基于人工神经网络的铣削磨损监测研究

Study of testing milling grind based on artificial neural networks

利用铣削刀具磨损多参量信号进行预处理及特征量提取,采用特征融合方法建立信号级、模型级、特征级和融合级层次结构实验方案,通过样本训练模糊小波神经网络逼近系统,建立刀具补偿系统的最优控制策略,从而对被检测对象进行有效的识别与估计。由实验结果对比可见,人工神经网络模型的预测精度基本在范围之内。实验表明该模型适用于切削条件下的铣刀磨损监控,可以较准确地监控铣刀的剧烈磨损。

The module based on artificial neural networks is used to establish system of testing milling grind. According to some cutting parameters, the grind＇s foretelling of milling tools in high - speed - steel milling tools hack faces grinding BP networks by experiments. The result of experiment shows that the module is fit for milling grind＇s testing under cutting condition, and it is helpful to control milling tools strong grindaccurately.