基于声发射技术及小波分析的砂轮钝化状态监测方法研究
|
摘要
磨削加工往往是精密零件制造过程中的最终工序,而砂轮在磨削加工过程中,又不可避免的出现钝化现象,这直接影响了砂轮磨削加工的效率与加工产品的质量。因此,必须对砂轮的钝化状态进行有效监测,以确定砂轮合理的修整周期,这对磨削加工成本的降低,加工效率与加工质量的提高,具有重要的现实意义。
通过分析研究陶瓷材料磨削时所产生的声发射信号,不难发现声发射信号随着砂轮的钝化程度的变化而变化。因此,可以选择将声发射信号作为砂轮钝化状态识别的监测信号。在分析比较了参数分析法、频谱分析法在处理声发射信号中的不足,并对小波分析方法在声发射信号处理中的应用进行了全面深入的研究之后,本文提出了一种基于声发射技术与小波分析技术相结合的砂轮钝化状态监测方法。
首先,本文探讨了各类噪声对声发射信号的影响,对各类噪声的频率进行辨别分析,并选择了合适的滤波器进行滤波,以最大限度地减少噪声对有用信息的干扰。同时,根据声发射信号的特点及小波分析理论,确定了声发射信号在小波分析中小波基的选取规则,并指出Coiflets小波、Daubechies小波和Symlets小波适合于声发射信号处理;确定了最大分解尺度选取规则,这对声发射信号的小波分析具有重要的指导作用,接着对信号进行多层分解。
其次,定义了小波能量系数的概念,提出了基于小波能量系数法的声发射信号特征分析方法,将砂轮各个钝化状态(初期、中期、严重)信号的小波能量系数分别提取,可以发现,砂轮各个钝化状态的小波能量系数具有很好的一致性,而各钝化状态之间的能量系数却大为不同。因此,可以将其作为砂轮钝化状态识别器的输入参数,来进行砂轮钝化状态的识别判断。
最后,本文采用改进的BP神经网络算法,建立了三层BP神经网络作为砂轮钝化状态识别器,通过试验,确定了神经网络的各种相关参数及最优的BP神经网络结构,并利用获得的样本信号对神经网络进行训练和仿真测试,得到了良好的识别效果(识别率达到90%左右),验证了此神经网络的可靠性与准确性,达到了预期目标。
本文的研究成果对于推动砂轮钝化状态监测技术的发展,提高磨削加工的质量与效率具有重要意义和实用价值。
Usually, the grinding machining is the final procedure in the process of the exact parts manufacture. The grinding wheel dull is ineluctable in the manufacture procedure, which influences the efficiency and the quality of products. Therefore, the dull state of the wheel should be monitoring effectively in order to make sure the dressing period of the wheel, which plays a very important role in debasing the machining cost, enhancing the work efficiency, promoting the machining quality.
Through AE signals analyzing, it can be found that the signal produced in engineering ceramic grinding varying with the grinding wheel dull state, so the AE signal is chose as the monitoring signal to monitor the grinding wheel dull state. The application of wavelet analysis on acoustic emission signal processing was studied in the thesis after analyzing the parameter analytical method and the frequency spectrum. The method based on AE technique and wavelet analysis was presented to monitor the wheel grinding dull state.
Firstly, the influence of the noises is discussed, and the frequencies of the noises are differentiated. The interference of the noises is decreased maximally by choosing the appropriate filter. The rules of how to select the suitable wavelets for acoustic emission signal processing were proposed based on the features of acoustic emission signal and the theory of the wavelet analysis, In terms of the rules, Daubechies wavelet, Symlets wavelet and Coiflets wavelet are regarded to be suitable for acoustic emission. And the maximum decomposition level of wavelet analysis was also confirmed. The research result above is important to use wavelet analysis for acoustic emission signal processing. And then the signals are decomposed to several layers.
Secondly, the concept of the wavelet energy coefficient is defined. The feature extraction method for acoustic emission signal based on wavelet energy coefficient is proposed. The wavelet energy coefficients of each dull condition are extracted. Experimental results showed that the wavelet energy coefficients are accordant to the dull state, and different among states. So the wavelet energy coefficients can be used as the input parameters of the dull state discriminator.
The final of this paper, the improved arithmetic of the BP neural network is chose, and the three-layer Back-Propagation artificial neural network is built up to recognize the grinding wheel dull state. Meanwhile, the chose correlative parameters are used to determine the optimized neural network structure. The sample signals are used to training and simulating. The simulation results indicate that the trained network can be used to recognize the wheel dull states effectively (the ratio of the recognition can achieve to 90% around), and achieve to anticipated goal.
The research results of the thesis have important significance and practical value to promote the development of technology of the monitoring grinding wheel dull state and to improve the quality and efficiency of the grinding process.
通过分析研究陶瓷材料磨削时所产生的声发射信号,不难发现声发射信号随着砂轮的钝化程度的变化而变化。因此,可以选择将声发射信号作为砂轮钝化状态识别的监测信号。在分析比较了参数分析法、频谱分析法在处理声发射信号中的不足,并对小波分析方法在声发射信号处理中的应用进行了全面深入的研究之后,本文提出了一种基于声发射技术与小波分析技术相结合的砂轮钝化状态监测方法。
首先,本文探讨了各类噪声对声发射信号的影响,对各类噪声的频率进行辨别分析,并选择了合适的滤波器进行滤波,以最大限度地减少噪声对有用信息的干扰。同时,根据声发射信号的特点及小波分析理论,确定了声发射信号在小波分析中小波基的选取规则,并指出Coiflets小波、Daubechies小波和Symlets小波适合于声发射信号处理;确定了最大分解尺度选取规则,这对声发射信号的小波分析具有重要的指导作用,接着对信号进行多层分解。
其次,定义了小波能量系数的概念,提出了基于小波能量系数法的声发射信号特征分析方法,将砂轮各个钝化状态(初期、中期、严重)信号的小波能量系数分别提取,可以发现,砂轮各个钝化状态的小波能量系数具有很好的一致性,而各钝化状态之间的能量系数却大为不同。因此,可以将其作为砂轮钝化状态识别器的输入参数,来进行砂轮钝化状态的识别判断。
最后,本文采用改进的BP神经网络算法,建立了三层BP神经网络作为砂轮钝化状态识别器,通过试验,确定了神经网络的各种相关参数及最优的BP神经网络结构,并利用获得的样本信号对神经网络进行训练和仿真测试,得到了良好的识别效果(识别率达到90%左右),验证了此神经网络的可靠性与准确性,达到了预期目标。
本文的研究成果对于推动砂轮钝化状态监测技术的发展,提高磨削加工的质量与效率具有重要意义和实用价值。
Usually, the grinding machining is the final procedure in the process of the exact parts manufacture. The grinding wheel dull is ineluctable in the manufacture procedure, which influences the efficiency and the quality of products. Therefore, the dull state of the wheel should be monitoring effectively in order to make sure the dressing period of the wheel, which plays a very important role in debasing the machining cost, enhancing the work efficiency, promoting the machining quality.
Through AE signals analyzing, it can be found that the signal produced in engineering ceramic grinding varying with the grinding wheel dull state, so the AE signal is chose as the monitoring signal to monitor the grinding wheel dull state. The application of wavelet analysis on acoustic emission signal processing was studied in the thesis after analyzing the parameter analytical method and the frequency spectrum. The method based on AE technique and wavelet analysis was presented to monitor the wheel grinding dull state.
Firstly, the influence of the noises is discussed, and the frequencies of the noises are differentiated. The interference of the noises is decreased maximally by choosing the appropriate filter. The rules of how to select the suitable wavelets for acoustic emission signal processing were proposed based on the features of acoustic emission signal and the theory of the wavelet analysis, In terms of the rules, Daubechies wavelet, Symlets wavelet and Coiflets wavelet are regarded to be suitable for acoustic emission. And the maximum decomposition level of wavelet analysis was also confirmed. The research result above is important to use wavelet analysis for acoustic emission signal processing. And then the signals are decomposed to several layers.
Secondly, the concept of the wavelet energy coefficient is defined. The feature extraction method for acoustic emission signal based on wavelet energy coefficient is proposed. The wavelet energy coefficients of each dull condition are extracted. Experimental results showed that the wavelet energy coefficients are accordant to the dull state, and different among states. So the wavelet energy coefficients can be used as the input parameters of the dull state discriminator.
The final of this paper, the improved arithmetic of the BP neural network is chose, and the three-layer Back-Propagation artificial neural network is built up to recognize the grinding wheel dull state. Meanwhile, the chose correlative parameters are used to determine the optimized neural network structure. The sample signals are used to training and simulating. The simulation results indicate that the trained network can be used to recognize the wheel dull states effectively (the ratio of the recognition can achieve to 90% around), and achieve to anticipated goal.
The research results of the thesis have important significance and practical value to promote the development of technology of the monitoring grinding wheel dull state and to improve the quality and efficiency of the grinding process.
引文
[1] Gorman M R. New Technology for Wave Based Acoustic Emission and Acoustic-Ultrasonic. American Society of Mechanical Engineers,Applied Mechanics Division, 1994, AMD 188 (6-11):47~59
[2] H Dunegan. Acoustic Emission:A Promising Technique. Report UCID-4643, Lawrence Radiation Laboratory, Livermore, California, December 9, 1963
[3]李波.工程陶瓷高效深磨声发射实验研究.[硕士学位论文].湖南:湖南大学,2006
[4]刘贵杰,巩亚东,等.磨削质量在线监测方法研究.金刚石与磨料磨具工程,2004, 143(5): 24~27
[5]刘志艳.用于磨削加工的声发射监测仪的研究.燕山大学学报,1999,23(3):240~241
[6]徐春广,王信义.磨削过程声发射信号传播特性研究.北京理工大学学报,1998 18(4): 415~419
[7]徐春广,王信义.流体声发射传感器实验性能研究.传感器技术,1998 17(3):5~10
[8]徐春广,王信义.流体声发射传感器波动特性理论研究.兵工学报, 199819(3):242~246
[9] Hundt W, Leuenberger D, Rehsteiner F. An approach to monitoring of the grinding process using acoustic emission (AE) technique. Annals of the CIRP, 1994, 43(1):295~298
[10] T W Hwang,E P Whitenton, N N Hsu, etal. Acoustic emission monitoring of high speed grinding of silicon nitride. Ultrasonics, 2000, 38:295~298
[11] Egon Susic. Characterization of the grinding process by acoustic emission. International Journal of Machine Tools & Manufacture, 2000, 40:225~238
[12] Qiang Liu, Xun Chen, Nabil Gindy. Fuzzy pattern recognition of AE signals for grinding burn. International Journal of Machine Tools&Manufacture,2005,45: 811~818
[13] Dr Amin A Mokbel, Dr T M A Maksoud. Monitoring of the condition of diamond grinding wheels using acoustic emission technique .Journal of Materials Processing Technology, 2000, 101: 292~297
[14] Ibrahin Nur Tansel, Christine Mekdeci.Detection of tool failure in end milling with wavelet transformations and neural networks. Int. J. Mach. Tools Manu fact. 1995,35(8):1137~1147.
[15] M.A. Mannan. Ashraf A.Kassim, Ma Jing. Application of image and sound analysis technique to monitor the condition of cutting tool. Pattern Recognition Letters , 2000, 969~979
[16]潘旭辉.基于切削加工声音信号的刀具状态监测技术基础研究.[硕士学位论文].江苏:南京航空航天大学,2006.3
[17]袁哲俊.精密和超精密加工技术的新进展.工具技术,2006,40(3):3~9
[18]谢秀娴.基于声发射和神经网络的刀具磨损状态识别研究.[硕士学位论文].四川:西南交通大学,2005
[19]郑建明,李言,黄玉美.刀具磨损切削力功率谱特征多分辨分析.机床与液压,2002 (4):189~191
[20] N. H. Abu-Zahra, G. Yu. Gradual wear monitoring of turning inserts using wavelet analysis of ultrasound waves. International Journal of Machine Tools & Manufacture 43. 2003, 337~343.
[21]喻俊馨,黄惟公,王计生.基于小波神经网络的刀具故障监测系统.新技术新工艺,2004 (8):19~21
[22] X. Li, Z. Yuan. Tool Wear Monitoring with Wavelet Packet Transform Fuzzy Clustering Method. Wear, 1998, 219(2): 145~154
[23]杨杰.声发射信号处理与分析技术的研究.[硕士学位论文].吉林:吉林大学2003
[24] D.W.Prine. Inspection of Nuclear Reactor Welding by Acoustic Emission. Welding Design and Fabrication. 1977,50(1):74.
[25] A. Rice. Characteristics of acoustic emission sensors employed for tool condition monitoring. Proceedings of the Seventh Workshop on supervising diagnostics of Machining Systems. Technical University of Wroclaw. CIRP, Darpacz. 1996:241~252.
[26] H. Vaidya. Pull-out and fragmentation in model fiber composites. J. Mater. Sci.1991, 26 2467～2476.
[27]张平.集成化声发射信号处理平台的研究.[博士学位论文].北京:清华大学,2002
[28]耿荣生.声发射信号处理和分析技术.无损检测,2002,24(1):23~28
[29] Xiaoli Li. A brief Revier:Acoustic Emission Method for Tool Wear Monitoring during Turning. International Journal of Machine Tool & Manufacture,2002,42(2):157~165
[30] Mossing J C,Tuthill T A. Reduced Interference Distributions for the Detection and Classification of Outside Sound Source Acoustic Emissions. In:IEEE International Conference on Acoustics,Speech, and Signal Processing. Atlanta,GA,USA:IEEE,1996.2758 ~2761
[31] A Terchi,Y H J Au. Acoustic Emission Signal Processing. Measurement and Control,2001,4(8):240~244
[32] E P Serrano,M A Fabio. Application of the Wavelet Transform to Acoustic Emission Signals Processing. IEEE Transactions on Signal Processing,1996,44(5):1270~1275
[33] Li X,Wu J.Wavelet Analysis of Acoustic Emission Signals in Boring.Journal of Engineering Manufacture,2000,214(5):412~424
[34] Hiroaki Suzuki , Tetsuo Kinjo, Mikio Takemoto. Fracture-Mode Determination of Glass-Fiber Composites by Various AE Processing. Progress in Acoustic Emission VIII, The Japanese Society for NDI,1996,47~52
[35] AG Beattic. Acoustic emission, principles and instrumentation. Acoustic Emission,1983, 2:95～128
[36]陈志奎.工程信号处理中的小波基和小波变换分析仪系统的研究:[博士学位论文].重庆:重庆大学测试中心,1998.24~27
[37] Javad Akbari, Yoshio Staito, adaaki Hanaoka. Effect of grinding parameters on acoustic emission signals while grinding ceramics. Journal of Materials Processing Technology, 1996, 62: 403~407
[38] Tian Y,Lewin P L,Davies A E. PD pattern identification using acoustic emission measurement and neural networks. IEE Conference Publication,1999,5(467):541~543
[39]沈功田,段庆儒,周裕峰.压力容器声发射信号人工神经网络模式识别方法研究.无损检测,2001,23(4):144~149
[40] ChengRunwei,T Tatsumi,G Mitsuo.AE Behaviors Evaluation with BP Neural Network. Computers & Industrial Engineering,1996,31(3-4):867~871 31~272
[41]李家林,董运朝.声发射源特性的神经网络模式识别研究.无损检测,2001,23(6):231~233
[42]张立明.人工神经网络的模型及其应用.上海:复旦大学出版社,1994,32~54
[43]阎平凡,张长水.人工神经网络与模拟进化计算.北京:清华大学出版社,2000,10~39,282~302
[2] H Dunegan. Acoustic Emission:A Promising Technique. Report UCID-4643, Lawrence Radiation Laboratory, Livermore, California, December 9, 1963
[3]李波.工程陶瓷高效深磨声发射实验研究.[硕士学位论文].湖南:湖南大学,2006
[4]刘贵杰,巩亚东,等.磨削质量在线监测方法研究.金刚石与磨料磨具工程,2004, 143(5): 24~27
[5]刘志艳.用于磨削加工的声发射监测仪的研究.燕山大学学报,1999,23(3):240~241
[6]徐春广,王信义.磨削过程声发射信号传播特性研究.北京理工大学学报,1998 18(4): 415~419
[7]徐春广,王信义.流体声发射传感器实验性能研究.传感器技术,1998 17(3):5~10
[8]徐春广,王信义.流体声发射传感器波动特性理论研究.兵工学报, 199819(3):242~246
[9] Hundt W, Leuenberger D, Rehsteiner F. An approach to monitoring of the grinding process using acoustic emission (AE) technique. Annals of the CIRP, 1994, 43(1):295~298
[10] T W Hwang,E P Whitenton, N N Hsu, etal. Acoustic emission monitoring of high speed grinding of silicon nitride. Ultrasonics, 2000, 38:295~298
[11] Egon Susic. Characterization of the grinding process by acoustic emission. International Journal of Machine Tools & Manufacture, 2000, 40:225~238
[12] Qiang Liu, Xun Chen, Nabil Gindy. Fuzzy pattern recognition of AE signals for grinding burn. International Journal of Machine Tools&Manufacture,2005,45: 811~818
[13] Dr Amin A Mokbel, Dr T M A Maksoud. Monitoring of the condition of diamond grinding wheels using acoustic emission technique .Journal of Materials Processing Technology, 2000, 101: 292~297
[14] Ibrahin Nur Tansel, Christine Mekdeci.Detection of tool failure in end milling with wavelet transformations and neural networks. Int. J. Mach. Tools Manu fact. 1995,35(8):1137~1147.
[15] M.A. Mannan. Ashraf A.Kassim, Ma Jing. Application of image and sound analysis technique to monitor the condition of cutting tool. Pattern Recognition Letters , 2000, 969~979
[16]潘旭辉.基于切削加工声音信号的刀具状态监测技术基础研究.[硕士学位论文].江苏:南京航空航天大学,2006.3
[17]袁哲俊.精密和超精密加工技术的新进展.工具技术,2006,40(3):3~9
[18]谢秀娴.基于声发射和神经网络的刀具磨损状态识别研究.[硕士学位论文].四川:西南交通大学,2005
[19]郑建明,李言,黄玉美.刀具磨损切削力功率谱特征多分辨分析.机床与液压,2002 (4):189~191
[20] N. H. Abu-Zahra, G. Yu. Gradual wear monitoring of turning inserts using wavelet analysis of ultrasound waves. International Journal of Machine Tools & Manufacture 43. 2003, 337~343.
[21]喻俊馨,黄惟公,王计生.基于小波神经网络的刀具故障监测系统.新技术新工艺,2004 (8):19~21
[22] X. Li, Z. Yuan. Tool Wear Monitoring with Wavelet Packet Transform Fuzzy Clustering Method. Wear, 1998, 219(2): 145~154
[23]杨杰.声发射信号处理与分析技术的研究.[硕士学位论文].吉林:吉林大学2003
[24] D.W.Prine. Inspection of Nuclear Reactor Welding by Acoustic Emission. Welding Design and Fabrication. 1977,50(1):74.
[25] A. Rice. Characteristics of acoustic emission sensors employed for tool condition monitoring. Proceedings of the Seventh Workshop on supervising diagnostics of Machining Systems. Technical University of Wroclaw. CIRP, Darpacz. 1996:241~252.
[26] H. Vaidya. Pull-out and fragmentation in model fiber composites. J. Mater. Sci.1991, 26 2467～2476.
[27]张平.集成化声发射信号处理平台的研究.[博士学位论文].北京:清华大学,2002
[28]耿荣生.声发射信号处理和分析技术.无损检测,2002,24(1):23~28
[29] Xiaoli Li. A brief Revier:Acoustic Emission Method for Tool Wear Monitoring during Turning. International Journal of Machine Tool & Manufacture,2002,42(2):157~165
[30] Mossing J C,Tuthill T A. Reduced Interference Distributions for the Detection and Classification of Outside Sound Source Acoustic Emissions. In:IEEE International Conference on Acoustics,Speech, and Signal Processing. Atlanta,GA,USA:IEEE,1996.2758 ~2761
[31] A Terchi,Y H J Au. Acoustic Emission Signal Processing. Measurement and Control,2001,4(8):240~244
[32] E P Serrano,M A Fabio. Application of the Wavelet Transform to Acoustic Emission Signals Processing. IEEE Transactions on Signal Processing,1996,44(5):1270~1275
[33] Li X,Wu J.Wavelet Analysis of Acoustic Emission Signals in Boring.Journal of Engineering Manufacture,2000,214(5):412~424
[34] Hiroaki Suzuki , Tetsuo Kinjo, Mikio Takemoto. Fracture-Mode Determination of Glass-Fiber Composites by Various AE Processing. Progress in Acoustic Emission VIII, The Japanese Society for NDI,1996,47~52
[35] AG Beattic. Acoustic emission, principles and instrumentation. Acoustic Emission,1983, 2:95～128
[36]陈志奎.工程信号处理中的小波基和小波变换分析仪系统的研究:[博士学位论文].重庆:重庆大学测试中心,1998.24~27
[37] Javad Akbari, Yoshio Staito, adaaki Hanaoka. Effect of grinding parameters on acoustic emission signals while grinding ceramics. Journal of Materials Processing Technology, 1996, 62: 403~407
[38] Tian Y,Lewin P L,Davies A E. PD pattern identification using acoustic emission measurement and neural networks. IEE Conference Publication,1999,5(467):541~543
[39]沈功田,段庆儒,周裕峰.压力容器声发射信号人工神经网络模式识别方法研究.无损检测,2001,23(4):144~149
[40] ChengRunwei,T Tatsumi,G Mitsuo.AE Behaviors Evaluation with BP Neural Network. Computers & Industrial Engineering,1996,31(3-4):867~871 31~272
[41]李家林,董运朝.声发射源特性的神经网络模式识别研究.无损检测,2001,23(6):231~233
[42]张立明.人工神经网络的模型及其应用.上海:复旦大学出版社,1994,32~54
[43]阎平凡,张长水.人工神经网络与模拟进化计算.北京:清华大学出版社,2000,10~39,282~302