基于进给伺服电流的铣削力预测模型研究
|
摘要
数控加工过程中,由于工件余量不均匀、材料硬度不等、刀具磨损变化等因素影响,一般选用保守的加工速度,机床应有效能不能充分发挥。对加工过程中的切削力进行实时监测并自适应调节,是提高加工效率的重要途径。
切削力直接测量一般采用三维测力仪,这种方法受条件限制,只适用于实验室研究,很难应用于实际加工过程。因此,本文提出了基于进给伺服电机电流的铣削力间接测量方法,并从以下几个方面进行了研究:
通过对铣削加工过程中铣削力传递和机电转换过程分析,得出铣削力和进给伺服电流这两个铣削过程的首尾参量之间存在联系。对切削实验中采集到的铣削力信号进行时域、频域分析,提取铣削力信号特征值。分析进给伺服电流信号时域、频域特性时,采用小波分析方法提取与铣削力频率相同的电流信号,以此信号幅值的平均值作为进给伺服电流信号表征铣削力的特征值。
针对进给伺服电流与铣削力的关系具有高度非线性的特点,采用神经网络BP算法,以进给电机电流信号的特征值、进给速度、主轴转速为输入,以铣削力为输出,通过大量实验样本的训练,建立了进给伺服电流和铣削力的关系模型,并通过优化神经网络参数,建立了一个符合要求的基于进给电机电流信号的铣削力预测神经网络模型。
最后,通过建立的铣削力和进给伺服电流检测实验平台,进行了几组典型的切削实验,对模型性能和预测精度进行了验证,并分析影响预测模型精度的主要因素。最终验证了模型的有效性。
During NC machining, the factors such as change in the depth of cut, material hardness variation and tool wear may result in the conservative feed rate selected and therefore insufficient use of NC machining tools. It is a important way to increase productivity by real-time monitoring and adaptive adjusting the machining process.
The method that the cutting force is measured directly by the 3-D dynamometer is only used in laboratory, can’t be used in industry. As a result, an indirect milling force measurement method based on feed motor current has been proposed. The dissertation focuses on the following research work.
Analyzing the transfer of milling force and the process of power conversion between mechanical and electrical power particularly, the relation of the milling force and feed motor current has been proved. Using the data from the experiments to analyze the milling force signal in both time and frequency domain, extracting the value of the milling force. Analyzing the feed motor current signal in both time and frequency domain, using the wavelet analysis method to extracting the current which has the same frequency as the milling force. The average value of this transformed current is used to be the value of feed motor current in token of milling force.
Because there is extraordinary non-linear relation between feed motor current and milling force, the BP arithmetic of artificial neural network method is used to establish the model of the relation. The value of feed motor current, feed speed and spindle rotate speed are used to be the input of the model, the milling force is used to be the output. After training by a lot of experimental data, optimizing the parameters of the net, a neural network model which is adapted to forecast milling force base on feed motor current has been established.
At last, using the experimental platform which is established to measure the milling force and feed motor current, taking several representative experiments to validate the capability and precision of this model, and analyzing the factors which have an effect on the model. Finally the validity of the prediction model has been proved.
切削力直接测量一般采用三维测力仪,这种方法受条件限制,只适用于实验室研究,很难应用于实际加工过程。因此,本文提出了基于进给伺服电机电流的铣削力间接测量方法,并从以下几个方面进行了研究:
通过对铣削加工过程中铣削力传递和机电转换过程分析,得出铣削力和进给伺服电流这两个铣削过程的首尾参量之间存在联系。对切削实验中采集到的铣削力信号进行时域、频域分析,提取铣削力信号特征值。分析进给伺服电流信号时域、频域特性时,采用小波分析方法提取与铣削力频率相同的电流信号,以此信号幅值的平均值作为进给伺服电流信号表征铣削力的特征值。
针对进给伺服电流与铣削力的关系具有高度非线性的特点,采用神经网络BP算法,以进给电机电流信号的特征值、进给速度、主轴转速为输入,以铣削力为输出,通过大量实验样本的训练,建立了进给伺服电流和铣削力的关系模型,并通过优化神经网络参数,建立了一个符合要求的基于进给电机电流信号的铣削力预测神经网络模型。
最后,通过建立的铣削力和进给伺服电流检测实验平台,进行了几组典型的切削实验,对模型性能和预测精度进行了验证,并分析影响预测模型精度的主要因素。最终验证了模型的有效性。
During NC machining, the factors such as change in the depth of cut, material hardness variation and tool wear may result in the conservative feed rate selected and therefore insufficient use of NC machining tools. It is a important way to increase productivity by real-time monitoring and adaptive adjusting the machining process.
The method that the cutting force is measured directly by the 3-D dynamometer is only used in laboratory, can’t be used in industry. As a result, an indirect milling force measurement method based on feed motor current has been proposed. The dissertation focuses on the following research work.
Analyzing the transfer of milling force and the process of power conversion between mechanical and electrical power particularly, the relation of the milling force and feed motor current has been proved. Using the data from the experiments to analyze the milling force signal in both time and frequency domain, extracting the value of the milling force. Analyzing the feed motor current signal in both time and frequency domain, using the wavelet analysis method to extracting the current which has the same frequency as the milling force. The average value of this transformed current is used to be the value of feed motor current in token of milling force.
Because there is extraordinary non-linear relation between feed motor current and milling force, the BP arithmetic of artificial neural network method is used to establish the model of the relation. The value of feed motor current, feed speed and spindle rotate speed are used to be the input of the model, the milling force is used to be the output. After training by a lot of experimental data, optimizing the parameters of the net, a neural network model which is adapted to forecast milling force base on feed motor current has been established.
At last, using the experimental platform which is established to measure the milling force and feed motor current, taking several representative experiments to validate the capability and precision of this model, and analyzing the factors which have an effect on the model. Finally the validity of the prediction model has been proved.
引文
[1]李圣怡,范大鹏.智能制造技术基础—智能控制理论、方法及应用.长沙:国防科技大学出版社,1995
[2]李小俚,董申.先进制造中的智能监控技术.北京:科学技术出版社,1999
[3]袁哲俊.金属切削实验技术.北京:机械工业出版社,1988
[4] K. Matsushima, P. Bertok. In Process Detection of Tool Breakage by Monitoring the Spindle Current of a Machine Tool. ASME J. of Measurement and Control for Batch Manufacturing, 1982: 145~154
[5] J. L. Stein, K. C. Shin. Current Monitoring of Field Controlled DC Spindle Drives. Journal of Dynamic System Measurement and Control, 1986, 108(12): 189~295
[6] M. A. Mannan, S. Broms. Monitorinf and Adaptive Control of Cutting Process by Means of Motor Power and Current Measurements. Annals of CIRP, 1989, 38(1): 347~350
[7] B. Y. Lee, Y. S. Tarng. Application of the Discrete Wavelet Transform to the Monitoring of Tool Failure in End Milling Using the Spindle Motor Current. Int. J. Adv. Manuf. Technl. 1999, 15: 238~243
[8] J. L. Stein, D. Colvin, G. Clever, and C.–H. Wang. Evaluation of DC Servo Machine Tool Feed Drives as Force Sensors[J]. ASME Transactions Journal of Dynamic Systems Measurement, and Control, 1986,108(4):279~288
[9] Y.Altintas, Prediction of Cutting Forces and Tool Breakage in Milling From Feed Driver Current Measurement [J]. ASME Transactions Journal of Engineering for Industry, 1992,114(4):286~392
[10] T.-Y.Kim and J.Kim. Adaptive Cutting Force Control for a Machining Center by Using Indirect Cutting Force Measurements [J]. International Journal of Machine Tools&Manufacture, 1996,36(8):925~937
[11] Y.-C.Chang, K.-T.Lee, and H.-y.Chuang. Cutting Force Estimation of Spindle Motor[J]. Journal of Control Systems Technology. 1995,3(2):145~152
[12]童强.智能自适应加工中铣削力间接测量的研究.[硕士学位论文].武汉:华中科技大学图书馆,2000
[13]李斌,贾瑜,吴波.伺服电流-铣削力关系间接建模的新方法.华中理工大学学报,2000,28(2): 51～53
[14]李水进.基于进给伺服电流的铣削力间接测量与恒进给力加工控制技术研究. [博士学位论文].武汉:华中科技大学图书馆,2001
[15]李曦.数控机床运动及加工过程中的非线性控制问题研究.[博士学位论文].武汉:华中科技大学图书馆,2002
[16] J W Sutherland, R E Devor. An improved method for cutting force and surface error prediction in flexible end milling system. Journal of Engineering for Industry,Transaction of the ASME,1986,108:269~279
[17]王素玉.高速立铣3Cr2Mo模具钢切削力建模及预测.山东大学学报,2006,26(1):1～5
[18]荆怀靖等.基于数字化加工的铣削力预测系统的开发.工具技术,2004,38(12):9～13
[19] X W Liu, K Cheng, D Webb, X C Luo. Prediction of cutting force distribution and its influence On dimensional accuracy in peripheral milling. Int.J.Mach.Tools& Manufact,2002,42:791~800
[20]郑治真著.小波变换及其MATLAB工具的应用.北京:地震出版社,2001.10
[21]高志,余啸海著.MATLAB小波分析工具箱原理与应用.北京:国防工业出版社,2004.1
[22]海金著;叶世伟等译.神经网络原理(第2版).北京:机械工业出版社,2004.1
[23]谷萩隆嗣等著.人工神经网络与模糊信号处理.北京:科学出版社,2003
[24]丛爽.面向MATLAB工具箱的神经网络理论与应用.合肥:中国科学技术大学出版社,2003
[25]陈虎,韩至骏.机床自适应控制的新发展.中国机械工程,1998(5):14～16
[26]陈统坚,王卫平,周泽华.铣削过程的约束型智能控制研究.华南理工大学学报,1994,8(4):90～96
[27]胡俊达,胡慧.自适应控制技术在数控与组合机床中的应用.组合机床与自动化加工技术,2004,11:73～75
[28] Tomizuka, M. S. J, Zhang, J. H. Oh, M. S. Chen. Model Reference Adaptive Control of the Milling Process. Control of Manufacturing Process and Robotic Systems, ASME, New York, 1983: 55~63
[29] Altinas Y. T. Yellowley, J. Tlusty. The Detection of Tool Breakage in Milling Operation. ASME J Eng for Industry, 1996, 110:271~277
[30] Y. Atlanta, C. C. H. Mo. Direct Adaptive Control of Milling Force. IEEE Trans.Automatic Control, 1990: 13~20
[31] S.J. Huang, C. Y. Shy, Fuzzy Logic for Constant Force Control of End Milling, IEEE Trans. on Industrial Electronic, 1999.46(1): 45~49
[32]王德斌.交流伺服进给系统及其数学模型的研究.机械制造与自动化,2006,35(1):86～88,91
[33]郑宝周.永磁同步电动机无传感器矢量控制技术研究.[硕士学位论文],郑州:郑州大学,2006
[34]龚云飞.交流永磁同步电机伺服系统的仿真及实现.[硕士学位论文],哈尔滨:哈尔滨工业大学,2006
[35]付宝琴,惠纪庄.数控铣削伺服系统的仿真建模与研究.西安公路交通大学学报,1997(1):79～82
[36]薛鸿健.切削加工智能监测技术与应用研究.[博士学位论文].武汉:华中科技大学图书馆,1996
[37]赵光宙,舒勤编.信号分析与处理.北京:机械工业出版社,2001
[38]刘波等著. MATLAB信号处理.北京:电子工业出版社,2006
[39]飞思科技产品研发中心.MATLAB6.5辅助小波分析与应用.北京:电子工业出版社,2003
[40]陈祥训.对几个小波基本概念的理解.电力系统自动化,2004(1):1～6
[41]刘顺兰,钱惠生,徐平原.Morlet小波变换在检测瞬时信号中的应用.杭州电子工业学院学报,1999(3):29～36
[42]孙勇,江南,陈南等.Morlet小波用于高分辨率的射频信号检测时的参数设置.现代电子技术,2007(6):184～187
[43]张波,李健君,李鸿超.基于Morlet小波带通滤波特性的振动系统频率识别.空军工程大学学报,2005(5):73～75
[44]卢新城等.基于Morlet小波的高分辨率信号频谱估计.武汉理工大学学报,2002(5):622～624,635
[45]李曦,李斌,李水进,周云飞.基于Morlet小波的机床伺服电流快速重构算法.机床与液压,2003(3):348～350
[46]楼顺天,施阳著.基于MATLAB的系统分析与设计—神经网络.西安:西安电子科技大学出版社
[47]飞思科技产品研发中心.MATLAB6.5辅助神经网络分析与设计.北京:电子工业出版社,2003
[2]李小俚,董申.先进制造中的智能监控技术.北京:科学技术出版社,1999
[3]袁哲俊.金属切削实验技术.北京:机械工业出版社,1988
[4] K. Matsushima, P. Bertok. In Process Detection of Tool Breakage by Monitoring the Spindle Current of a Machine Tool. ASME J. of Measurement and Control for Batch Manufacturing, 1982: 145~154
[5] J. L. Stein, K. C. Shin. Current Monitoring of Field Controlled DC Spindle Drives. Journal of Dynamic System Measurement and Control, 1986, 108(12): 189~295
[6] M. A. Mannan, S. Broms. Monitorinf and Adaptive Control of Cutting Process by Means of Motor Power and Current Measurements. Annals of CIRP, 1989, 38(1): 347~350
[7] B. Y. Lee, Y. S. Tarng. Application of the Discrete Wavelet Transform to the Monitoring of Tool Failure in End Milling Using the Spindle Motor Current. Int. J. Adv. Manuf. Technl. 1999, 15: 238~243
[8] J. L. Stein, D. Colvin, G. Clever, and C.–H. Wang. Evaluation of DC Servo Machine Tool Feed Drives as Force Sensors[J]. ASME Transactions Journal of Dynamic Systems Measurement, and Control, 1986,108(4):279~288
[9] Y.Altintas, Prediction of Cutting Forces and Tool Breakage in Milling From Feed Driver Current Measurement [J]. ASME Transactions Journal of Engineering for Industry, 1992,114(4):286~392
[10] T.-Y.Kim and J.Kim. Adaptive Cutting Force Control for a Machining Center by Using Indirect Cutting Force Measurements [J]. International Journal of Machine Tools&Manufacture, 1996,36(8):925~937
[11] Y.-C.Chang, K.-T.Lee, and H.-y.Chuang. Cutting Force Estimation of Spindle Motor[J]. Journal of Control Systems Technology. 1995,3(2):145~152
[12]童强.智能自适应加工中铣削力间接测量的研究.[硕士学位论文].武汉:华中科技大学图书馆,2000
[13]李斌,贾瑜,吴波.伺服电流-铣削力关系间接建模的新方法.华中理工大学学报,2000,28(2): 51～53
[14]李水进.基于进给伺服电流的铣削力间接测量与恒进给力加工控制技术研究. [博士学位论文].武汉:华中科技大学图书馆,2001
[15]李曦.数控机床运动及加工过程中的非线性控制问题研究.[博士学位论文].武汉:华中科技大学图书馆,2002
[16] J W Sutherland, R E Devor. An improved method for cutting force and surface error prediction in flexible end milling system. Journal of Engineering for Industry,Transaction of the ASME,1986,108:269~279
[17]王素玉.高速立铣3Cr2Mo模具钢切削力建模及预测.山东大学学报,2006,26(1):1～5
[18]荆怀靖等.基于数字化加工的铣削力预测系统的开发.工具技术,2004,38(12):9～13
[19] X W Liu, K Cheng, D Webb, X C Luo. Prediction of cutting force distribution and its influence On dimensional accuracy in peripheral milling. Int.J.Mach.Tools& Manufact,2002,42:791~800
[20]郑治真著.小波变换及其MATLAB工具的应用.北京:地震出版社,2001.10
[21]高志,余啸海著.MATLAB小波分析工具箱原理与应用.北京:国防工业出版社,2004.1
[22]海金著;叶世伟等译.神经网络原理(第2版).北京:机械工业出版社,2004.1
[23]谷萩隆嗣等著.人工神经网络与模糊信号处理.北京:科学出版社,2003
[24]丛爽.面向MATLAB工具箱的神经网络理论与应用.合肥:中国科学技术大学出版社,2003
[25]陈虎,韩至骏.机床自适应控制的新发展.中国机械工程,1998(5):14～16
[26]陈统坚,王卫平,周泽华.铣削过程的约束型智能控制研究.华南理工大学学报,1994,8(4):90～96
[27]胡俊达,胡慧.自适应控制技术在数控与组合机床中的应用.组合机床与自动化加工技术,2004,11:73～75
[28] Tomizuka, M. S. J, Zhang, J. H. Oh, M. S. Chen. Model Reference Adaptive Control of the Milling Process. Control of Manufacturing Process and Robotic Systems, ASME, New York, 1983: 55~63
[29] Altinas Y. T. Yellowley, J. Tlusty. The Detection of Tool Breakage in Milling Operation. ASME J Eng for Industry, 1996, 110:271~277
[30] Y. Atlanta, C. C. H. Mo. Direct Adaptive Control of Milling Force. IEEE Trans.Automatic Control, 1990: 13~20
[31] S.J. Huang, C. Y. Shy, Fuzzy Logic for Constant Force Control of End Milling, IEEE Trans. on Industrial Electronic, 1999.46(1): 45~49
[32]王德斌.交流伺服进给系统及其数学模型的研究.机械制造与自动化,2006,35(1):86～88,91
[33]郑宝周.永磁同步电动机无传感器矢量控制技术研究.[硕士学位论文],郑州:郑州大学,2006
[34]龚云飞.交流永磁同步电机伺服系统的仿真及实现.[硕士学位论文],哈尔滨:哈尔滨工业大学,2006
[35]付宝琴,惠纪庄.数控铣削伺服系统的仿真建模与研究.西安公路交通大学学报,1997(1):79～82
[36]薛鸿健.切削加工智能监测技术与应用研究.[博士学位论文].武汉:华中科技大学图书馆,1996
[37]赵光宙,舒勤编.信号分析与处理.北京:机械工业出版社,2001
[38]刘波等著. MATLAB信号处理.北京:电子工业出版社,2006
[39]飞思科技产品研发中心.MATLAB6.5辅助小波分析与应用.北京:电子工业出版社,2003
[40]陈祥训.对几个小波基本概念的理解.电力系统自动化,2004(1):1～6
[41]刘顺兰,钱惠生,徐平原.Morlet小波变换在检测瞬时信号中的应用.杭州电子工业学院学报,1999(3):29～36
[42]孙勇,江南,陈南等.Morlet小波用于高分辨率的射频信号检测时的参数设置.现代电子技术,2007(6):184～187
[43]张波,李健君,李鸿超.基于Morlet小波带通滤波特性的振动系统频率识别.空军工程大学学报,2005(5):73～75
[44]卢新城等.基于Morlet小波的高分辨率信号频谱估计.武汉理工大学学报,2002(5):622～624,635
[45]李曦,李斌,李水进,周云飞.基于Morlet小波的机床伺服电流快速重构算法.机床与液压,2003(3):348～350
[46]楼顺天,施阳著.基于MATLAB的系统分析与设计—神经网络.西安:西安电子科技大学出版社
[47]飞思科技产品研发中心.MATLAB6.5辅助神经网络分析与设计.北京:电子工业出版社,2003