZ ceramic grinding based on correlation analysis and CNN-BiLSTM neural network

Abstract

Partially stabilized zirconia (PSZ) ceramics are widely used in aerospace industry and other fields due to their superior performance. Surface roughness is a key indicator for evaluating the grinding level of PSZ ceramics. In order to reduce the prediction error of grinding surface roughness, an acoustic emission prediction model for PSZ ceramic grinding surface roughness based on correlation analysis and convolution-bidirectional long short term memory neural network (CNN-BiLSTM) was proposed. By analyzing the correlation between the eigenvalues of grinding acoustic emission signals and the grinding surface roughness values, the most relevant frequency bands and feature matrices between the grinding acoustic emission signals and the grinding surface roughness were selected as the input parameters of the CNN-BiLSTM neural network to reduce the error of acoustic emission prediction of grinding surface roughness. The research results show that the average prediction error of PSZ ceramic grinding surface roughness based on correlation analysis and CNN-BiLSTM neural network is less than 3.92%.