Cutter wear prediction method based on multi-scale convolutional neural network

Abstract

The invention provides a tool wear prediction method based on a multi-scale convolutional neural network, and solves the problems that an existing tool wear prediction method depends on artificial feature extraction, the model generalization ability is insufficient, and the application scene is limited. According to the method, the essential features of the data are extracted from the original data by using deep learning, the mapping relation between the features and the target is automatically established, and the mapping relation is combined with an artificial feature extraction method, so that the accuracy of the model is improved; according to the method, features are better represented under the condition that original signal information is not lost, an enhanced multi-scale CNN structure is designed and is applied to a wavelet scale map after transformation, and wear features of a cutter are automatically learned from original data; and finally, the artificial features and the automatic features are combined to form a mixed feature vector, and fusion of the automatic features and the artificial features is realized by using a full-connection neural network so as to realize prediction of tool wear.

Description

technical field [0001] The invention belongs to the technical field of tool wear evaluation in mechanical processing, and in particular relates to a tool wear prediction method based on a multi-scale convolutional neural network. Background technique [0002] Under the new production mode, the tool wear prediction system, as an indispensable and important part of the automatic and intelligent machining process, is increasingly valued by researchers and engineers. Accurately judging the tool wear state in the machining process is of great significance for ensuring the quality of workpieces, improving machining efficiency, and promoting automated and intelligent machining. [0003] The existing tool wear prediction methods mainly include: traditional data-driven tool wear prediction modeling method and tool wear prediction modeling method based on deep learning. Among them, traditional data-driven tool wear prediction modeling methods such as fuzzy clustering, support vector ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problems that existing tool wear prediction methods rely on manual feature extraction, insufficient model generalization ability, and limited application scenarios, and provide a tool wear prediction method based on multi-scale convolutional neural network

Images