Power Transformer Condition Assessment by On-Line Monitoring and Trend Analysis

Power transformer condition assessment is the basi of condition-based maintenance and plays a very important role in carrying assets management and risk assessment for power system. As the heart of transmission and transformation system, power transformer’s condition has close relation with many factors, such as the production quality of itself, operation environment and so on. This paper analyzes some variables which could represent the transformer’s condition, including the data from preventative test, factory test and commission test, diagnostic test, on-line detection and monitoring, routine inspection, operation history, family defection and environment information. Some parameters which have representation are chosen as the condition variables. Artificial neuron network (ANN) and Dempster-Shafer (D-S) evidence theory are adopted to form the multiparameter information fusion condition assessment model. Besides the static value, the change trends of some parameters are considered as an independent evidence section. Cooperating with on-line monitoring, it will be helpful to improve the accuracy and efficiency of condition assessment.