Research On The Real-time Pricing In Smart Grid Based On Machine Learning

With the development of economy and society,the power industry runs through all areas and the demand of electricity is increasing.This brings a great challenge to the stable operation of power grid and other related works.While the rapid development of digital and smart power grid enabling a lot of historical operating data and real time data were acquired.Under the background of big data,machine learning techniques can get as much as possible implicit information of the system.The real-time pricing strategy based on demand response is one of the essential technologies,which can achieve validity and reliability of the smart grid.Locational marginal price is widely used both in the spot pricing of real-time market and in the day-ahead market.Based on the real data of U.S.PJM electrical market,analyzing user's response to the price,and realizing the game real-time pricing model of the supply and demand in the smart electrical market under the LMP background.Achieving the optimal electricity pricing strategy under the known demand——the forecast of LMP determination.Using autoencoder to train the neural network from the structure of the power system,then constructing the determination predicting model of LMP.The effectiveness of the proposed approach is examined on the data of PJM power markets in U.S.,and simulation results are compared with that of conventional back-propagation(BP)neural network and LMP of real-time markets in PJM power market.The approach is found computationally efficient.Learning the changes of the user's demand under the known price——the price elasticity of the demand.The rational analysis of the user's response to the RTP is significant for the RTP and DR strategies.Therefore,it was proposed a regression model to learn the price elasticity of electricity demand based on the demand response model.Experimental results show that compared with the content PED obtained by tradition questionnaire,the learned PED can effectively fit the demand,and the regression model can overcome the problems both in time and space and can achieve the analysis of user's demand efficiently,provide decision support for the real-time pricing strategy.Achieving the model of game price,it is proposed the pricing model by game theory based on the LMP prediction model and the price elasticity of demand obtained before,which can build a fair and effective real-time pricing strategy.This will increase consumer's satisfaction,reduce the cost of electricity,and further improve the status of power market.It also has important significance for the enterprises to win the market,reduce the cost of production,strive for adequate space to survive and develop,improve performance and core competitiveness of the smart grid and especially achieve sustainable development.