Circuit breaker fault diagnosis method based on Bayesian network

Abstract

The invention discloses a circuit breaker fault diagnosis method based on a Bayesian network. A Bayesian network diagnosis model used for circuit breaker fault diagnosis is built on an open source data mining KNIME platform according to the structure and fault characteristics of circuit breakers. The convergence, efficiency and accuracy of the diagnosis model method are verified through simulation experiment of real data of a large number of projects. A Bayesian network algorithm is adopted in the invention. The causality and uncertainty of voltage, current, insulation resistance and other variable data in the event of failure of a circuit breaker are taken into full consideration. The convergence and accuracy of the fault diagnosis result are improved greatly.

Description

technical field [0001] The invention relates to a circuit breaker fault diagnosis method, in particular to a circuit breaker fault diagnosis method based on a Bayesian network, which belongs to the field of electric equipment safety monitoring. Background technique [0002] In recent years, various mining algorithms have been widely used in equipment fault diagnosis and achieved good results. For example, various types of expert diagnostic systems, neural network diagnostic systems and diagnostic systems based on fuzzy theory. At present, the circuit breaker fault diagnosis method based on neural network algorithm is very common and effective. However, it also has many defects: first, when the number of training samples is large and the relationship between input and output is complex, its convergence speed becomes slow, or even does not converge; second, when the input feature vector dimension is large, its network performance poor performance. [0003] Bayesian network ...

AI Technical Summary

Problems solved by technology

However, it also has many defects: first, when the number of training samples is large and the relationship between input and output is complex, its convergence speed becomes slow, or even does not converge; second, when the input feature vector dimension is large, its network performance poor performance

However, when the number of training samples is large and the input-output relationship is complex, its convergence speed is very slow, or even does not converge; and when it is performing fault diagnosis, either the accuracy of the diagnosis result is low, or the diagnosis time is long

Images