Short-Term Recognition Method of Transformer Fault Based on Entropy Weight Fusion Algorithm and Weighted Support Vector Machine
The short-term recognition method of transformer fault based on entropy weight fusion algorithm and weighted support vector machine is studied to improve the short-term recognition effect of transformer fault.The passive RFID sensor is used to collect the short-term data of the transformer,and the logarithmic standardized processing of the collected short-term data is performed.The entropy-weight fusion algorithm is used to assign weights to short-term data after logarithmic processing and the corresponding data are input into support vector machine to establish weighted support vector machine.The bat algorithm is improved by introducing self-learning factor and proportional weight coefficient.The parameters of weighted support vector machine are optimized by using improved bat algorithm.After parameter optimization,the weighted support vector machine is used to input the weighted short-term data and output the short-term recognition results of transformer faults.Experimental results show that the proposed method can effectively collect transformer data of short-term faults,process the collected data reasonably logarithmically,and make the data distribution more uniform.The proposed method can accurately identify transformer short-term faults.In different fault scenarios,the average absolute percentage error of the method is low,and it has a high short-term transformer fault identification accuracy.
万方数据
维普
