Research on Extraction and Integration Method of Core Feature of Power Transformer under Incomplete Information

Abstract

Background: The digital transformation strategy puts forward higher requirements for the information support system of the power system. Also, it points out the direction for the operation and maintenance of essential equipment such as large power transformers. However, the complex operating environment of power transformers poses a challenge to the digital operation and maintenance of equipment. Objective: Complex power transformer monitoring data is characterized by a wide range of sources, complex structures, incomplete information and unfocused distribution. This paper discussed and studied the Extraction and Integration Method of Core Feature of Power Transformer under Incomplete Information to fully exploit the potential value of power data and realize data collection and condition assessment of the equipment. Method: First, the power transformer holographic data perception technology and its analysis methods were explained from the perspective of multiple parameters; Then, based on actual cases, an information decision table based on incomplete information of power transformers based on equipment components and fault types was established to obtain the core feature of the power transformer itself; Secondly, by analyzing the distribution of the core feature index, a model for multi-task data integration and fusion in a cloud-edge-end architecture was proposed, and the integration method for multi-source data transfer was elaborated. Results: Finally, the feasibility of the proposed method was verified through simulation analysis of a calculation example. Conclusion: The method described in this article has a certain reference value for evaluating complex equipment that is mainly data-driven but lacks more serious information in the power grid digital transformation process.