Research On Prediction Method Of Transformer Top-oil Temperature Based On Assisting Dispatchers In Decision-making

The oil immersed transformer is one of the most important equipment in power system and it plays an important role in the conversion of electric energy.The damage of transformer will bring huge economic losses,and affect the safe and stable operation of the power grid.Most of the damages of the transformers are due to the loss of their proper insulation,while the top oil temperatures are the important factors to characterize the transformer insulation ability.The top oil temperature beyond the limit will directly restrict the load capacity and service life of the transformer especially during the peak summer and winter.Therefore,it is necessary to calculate and predict the top oil temperature in order to accurately predict the running state of the transformer and ensure the safe and stable operation of the transformer.In this paper,the calculation model and the prediction method of the top oil temperature of the transformer and the assistant dispatcher decision-making are studied in three aspects.The main contents are as follows:First of all,this paper analyzes the thermal characteristics of the oil immersed transformer deeply,describes the process of transformer oil flow and the theory of heat transfer in detail,and draws the transformer heating curve and cooling curve.Then,we introduce the common and classic method to calculate the transformer top oil temperature.Pointing out the shortcomings,we introduce the correction term which represents the air humidity and we put forward an improved calculation model of transformer top oil temperature.Under the same input conditions,the prediction results of the IEEE model and the improved model are compared.The results show that the improved model has higher prediction accuracy.Secondly,according to the actual situation of transformer online monitoring,in order to predict the transformer top oil temperature better,selecting the load current,ambient temperature,air humidity,active power,reactive power and a time before the top oil temperature as the characteristic variables,this paper establishes a prediction model of the transformer top oil temperature based on Elman neural network.The Elman network prediction model is applied to two parallel transformers and compared with the predicted values of the BP neural network model.The results show that the predicted value of the Elman network model is in better agreement with the measured value,and it has a better prediction effect on the top oil temperature.Finally,according to the actual situation of transformer fault in transformer substation in Sichuan during the peak period of power consumption,this paper discusses the fault type and common abnormal condition of transformer on-line monitoring.We select a transformer fault condition as the research object and we use the Elman neural network prediction model to assist the dispatcher.This paper analyzes the concrete process and the realization method in detail and gives the suggestion of assistant decision.