Anasis And Decision For Electricity Retail Companies Purchasing Behavior Under Electricity Market Environment

This subject is one of the main research contents of the project “Development of Power Demand Side Management Platform” of Shanxi Fengxing Measurement and Control Co.,Ltd.The release of the “No.9 Document” of the electricity mark marks the official opening of retail electricity market reform,as the background,a large number of electricity retailer companies have been established nationwide.While meeting the opportunities brought by the reform,the electricity retailer companies are also facing various risks and challenges.Therefore,the analysis and decision for retailers purchasing behavior under the electricity market environment is a new subject to be studied.On the basis of the overall analysis of the market trading mechanism and operation mode of the electricity retailer companies,this paper proposes a multimodel forecasting method for the electricity consumption of the retailers' customers based on the analysis of the electricity consumption mode.Based on the current electricity reform policy and market trading scheme,proposed an optimization model for the electricity purchasing strategy of the electricity retailer companies in a multi-period and multi-market environment.The main research contents are as follows:(1)Firstly,it introduces the reform process of electricity marketization and various market trading mechanisms,including medium and long-term trading mechanism,spot trading mechanism and market convergence mechanism.Secondly,it summarizes the types and operating modes of existing electricity retailer companies.Based on this,the paper analyzes the market risks and analytical methods faced by retailers as a new market entity.(2)Propose a multi-model forecasting method of electricity consumption for electricity retailer companies.This method uses data mining as a technical means to analyze the traditional FCM algorithm and point out an improved strategy to adaptively determine the required parameters and achieve optimal clustering results.Similarity calculation method of users is used to judge the user's electricity consumption mode.According to the changing trend of different electricity consumption modes,such as non-stationary random fluctuation,periodic fluctuation and trend change,forecasting models are proposed.The validity of this method is verified by an example analysis,which avoids the problem of low forecasting accuracy caused by the difference of electricity consumption behavior.The forecasting results lay a good foundation for the electricity purchasing strategy of electricity retailer companies in the medium and long-term trading mechanism.(3)As the current electricity market construction in China is in the initial stage of the monthly and other medium and long-term electricity trading,based on the existing mechanism,considering bilateral negotiation,centralized bidding,contract transfer and biased electricity quantity assessment,an optimal decisionmaking model of electricity purchasing strategy is constructed,which aims at maximizing the profit of electricity retailer companies.An Adaptive Bacterial Foraging Optimization-Particle Swarm Optimization(ABFO-PSO)algorithm is proposed to introduced the adaptive step size bacterial foraging process into the traditional Particle Swarm Optimization algorithm,which improves the local search accuracy of the algorithm.The analysis of the example shows that considering the types of transaction comprehensively and rationally distributing the proportion of electricity purchased in different markets are beneficial for the electricity retailer companies to optimize revenue and reduce deviation assessment.(4)In view of the fact that there is no mature reference experience for the construction mechanism of the spot market in other foreign countries,ensuring the healthy and stable operation of the spot trading market is undoubtedly the key issue facing the round of electricity reform.Starting from the operation mode and risk analysis of the electricity retailer companies,the electricity purchase model is constructed including the spot market transaction including medium and longterm electricity decomposition,the distributed electricity purchase transaction and the adjustable load transaction,which takes full account of uncertainties such as user-side load,distributed generation and spot market price,takes Conditional Value at Risk(CVaR)as a risk measurement method,and takes the profit and customer satisfaction as the objective of dynamic electricity purchase optimization decision-making model.Among them,the Latin Hypercube Sampling,Cholesky decomposition and scene reduction techniques are used to describe the uncertainties,which fully cover the sample search space and avoid the explosion of scene dimension.An Multi Objective Self-Adaptive Differential Evolution(MOSADE)is used to solve the problem,which improves the evolution direction of the algorithm and achieves the preservation and ordering of nondominant solutions.The expected transaction volume of electricity from different purchasing sources is obtained,which is helpful for the retailers to better adapt to the new market environment and win the market initiative.