Overview of Transformer State Evaluation and Fault Diagnosis Research

Abstract: The state assessment and fault diagnosis of power transformers provide important guarantees for the safe and stable operation of equipment. In the context of the widespread application of big data in power, the structure of smart grids is rapidly constructed, and the status data of power equipment presents characteristics such as large quantity and multiple types. Therefore, transformer status evaluation and fault diagnosis algorithms are gradually transitioning from threshold judgment methods to machine learning algorithms. The author of this article summarizes the methods used in transformer monitoring research both domestically and internationally in recent years; This paper provides an overview of the current research status in the field of transformer state assessment and fault diagnosis, and introduces the relevant principles of commonly used algorithms, including fuzzy theory, set pair analysis, traditional machine learning algorithms, prediction algorithms, and deep machine learning algorithms; Analyzed the urgent problems that need to be solved in this field and provided prospects for future research directions.

Keywords: power transformer; artificial intelligence; Status monitoring; State assessment; fault diagnosis