Tool Wear Monitoring Method Based on Multi-scale Deep Convolutional Recurrent Neural Network

Abstract

The invention discloses a tool wear monitoring method based on a multi-scale deep convolutional cyclic neural network. The method includes the following steps: preprocessing data based on tool data measured by multiple sensors, constructing an input matrix; constructing a multi-scale convolutional neural network, Obtain multi-scale features, and each branch of the multi-scale convolutional neural network performs feature fusion based on the output of the maximum pooling, and finally obtains multi-scale features; constructs a deep loop GRU network to extract features and representations of different time scales, deep loop The GRU network includes the first layer GRU network and the second layer GRU network. The number of units in the second layer GRU network is more than that of the first layer GRU network. The multi-scale features are processed by the deep cycle GRU network to obtain features of different time scales. and representation; construct a fully-connected layer based on features, and map the features to the sample label space; construct a linear regression layer based on the output of the fully-connected layer to obtain the amount of tool wear.

Description

technical field [0001] The invention belongs to the field of cutting tools, in particular to a method for monitoring tool wear based on a multi-scale deep convolutional cyclic neural network. Background technique [0002] During machining, tool wear will reduce the dimensional accuracy and surface integrity of the part. When the tool wear is severe, it may even cause tool damage, resulting in the scrapping of the workpiece and damage to the machine tool. Specifically, the state of the tool has an important influence on the cutting deformation process, and tool wear will not only cause a decrease in the dimensional accuracy of the part, but also deteriorate the surface quality of the part. When the tool enters the severe wear stage, the cutting force increases sharply, the friction between the tool flank and the processed surface of the workpiece intensifies, and the surface temperature of the workpiece rises significantly, causing thermoplastic deformation on the surface of ...

AI Technical Summary

Problems solved by technology

[0004] However, indirect methods also have corresponding disadvantages: most indirect methods for tool wear monitoring are feature-driven

Second, the extraction of these features depends on experience and professional knowledge, which is difficult for non-professionals to complete, making the model not universally applicable; third, feature extraction is difficult to achieve under complex working conditions and strong noise environments; fourth, feature extraction is a Time-consuming and labor-intensive process, making it difficult to improve the efficiency of tool life monitoring

Images