基于深度残差网络的刀具磨损量预测方法研究

数控机床刀具磨损将直接影响着加工产品的精度,导致产品质量下降.在自动化程度越来越高的数控加工中,监测刀具状态变得更加困难.为了保证产品质量,快速、精确的预测刀具磨损量,本文提出基于深度残差神经网络的多传感器刀具磨损量预测方法,首先,该方法提取振动、切削力和声发射传感器信号的时域特征,然后,利用深度残差网络对时域特征进行...

Research on Tool Wear Forecast Method Based on Deep Residual Network

Tool wear of CNC machine tools will directly affect the accuracy of processed products,resulting in a decline in product quality.In the increasingly automated CNC machining,monitoring the tool status becomes more difficult.In order to ensure product quality and predict tool wear quickly and accurately,this paper proposes a multi-sensor tool wear prediction method based on deep residual neural network.First,the method extracts the time domain characteristics of vibration,cutting force and acoustic emission sensor signals.Then,the deep residual network is used for supervised learning and training of temporal features,and finally,the trained model is tested on the test data.The model proposed in this paper is validated through experiments.The absolute mean value(MAE)of the model evaluation index is about 1.51×10-3mm,which has high prediction accuracy.