Fault detection method and system

Abstract

The invention relates to a fault detection method and system. The fault detection method comprises a first step of inputting fault phenomena; a second step of fully converging fault phenomenon nodes influencing each other and directly associated with fault conclusions, and establishing a Naive Bayesian network; a third step of utilizing a Bayesian formula to detect the credibility grading of all fault conclusions and the fault phenomena; and a fourth step of sequencing all the fault conclusions from high to low according to credibility. The faults do not need to be judged by establishing a directed acyclic graph, fault detection possibility is improved, and the fault detection method and system are particularly suitable for detecting the faults with a plurality of mutually influenced phenomena.

Description

technical field [0001] The invention relates to the field of fault detection, in particular to a fault detection method and system. Background technique [0002] In order to perform fault detection, expert systems are often provided in the prior art to help people perform some detections. The expert system in the prior art is based on the correlation between phenomenon A and conclusion B, using the Bayesian formula (Formula 1) to calculate the correlation probability, and then sorting. [0003] The Bayes formula is defined as follows: [0004] P ( B i / A ) = P ( A / B i ) · P ( B ...

AI Technical Summary

Problems solved by technology

However, many relationship structures are complex, and it is not easy to establish a DAG structure, because Bayesian networks are more complex than Naive Bayesian, and it is extremely difficult to construct and train a good Bayesian network

At the same time, the relationship between many factors, such as the above-mentioned body vibration and directional control difficulties, the mutual influence relationship will form a loop, which cannot meet the DAG

The above problems have caused the current expert system design methods and algorithms to be complex, computer implementation is complicated, and the scope of use is limited.

Images