Trilayer Stackelberg Game Approach for Robustly Power Management in Community Grids

Multiuncertainties in community grids bring enormous challenges to system scheduling. To reply to renewable uncertainties on the prosumer side, this article proposes a trilayer stackelberg game (SG) based robust power management method for a typical community grid. Firstly, in the upper-layer, the supplier acts as a leader to determine the trading prices according to prosumers' power demands in the normal scenario. The scheduling of prosumers before and after uncertainties is formulated as an inner-loop SG model in the middle- and lower-layers, and a linear robust counterpart model is developed to address the nonlinearity and nonconvexity of traditional robust models, thereby ensuring the convergence of the robust equilibrium solution. Second, to solve the trilayer SG model effectively, the existence and uniqueness of the equilibrium solution are proved based on a multistep backward induction method. Third, a novel two-stage distributed iterative algorithm is further exploited to avoid the nested iterations in existing solution methods. Furthermore, the possible oscillation in iterative optimization is also prevented to improve the computational efficiency. Case studies illustrate the effectiveness of the proposed trilayer SG model, and the developed distributed algorithm requires fewer iterations and less solution time comparing to other algorithms. Besides, this method shows better scalability and practicality with the expansion of the grid-scale.