Grinding roughness prediction method based on multivariate nonlinear fitting and BP neural network

Abstract

A grinding roughness prediction method based on multivariate nonlinear fitting and a BP neural network comprises the following steps: 1) recording the grinding parameter values of the rotating speed, the feeding speed and the grinding depth of a grinding wheel, and collecting the grinding roughness values under corresponding grinding parameters; 2) establishing a multivariate nonlinear fitting relation function model of the grinding roughness and the grinding parameters; 3) converting the multivariate nonlinear function model of the grinding roughness and the grinding parameters into a multivariate linear function model; 4) preprocessing the grinding roughness data and the grinding parameter data; 5) establishing a BP neural network based on the multivariate linear function model structure, and performing iterating through a gradient descent method to obtain multivariate linear function model parameters; 6) inversing the multivariate linear function model into a multivariate nonlinear function model, and predicting the grinding roughness value. The BP neural network is used for training to obtain multivariate nonlinear fitting function parameters, the grinding roughness is predicted according to the multivariate nonlinear function, the accuracy is higher, and the reliability is better.

Description

technical field [0001] The invention belongs to the field of fine machining and ultra-precision machining. Aiming at the problem that the surface processing quality is difficult to evaluate and guarantee during the use of grinding machines, a grinding roughness prediction method based on multivariate nonlinear fitting and BP neural network is proposed. Background technique [0002] As a traditional processing method of finishing and ultra-precision processing, grinding is widely used in industry, military and daily life. Due to the complexity of the grinding mechanism, it is difficult to evaluate and guarantee the quality of the grinding process, so the prediction of the grinding roughness is a realistic problem that must be solved in the grinding process of the grinding machine. [0003] Aiming at the prediction of roughness in the grinding process, many scholars at home and abroad have established a theoretical prediction model of grinding roughness and achieved certain r...

AI Technical Summary

Problems solved by technology

However, in the past, the theoretical prediction model only used a certain method to study the grinding roughness prediction model, and there were few researches on the combination of multivariate nonlinear fitting and BP neural network for grinding roughness prediction.

Images