Research On Regulation Strategy Of Demand Response Resources For Load Aggregator

Under the circumstances of electricity marketization,the load aggregator(LA)plays a coordinating role between the independent system operator and retail customers.LAs can evaluate the regulatory value of small or medium users based on complete technical methods to achieve an efficient integration of discrete demand response resources,which brings positive effect to the promotion of intelligent energy use,even the sustainable development of power industry and the society.As long as the LA participates in the competition on behalf of retail customers,both market volatility and user-side uncertainties have a negative impact on the realization of the profit targets.The scientific design of demand response mechanism is the basis for the business operations and the risk control.China is still in the exploration period of power system reform.It is of great significance to take the LA as the research object and further explore the regulation strategy of demand response resources under the market structure.In general,this paper lays emphasis on the approach to consider various factors comprehensively,including users’ power consumption behavior,willingness to participate,demand response capacity,incentive cost and the changes of electricity market price.Afterwards,differentiated incentive policies that adapt to each customer group are formulated,aiming at reducing market risks faced by the LA and maximizing its overall benefits.The main work includes the following aspects:(1)Starting with the relavant concepts of the LA,its business scope and the basic characteristics of various types of resources are introduced.Then,typical scenarios to implement demand response are analyzed and compared from the perspectives of bidding in the main energy market,bidding in the ancillary service market and dealing with the fluctuations of electricity price in the energy market.What’s more,fundamental elements that influence the LA’s decisions are summarized,including the user-side electricity price mechanism,the forecasting of electricity price and the calculation of customer baseline load.(2)A method for creating the residential users’ power consumption portraits is proposed,which involves three dimensions: the basic attribute,the power consumption behavior characteristics and the willingness to participate.A quantitative index reflecting typical characteristics of the load curve is established.As well,details of each index’s calculation formula and weight allocation method of are given.Then,the process of selecting target users for demand response project based on TOPSIS method is proposed further.The comprehensive evaluation index of sample users in a certain area is obtained from the example analysis.Ulteriorly,the principles and calculation steps of the iterative self-organizing data analysis algorithm(ISODATA)are introduced.Based on ISODATA,the baseline load of users selected by the comprehensive evaluation results are clustered into different types.(3)Taking user-side uncertainty and different comfort requirements into account,with economic benefits and comfort effects considered,the stochastic programming theory is used to establish the lower rational user response model.Then,an upper load aggregator decisionmaking model with the goal of maximizing comprehensive benefits is proposed.Variables are passed between the upper and lower layers to complete information interaction,forming a bilevel optimization model of LA day-ahead demand response strategy coping with the fluctuation of market electricity price.Afterwards,the calculation formula and iterative process of the differential evolution(DE)algorithm are given.Through the simulation of two cases,behavior distinction between different types of residential users in the demand response project is analyzed.Based on DE algorithm,each type of users’ optimal compensation standard is calculated.At the same time,the benefit of LA before and after demand response project are compared.Finally,the stability of the model is verified by the sensitivity analysis as the key parameters change.