基于NARX神经网络的材料电磁脉冲响应预测
详细信息 查看官网全文
- 英文论文题名：Prediction of Response of Material against Electromagnetic Pulse Based on NARX Neural Network
- 论文作者：陈翔 ; 韩慧 ; 陈永光 ; 郝晓军
- 英文论文作者：CHEN Xiang ; HAO Xiao-jun ; Chen Yong-guang ; Han Hui ; State Key Laboratory of Complex Electromagnetic Effects on Electronics and Information System ; Beijing Institute of Tracking and Telecommunications Technology
- 年：2015
- 作者机构：电子信息系统复杂电磁环境效应国家重点实验室;北京跟踪与通信技术研究所;
- 论文关键词：NARX神经网络 ; 电磁脉冲 ; 屏蔽效能 ; 时域响应 ; 最小相位法
- 英文论文关键词：NARX neural network ; electromagnetic pulse ; shielding effectiveness ; time-domain response ; minimal-phase
- 会议召开时间：2015-08-12
- 会议录名称：静电放电：从地面新技术应用到空间卫星安全防护—中国物理学会第二十届全国静电学术会议论文集
- 语种：中文
- 分类号：TN03;TP183
- 学会代码：WLJD
- 会议名称：静电放电 ; 从地面新技术应用到空间卫星安全防护—中国物理学会第二十届全国静电学术会议
- 会议地点：中国河北石家庄
- 主办单位：中国物理学会静电专业委员会
- 学会名称：中国物理学会静电专业委员会
- 页数：7
- 文件大小：718k
- 原文格式：D
- 会议级别：全国
摘要
为了准确估计材料对于强电磁脉冲的响应,材料屏蔽前后的时域响应波形被看作黑箱系统的输入-输出时间序列。将不同类型、峰值的电磁脉冲激励下,加载材料前后的时域波形分别拼接起来作为建模数据,并将建模数据被分成训练数据和验证数据,另取额外的输入-输出数据用作测试数据,通过训练NARX神经网络直接建立了材料电磁脉冲响应的时域模型。即使对于宏观电磁参数可能受电压影响的材料,该时域模型也能够在一定幅值和上升时间范围内估计材料的电磁脉冲响应。而且,所得结果比由幅频谱估计时域响应的方法更为准确。
In order to accurately estimate the response of the materials against the strong electromagnetic pulse,time-domain waveforms when loaded and unloaded materials are seen as an input-output time series of black-box system.Different types of EMP data when loaded and unloaded materials were respectively combined as modeling data.The modeling data are divided into training data and validation data,and another input-output data are utilized to as testing data.Time-domain model of the response of materials against EMP was directly established by training the NARX neural network.Even when the macroscopic electromagnetic parameters of material are affected by the voltage,the time-domain model is also able to estimate the response of materials against EMP within a certain range of amplitude and rise time.And the predicted results are more accurate than the method of estimating the time-domain response by the amplitude spectrum.
In order to accurately estimate the response of the materials against the strong electromagnetic pulse,time-domain waveforms when loaded and unloaded materials are seen as an input-output time series of black-box system.Different types of EMP data when loaded and unloaded materials were respectively combined as modeling data.The modeling data are divided into training data and validation data,and another input-output data are utilized to as testing data.Time-domain model of the response of materials against EMP was directly established by training the NARX neural network.Even when the macroscopic electromagnetic parameters of material are affected by the voltage,the time-domain model is also able to estimate the response of materials against EMP within a certain range of amplitude and rise time.And the predicted results are more accurate than the method of estimating the time-domain response by the amplitude spectrum.
引文
[1]HAYES M H,LIM J S,OPPENHEIM A V.Signal reconstruction from phase or magnitude[J].IEEETrans.Acoustics,Speech,Signal Processing,1980,28:672-680.
[2]TESCHE F M.On the use of Hibert transform for processing measure CW data[J].IEEE Transactions on Electromagnetic Compatibility,1992,34(3):259-266.
[3]石立华,周璧华,陈彬,等.基于幅-频曲线的系统时域响应特性评价方法[J].电波科学学报,2000,15(4):467-471.SHI Lihua,ZHOU Bihua,CHEN Bin,et al.Time-doamin characterization of a system based on the magnitude of its frequency response[J].Chinese Journal of Radio Science,2000,15(4):467-471.
[4]谢彦召,王赞基,王群书,等.基于频域幅度谱数据重建电磁脉冲时域波形[J].强激光与粒子束,2004,16(3):320-324.XIE Yanzhao,WANG Zanji,WANG Qunshu,et al.Reconstruction of electromagnetic pulse waveform based on the amplitude spectrum data[J].High Power Laser and Particle Beams,2004,16(3):320-324.
[5]谢彦召,刘顺坤,孙蓓云,等.电磁脉冲传感器的时域和频域标定方法及其等效性[J].核电子学与探测技术,2004,24:394-399.Xie Yanzhao,LIU Shunkun,SUN Beiyun,et al.Methodology of time-domain&frequency-domain calibration and equivalence for EMP sensor[J].Nuclear Electronics&Detection Technology,2004,24(4):395-399.
[6]曹景阳,谢树果,苏东林.基于最小相位法重建电磁脉冲时域波形[J].电波科学学报,2011,26(6):1102-1106.CAO Jingyang,XIE Shuguo,SU Donglin.Application of minimum phase method in a pulse measurement[J].Chinese Journal of Radio Science,2011,26(6):1102-1106.
[7]X CHEN,Y.G.CHEN,M WEI,M CUI.Broadband coaxial holder with continuous-conductor used for shielding effectiveness of materials against electromagnetic pulse[J].Electronics Letters,2013,49(8):532-534.
[8]Stephen J.Norton,I.J.Won.Identification of Buried Unexploded Ordnance from Broadband Electromagnetic Induction Data[J].IEEETransaction on Geoscience and Remote Sensing,2001,39(10):2253-2261.
[9]杨伟斌,吴光强,秦大同,等.人工神经网络的各参数对系统辨识精度的影响分析及各参数的确定方法[J].机械工程学报,2006,42(7):217-221,226.YANG Weibin,WU Guangqiang,QIN Datong,et al.Influence and analysis of artificial neural networks’s parameters on system identification accuracy and determination method of parameters[J].Chinese of journal of mechanical engineering,2006,42(7):217-221,226.
[10]K.Funahashi,Y.Nakamura.Approximation of Dynamical Systems by Continuous Time Recurrent Neural Networks[J].Neural Networks,1993,6(6):801-806.
[11]L.K.Li.Approximation Theory and Recurrent Networks[C].International Joint Conference on Neural Networks,1992,266-271.
[12]Kumpati S.Narendra,Kannan Parthasarathy.Identification and Control of Dynamical Systems Using Neural Networks[J].IEEE Transactions on Neural Networks,1990,1(1):4-27.
[13]Yonghua Fang,Mustapha C.E.Yagoub,Fang Wang,et al.a new macromodeling approach for nonlinear microwave circuits based on recurrent neural networks[J].IEEE Transactions on Microwave Theory and Techniques,2000,48(12):2235-2344.
[14]Jianjun Xu,Mustapha C.E.Yagoub,Qijun Zhang.Neural-Based Dynamic Modeling of Nonlinear
[15]Microwave Circuits[J].IEEE Transactions on Microwave Theory and Techniques,2002,50(12):2769-2780.
[16]Yi Cao,Runtao Ding,Qijun Zhang.State-Space Dynamic Neural Network Technique for High-Speed IC Applications:Modeling and Stability Analysis[J].IEEE Transactions on Microwave Theory and Techniques,2006,54(6):2398-2409.
[17]Hitaish Sharma,Qijun Zhang.Automated Time Domain Modeling of Linear and Nonlinear Microwave Circuits Using Recurrent Neural Networks[J].International Journal of RF and Mircrowave Computer-Aided Engineering,2008,18(3):195-208.
[18]Yi Cao,Qijun Zhang.A New Training Approach for Robust Recurrent Neural-Network Modeling of Nonlinear Circuits[J].IEEE Transactions on Microwave Theory and Techniques,2009,57(6):1539-1553.
[19]刘亚秋,马广富,石忠.NARX网络在自适应逆控制动态系统辨识中的应用[J].哈尔滨工业大学学报,2005,37(2):173-176.LIU Ya-qiu,MA Guang-fu,SHI Zhong.NARX network for dynamic system identification in adaptive inverse control[J].Journal of HARBIN Institute of technology,2005,37(2):173-176.
[20]蔡磊,马淑英,蔡红涛,等.利用NARX神经网络由IMF与太阳风预测暴时SYM-H指数[J].中国科学,2010,40(1):77-84.Cai L,Ma S Y,Cai H T,et al.Prediction of SYM-Hindex by NARX neural network from IMF and solar wind data.Sci China Ser E-Tech Sci,2009,52(10):2877-2885.
[21]吴启蒙,魏明,张希军,等.瞬态抑制二极管电磁脉冲响应建模[J].强激光与粒子束,2013,25(3):799-804.WU Qimeng,Wei Ming,Zhang Xijun,Fan Gaohui.Electromagnetic pulse response modeling of transient voltage suppressor[J].High power laser and particle beams,2013,25(3):799-804.
[22]Asriel U.Levin,Kumpati S.Narendra.Control of nonlinear dynamical systems using neural networks-Part II:observability,identification,and control[J].IEEE Transactions on Neural Networks,1996,7(1):30-42.
[23]Simon Haykin.Neural Networks:A Comprehensive Foundation.2nd ed[M].New Jersey:Prentice-Hall Inc.,1999.
[24]邱关源.电路[M].北京:高等教育出版社,1999.Qiu Guanyuan.Electric circuit[M].Beijing:Higher Education Press,1999.
[2]TESCHE F M.On the use of Hibert transform for processing measure CW data[J].IEEE Transactions on Electromagnetic Compatibility,1992,34(3):259-266.
[3]石立华,周璧华,陈彬,等.基于幅-频曲线的系统时域响应特性评价方法[J].电波科学学报,2000,15(4):467-471.SHI Lihua,ZHOU Bihua,CHEN Bin,et al.Time-doamin characterization of a system based on the magnitude of its frequency response[J].Chinese Journal of Radio Science,2000,15(4):467-471.
[4]谢彦召,王赞基,王群书,等.基于频域幅度谱数据重建电磁脉冲时域波形[J].强激光与粒子束,2004,16(3):320-324.XIE Yanzhao,WANG Zanji,WANG Qunshu,et al.Reconstruction of electromagnetic pulse waveform based on the amplitude spectrum data[J].High Power Laser and Particle Beams,2004,16(3):320-324.
[5]谢彦召,刘顺坤,孙蓓云,等.电磁脉冲传感器的时域和频域标定方法及其等效性[J].核电子学与探测技术,2004,24:394-399.Xie Yanzhao,LIU Shunkun,SUN Beiyun,et al.Methodology of time-domain&frequency-domain calibration and equivalence for EMP sensor[J].Nuclear Electronics&Detection Technology,2004,24(4):395-399.
[6]曹景阳,谢树果,苏东林.基于最小相位法重建电磁脉冲时域波形[J].电波科学学报,2011,26(6):1102-1106.CAO Jingyang,XIE Shuguo,SU Donglin.Application of minimum phase method in a pulse measurement[J].Chinese Journal of Radio Science,2011,26(6):1102-1106.
[7]X CHEN,Y.G.CHEN,M WEI,M CUI.Broadband coaxial holder with continuous-conductor used for shielding effectiveness of materials against electromagnetic pulse[J].Electronics Letters,2013,49(8):532-534.
[8]Stephen J.Norton,I.J.Won.Identification of Buried Unexploded Ordnance from Broadband Electromagnetic Induction Data[J].IEEETransaction on Geoscience and Remote Sensing,2001,39(10):2253-2261.
[9]杨伟斌,吴光强,秦大同,等.人工神经网络的各参数对系统辨识精度的影响分析及各参数的确定方法[J].机械工程学报,2006,42(7):217-221,226.YANG Weibin,WU Guangqiang,QIN Datong,et al.Influence and analysis of artificial neural networks’s parameters on system identification accuracy and determination method of parameters[J].Chinese of journal of mechanical engineering,2006,42(7):217-221,226.
[10]K.Funahashi,Y.Nakamura.Approximation of Dynamical Systems by Continuous Time Recurrent Neural Networks[J].Neural Networks,1993,6(6):801-806.
[11]L.K.Li.Approximation Theory and Recurrent Networks[C].International Joint Conference on Neural Networks,1992,266-271.
[12]Kumpati S.Narendra,Kannan Parthasarathy.Identification and Control of Dynamical Systems Using Neural Networks[J].IEEE Transactions on Neural Networks,1990,1(1):4-27.
[13]Yonghua Fang,Mustapha C.E.Yagoub,Fang Wang,et al.a new macromodeling approach for nonlinear microwave circuits based on recurrent neural networks[J].IEEE Transactions on Microwave Theory and Techniques,2000,48(12):2235-2344.
[14]Jianjun Xu,Mustapha C.E.Yagoub,Qijun Zhang.Neural-Based Dynamic Modeling of Nonlinear
[15]Microwave Circuits[J].IEEE Transactions on Microwave Theory and Techniques,2002,50(12):2769-2780.
[16]Yi Cao,Runtao Ding,Qijun Zhang.State-Space Dynamic Neural Network Technique for High-Speed IC Applications:Modeling and Stability Analysis[J].IEEE Transactions on Microwave Theory and Techniques,2006,54(6):2398-2409.
[17]Hitaish Sharma,Qijun Zhang.Automated Time Domain Modeling of Linear and Nonlinear Microwave Circuits Using Recurrent Neural Networks[J].International Journal of RF and Mircrowave Computer-Aided Engineering,2008,18(3):195-208.
[18]Yi Cao,Qijun Zhang.A New Training Approach for Robust Recurrent Neural-Network Modeling of Nonlinear Circuits[J].IEEE Transactions on Microwave Theory and Techniques,2009,57(6):1539-1553.
[19]刘亚秋,马广富,石忠.NARX网络在自适应逆控制动态系统辨识中的应用[J].哈尔滨工业大学学报,2005,37(2):173-176.LIU Ya-qiu,MA Guang-fu,SHI Zhong.NARX network for dynamic system identification in adaptive inverse control[J].Journal of HARBIN Institute of technology,2005,37(2):173-176.
[20]蔡磊,马淑英,蔡红涛,等.利用NARX神经网络由IMF与太阳风预测暴时SYM-H指数[J].中国科学,2010,40(1):77-84.Cai L,Ma S Y,Cai H T,et al.Prediction of SYM-Hindex by NARX neural network from IMF and solar wind data.Sci China Ser E-Tech Sci,2009,52(10):2877-2885.
[21]吴启蒙,魏明,张希军,等.瞬态抑制二极管电磁脉冲响应建模[J].强激光与粒子束,2013,25(3):799-804.WU Qimeng,Wei Ming,Zhang Xijun,Fan Gaohui.Electromagnetic pulse response modeling of transient voltage suppressor[J].High power laser and particle beams,2013,25(3):799-804.
[22]Asriel U.Levin,Kumpati S.Narendra.Control of nonlinear dynamical systems using neural networks-Part II:observability,identification,and control[J].IEEE Transactions on Neural Networks,1996,7(1):30-42.
[23]Simon Haykin.Neural Networks:A Comprehensive Foundation.2nd ed[M].New Jersey:Prentice-Hall Inc.,1999.
[24]邱关源.电路[M].北京:高等教育出版社,1999.Qiu Guanyuan.Electric circuit[M].Beijing:Higher Education Press,1999.