Abstract: In the smart grid environment, load prediction is introduced based on the real-time price Stackelberg game model. In order to match the real-time load and the predicted load, the master-slave game model between the seller and the user is designed. By establishing the real-time pricing mechanism under load forecast update, the optimal real-time power price and optimal power consumption behavior of both sides are obtained. By integrating the load time series under the equalization state of the day-ahead real-time electricity price mechanism into the power supplier power price weight time series vector, the further optimized load-balancing time series under the day-ahead real-time electricity price mechanism is obtained, finally forming a closed-loop that is constantly advancing and optimizing. At the same time, the matching degree evaluation index and judgment criterion of real-time load and predictive load sequence are presented. Through numerical simulation analysis and comparison with the unoptimized real-time pricing mechanism, it is found that the proposed real-time pricing mechanism based on load prediction and update can significantly reduce the power consumption cost of power users while improving the operation efficiency of the power grid.

Key words: smart grid, real-time pricing, load forecasting, Stackelberg game