Transformer fault diagnosis method considering operation state grade

Abstract

The invention discloses a transformer fault diagnosis method considering an operation state grade. The method comprises the following steps: collecting five typical fault data of a transformer, constructing a fault diagnosis sample data set, and performing tagging processing; extracting seven gas contents and 19 gas ratios of the transformer oil chromatographic data as fault features, and performing feature dimension reduction fusion on the preprocessed 26-dimensional fault features by using a principal component analysis (PCA) method; constructing an adaptive relevance vector machine fault diagnosis model based on a particle swarm algorithm and K-fold cross validation; and when the operation state level of the transformer is evaluated as a serious fault, using the fault diagnosis model for diagnosis. According to the fault diagnosis model constructed by the invention, automatic optimization of kernel parameters is realized in a model training process, and compared with a multi-stage binary classifier, the structure is simple. Through fault diagnosis considering the operation state level, the error diagnosis risk of real-time online fault diagnosis of a traditional transformer canbe reduced.

Description

technical field [0001] The invention relates to the technical field of transformer fault diagnosis, in particular to a transformer fault diagnosis method considering the level of operating status. Background technique [0002] As one of the key pivotal equipment in the power system, the transformer runs continuously in the power system, and its own operating state directly affects the safe and stable operation of the entire power system. When the transformer has serious operating conditions and effective measures are not taken in time, it may cause fire, explosion, and even large-scale power outages. Therefore, when a transformer fails, it is of great significance to accurately determine the type of transformer fault through fault diagnosis technology, which can facilitate the staff to take corresponding maintenance measures in time, reduce the loss of faults, and avoid the expansion of accidents. [0003] Due to the high reliability of transformer operation, transformer fa...

AI Technical Summary

Problems solved by technology

Obviously, this diagnosis method does not meet the actual requirements of transformer real-time online fault diagnosis, and the imbalance between normal and fault data may lead to serious misdiagnosis and missed diagnosis.

In addition, the traditional fault diagnosis method directly divides the transformer into normal and fault types. The standard of this classification method is too rough and ignores the sub-health state (general defect) of the transformer, which has great ambiguity and uncertainty.

When the transformer has general defects, it is difficult to obtain effective diagnosis results by traditional diagnosis methods, which brings safety hazards to the reliable operation of the transformer

Images