Research On The Prediction Method Of Pollution Flashover Of The Insulator In The Contact Network

In recent years,with the development of Chinese economy and society,the reliability of power supply system for overhead contact system of electrified railway is required more and more highly.However,various industrial and agricultural pollutants in our country cause the increasing severity of the atmospheric pollution,and fouling rate of insulator surface also rises year by year.As a result,the accidents of catenary pollution flashover occur frequently,which seriously threatens the safety of railway transportation and to some extent caused large economic loss.Thus it can be seen that it is necessary to further research the prediction methods of catenary pollution flashover.The two necessary conditions of pollution flashover occurring in the running insulator are the enough degree of pollution accumulation and the suitable meteorological conditions.According to that,a forecasting model of insulator's pollution grade is put forward based on artificial shoal-BP neural network,and on the basis of this model builds probability prediction methods of pollution flash over of catenary insulator and lines,which can not only timely grasp the pollution degree of the insulator to scientifically arrange the cleaning work,but also offer guidance basis to the establishment of effective mechanism of anti-pollution flashover.Firstly,the influencing process of meteorological factors was analyzed such as rainfall,wind speed,relative humidity,and dust-fall volume on the pollution degree of catenary insulator,then introduces the theoretical basis of the BP neural network and artificial fish algorithm,and the based forecast model of insulator's pollution grade was put forward based on artificial shoal-BP neural network.Selecting the four representative meteorological factors of monthly precipitation,monthly average wind speed,monthly average relative humidity and monthly dust-fall volume,and shape coefficient,running time(monthly)of insulator as inputting characteristic quantity,the data of six characteristic quantity last month was input to predict insulator pollution level of the current month.Afterwards the thesis builds probability forecasting model of pollution flashover based on insulator pollution level,and inputs salt density corresponding to the predicted pollution level and forecasts data of severe weather in current month into the probability prediction model of pollution flashover,so as to predict the pollution flashover probability of catenary insulator and lines in current month,then implements graded early warning of the predicted results,and establishes the responding short-term mechanism of pollution flashover forecast to prevent pollution flashover accidents as far as possible.Finally the validity and the correctness of forecasting model was verified through the sample calculation.The verification result shows that the methods proposed in the thesis can achieve effective prediction of pollution flashover of insulator and catenary lines,which play a role of auxiliary reference in evaluation of catenary line reliability and anti-pollution flashover work,to a certain extent,to avoid the waste of manpower,material and financial resources.