A multi-layer power transformer life span evaluating decision model based on information fusion

The operational reliability of power transformers directly affects the stable and reliable power supply of the power system. According to the available data such as condition assessment, routine test results and maintenance records, a hierarchical architecture and information fusion technology based power transformers life estimation model is presented. Firstly, the model takes advantage of the fault mechanism and the knowledge fitting laws of export system to analyze association property between characteristic parameters and aging information. Then it collects the information representing the insulation aging, reclassify them into several types, and build up a multi-objective evaluating model of power transformers life estimation. After the hierarchy of model is confirmed, we fulfill single characteristic parameter based life estimation and probability density of failure by means of threshold value diagnosis, fuzzy inferences, expert knowledge system, etc. The value of parameters, illation process and evaluation results of the estimation are coded and saved in the database of expert system and enrich the expert knowledge base through deep data mining technology. According to the established evaluation hierarchical architecture, single characteristic parameter models are integrated into the evaluating index system and multi-objective model by analytic hierarchy process (AHP), which can determine the weights at all levels. The multi-parameter power transformer life span evaluating decision model based on information fusion is established so far. The lifetime characteristic parameters from all kinds of sources are integrated to serve as the base layer of the multi-layer decision model, which enhances the efficiency and reliability of this model. It provides a reference for drafting optimal maintenance scheme, achieving the aim of condition based maintenance and power utilities life prediction.