贝叶斯网络在飞机故障诊断与维修优化中的应用
|
摘要
在高技术条件下的局部战争中，战时维修的关键是时间，要求能快速及时地修复飞机投入战斗，以增加出动次数，这样就对飞机的故障诊断与维修保障的要求越来越高。机务维修人员就要从维修技术和手段上进一步提高，使其能具备快速诊断与维修能力，来降低故障诊断和维修所需要的时间。本文就是在军队“十五”国防预研项目“基于信息融合的装备快速故障诊断技术”的支持下，从维修技术和手段上能进一步得到提高进行研究的。
现在，世界各国开展大量的故障诊断的研究工作，通过各种途径来诊断故障。本文利用贝叶斯网络在故障诊断和维修优化中的优势，针对飞机故障诊断中的一些问题，将贝叶斯网络的相关理论研究应用到具体的飞机地面电源车和发动机的故障诊断与维修实践中，提高了对电源设备的维修能力和发动机故障诊断速度，重点是降低维修时间。论文主要完成的工作包括：
(1)在分析目前故障诊断与维修面临的主要问题的基础上，引入了基于贝叶斯网络的故障诊断与维修优化方法；通过对故障诊断问题的一般描述、对诊断贝叶斯网络的表达方式和数学描述进行的介绍，对贝叶斯网络故障诊断系统进行了分析和设计。
(2)目前贝叶斯网络在故障诊断领域进行的应用，重点是贝叶斯网络诊断模型的建造。本文对飞机地面电源车和发动机进行了贝叶斯网络模型建造。
(3)平时是凭机务人员的维修经验来进行工作的，很少把以往发生的故障进行分类和去寻找故障间的关联关系，本文的一个主要创新点就是对平时飞机的故障诊断与维修技术和手段进行改进，对以往发生的故障进行分类总结和找故障间的关联关系，把故障诊断与维修交替进行。把这样故障诊断和维修方法应用到飞机地面电源车和发动机故障诊断的实践中，降低了故障诊断和维修所需要的时间。
综上所述，本文开发的贝叶斯网络故障诊断系统，实现了一个将贝叶斯网络从理论研究过渡到飞机故障诊断和维修实践的应用。
In high technique war, the key of wartime maintain is time. Itrequests to quickly and in time repair airplanes which will be devotedfight. So the request of the trouble examining and maintaining of airplanebecome more and more high. The research of this thesis is just supportedby the troops "15 national defenses prepare projects-quickly equipmentsbreaksdown examining technique by information fusing",researching aimproved trouble examining and maintaining technique.
Now, the international communitys have invested a lot of money inbreakdown examining and a great number of breakdown examining system hadbeen invented. In this thesis, we apply Bayesian network to resolve theairplane breakdown diagnosis and maintaining, improving the maintainingability and examing speed. The works of the thesis mainly include:
(1) based on the analysis of the currently main problems of breakdownexamining and maintaining, we research the breakdown examining andmaintaining based on Bayesian network, and analyse and design the Bayesiannetwork breakdown examining system.
(2) Currently, the main problem of Bayesian network in breakdownexamining is how to construct the Bayesian network examining model. Thisthesis constructs the power supply car and launched machine Bayesiannetwork model.
(3) Usually, the maintaining work is only based on the aircraft crewexperience. It seldom classifies the breakdowns into different classes andfinds the relationship between these breakdowns, ln this thesis, we porposea method to improve the breakdown examining and maintaining technology. Thisapproach classifies these breaddown which former occur and analyses therelationship between those breaddowns, consequently implementingbreaddowns examining and maintaining in turn, reducing the time inbreakdown examining and maintaining.
As describing above, In this thesis, we persent a Bayesian networkbreakdown examining system. This system apply the Bayesian network theories to resolve the airplane breakdown examining and maintaining.
现在，世界各国开展大量的故障诊断的研究工作，通过各种途径来诊断故障。本文利用贝叶斯网络在故障诊断和维修优化中的优势，针对飞机故障诊断中的一些问题，将贝叶斯网络的相关理论研究应用到具体的飞机地面电源车和发动机的故障诊断与维修实践中，提高了对电源设备的维修能力和发动机故障诊断速度，重点是降低维修时间。论文主要完成的工作包括：
(1)在分析目前故障诊断与维修面临的主要问题的基础上，引入了基于贝叶斯网络的故障诊断与维修优化方法；通过对故障诊断问题的一般描述、对诊断贝叶斯网络的表达方式和数学描述进行的介绍，对贝叶斯网络故障诊断系统进行了分析和设计。
(2)目前贝叶斯网络在故障诊断领域进行的应用，重点是贝叶斯网络诊断模型的建造。本文对飞机地面电源车和发动机进行了贝叶斯网络模型建造。
(3)平时是凭机务人员的维修经验来进行工作的，很少把以往发生的故障进行分类和去寻找故障间的关联关系，本文的一个主要创新点就是对平时飞机的故障诊断与维修技术和手段进行改进，对以往发生的故障进行分类总结和找故障间的关联关系，把故障诊断与维修交替进行。把这样故障诊断和维修方法应用到飞机地面电源车和发动机故障诊断的实践中，降低了故障诊断和维修所需要的时间。
综上所述，本文开发的贝叶斯网络故障诊断系统，实现了一个将贝叶斯网络从理论研究过渡到飞机故障诊断和维修实践的应用。
In high technique war, the key of wartime maintain is time. Itrequests to quickly and in time repair airplanes which will be devotedfight. So the request of the trouble examining and maintaining of airplanebecome more and more high. The research of this thesis is just supportedby the troops "15 national defenses prepare projects-quickly equipmentsbreaksdown examining technique by information fusing",researching aimproved trouble examining and maintaining technique.
Now, the international communitys have invested a lot of money inbreakdown examining and a great number of breakdown examining system hadbeen invented. In this thesis, we apply Bayesian network to resolve theairplane breakdown diagnosis and maintaining, improving the maintainingability and examing speed. The works of the thesis mainly include:
(1) based on the analysis of the currently main problems of breakdownexamining and maintaining, we research the breakdown examining andmaintaining based on Bayesian network, and analyse and design the Bayesiannetwork breakdown examining system.
(2) Currently, the main problem of Bayesian network in breakdownexamining is how to construct the Bayesian network examining model. Thisthesis constructs the power supply car and launched machine Bayesiannetwork model.
(3) Usually, the maintaining work is only based on the aircraft crewexperience. It seldom classifies the breakdowns into different classes andfinds the relationship between these breakdowns, ln this thesis, we porposea method to improve the breakdown examining and maintaining technology. Thisapproach classifies these breaddown which former occur and analyses therelationship between those breaddowns, consequently implementingbreaddowns examining and maintaining in turn, reducing the time inbreakdown examining and maintaining.
As describing above, In this thesis, we persent a Bayesian networkbreakdown examining system. This system apply the Bayesian network theories to resolve the airplane breakdown examining and maintaining.
引文
[1] 吴今培，肖健华，智能故障诊断与专家系统。北京：科学出版社，1997
[2] 朱时银等，C++Buiider5编程实例与技巧。北京：机械工业出版社，2001
[3] 李凡，人工智能中的不确定性。北京：气象出版社，1992
[4] 张志涌等，精通MATLAB 6.5版。北京：北京航空航天大学出版社，2003
[5] GeigerD, HeckermanD. A Characterization of The Dirichlet Distribution with Applicable to learning Bayesian Networks. Proceedings of Eleventh Conference on Uncertainty in Artificial Intelligence, Montreal, QU, 1995: 196-207
[6] 张安华，设备故障诊断中的信息融合技术，机械科学与技术，1997，
[7] 史忠植，知识发现。北京：清华大学出版社，2002
[8] Dasarathy B V. Sensor Fusion Potential Exploitation-Innovative Architecture and Illustrative Application. Proceedings of the IEEE, 1997, 85(1):29-36
[9] 金星，洪延姬，武江涛，沈怀荣，基于故障树的智能型故障诊断系统。宇航学报，2001，22(3)：111—114
[10] 李俭川，贝叶斯网络故障诊断与维修决策方法及应用研究：[博士学位论文]。长沙：国防科技大学，2002
[11] C++Buiider与Matlab混合编程Http://cdf.51.net/software/softeng/developmentc++ma, html
[12] 高文胜，钱政，严璋．基于决策树神经网络模型的电力变压器故障诊断方法。西安交通大学学报，1999，33(6)：23-29
[13] 徐章遂，故障信息诊断原理与应用[M]，北京，国防工业出版社2000
[14] 李国华，机械故障诊断[M]，北京，化学工业出版社1999
[15] 赵不贿，一种基于模型模拟电路模糊故障诊断方法。电子测量技术，2001
[16] 关惠玲，设备故障诊断专家系统原理与实践，北京：机械工业出版社2000
[17] 陈长征，刘强，概率因果网络在汽轮机故障诊断中的应用，中国电机工程学报，2001，21(3)：78-81
[18] 杨莉，尚勇，周跃峰，严璋。基于概率推理和模糊数学的变压器综合故障诊断模型，中国电机工程学报，2000，20(7)：19-23
[19] 蔡自兴，徐光佑，人工智能及其应用，北京：清华大学出版社，1996
[20] AgreGDiagnostic Bayesian Networks. Computers and Artificial Intelligence, 1996, 16(1):41-56
[21] 张子丘，王建平，装备技术保障概论，军事科学出版社，2001年7月
[22] 严仰光主编，航空航天器供电系统，北京：航空工业出版社，1995
[23] 李长云主编，飞机机体修理，中国人民解放军空军装备部2004年3月
[24] 甘茂治等著，军用装备维修工程学，北京：国防工业出版社，1999
[25] Romessis C, Stamatis A, Mathioudakis K. Setting up a BeliefNetwork for Turbofan Diagnosis With the Aid of an Engine Performance Model. ISABE 2001:1032～1040
[26] Dechter R. Bucket Elimination:Aunifying Framework for Probabilistic Inference. Proc. 12th Conference on Uncertainty in Artificial Intelligence. Porland, OR, 1996:211-219
[27] Przytula K, Hagen F, Yung K. Bayesian Networks for SateIIite Payload Testing, Proceeding of the Forty-Fourth SPIE, Denver, July I 999
[28] Skaanning C, Jensen F V, Kjxrulff Uo Printer Troubleshooting Using Bayesian Networks. Industrial and Engineering Applications of Artificial Intelligence and Expert Systems (IEA/AIE) 2000, New Orleans, USA, June, 2000
[29] Hansen A F. Bayesian Networks as a Decision Support Tool in Marine Applications. PhD Thesis, Department of Naval Architecture and Offshore Engineering, Technical University of Denmark, December, 2000
[30] 李占山，基于模型的诊断问题分解及其算法[J]，计算机学报，2003
[31] Phillip J. Brooke, Richard F. Paige. Fault trees for security system design and analysis[J]. Computer&Security. 2003 (3)
[2] 朱时银等，C++Buiider5编程实例与技巧。北京：机械工业出版社，2001
[3] 李凡，人工智能中的不确定性。北京：气象出版社，1992
[4] 张志涌等，精通MATLAB 6.5版。北京：北京航空航天大学出版社，2003
[5] GeigerD, HeckermanD. A Characterization of The Dirichlet Distribution with Applicable to learning Bayesian Networks. Proceedings of Eleventh Conference on Uncertainty in Artificial Intelligence, Montreal, QU, 1995: 196-207
[6] 张安华，设备故障诊断中的信息融合技术，机械科学与技术，1997，
[7] 史忠植，知识发现。北京：清华大学出版社，2002
[8] Dasarathy B V. Sensor Fusion Potential Exploitation-Innovative Architecture and Illustrative Application. Proceedings of the IEEE, 1997, 85(1):29-36
[9] 金星，洪延姬，武江涛，沈怀荣，基于故障树的智能型故障诊断系统。宇航学报，2001，22(3)：111—114
[10] 李俭川，贝叶斯网络故障诊断与维修决策方法及应用研究：[博士学位论文]。长沙：国防科技大学，2002
[11] C++Buiider与Matlab混合编程Http://cdf.51.net/software/softeng/developmentc++ma, html
[12] 高文胜，钱政，严璋．基于决策树神经网络模型的电力变压器故障诊断方法。西安交通大学学报，1999，33(6)：23-29
[13] 徐章遂，故障信息诊断原理与应用[M]，北京，国防工业出版社2000
[14] 李国华，机械故障诊断[M]，北京，化学工业出版社1999
[15] 赵不贿，一种基于模型模拟电路模糊故障诊断方法。电子测量技术，2001
[16] 关惠玲，设备故障诊断专家系统原理与实践，北京：机械工业出版社2000
[17] 陈长征，刘强，概率因果网络在汽轮机故障诊断中的应用，中国电机工程学报，2001，21(3)：78-81
[18] 杨莉，尚勇，周跃峰，严璋。基于概率推理和模糊数学的变压器综合故障诊断模型，中国电机工程学报，2000，20(7)：19-23
[19] 蔡自兴，徐光佑，人工智能及其应用，北京：清华大学出版社，1996
[20] AgreGDiagnostic Bayesian Networks. Computers and Artificial Intelligence, 1996, 16(1):41-56
[21] 张子丘，王建平，装备技术保障概论，军事科学出版社，2001年7月
[22] 严仰光主编，航空航天器供电系统，北京：航空工业出版社，1995
[23] 李长云主编，飞机机体修理，中国人民解放军空军装备部2004年3月
[24] 甘茂治等著，军用装备维修工程学，北京：国防工业出版社，1999
[25] Romessis C, Stamatis A, Mathioudakis K. Setting up a BeliefNetwork for Turbofan Diagnosis With the Aid of an Engine Performance Model. ISABE 2001:1032～1040
[26] Dechter R. Bucket Elimination:Aunifying Framework for Probabilistic Inference. Proc. 12th Conference on Uncertainty in Artificial Intelligence. Porland, OR, 1996:211-219
[27] Przytula K, Hagen F, Yung K. Bayesian Networks for SateIIite Payload Testing, Proceeding of the Forty-Fourth SPIE, Denver, July I 999
[28] Skaanning C, Jensen F V, Kjxrulff Uo Printer Troubleshooting Using Bayesian Networks. Industrial and Engineering Applications of Artificial Intelligence and Expert Systems (IEA/AIE) 2000, New Orleans, USA, June, 2000
[29] Hansen A F. Bayesian Networks as a Decision Support Tool in Marine Applications. PhD Thesis, Department of Naval Architecture and Offshore Engineering, Technical University of Denmark, December, 2000
[30] 李占山，基于模型的诊断问题分解及其算法[J]，计算机学报，2003
[31] Phillip J. Brooke, Richard F. Paige. Fault trees for security system design and analysis[J]. Computer&Security. 2003 (3)