{
"title": "Bond Graph and Bayesian Networks for Reliable Diagnosis",
"authors": "Abdelaziz Zaidi, Belkacem Ould Bouamama, Moncef Tagina",
"country": null,
"institution": null,
"volume": "60",
"journal": "International Journal of Electrical, Computer, Energetic, Electronic and Communication Engineering",
"pagesStart": 1750,
"pagesEnd": 1764,
"ISSN": "1307-6892",
"URL": "http:\/\/waset.org\/publications\/8968",
"abstract": "Bond Graph as a unified multidisciplinary tool is widely\nused not only for dynamic modelling but also for Fault Detection and\nIsolation because of its structural and causal proprieties. A binary\nFault Signature Matrix is systematically generated but to make the\nfinal binary decision is not always feasible because of the problems\nrevealed by such method. The purpose of this paper is introducing a\nmethodology for the improvement of the classical binary method of\ndecision-making, so that the unknown and identical failure signatures\ncan be treated to improve the robustness. This approach consists of\nassociating the evaluated residuals and the components reliability data\nto build a Hybrid Bayesian Network. This network is used in two\ndistinct inference procedures: one for the continuous part and the\nother for the discrete part. The continuous nodes of the network are\nthe prior probabilities of the components failures, which are used by\nthe inference procedure on the discrete part to compute the posterior\nprobabilities of the failures. The developed methodology is applied\nto a real steam generator pilot process.",
"references": null,
"publisher": "World Academy of Science, Engineering and Technology",
"index": "International Science Index 60, 2011"
}