Method For Forming Retail Electricity Price Package Based On Electricity Price Forecast

In the future power retail market,users will have the right to select and replace sellers.In order to avoid risks and increase market share,retail electric providers need to provide scientific and reasonable power sales packages to terminal power users.At the same time,under the background of new electricity reform,the change of the market trading mode,the electricity price formation and transmission mechanism make the sales price affected by more market information.It is necessary to establish a more accurate electricity price forecasting model to provide market participants with some support,in order to grasp the market information and decision-making behavior in the electricity sales market.As the basis for the cost calculation of the retail electric provider,the sales price forecast is also the key link in the design optimization of the retail electricity price package.Based on the accuracy requirements of electricity price forecasting and the optimization requirements of retail electricity packages,the paper first proposed an Elman neural network prediction model that taken into account the timing adjustment,which was used to predict the electricity sales price.Then based on the predicted electricity sales price,optimize the design of the electricity sales package,combined with the consumer reference price decision and user stickiness.Firstly,based on the many relevant literatures reviewed,the paper analyzed and summarized the structure and mode of electricity price forecasting technology and retail electricity price package in the existing electricity market environment at home and abroad,and taken Texas's electricity market as an example compares the electricity price level in multi-type electricity market,and then the factors affecting the sales price were analyzed,the basis for the formation of the retail electricity price package was studied;Secondly,based on the analysis of the factors affecting the sales price,this paper proposed an Elman neural network price prediction model that taken into account the timing adjustment.This method used FastICA principal component analysis to extract important information about the influence of electricity price,based on the minimum intra-class distance and maximum extra-class distance for load characteristics clustering analysis.A training method that taken timing adjustment was introduced in the training algorithm of neural network to improve the accuracy of the prediction algorithm.Finally.based on the sales price forecast results and the establishment of multi-model electric energy cost accounting model of the retail electric provider,combined with the existing foreign electricity sales package structure.the new electric power reform background and the consumer reference price decision and user stickiness,the optimized design of the electricity sales packages were proposed.The method taken the total utility function of the retail electric provider maximum as the goal,based on the influence of the user reference price on the consumption decision,considering the preference behavior of the user consumption and the influence on the price acceptance degree,and then the individual user stickiness function,user group decision matrix and assimilation coefficient model that influences the objective function were established successively.The improved cross particle swarm optimization algorithm was used to store the multiple optimal package prices by using the cell array as a unit for storing multiple packages.From the perspective of effectiveness and correctness,the sales price forecasting model and the formation and optimization methods of the retail electricity price package were verified and analyzed.