基于小波神经网络PID控制的主动夹具铣削颤振研究  

Research on Milling Chatter of Active Fixture Based on Wavelet Neural Network PID Control

作　　者：李香服[1] 张文灼[1] 温彬彬 刘亚川 LI Xiang-fu;ZHANG Wen-zhuo;WEN Bin-bin;LIU Ya-chuan(Intelligent Manufacturing Institute,Hebei College of Industry and Technology,Shijiazhuang 050000,China;College of career Technology,Hebei Normal University,Shijiazhuang 050031,China)

机构地区：[1]河北工业职业技术学院智能制造系,石家庄050000 [2]河北师范大学职业技术学院,石家庄050031

出　　处：《组合机床与自动化加工技术》2021年第1期49-52,56,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：河北省高等学校科学技术研究项目(ZD2019125);河北工业职业技术学院博士基金(ZRB201903)。

摘　　要：为了抑制铣削过程中产生的颤振,提高铣削加工过程中零部件表面质量。设计了小波神经网络PID控制方法,并对控制效果进行仿真。采用时域数值法对动态铣削过程中离散时间进行求解,利用小波神经网络PID控制方法对铣削过程进行控制。通过仿真和实验对铣削金属表面粗糙度进行测量,并且与增量式PID控制系统进行比较和分析。结果显示,采用增量式PID控制方法,铣削力和铣削深度实际值与理论值存在较大误差;采用小波神经网络PID控制方法,铣削力和铣削深度实际值与理论值存在较小误差。采用小波神经网络PID控制方法,可以提高铣削参数控制精度,减少铣削过程中对颤振的影响,提高铣削零部件表面质量。In order to suppress the chatter and improve the surface quality of parts in milling process.In this paper,the wavelet neural network PID control method is designed and the control effect is simulated.The discrete time of dynamic milling process is solved by time domain numerical method,and the milling process is controlled by wavelet neural network PID control method.The surface roughness of milled metal is measured by simulation and experiment,and compared with incremental PID control system.The results show that the actual value of milling force and milling depth has a large error compared with the theoretical value with the incremental PID control method,and the actual value of milling force and milling depth has a small error compared with the theoretical value with the wavelet neural network PID control method.Using wavelet neural network PID control method can improve the accuracy of milling parameters control,reduce the impact of chatter in milling process,and improve the surface quality of milling parts.

关 键 词：小波神经网络 增量式PID控制 铣削力 铣削深度 仿真 

分 类 号：TH164[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]