Research On Model Optimization Of Distribution Network Reliability Forecast

The distribution network is a bridge connecting the transmission system and users,and plays an important role in the distribution of electrical energy in the power network.The reliability of the distribution network is becoming increasingly important in the planning and operation of power systems and management services.At present,there is still little research on the reliability prediction of distribution networks at home and abroad.Studying the reliability prediction of distribution network can get the development trend of reliability,and provide a reliable basis for grid companies to improve the structure of distribution network,formulate reliable electricity prices and investment plans.This paper first uses the time series method to make predictions,and selects the system average power supply availability rate as the prediction target from the index system.The average power supply availability index of the system is monotonically increasing and the upper limit of the index value is 1,in order to establish a gray prediction model based on data transformation to predict the reliability of the distribution network.First,data transformation is performed on the reliability index of the distribution network.Then,the transformed data is input into the gray model to obtain the predicted value.Finally,inversely transform the obtained predicted value to obtain the true predicted value of the average power supply availability of the system.Analysis of the statistical data of a domestic distribution network:The gray prediction model has a mean square error of 0.00277 for the distribution network reliability prediction,and solves the problem that the index used in the gray model prediction value will exceed the upper limit 1.The reliability of the distribution network is mainly determined by factors such as the network structure,power equipment,and operating characteristics of the distribution network.In order to obtain more accurate prediction results,it is necessary to screen out the key factors from the reliability influencing factors.Firstly,the factors affecting the reliability of the distribution network are identified by the grey correlation analysis method,and then the factors are simplified by the correlation coefficient matrix analysis method to obtain the key influencing factors of the reliability indicators of the distribution network.Based on the analysis of influencing factors,a neural network prediction method for distribution network reliability under random failure is proposed,and the model is optimized by improved particle swarm optimization.Establish a generalized regression neural network prediction model for distribution network reliability optimized by improved particle swarm optimization algorithm,using key influencing factors of distribution network reliability as model input,and the number of users under random failure as power outage as model output.The parameters of the neural network model are optimized,and the optimized neural network can realize the distribution network reliability index prediction under random failure.Analysis through calculation examples:under random failure,the maximum error percentage predicted by the model is 6.8%,and the average error percentage is 4.2%,which has a good prediction effect.Figure[16]table[10]reference[73]...