Multiple sectioned Bayesian network-based electronic circuit fault diagnosis method

Abstract

The invention relates to a multiple sectioned Bayesian network-based electronic circuit fault diagnosis method. Common electronic circuit fault diagnosis methods include a fuzzy set fault dictionary method, a neural network approach, a Bayesian network method and the like, and have low fault resolution, interpretability and real-time property. The method comprises the following steps of: setting two adjacent fault diagnosis reasoning credibility threshold parameters, and determining the number of intelligent agents; obtaining Bayesian subnetwork structures, mapping a fault cause source to each Bayesian subnetwork, and learning credibility condition probability parameters among nodes of a Bayesian subnetwork model by using an expectation-maximization (EM) algorithm; using nodes corresponding to overlapped signals as overlapped subareas of the network to form a complete multiple sectioned Bayesian network (MSBN) so as to construct a linked junction forest; and inputting respective k target characteristic signals serving as observation evidence into each Bayesian subnetwork. A spatial multi-source information fusion method is adopted, the fault diagnosis capacity of a system is improved, the method is suitable for complicated and uncertain systems, and the fault diagnosis accuracy and speed are greatly improved.

Description

technical field [0001] The invention relates to a method for diagnosing electronic circuit faults, in particular to a method for diagnosing electronic circuit faults based on a multi-module Bayesian network. Background technique [0002] In recent years, with the rapid development of electronic circuit design and manufacturing technology, various functional systems have emerged one after another. However, progress in fault detection and diagnosis of electronic circuits has been relatively slow, and the ability to design complex circuits has far exceeded the ability to detect and repair faults. According to the report of the U.S. military department, digital circuit boards account for 80% of electronic equipment, and analog circuit boards account for 20%. The reliability of electronic circuits determines the reliability of the entire system. However, for large and complex circuit system fault diagnosis, often due to some uncertain factors, the system presents uncertainty. ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an electronic circuit fault diagnosis method based on a multi-module Bayesian network that can overcome the shortcomings of low resolution, difficult model interpretation, and poor real-time performance in the fault diagnosis technology of existing electronic circuits.

Images