免疫前馈神经网络在传感器故障诊断中的应用
|
摘要
随着现代工业及科学技术的迅速发展,设备的安全性、可靠性和有效性变得越来越重要和突出,传统的故障诊断技术已不能满足需求,而人工神经网络(Artificial Neural Network,ANN)的迅速发展为故障诊断领域开辟了一条新的途径。但是针对具体问题来设计神经网络是一个很复杂的问题,网络的结构(隐含层数和各层神经元个数)、各层的激活函数和学习训练方法等往往需要根据设计者的经验和多次实验来确定,这样导致得到的神经网络并不是最优的,影响了故障诊断的性能。
免疫算法(Immune Algorithm, IA)作为一种抽取和反映生物机体免疫系统特点的优化算法。它通过对抗体交叉、变异的进化操作和基于抗体浓度的调节操作,使抗体不断优化,从而找到最佳抗体。本文运用ANN的方法对水质监测系统进行故障诊断,并且引入免疫算法来优化神经网络,更好的满足实际应用需要,提高故障诊断水平。
本文对水质监测系统的水质参数测量原理进行分析,建立了水质故障诊断的动态模型。选用反向传播(Back Propagation,BP)网络进行水质监测系统的故障诊断。提出用免疫算法优化BP网络,将网络结构、激活函数和训练方法等编码作为个体,进行免疫操作,得到最优或次优解,克服了网络结构、激活函数和训练方法的确定没有可循规则的弊端。以PC机和PCI9112数据采集卡为硬件基础,虚拟仪器Labwindows/CVI为软件开发平台,采用在Labwindows/CVI环境下调用Matlab程序的方式实现免疫算法优化BP网络的软件设计,完成水质监测系统故障诊断的界面,功能包括神经网络离线训练学习和在线故障诊断。最后对免疫算法优化的神经网络和经验得到的神经网络进行性能比较和分析。结果表明,在收敛速度和最小均方误差方面均优于经验得到的神经网络。
With the rapid development of modern industry and technology, the security, reliability and validity become more and more important. While traditional fault diagnosis can not meet the requirement, a new method is found in the field of fault diagnosis when Artificial Neural Network (ANN) is prompt progressing. But it is very complex to design neural network for the specifical problem. The network structure, activation function and training method are usually acquired according to the designer’s experience and many tries, it makes that the solution is not optimum and reduces the performance.
Immune Algorithm (IA) is a kind of optimization algorithms which comes from immune system’s characteristic of biologic system. It makes antibody optimize continually by using cross and mutation, adjusting based on antibody concentration, and then gets the optimum antibody. So in this paper, the ANN, which is optimized by IA, is used in the fault diagnosis of water-quality monitor to get better for practical application and raise the level of fault diagnosis.
The paper studies the measuring principle of water-quality parameters in the monitor system, and establishes the dynamic model of fault diagnosis about water-quality. Back Propagation (BP) network is used in the fault diagnosis of water-quality monitor. A method is presented to optimize BP network by IA, in which the network structure, activation function and training method are encoded as an individual, in the purpose of optimum solution. The problem has been availably solved that there isn’t a guided rules to specify the network structure, activation function and training method. Hardware in the project is based on computer and PCI 9112 DAQ card, and software is based on virtual instrument Labwindows/CVI. The software design is finished for optimal BP via IA in Labwindows/ CVI by calling ANN function of Matlab. The interface of water-quality monitoring system is accomplished with the function of ANN training off line and fault diagnosis on line. Finally, the performance is compared and analyzed between optimized network by IA and experiential network. The results show that the convergence rate and the mean squared-error are both better than the experiential network.
免疫算法(Immune Algorithm, IA)作为一种抽取和反映生物机体免疫系统特点的优化算法。它通过对抗体交叉、变异的进化操作和基于抗体浓度的调节操作,使抗体不断优化,从而找到最佳抗体。本文运用ANN的方法对水质监测系统进行故障诊断,并且引入免疫算法来优化神经网络,更好的满足实际应用需要,提高故障诊断水平。
本文对水质监测系统的水质参数测量原理进行分析,建立了水质故障诊断的动态模型。选用反向传播(Back Propagation,BP)网络进行水质监测系统的故障诊断。提出用免疫算法优化BP网络,将网络结构、激活函数和训练方法等编码作为个体,进行免疫操作,得到最优或次优解,克服了网络结构、激活函数和训练方法的确定没有可循规则的弊端。以PC机和PCI9112数据采集卡为硬件基础,虚拟仪器Labwindows/CVI为软件开发平台,采用在Labwindows/CVI环境下调用Matlab程序的方式实现免疫算法优化BP网络的软件设计,完成水质监测系统故障诊断的界面,功能包括神经网络离线训练学习和在线故障诊断。最后对免疫算法优化的神经网络和经验得到的神经网络进行性能比较和分析。结果表明,在收敛速度和最小均方误差方面均优于经验得到的神经网络。
With the rapid development of modern industry and technology, the security, reliability and validity become more and more important. While traditional fault diagnosis can not meet the requirement, a new method is found in the field of fault diagnosis when Artificial Neural Network (ANN) is prompt progressing. But it is very complex to design neural network for the specifical problem. The network structure, activation function and training method are usually acquired according to the designer’s experience and many tries, it makes that the solution is not optimum and reduces the performance.
Immune Algorithm (IA) is a kind of optimization algorithms which comes from immune system’s characteristic of biologic system. It makes antibody optimize continually by using cross and mutation, adjusting based on antibody concentration, and then gets the optimum antibody. So in this paper, the ANN, which is optimized by IA, is used in the fault diagnosis of water-quality monitor to get better for practical application and raise the level of fault diagnosis.
The paper studies the measuring principle of water-quality parameters in the monitor system, and establishes the dynamic model of fault diagnosis about water-quality. Back Propagation (BP) network is used in the fault diagnosis of water-quality monitor. A method is presented to optimize BP network by IA, in which the network structure, activation function and training method are encoded as an individual, in the purpose of optimum solution. The problem has been availably solved that there isn’t a guided rules to specify the network structure, activation function and training method. Hardware in the project is based on computer and PCI 9112 DAQ card, and software is based on virtual instrument Labwindows/CVI. The software design is finished for optimal BP via IA in Labwindows/ CVI by calling ANN function of Matlab. The interface of water-quality monitoring system is accomplished with the function of ANN training off line and fault diagnosis on line. Finally, the performance is compared and analyzed between optimized network by IA and experiential network. The results show that the convergence rate and the mean squared-error are both better than the experiential network.
引文
[1] 孟传伟,陈辉堂.控制系统的解析冗余故障诊断方法综述[J].浙江大学学报(增刊),1998:319-329
[2] 周东华,叶昊,王贵增.基于观测器方法的故障诊断技术若干重要问题的探讨[J].自动化学报,1998,24(3):338-344
[3] 周东华,孙优贤.控制系统的故障检测与诊断技术[M].清华大学出版社,1994
[4] 薛飞.控制系统综合故障诊断方法研究[M].北京理工大学,2000
[5] 徐常胜.基于神经网络和专家系统的故障诊断[J].控制与决策, 1995,10(4):342-346
[6] 王镛根,张学峰.故障诊断的神经网络多重模型自适应方法[J].航空动力学报,1997,12(2):152-155
[7] 程 瑞 琪 . 人 工 神 经 网 络 在 设 备 故 障 诊 断 中 的 应 用 [J]. 机 械 研 究 与 运用,1999,12(1):13-14
[8] 张定会.基于知识和神经网络相结合的实时故障诊断专家系统[J].自动化仪表,2000,21(7):11-15
[9] 张 建 华 , 张 俊 华 , 候 国 莲 . 神 经 网 络 在 故 障 诊 断 中 的 应 用 [J], 电 力 学报,1998,13(3):163-168
[10] 房方 , 魏乐 . 传感器故障的神经网络信息融合诊断方法 [J], 传感技术学报,2000,13(5):50-52
[11] 国家环境保护总局.水和废水监测分析方法[M].中国环境科学出版社,2002.8
[12] 日本工业用水协会.水质试验法[M].中国环境科学出版社,1990
[13] 陈小平,陈红仙.水电导率的双频测定方法[J].仪器仪表学报,2006,27 (5):520-522
[14] 任敏,王万良.基于神经网络的污水处理软测量系统的研究[J].自动化仪表,2001,22(10):8-9
[15] 雷少刚.基于 AD590 组成的温度测量电路及应用[J].西安航空技术高等专科学校学报,2007,25(3):21-22
[16] 胡守仁.神经网络导论[M].国防科技大学出版社,1993
[17] 任妮.改进 L-M 型 BP 模型的降雨径流预报[D].大连理工大学,2006
[18] 魏海坤.神经网络结构设计的理论与方法[M].国防工业出版社,2005
[19] Ampazis:Perantonis,S.J.:Levenberg-Marquard algorithm with adaptive momentum for the efficient training of feed forward networks Neural Network,2000.UCNN 2000, Proceedings of IEEE International Joint Conference on,Volume:1, 2000,Page(s):518-523 vol.2
[20] Chua Boon Lay:Khalid,M.:Yusof,R;An enhanced intelligent database engine by neural network and data mining: TENCON 2000.Proceeding,Volume: 2,2000, Page(s):518-523 vol.2
[21] 唐晓萍.数据挖掘与知识发现综述[J].电脑开发与应用,2002,15(4):31-35
[22] 韩家纬,MICHELINK.数据挖掘:概念与技术[M].机械工业出版社,2001
[23] 飞思科技产品研发中心.神经网络理论与 MATLAB7 实现[M].电子工业出版社,2005:234-240
[24] 刘超,陈小平.免疫前馈神经网络在电力负荷预报中的应用[J].江苏电器(增刊),2007:34-28
[25] De CASTROAN,Von ZUBEN FJ.Learnting and optimization using the clonal selection principle[J] .IEEE Trans on Evolutionary Computation, Special Issue on Artificial Immune Systems,2002,6(3):239-251
[26] CHUN Jang Sung, JUNG Hyun Kyo.HAHN Song Yop.A study on Comparison of Optimization Performance between Immune Algorithm and other Heuristic Algorithm[J].IEEE Trans on Magnetics,1998,34(5):2972-2975
[27] CHUN Jang Sung,LIM Jeong Pil,YOON Joong Suk.Optimal Design of Synchronous Motor with Parameter Correction using Immune Algorithm[J].IEEE Trans on Energy Conversion,1999,14(3):610-615
[28] 杨唐胜,陈文清.一种与遗传算法类似的人工免疫算法.武汉理工大学学报,2005,27(10):106-108
[29] Forrest S,Perelsom A,Cherukuri.Self-nonself discrimination in a computer [A].In: Proceedings of 1994 IEEE Computer Society Symposium on Research in Securityand Privacy[C].Los Almitos,CA,USA,IEEE Computer Society,994:202-212
[30] Castro LN,Von Zuben FJ.Data Mining: a Heuristic Approach[M].USA,Idea Group Publishing,2001:231-259
[31] 熊和金,陈德军.智能信息处理[M].国防工业出版社,2006,203-215
[32] 陈小平, 赵鹤鸣. FIR 滤波器设计:基于免疫算法的频率抽样技术[J]. 电子与信息学报,2000,24(11):1723-1727
[33] 丛爽.面向 Matlab 工具箱的神经网络理论与应用[M].中国科学技术大学出版社,1998
[34] 楼顺天,施阳.基于 MATLAB 的系统分析与设计—神经网络[M].西安电子科技大学出版社,2000
[35] 孙晓云,郭立炜.基于 LabWindows/CVI 的虚拟仪器设计与应用[M].电子工业出版社,2005
[36] 刘君华.基于 LabWindows/CVI 的虚拟仪器设计[M].电子工业出版社,2003
[2] 周东华,叶昊,王贵增.基于观测器方法的故障诊断技术若干重要问题的探讨[J].自动化学报,1998,24(3):338-344
[3] 周东华,孙优贤.控制系统的故障检测与诊断技术[M].清华大学出版社,1994
[4] 薛飞.控制系统综合故障诊断方法研究[M].北京理工大学,2000
[5] 徐常胜.基于神经网络和专家系统的故障诊断[J].控制与决策, 1995,10(4):342-346
[6] 王镛根,张学峰.故障诊断的神经网络多重模型自适应方法[J].航空动力学报,1997,12(2):152-155
[7] 程 瑞 琪 . 人 工 神 经 网 络 在 设 备 故 障 诊 断 中 的 应 用 [J]. 机 械 研 究 与 运用,1999,12(1):13-14
[8] 张定会.基于知识和神经网络相结合的实时故障诊断专家系统[J].自动化仪表,2000,21(7):11-15
[9] 张 建 华 , 张 俊 华 , 候 国 莲 . 神 经 网 络 在 故 障 诊 断 中 的 应 用 [J], 电 力 学报,1998,13(3):163-168
[10] 房方 , 魏乐 . 传感器故障的神经网络信息融合诊断方法 [J], 传感技术学报,2000,13(5):50-52
[11] 国家环境保护总局.水和废水监测分析方法[M].中国环境科学出版社,2002.8
[12] 日本工业用水协会.水质试验法[M].中国环境科学出版社,1990
[13] 陈小平,陈红仙.水电导率的双频测定方法[J].仪器仪表学报,2006,27 (5):520-522
[14] 任敏,王万良.基于神经网络的污水处理软测量系统的研究[J].自动化仪表,2001,22(10):8-9
[15] 雷少刚.基于 AD590 组成的温度测量电路及应用[J].西安航空技术高等专科学校学报,2007,25(3):21-22
[16] 胡守仁.神经网络导论[M].国防科技大学出版社,1993
[17] 任妮.改进 L-M 型 BP 模型的降雨径流预报[D].大连理工大学,2006
[18] 魏海坤.神经网络结构设计的理论与方法[M].国防工业出版社,2005
[19] Ampazis:Perantonis,S.J.:Levenberg-Marquard algorithm with adaptive momentum for the efficient training of feed forward networks Neural Network,2000.UCNN 2000, Proceedings of IEEE International Joint Conference on,Volume:1, 2000,Page(s):518-523 vol.2
[20] Chua Boon Lay:Khalid,M.:Yusof,R;An enhanced intelligent database engine by neural network and data mining: TENCON 2000.Proceeding,Volume: 2,2000, Page(s):518-523 vol.2
[21] 唐晓萍.数据挖掘与知识发现综述[J].电脑开发与应用,2002,15(4):31-35
[22] 韩家纬,MICHELINK.数据挖掘:概念与技术[M].机械工业出版社,2001
[23] 飞思科技产品研发中心.神经网络理论与 MATLAB7 实现[M].电子工业出版社,2005:234-240
[24] 刘超,陈小平.免疫前馈神经网络在电力负荷预报中的应用[J].江苏电器(增刊),2007:34-28
[25] De CASTROAN,Von ZUBEN FJ.Learnting and optimization using the clonal selection principle[J] .IEEE Trans on Evolutionary Computation, Special Issue on Artificial Immune Systems,2002,6(3):239-251
[26] CHUN Jang Sung, JUNG Hyun Kyo.HAHN Song Yop.A study on Comparison of Optimization Performance between Immune Algorithm and other Heuristic Algorithm[J].IEEE Trans on Magnetics,1998,34(5):2972-2975
[27] CHUN Jang Sung,LIM Jeong Pil,YOON Joong Suk.Optimal Design of Synchronous Motor with Parameter Correction using Immune Algorithm[J].IEEE Trans on Energy Conversion,1999,14(3):610-615
[28] 杨唐胜,陈文清.一种与遗传算法类似的人工免疫算法.武汉理工大学学报,2005,27(10):106-108
[29] Forrest S,Perelsom A,Cherukuri.Self-nonself discrimination in a computer [A].In: Proceedings of 1994 IEEE Computer Society Symposium on Research in Securityand Privacy[C].Los Almitos,CA,USA,IEEE Computer Society,994:202-212
[30] Castro LN,Von Zuben FJ.Data Mining: a Heuristic Approach[M].USA,Idea Group Publishing,2001:231-259
[31] 熊和金,陈德军.智能信息处理[M].国防工业出版社,2006,203-215
[32] 陈小平, 赵鹤鸣. FIR 滤波器设计:基于免疫算法的频率抽样技术[J]. 电子与信息学报,2000,24(11):1723-1727
[33] 丛爽.面向 Matlab 工具箱的神经网络理论与应用[M].中国科学技术大学出版社,1998
[34] 楼顺天,施阳.基于 MATLAB 的系统分析与设计—神经网络[M].西安电子科技大学出版社,2000
[35] 孙晓云,郭立炜.基于 LabWindows/CVI 的虚拟仪器设计与应用[M].电子工业出版社,2005
[36] 刘君华.基于 LabWindows/CVI 的虚拟仪器设计[M].电子工业出版社,2003