A method and apparatus for predicting surface roughness of liquid jet polishing

Abstract

The application provides a liquid jet polishing surface roughness prediction method and device, comprising: obtaining sample processing parameter input data set and roughness output data set of liquid jet polishing experiment; determining network architecture of neural network based on sample processing parameter data set and roughness output data set to construct initial roughness prediction model; processing the initial roughness prediction model according to quantum particle swarm optimization algorithm to obtain optimal model parameter; updating the initial roughness prediction model according to the optimal model parameter to obtain target roughness prediction model; and predicting surface roughness of target liquid jet polishing object based on the target roughness prediction model. The application constructs the target roughness prediction model through the sample processing parameter input data set and the roughness output data set, and evaluates through the target roughness prediction model set, so as to solve the problem of high production and detection cost and material waste caused by measuring with instruments after processing.

Images