A Transformer Fault Diagnosis Method Based on rbpnn

Abstract

The invention discloses a transformer fault diagnosis method based on RBPNN, which belongs to the technical field of monitoring and diagnosis of intelligent substation equipment. The method includes: step 1, collecting the concentration data of transformer fault characteristic gases as historical data, and dividing the historical data into training samples and test samples after normalization processing, wherein the fault characteristic gases include hydrogen, methane, ethane, ethylene and acetylene ; Step two, use the concentration of the fault characteristic gas as the input layer, and the fault type as the output layer to establish the RBPNN model; step three, use the training samples to train the model, and combine the PSO algorithm to obtain the optimal RBPNN model; step four, test The samples are input into the optimized RBPNN model to obtain the predicted fault type. The invention introduces the RBPNN model, greatly improves the fault prediction accuracy and convergence speed after transformation and optimization, has high stability, and provides a new way for transformer fault diagnosis.

Description

technical field [0001] The invention relates to a transformer fault diagnosis method based on RBPNN, in particular to a diagnosis method of power transformer discharge and overheat faults, and belongs to the technical field of intelligent substation equipment monitoring and diagnosis. Background technique [0002] With the continuous expansion of the power grid capacity, the core equipment of the power system, the power transformer, has an increasing incidence of internal failures. Therefore, for the safe operation of the entire power grid, monitoring and judging the operating status of the transformer and its early latent faults has attracted great attention from the relevant power system departments. [0003] At present, Dissolved Gas Analysis (DGA) in oil is one of the most common and effective methods for transformer internal fault diagnosis, and the inherent disadvantage of its three-ratio method is that the coding boundary is too absolute and the coding is incomplete. ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art, provide a transformer fault diagnosis method based on RBPNN, introduce the RBPNN model, perform fault prediction after simplification and optimization, and solve the problem of slow convergence speed of fault diagnosis methods in the prior art. Technical issues with low accuracy

Images