Research On Load Forecasting Of Active Distribution System Base On The Effect Of Demand Response And Electricity Market

With the increasing integration of distributed generation and new load such as electric vehicle on the consumer side,distribution network is developing from traditional distribution network to the smart distribution network and active distribution network.Under the development trend of power sale market gradually opening,guided by market incentives,the consumer will be using the generalized demand response resources to participate in the market,making the consumer power consumption mode diversified,and the consumer load regularity changing.Therefore,it is necessary to study theory and method of active distribution network load forecasting base on the effect of demand response and power sale market.This paper studied the composition architecture and operating system of the active distribution system.Based on the analysis of characteristics of various generalized demand response resources,and combining the features such as multi-subject and mode diversification of future power sale market,the three layers architecture and operation organization system of active distribution system contains power sale market,aggregation agents,response consumers are established,and the function,profit model and synergetic mechanism of main entities body in the system are analyzed and designed.The load characteristics and influence factors are studied.The load characteristics of the active distribution system and polymerization agent are analyzed from indicators such as the load peak valley difference,cyclical volatility.In the new operation environment,the response electricity price,the response signal and the user response model become the influence factors of active distribution load.The mechanism,classification definition and performance characteristics of different influence factors are analyzed to provide basis for active power load forecasting.The active distribution network short-term load forecasting considering dynamic electricity price and demand response each other are presented.On the basis of the elastic response matrix which can accurately identify response model of consumers,the responsive elastic matrix based load forecasting model,which is optimized by genetic algorithm,is proposed to forecast load of active distribution systems.This presented method and the load forecasting model is applied in an example distribution system,the results demonstrate the effectiveness of the method and the model.In the end,this paper analyzes several performance indexes which may have influence on forecasting accuracy,and presents parameter selection methods for high accurate forecasting.In view of the active distribution network operators can obtain user response plan information by security checking of market transactions,this paper proposes a combination forecast method considering demand response,Which by applying stabilization with the response signal and wavelet decomposition process for active distribution system load,getting low correlation load part based seasonal and the load dominated by demand response,and using time series analysis model and support vector regression model to forecast the two parts,respectively.By the example calculation,compared with the traditional load forecasting methods,the proposed model is effectiveness and has advantages in deal with strong volatility of the active distribution network load.