人工神经网络在外圆断续车削中的应用研究
|
摘要
在圆锯片基体外圆断续车削过程中,刀具磨损是影响机床工作效率的主要因素之一。通过Deform软件对圆锯片基体外圆的断续车削过程进行了仿真并提取出了切向力Fx、进给力F y、背向力Fz及刀具温度随时间的变化信息,利用时域分析法和频域分析法分别对采集到的切向力Fx、进给力F y、背向力Fz随时间的变化信息及刀具温度随时间的变化信息进行了相应的分析,并分别采用BP神经网络和RBF神经网络对能反映刀具磨损状态的时域、频域特征量进行了智能识别比较研究。结果显示,RBF神经网络性能表现要更好一些。
Tool wear is the main factor that affects the working efficiency of the machine tool in the saw blade excircle intermittent turning process. The process of intermittent turning circular saw blade is simulated by using Deform-3 D software, and the change information of force(Tangential force Fx, feed force F y, back force Fz)with time and the change of tool temperature with time are extracted. Using time domain analysis and frequency domain analysis method, the information of the tangential force, feed force, the change of back force with time and the change of tool temperature with time are analyzed. And Comparison of intelligent recognition using BP neural network and RBF neural network are carried out respectively for the characteristic quantity in time domain and frequency domain of data which are extracted from the raw data and can reflect the tool wear state. The results show that the performance of RBF neural network is better.
Tool wear is the main factor that affects the working efficiency of the machine tool in the saw blade excircle intermittent turning process. The process of intermittent turning circular saw blade is simulated by using Deform-3 D software, and the change information of force(Tangential force Fx, feed force F y, back force Fz)with time and the change of tool temperature with time are extracted. Using time domain analysis and frequency domain analysis method, the information of the tangential force, feed force, the change of back force with time and the change of tool temperature with time are analyzed. And Comparison of intelligent recognition using BP neural network and RBF neural network are carried out respectively for the characteristic quantity in time domain and frequency domain of data which are extracted from the raw data and can reflect the tool wear state. The results show that the performance of RBF neural network is better.
引文
[1]陈士军.开放式数控系统中在机检测运动控制及检测精度研究[D].天津:天津大学,2009:61-74.
[2]王磊,黄志伟.带磨损补偿功能的刀具状态在线检测技术研究[J].工具技术,2010(44):103-104.
[3]计时鸣,张宪,张利,等.计算机视觉在刀具状态监测中的应用[J].浙江工业大学学报,2002(4):143-148.
[4]朱大奇,史慧.人工神经网络原理及应用[M].北京:科学出版社,2006:36-45.
[5]胡金滨,唐旭清.人工神经网络的BP算法及其应用[J].信息技术,2004,28(4):1-4.
[6]王阳萍,朱正平.MATLAB在RBF径向基神经网络仿真应用[J].甘肃科技,2004,20(10):54-55.
[7]陈伟,马如雄,郝艳红.基于MATLAB的BP人工神经网络设计[J].电脑学习,2005(4):30-31.
[8]智会强,牛坤,田亮,等.BP网络和RBF网络在函数逼近领域内的比较研究[J].科技通报,2005,21(2):193-197.
[9]桂现才.BP神经网络在MATLAB上的实现与应用[J].湛江师范学院学报,2004,25(3):79-83.
[10]董长虹.MATLAB神经网络与应用[M].北京:国防工业出版社,2005.
[11]叶健,葛林东,吴月娴,一种优化的RBF神经网络在调控识别中的应用[J].自动化学报,2007,33(6):652-654.
[12]臧小刚,宫新保,常成,等.一种基于免疫系统的RBF网络在线训练方法[J].电子学报,2008,36(7):1396-1400.
[13]乔俊飞,韩红桂.RBF神经网络的结构动态优化设计[J].自动化学报,2010,36(6):865-872.
[14]陆剑中,孙家宁.金属切削原理与刀具[M].北京:机械工业出版社,2003:200-202.
[15]莫国瑞,刘开第.函数逼近论方法[M].北京:科学出版社,2003:171-177.
[2]王磊,黄志伟.带磨损补偿功能的刀具状态在线检测技术研究[J].工具技术,2010(44):103-104.
[3]计时鸣,张宪,张利,等.计算机视觉在刀具状态监测中的应用[J].浙江工业大学学报,2002(4):143-148.
[4]朱大奇,史慧.人工神经网络原理及应用[M].北京:科学出版社,2006:36-45.
[5]胡金滨,唐旭清.人工神经网络的BP算法及其应用[J].信息技术,2004,28(4):1-4.
[6]王阳萍,朱正平.MATLAB在RBF径向基神经网络仿真应用[J].甘肃科技,2004,20(10):54-55.
[7]陈伟,马如雄,郝艳红.基于MATLAB的BP人工神经网络设计[J].电脑学习,2005(4):30-31.
[8]智会强,牛坤,田亮,等.BP网络和RBF网络在函数逼近领域内的比较研究[J].科技通报,2005,21(2):193-197.
[9]桂现才.BP神经网络在MATLAB上的实现与应用[J].湛江师范学院学报,2004,25(3):79-83.
[10]董长虹.MATLAB神经网络与应用[M].北京:国防工业出版社,2005.
[11]叶健,葛林东,吴月娴,一种优化的RBF神经网络在调控识别中的应用[J].自动化学报,2007,33(6):652-654.
[12]臧小刚,宫新保,常成,等.一种基于免疫系统的RBF网络在线训练方法[J].电子学报,2008,36(7):1396-1400.
[13]乔俊飞,韩红桂.RBF神经网络的结构动态优化设计[J].自动化学报,2010,36(6):865-872.
[14]陆剑中,孙家宁.金属切削原理与刀具[M].北京:机械工业出版社,2003:200-202.
[15]莫国瑞,刘开第.函数逼近论方法[M].北京:科学出版社,2003:171-177.