Transformer fault diagnosis and health assessment system based on XGBoost algorithm and method thereof

Abstract

The invention provides a transformer fault diagnosis and health assessment system based on an XGBoost algorithm and a method thereof, and the system comprises the following modules: a fault diagnosis model module which defines a fault state from the two perspectives of fault performance and fault degree, enriches a fault set, and builds a fault diagnosis model based on the XGBoost algorithm; and a health assessment model module which is used for establishing a state interval through KMeans clustering, enabling an output score to have distinction degree, establishing a health assessment model based on an XGBoost algorithm, giving a score to the transformer between a normal state and a fault state, and presenting a transformer health assessment result through the state interval. According to the method, the XGBoost algorithm is used, fault diagnosis is accurate and reliable, health assessment is objective and reasonable, the feasibility and effectiveness are high, and the method has guiding significance for state assessment of the power transformer; and the power transformer is especially suitable for a power system with high requirements for reliability and safety of the power transformer.

Description

technical field [0001] The invention relates to the technical field of transformer fault diagnosis and health assessment, in particular to a transformer fault diagnosis and health assessment system and method based on an XGBoost algorithm. Background technique [0002] As one of the most important electrical equipment in the power system, the health of the power transformer is directly related to the safe and stable operation of the power grid. At present, the power grid company mainly adopts the strategy of regular maintenance for power transformers, and there are problems of too frequent maintenance and untimely maintenance. Compared with the regular maintenance strategy, the annual maintenance cost can be reduced by 25-50% and the outage time can be reduced by 75% by implementing the condition assessment of the transformer. It can be seen that the state assessment of power transformers can not only detect faults in time, improve the safety and reliability of power system...

AI Technical Summary

Problems solved by technology

[0005] In view of the related technologies mentioned above, the inventor believes that for the existing fault diagnosis methods: the traditional machine learning method is difficult to further improve the diagnostic accuracy due to its limited ability to fit functions; the deep learning method has strict requirements on the quantity and quality of data , the training cost is high, and it is prone to overfitting

For the existing health assessment methods: the scores and corresponding weights of each influencing factor are assigned by individuals based on experience, so they are more dominated by subjective factors and not objective and precise enough

Images