Study On Condition Assessment And Fault Diagnosis Model Of Transformer

The key equipment of power system is transformer,the continuous and reliable operation of transformer is the basis of ensuring the normal operation of power system.Departments collect multi-dimensional and multi-source heterogeneous parameter data in the process of transformer operation and maintenance.The parameter data can directly reflect the current operating state of transformer equipment.With the development of smart grid,the rapid increase in the scale of parametric data increases the difficulty of data processing and analysis.Aiming at the problems existing in transformer equipment condition assessment and fault diagnosis,this paper carries out research on the basis of summarizing the existing theoretical research.Firstly,the key parameter system is established.The quantitative parameters and qualitative indicators are effectively combined,and the data that can fully characterize the operation state of the transformer is incorporated into the system,and the influence of external factors is fully considered.An improved parameter screening algorithm based on kernel principal component analysis(KPCA)is proposed to mine the effective information of data and establish the key parameter system for transformer condition evaluation and fault diagnosis.Then the transformer condition evaluation model is built.An improved GWO-MCSVM model was proposed,and the parameters of the model were optimized by training samples to remove the subjectivity of expert evaluation to the greatest extent and improve the accuracy of evaluation results.The evaluation performance of the model proposed in this paper is verified by several groups of measured samples.Dissolved gas in oil data(DGA)can be used for transformer fault diagnosis,which can effectively find the latent fault of transformer and improve the maintenance level of transformer.The fault diagnosis model is built on the basis of the improved GWO-MCSVM model,and different types of parameter data are preprocessed to make the model have good generalization performance.Finally,the software platform of transformer condition evaluation and fault diagnosis is developed.The above algorithm and model are programmed,and the functions of key parameter system establishment,transformer status evaluation and transformer fault diagnosis are realized in the software,so as to complete the task of all-round monitoring of the actual transformer in operation.