Study On The Comprehensive Assessment Model Of Transformer Operation Condition

As the key apparatus in electric power system, the operating reliability of large power transformer has deep influences on security and stability of power systems. With technology development, condition based maintenance, as a maintenance mode which can save maintenance cost compared with time based maintenance, shorten outage time, extend equipment lifespan, has become a focus of research. So the condition assessment on operating transformer is of great importance.Based on the analysis of transformer condition information, an index system of transformer condition assessment with practicability and reliability has been built. The static weight of each index has been calculated by Analytic Hierarchy Process with Group Decision Making, considering subjective weight and objective weight of experts at the same time. Then the dynamic weight was determined through the change of test data using information entropy idea, to revise the static weight obtained above, which makes the weight distribution of index more objective and reasonable. A hierarchical model was adopted, in order to avoid the over complexity of model while using the transformer condition information adequately. The index of model can be added or removed according to the practical situation, which makes the model has good expansibility and flexibility. The evaluation method combined weighted average technique and evidence theory was adopted. The valuation of basic probability assignments can be sounder by using fuzzy membership functions. According to actual situation studied, reliability factorα, as the index weight, was introduced to evidence theory, so that the different importance of each evidence can be shown and the evidence conflict be avoided.The effectiveness of this assessment model is examined using examples. It helps to establish maintenance scheme of transformers, and gives the economic foundation for implementing condition based maintenance of power transformers.