Polymorphic failure tree-based bayesian network failure prediction method

• Main Authors: LI SHUMIN, WANG NING, SI WEITAO, ZHANG SHUAI, CAI ZHIQIANG, SI SHUBIN
• Format: Patent
• Language: Chinese; English
• Published: 24.09.2014
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a polymorphic failure tree-based bayesian network failure prediction method for solving the technical problem that the application range is small caused by difficult modeling in the existing bayesian network failure prediction method. The technical scheme is that according to the method, firstly a system failure mode to be predicted is determined, then prior probability distribution of corresponding root nodes in an FPBN (failure prediction bayesian network) is determined according to known prior probabilities of all bottom events of a failure tree. For a failure detection event existing in the polymorphic failure tree, related nodes in an FPBN model are further corrected, and the prior probability and conditional probability distribution of corresponding nodes are updated; finally, the probability of each state of the failure mode to be predicted is solved based on a built failure prediction model, and failure prediction is performed. According to the method, expanding is performed ba