Method, system and computer program product for automatic generation of Bayesian networks from system reliability models

摘要

A method, apparatus and computer program product for the conversion of at least one reliability model of a technical system to a Bayesian network model for assisting in the system's failure diagnostics, has the steps of creating a structure of a Bayesian network using information from at least one reliability model of the technical system, creating parameters of the Bayesian network using information from the reliability model of the technical system, the Bayesian network model having a plurality of observation nodes, obtaining information about the plurality of observation nodes from a list of observations that augments information contained in the reliability model of the technical system, and inserting the observation nodes into the created structure of the Bayesian network.

主权项

1. A method of conversion of a reliability model of a technical system to a Bayesian network model for assisting in the technical system's failure diagnostics, comprising the steps of:
creating a structure of a Bayesian network using information from at least one reliability model of the technical system defining multiple states, including converting each reliability model into an individual node of the Beyesian network; creating parameters of the Bayesian network using information from the reliability model of the technical system; and creating a diagnostic Bayesian network based on the Bayesian network and a list of observations by:
for at least one individual node of the Beyesian network corresponding to a converted reliability model of the technical system, accessing a list of observations that augments information contained in the respective reliability model of the technical system, the list of observations defining an observed probability of failure for each of the multiple states of the respective reliability model; andgenerating an observation node based on the accessed list of observations, andinserting the generated observation node into the created structure of the Bayesian network by linking the observation node to one or more other nodes of the Bayesian network.