Turning tool wear determination method based on adversarial neural network

Abstract

The invention discloses a turning tool wear determination method based on an adversarial neural network. The method comprises the following steps that main shaft current signal data in a numerical control machine tool are acquired; the data are pretreated; an alarm region is defined by using a priori knowledge; a data sample is enhanced by using the adversarial neural network; a CNN model is trained by using the current signal data in the enhanced data sample, the CNN model parameters are adjusted, and the tool wear is divided into primary wear, normal wear, the alarming region and severe wear; and the to-be-tested current signal data are input into the trained CNN model, and a classification result is obtained. According to the turning tool, the three categories of the tool are changed into four categories, and the imbalance problem of the data samples is processed by the adversarial neural network, so that the tool state can be effectively detected and predicted, the damage of the tool is early-warned, the increment of the rejection rate caused by reduction of the machining precision of the tool entering the severe wear stage is avoided, and the reliability of the machine tool isimproved.

Description

Technical field [0001] The invention relates to the field of numerical control machine tool tools, in particular to a method for determining the wear of a turning tool based on an anti-neural network. Background technique [0002] With the introduction of "Industry 4.0", the manufacturing industry is booming in our country, and a large number of mechanical processing and mechanical manufacturing are widely demanded. CNC machine tools are called "industrial mother machines" and play a decisive role in the manufacture of major machine parts. According to the current trend, the precision requirements of machine processing are getting higher and higher, and the requirements for the quality of turning tools in machine tools are also getting higher and higher. [0003] From the perspective of processing technology, on the one hand, the use and demand of turning tools are increasing, on the other hand, the working environment of turning tools is relatively harsh, and the processing stren...

AI Technical Summary

Problems solved by technology

However, the fault diagnosis methods of classical deep learning models usually only consider the case of sufficient number of samples.

In this way, the traditional deep learning model is difficult to develop, and the diagnosis effect is not ideal

Images