Transformer Fault Comprehensive Diagnosis Method Based on Improved HPO-ELM and D-S Evidence Theory Fusion

Abstract

The existing transformer fault diagnosis methods have the problems of single information source and low diagnostic accuracy, making it difficult to make accurate and comprehensive judgments on the actual situation of transformers. On the basis of multi-source information fusion in power transformers, this paper proposes an improved fault diagnosis method that combines HPO-ELM and D-S evidence theory. optimize the output of HPO-ELM through mapping to obtain probability outputs for different labels, and then use evidence theory to fuse the probability allocation matrix. The diagnostic methods were compared and analyzed through experiments, and the algorithm was tested and analyzed using an actual example of a transformer, verifying the effectiveness and accuracy of the proposed method in this paper.