Evaluation Of Transformer Insulation Aging Status Based On Multi-parameter Fusion

As the core equipment of the power system,the health status of the transformer will directly affect the safe operation of the entire power system.Hunan’s Minfeng,Yuntian,Gangshi and other ultra-high voltage substations have been in operation for more than 30 years.According to the test and operation statistics of the operation and maintenance department,some transformers and high-resistance insulation are aging,which brings security risks to the operation of the transformer.This paper carries out insulation aging diagnosis research on old transformers,so as to correctly judge the insulation aging status of transformers and maintain them in time,which has important engineering significance for the safe operation of transformers and even power systems.Considering that transformer oil chromatographic test is a routine test,it is easy to implement in engineering,and old transformers have nearly 20 years of oil chromatographic data,transformer insulation aging diagnosis based on oil chromatographic data has engineering application value.Firstly,a transformer insulation aging simulation test platform was built,electrical and thermal aging methods were used to conduct aging simulation tests on oil-paper insulation,and the dissolved gases CO,CO2 and furfural content in oil of oil-paper at different aging times and different aging degrees were tested and analyzed.It is confirmed that CO and CO2 in oil are the characteristic quantities of oil-paper insulation aging,and their gas production law can reflect the aging process of transformer insulation.Secondly,improve the transformer insulation aging diagnosis method based on oil chromatography data,namely CO and CO2 gas production rate Rco,RCO2.Taking five characteristic parameters such as CO,CO2,CO/CO2,RCO,and RCO2,the fuzzy analytic hierarchy process is used to comprehensively judge the aging state of oil-paper insulation,which reduces the influence of CO and CO2 test errors and greatly improves the accuracy of diagnosis.In order to reduce the influence of the test environment factors of the oil chromatography data and the dispersion of the data on the calculation of the gas production rate,the wavelet multi-scale analysis was used to process the transformer CO and CO2 test data series.At scale 3,the smoothed component of the data series is extracted,that is,the trend of the data series reflecting the insulation aging process.The smooth component is used to calculate the gas production rate of CO and CO2,which can eliminate the influence of test errors on the gas production rate and improve the correct rate of insulation aging diagnosis.Given the problems of single characteristic diagnosis methods,in order to improve the accuracy of transformer insulation aging diagnosis,this article selects the oil chromatographic data CO,CO2 and CO2/CO,Rco,RCO2 and furfural in the oil,water,acid value,and many other feature parameters,then the D-S evidence theory is used to fuse the data of multiple parameters to diagnose the transformer insulation aging state comprehensively.Finally,the corresponding calculation software is developed by Matlab platform to realize the diagnosis algorithm of transformer oilpaper insulation aging.The calculation results show that compared with the diagnosis method based on single characteristic quantity,the diagnosis accuracy of transformer insulation aging based on multi-parameter fusion is improved greatly,and it is more conducive to transformer field test,and has higher engineering application value.