Research On PD Characteristics And Type Identification Of Transformer Based On Single Wave Characteristics

With the continuous development of power system,high-voltage and ULTRA-high-voltage DC transmission technology can deliver electric energy over long distances and in large capacity,and its advantages are becoming more and more prominent.The rapid development of this technology makes the power system more demanding on power equipment.In the power system,there are transformer,cable and other power equipment,transformer as one of the important equipment in these facilities,its fault diagnosis technology is very important in the power grid application and manufacturing industry.Before transformer failure,there is pulse current,local heating or gas generation.It is because of these phenomena that some data can be measured by instruments,which not only reflect the insulation state of equipment,but also can be used as the characteristic quantity of fault analysis.At the same time,these characteristic quantities can be used to distinguish fault types by computers.According to the mechanism of partial discharge in transformer,the measurement,feature extraction and defect identification of partial discharge signal are studied in this paper.The main research work is as follows:This paper mainly studies the partial discharge of internal defects under ac voltage during transformer operation.The distribution of electric field before discharge is analyzed.Firstly,according to the characteristics of the insulation results of the transformer pressure,several common defects within the transformer are analyzed,and then four kinds of defects are made: air gap defect,suspension defect,tip defect and surface defect.Finally,the characteristics of electric field distribution of these four kinds of defects are analyzed by using the finite element simulation software COMSOL.In this paper,a partial discharge test platform is built,and a special power frequency transformer is used in the experiment.The transformer can automatically switch defects;During the compression test,it can ensure the safety and accuracy of the experiment.Meanwhile,high-frequency current sensor is used in the experiment to measure more abundant experimental data.Through the measured experimental data,namely the single baud property of pd(characteristic quantity),then use MATLAB software analysis and processing.Finally,the improved BA optimized LS-SVM algorithm,LS-SVM algorithm,traditional BP algorithm and PSO optimized BP algorithm were used to identify various discharge types,and the software was used for statistics.