A systematic method for rational definition of plant diagnostic symptoms by self-organizing neural networks

Abstract A method for evaluation of feature representations and definition of appropriate symptoms for diagnosis of large-scale artifacts is proposed in this paper. The central idea is the extraction of diagnostic information in symptoms obtained by a feature representation through automated categorization. Each possible feature representation is regarded as a feature vector in a specific parameter space. The Kohonen self-organizing network technique was applied to the feature vectors in order to obtain the optimal number of categories. Useful evaluation measures for the rational definition of symptoms were derived from the results of the categorization. By using these measures in evaluation processes, an appropriate set of feature representations can be implemented in a diagnosis system. The performance of the proposed method was evaluated through numerical experiments with a nuclear power plant simulator.

[1]  Teuvo Kohonen,et al.  The self-organizing map , 1990 .

[2]  James C. Bezdek,et al.  Generalized clustering networks and Kohonen's self-organizing scheme , 1993, IEEE Trans. Neural Networks.

[3]  Kazuo Yoshida,et al.  Statistical method application to knowledge base building for reactor accident diagnostic system , 1989 .

[4]  Erkki Oja,et al.  Rival penalized competitive learning for clustering analysis, RBF net, and curve detection , 1993, IEEE Trans. Neural Networks.

[5]  J. Joutsensalo,et al.  Nonlinear data compression and representation by combining self-organizing map and subspace rule , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[6]  M. Minsky.  The Society of Mind , 1986 .

[7]  Hiroshi Furukawa,et al.  Information Diversification for Improving Diagnostic Reliability in Complex Dynamic Processes , 1993, Robotics, Mechatronics and Manufacturing Systems.

[8]  L. Beltracchi,et al.  A strategy for minimizing common mode human error in executing critical functions and tasks , 1992 .

[9]  Poong Hyun Seong,et al.  Incipient multiple fault diagnosis in real time with application to large-scale systems , 1994 .

[10]  Steven X. Ding,et al.  Frequency domain approach to optimally robust residual generation and evaluation for model-based fault diagnosis , 1994, Autom..

[11]  J. Lampinen,et al.  Unsupervised Learning for Feature Extraction , 1994 .

[12]  Enrico Zio,et al.  Fault Diagnosis Via Neural Networks: The Boltzmann Machine , 1994 .

[13]  Jacek M. Zurada,et al.  Computational Intelligence: Imitating Life , 1994 .

[14]  Alexander G. Parlos,et al.  Incipient fault detection and identification in process systems using accelerated neural network learning , 1994 .

[15]  Jie Chen,et al.  Optimal unknown input distribution matrix selection in robust fault diagnosis , 1993, Autom..