基于离散小波变换与遗传BP神经网络的木工刀具磨损状态监测  

Woodworking tool wear condition monitoring based on discrete wavelet transformation and genetic algorithm-BP neural network

作　　者：董伟航 胡勇 田广军 邱学海 郭晓磊 DONG Weihang;HU yong;TIAN Guangjun;QIU Xuehai;GUO Xiaolei(College of Materials Science and Engineering,Nanjing Forestry University,Nanjing 210037,Jiangsu,China;Bosun Prewi(Shanghai)Tool System Co.,Ltd,Shanghai 201316,China)

机构地区：[1]南京林业大学材料科学与工程学院,江苏南京210037 [2]博深普锐高(上海)工具有限公司,上海201316

出　　处：《中南林业科技大学学报》2021年第6期157-166,共10页Journal of Central South University of Forestry & Technology

基　　金：江苏省研究生科研与实践创新计划项目(SJCX20_0273);国家自然科学基金项目(31971594)。

摘　　要：【目的】为解决木质家具生产过程中木工刀具磨损造成的加工质量下降和生产成本升高的问题,需要对生产过程中的木工刀具磨损状态进行精确监测。【方法】提出了一种基于离散小波变换与遗传BP神经网络的木工刀具磨损状态监测方法。通过接入机床控制箱的功率传感器采集不同主轴转速、铣削深度和刀具磨损状态下的机床主轴功率信号,使用离散小波变换提取主轴功率信号的近似系数,将所提取的近似系数、主轴转速、铣削深度作为输入向量,刀具磨损作为输出向量,建立样本数据集,并将样本数据集输入BP神经网络中进行木工刀具磨损状态监测模型训练,同时使用遗传算法对BP神经网络的阈值和权值进行优化,实现对不同铣削条件下的木工刀具磨损状态进行精确监测。【结果】离散小波变换所提取主轴信号的近似系数能明显反映木工刀具磨损状态变化;在使用相同的样本数据集与遗传算法参数时,使用遗传BP神经网络所建立的木工刀具磨损状态监测模型的准确度可以达到100%,优于使用遗传概率神经网络建立监测模型的准确度。【结论】即使在样本数据集选取不佳时,本研究提出的监测方法仍然能对不同铣削条件下的木工刀具磨损状态进行精准监测,可以用于木质家具实际生产,达到提高木质家具加工质量、降低生产成本的目的。【Objective】Tool wear conditions monitoring is an important mechanical processing system that can improve the processing quality of wooden furniture and reduce industrial production costs.An appropriate signal,feature extraction method and model establishment method can effectively improve the accuracy of tool wear monitoring and reduce production costs.In order to solve the problem of wood furniture quality deterioration and high production cost caused by wood cutting tool wear in the process of wood furniture production.【Method】An effective method based on discrete wavelet transformation and genetic Algorithm-BP neural network was proposed to monitor the wear conditions of woodworking tools.The spindle power signals under different spindle speeds,depths of milling and tool wear conditions were collected by the power sensors connected to the machine tool control box.Based on feature extraction method,the approximate coefficients of spindle power signal were extracted by Discrete Wavelet Transformation.Then the extracted approximate coefficients,spindle speed and depth of milling and tool wear conditions were taken as samples to train the woodworking tool wear conditions monitoring model.Meanwhile,threshold and weight of BP neural network were optimized by genetic algorithm.【Result】The accuracy of monitoring model established by the genetic algorithm-BP neural network can reach 100%,which is better than model established by the genetic algorithm-probabilistic neural network.【Conclusion】Thus,the monitoring method proposed in this work can accurately monitor the wear conditions of woodworking tools with different milling parameters,which can achieve the purpose of improving the processing quality of wooden furniture and reducing production costs.

关 键 词：木工刀具磨损状态监测 铣削参数 主轴功率 离散小波变换 遗传算法 BP神经网络 

分 类 号：S784[农业科学—林学]