Fault Pattern Recognition Based on Internal Pressure Distribution of Oil-immersed Transformer

Abstract

The safe and stable operation of power transformer is the key link to ensure the reliable power supply of power system. Once an explosion accident occurs, it will bring huge economic and social losses to power grid companies. To improve the anti-explosion performance of oil-immersed power transformers, it is necessary to study the analysis method of internal fault pressure characteristics of transformers, and on this basis to study the non-electricity protection method of transformers and design the anti-explosion structure of transformer tank. Therefore, based on the simulation model of internal fault pressure characteristics of transformer, the mapping relationship between fault feature and internal pressure of oil tank is constructed by BP neural network, and the fault pattern recognition algorithm of transformer based on pressure is established by random forest model.