Model And Application Of Electricity Procurement And Sale Strategies For Electricity Retailers Considering Demand Response And Price Uncertainties

In recent years,with the deepening of China's power system reform,the electricity retailers have been established in various provinces.As the main participant in the electricity market,the main business of the electricity retailers is to procure electricity from the electricity market and sell it to the users.Therefore,the decision of the proportion of electricity procurement from different markets and the choice of the electricity sale scheme are the core issues in the decision of the electricity retailers.However,with the deepening of the reform,the uncertainty of the pool price increases.How to accurately assess the uncertainty of the price and make the best decision will be the key problem to be considered in the operation of the electricity retailers.In addition,in the market environment,in order to reduce the cost of electricity consumption,the users will actively adjust the use of electricity according to the demand response,which will directly affect the procure and sale of electricity retailers.This paper focuses on the uncertainty of pool price and users' demand response,and studies the decision-making of purchasing and selling electricity by the electricity retailers.The main research contents are as follows:The electricity retailers mainly procure electricity through forward contract market,pool market and self-production facilities.The pool price has the greatest impact on the decision making of the electricity retailers.In order to reduce the risk of electricity procure and sale decision,the electricity price of the future pool market must be accurately predicted.Therefore,based on the characteristics of the pool price in the electricity market,this paper chooses the radial basis function neural network based on deterministic annealing clustering to establish the prediction model of pool price.In order to improve the prediction accuracy,the hidden layer function center of the radial basis function neural network is selected by deterministic annealing clustering,and the shortcoming of the traditional K-means clustering algorithm is overcome.And through the correlation analysis,we studied the selection of network input samples.In the example analysis,the pool price forecasting based on PJM data shows that the proposed model has better prediction accuracy compared with the traditional method.In order to calculate the influence of the pool price uncertainty on the procure and sale decision of the electricity retailers,the pool electricity price predicted by the radial basis function neural network model based on the deterministic annealing clustering is used as the mean value,and the price fluctuation interval is constructed on the basis of the maximum relative forecast error.At the same time,for the forward contract market and the self-production facilities,the long-term bilateral contract electricity procure cost model with cross boundary punishment and the cost model of the three-stage production cost are set up respectively.Because of the uncertainty of pool price,it is difficult to solve the procure cost model in the pool market directly,so the robust optimization method is used to transform the model into a robust decision model.The example analysis shows that when the electricity retailers adopt a more robust purchasing strategy,more electricity should be procured from the bilateral contract market and the self-production facilities to avoid the price risk in the pool market.In the electricity market environment,the users,as a participant in the market,will be directly affected by the sale price and incentive mechanism of the electricity retailers,and the users' electricity consumption will in turn affect the procure and sale decision of the electricity retailers.In order to consider the influence of users' demand response on the procure and sale decision of electricity retailers,this paper considers the time of use and interruptible load contract,and construct the users' utility model through the consumer surplus of the users and the profit of transfer and interruption load.Further,we set up a bi-level model for the procure and sale of electricity retailers that consider the users' demand response and the uncertainty of the pool price.The upper level of the model is aimed at maximizing the profit of electricity retailers,and the lower level is aimed at the maximization of the utility of the user.The data of a typical week in the PJM market in the United States is selected to carry out an example analysis.The results show that a better decision on the procure and sale of electricity retailers can be given considering the user's demand response.