Reliability Prediction Methods Of Urban Distribution Network

With the rapid development of economy in our country,distribution network has got more and more attention.Alone with the quality of life is guaranteed,people also put forward higher expectation on the quality of power distribution network.The urban power distribution reliability has become one of the important factors to consider a city's development.Therefore,forecasting the reliability of urban power supply has important guiding significance for the operation and maintenance of the power system.Aiming at the problem of power supply reliability prediction in distribution network,some new modeling methods are proposed,which provides new ideas for accurately forecasting the reliability index of power supply in urban distribution network.The main content includes three aspects:(1)Considering that the traditional method and the statistical prediction method can not consider the correlation between the line failure rate and the duration of power outage,a method of power supply reliability prediction based on non-parametric kernel density estimation and Copula function is proposed.The method considers the correlation between line average failure rate and line average maintenance time,and obtains the edge distribution of the two by non-parametric kernel density estimation.Copula function is used to establish the correlation model of line average failure rate and average maintenance time,and then predict the future power supply reliability index.(2)In order to overcome the shortcoming of the traditional BP neural network and SVM neural network in the power system reliability forecasting.A power supply reliability prediction algorithm of distribution network based on PSO-BFO-KELM is proposed.The kernel limit learning machine model,which can avoid the linear inseparable situation of the traditional extreme learning machine when dealing with low dimensional data.The PSO-BFO algorithm is used to optimize the penalty coefficient and width of kernel function,which reduces the human influence.The simulation results show that the method is feasible.(3)Because the single prediction model has some limitations,a combination forecasting model of power supply reliability based on VC weight algorithm is proposed.According to the advantages and disadvantages of each algorithm to improve the prediction accuracy.Taking some local data as an example,the rationality and reliability of the combined algorithm are verified by Matlab simulation.This study,on the one hand,supplements the deficiencies in the reliability prediction of distribution network,on the other hand,it provides technical support for improving the prediction accuracy of power supply reliability and determining the future investment direction of power grid,which has more important academic research significance and practical application value.