Comprehensive Evaluation Of Operation State For Power Transformer

In power system,transformer is the pivotal equipment of transmission and transformation system.The operation of safe and stable state determines whether the whole power system can operate normally.The evaluation of operation status can be applied to subsequent maintenance work and life prediction,then the traditional periodic maintenance and post-accident maintenance will be transformed into condition-based maintenance,which is the focus of current research on electrical equipment.For this reason,this paper regards the operation state evaluation of transformer as the research content,focusing on the following work:Firstly,aiming at the selection of evaluation indicators,it analyzed the characteristic quantities of common evaluation indicators,and collected the operation and test data of the index quantities.Because of the complexity of the number of indicators,the correlation between the number of indicators and fault types is studied.The confidence degree of state index and fault type after c.orrelation degree calculation is taken as the criterion to determine the key evaluation index,then the key index system of evaluation is constructed.Secondly,the weights of key indicators are allocated.Aiming at the complexity of transformer,based on the single weight determination methods such as the improvement of analytic hierarchy process and entropy method,a minimum deviation fitting method is p.roposed to synthesize the weights and obtain the combined weights of each key index.Thus,the empowerment of weights is more objective.Finally,in view of the limitation of index information acquisition and the absoluteness of boundary level,an evaluation method based on normal cloud model is adopted to establish the evaluation model.The index data are expanded appropriately and the cloud of data is established,then extend the fuzzy state boundary level to establish the hierarchical cloud.After the expansion of cloud model,it exists the occurrence probability of data.Therefore,using the different probability of cloud droplet occurrence,the correlation degree between data and evaluation grade is obtained,and the evaluation matrix is obtained.Based on the theory of cloud and membership function,the results of transformer state grade are evaluated,and the accurate evaluation of transformer state is realized.Through collecting and analyzing the state index data of a 330 kV transformer in Ningxia,the key indexes are reasonably weighted,and applying the evaluation model proposed in this paper,it verifies the accuracy in state evaluation.Compared with other methods,it has better universality in this kind of problems.