Surface roughness online prediction method based on fuzzy neural network and principal component analysis

Abstract

The invention relates to a surface roughness prediction method based on a fuzzy neural network and principal component analysis, and the method comprises the steps: firstly, collecting acoustic emission and vibration signals in a grinding process, extracting related time domain features, frequency domain features and wavelet packet feature parameters, and carrying out the dimension reduction of a feature quantity through the principal component analysis, and optimizing the feature value; then constructing a surface roughness fuzzy neural network prediction model, and taking the signal characteristic quantity and the surface roughness as input and output of a fuzzy neural network; and finally, training the surface roughness fuzzy neural network prediction model, and verifying the surface roughness prediction precision. Modeling is carried out by utilizing the characteristic values of the acoustic emission signals and the vibration signals and a principal component dimension reduction method, and the level of an online surface roughness prediction technology can be improved.

Description

technical field [0001] The invention relates to an on-line detection method of workpiece surface roughness in the grinding process, in particular to an on-line prediction method of surface roughness based on fuzzy neural network and principal component analysis. Background technique [0002] There are many factors that affect the surface roughness of the workpiece during the grinding process, including feed speed, grinding amount, grinding wheel speed, grinding wheel dressing status, workpiece material composition and some uncertain factors in the processing process, etc. The grinding process is dynamic, and many factors affecting the surface roughness such as grinding amount, workpiece material, and grinding vibration are constantly changing. The monitoring process of the workpiece surface roughness affects the processing quality and processing efficiency. If the real-time detection of the surface roughness of the workpiece can be realized, and the processing parameters in ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem of large errors in the characteristic parameters of manual experience selection in the process of surface roughness modeling and prediction, the present invention proposes a surface roughness prediction method based on fuzzy neural network and principal component analysis, which is used to improve the surface roughness of the workpiece during the grinding process. degree of recognition accuracy

Images