基于本体的机械故障诊断贝叶斯网络

中国机械工程 ›› 2013, Vol. 24 ›› Issue (09): 1195-1200,1209.

基于本体的机械故障诊断贝叶斯网络

秦大力1,2;于德介1

1.湖南大学汽车车身先进设计制造国家重点实验室，长沙，410082
2.湖南农业大学，长沙，410128

出版日期:2013-05-10 发布日期:2013-05-16
基金资助:
国家高技术研究发展计划(863计划)资助项目(2009AA04Z414)；教育部长江学者与创新团队发展计划资助项目(531105050037)；广东省省部产学研结合项目（2009B090300312）
National High-tech R&D Program of China (863Program) （No. 2009AA04Z414）;
Supported by Program for Changjiang Scholars and Innovative Research Team in University（No. 531105050037）

Ontology-based Diagnostic Bayesian Networks for Mechanical Fault Diagnosis

Qin Dali1,2;Yu Dejie1

1.State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,Hunan University,Changsha,410082
2.Hunan Agriculture University,Changsha,410128

Online:2013-05-10 Published:2013-05-16
Supported by:
National High-tech R&D Program of China (863Program) （No. 2009AA04Z414）;
Supported by Program for Changjiang Scholars and Innovative Research Team in University（No. 531105050037）

摘要/Abstract

摘要：

针对机械设备维护与故障诊断过程中的不确定性，提出了一种将本体语义表示与贝叶斯网络相结合的故障概率推理模型。从异构多源的维护诊断信息和非结构化的专家经验知识出发，建立语义知识模型并进行概率扩展。利用贝叶斯分类器实现异常工况识别，给出了基于最大可能解释(MPE)的故障概率推理算法，从而根据运行工况、故障征兆和证据信息推理获得故障诊断解释。将本体语义描述的精确性和贝叶斯网络的概率推理能力相结合，既实现了诊断领域知识的形式化描述与共享，又能在一定程度上消除诊断过程的不确定性。某凉水塔风机转子典型故障诊断实例表明，该模型具有较好的故障识别效果。

关键词: 故障诊断, 本体建模, 贝叶斯网络, 概率推理

Abstract:

In view of the uncertainty of mechanical plant maintenance and fault diagnosis process, a hybrid fault reasoning model, which combined ontology semantic representation and Bayesian networks, was proposed. By extracting multi-source heterogeneous diagnostic information and non-structured experts' knowledge, a diagnostic ontology was constructed with probabilistic extension. Abnormal working conditions were identified using Bayesian classifier, and a fault probabilistic inference algorithm based on MPE(most probable explanation) was given, consequently diagnostic explanations could be obtained by the algorithm in accordance with operating conditions, symptoms and evidences. Combining the accuracy of semantic description and the probabilistic inference ability of Bayesian networks, this model realizes formalism and sharing of diagnostic domain knowledge, and eliminates uncertainty within the diagnostic processes to some extents. A typical rotor's fault diagnosis instances in cooling tower fan demonstrates that the proposed model has preferable fault recognition results.

Key words: fault diagnosis, ontology modeling, Bayesian network, probabilistic inference

中图分类号:

TP277

秦大力1, 2, 于德介1. 基于本体的机械故障诊断贝叶斯网络[J]. 中国机械工程, 2013, 24(09): 1195-1200,1209.

QIN Da-Li-1, 2, XU De-Ge-1. Ontology-based Diagnostic Bayesian Networks for Mechanical Fault Diagnosis[J]. China Mechanical Engineering, 2013, 24(09): 1195-1200,1209.

参考文献

[1]Lu Bin, Durocher D B, Stemper P. Predictive Maintenance Techniques[J]．Industry Applications Magazine, IEEE, 2009, 15(6): 52-60.
[2]Laskey K B. Model Uncertainty: Theory and Practical Implications[J]. Systems, Man and Cybernetics, Part A: Systems and Humans, 1996, 26(3): 340-348.
[3]Matsokis A, Kiritsis D. An Ontology-based Approach for Product Lifecycle Management[J]. Computers in Industry, 2010,61(8): 787-797.
[4]Garcia-Crespo A, Ruiz-Mezcua B, Lopez-Cuadrado J L, et al. Conceptual Model for Semantic Representation of Industrial Manufacturing Processes[J]. Computers in Industry, 2010, 61(7): 595-612.
[5]Yang S C, Patil L, Dutta D. Function Semantic Representation (FSR): a Rule-Based Ontology for Product Functions[J]. Journal of Computing and Information Science in Engineering, 2010, 10(3): 031001-8.
[6]雷琦, 刘飞, 王琦峰. 基于语义的协同制造执行平台集成框架研究[J]. 中国机械工程, 2008, 19(17): 2050-2054.
Lei Qi, Liu Fei, Wang Qifeng. Semantic-based Integration Framework for Collaborative Manufacturing Execution Platform[J]. China Mechanical Engineering, 2008, 19(17): 2050-2054.
[7]McNaught K, Chan A. Bayesian Networks in Manufacturing[J]. Journal of Manufacturing Technology Management, 2011, 22(6): 734-747.
[8]Agre G. Diagnostic Bayesian Networks[J]. Computers and Artificial Intelligence, 1996, 16(1): 1-21.
[9]Gilabert E, Arnaiz A. Intelligent Automation Systems for Predictive Maintenance: a Case Study[J]. Robotics and Computer-integrated Manufacturing, 2006, 22(5/6): 543-549.
[10]姜万录, 刘思远. 多特征信息融合的贝叶斯网络故障诊断方法研究[J]. 中国机械工程, 2010, 21(8): 940-945.
Jiang Wanlu, Liu Siyuan. Fault Diagnosis Approach Study of Bayesian Networks Based on Multi-characteristic Information Fusion[J]. China Mechanical Engineering, 2010, 21(8): 940-945.
[11]Darwiche A. Bayesian Networks[J]. Communications of the ACM, 2010, 53(12): 80-90.
[12]Duda R O, Hart P E, Stork D G. Pattern Classification[M]. New York: John Wiley & Sons, 2001.
[13]Cooper G F. The Computational Complexity of Probabilistic Inference Using Bayesian Belief Networks[J]. Artificial Intelligence, 1990, 42(2/3): 393-405.
[14]Gennari J H, Musen M A, Fergerson R W, et al. The Evolution of Protege: an Environment for Knowledge-based Systems Development[J]. International Journal of Human-computer Studies, 2003, 58(1):89-123.
[15]Carroll J J, Dickinson I, Dollin C. Jena: Implementing the Semantic Web Recommendations[C]//Association for Computing Machinery. Proceedings of the 13th International WWW Conference on Alternate Track Papers and Posters. New York: ACM, 2004: 74-83.
[16]Decision Systems Laboratory. GeNIe Documentation. Pittsburgh, DSL: University of Pittsburgh, [2012-02-16]. http://genie.sis.pitt.edu/wiki/GeNIe_Documentation.