基于带阈值模块一维残差网络的刀具磨损监测方法

基于带阈值模块一维残差网络和双向长短期记忆网络,设计了刀具磨损监测模型和预测模型。将传感器信号经过小波分解后输入监测模型中,阈值模块自动选择阈值对信号降噪,残差模块提取信号特征,然后输出刀具磨损监测值,再将其输入到预测模型中获得刀具磨损预测值。实验证明:该监测模型与不带阈值模块的一维残差网络模型和卷积神经网络模型进行了对比,监测准确率分别提高了0.327%和1.697%;预测模型的预测效果较好,绝对误差仅为0.023。

Tool Wear Monitoring Method Based on One-Dimensional Residual Network with Threshold Module

A tool wear monitoring model based on one-dimensional residual network with threshold module and a tool wear prediction model based on bidirectional long short-term memory network are designed. The sensor signal is input into the monitoring model after wavelet decomposition, the threshold module automatically selects the threshold to reduce the noise of the signal and the residual module extracts the signal characteristics, then outputs the tool wear monitoring value, and inputs it into the prediction model to obtain the tool wear prediction value. The experimental results show that the monitoring accuracy of this monitoring model is improved by 0.327% and 1.697% respectively compared with the one-dimensional residual network model without threshold module and convolution neural network model; the prediction effect of the prediction model is good, and the absolute error is only 0.024.