تدوين يك الگوريتم هوشمند به منظور تخمين عمر باقيمانده ترانسفورمرهاي قدرت

چكيده انگليسي :

102 Development of an Intelligent Algorithm for Estimating the Remnant Life of Power Transformers Mohammad Aboohosseini m aboohosseini@ec iut ac ir Date of Submission 2012 03 03 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language FarsiSupervisor Mehdi Moallem moallem@cc iut ac irAbstractTransformers are one of the most expensive components of power systems and their sudden departure fromthe circuit will cause great damage Therefore it is essential to evaluate and estimate their remaining life Lifeof a transformer is estimated at the design stage but due to different climatic and operating conditions it isdifferent with its real life Transformer s life is proportional to its insulation life in transformers insulationsare divided into two types of solid and liquid insulations Transformer insulating oil can be refined andreplaced if it is old or degraded but cellulose insulation normally can not be replaced and repaired For thisreason the life of transformers are equal to the life of their cellulosic insulation life The cellulosicinsulations in the transformers are also divided into two types the pressboards and paper insulations Thepapers have less strength than pressboards so their life is shorter and is equal to transformer s life The bestway for estimating the remnant life of these paper insulations is to sample the insulations and to have degreeof polymerization experiment on them But it is required to turn off and open the operating transformers thatusually it is not possible For this reason many researches have been done to find the aging and cellulosedegrading effects in the transformer oil They have found furfural as the best material for describing the agingof cellulose papers Then some methods for estimating the remaining life of the cellulosic insulations basedon the oil experiments in acceleration aging have proposed but these methods because of being justexperimental are not accurate and do not have a suitable answer for real transformers For these reasons inthis thesis for estimating the transformer remnant life a combination of methods is proposed The aim of this thesis is to develop an algorithm for estimating the remnant life of power transformers based on the field experiments on the transformers Toward this goal four methods are used These methodsare as following structural artificial neural network life estimation from the top oil temperature using thetransformer health index and life estimation from the furfural concentration dissolved in oil These fourmethods are needed to perform various experiments and measurements These experiments are chosen byassessing the elements that have effects on the life of transformer insulations and experiments that are usefulfor measuring these effects Finally an algorithm that uses a fuzzy system is designed This algorithmestimates the remnant life of transformers by inserting the four methods results as its inputs The final resultsfrom this algorithm are compared with the other usual methods based on experimental works In addition these results are compared with the measured ones from the real transformers The results of this comparisonshow that the algorithm can estimate the remnant life of transformers properly except in the cases that agingis not significant Furthermore in this thesis some samples of oil and paper insulations of some powertransformers have been taken and these samples have been investigated in the chemistry labratory Keywords Transformer Estimating Life Furfural Degree of Polymerization Neural Network HealthIndex HotSpot temperature

ابوحسيني، محمد

تدوين يك الگوريتم هوشمند به منظور تخمين عمر باقيمانده ترانسفورمرهاي قدرت

تست فور فورال , درجه پليمريزاسيون , دماي نقطه داغ , شاخص سلامتي