Distributed Market Clearing with Wind Generation and Large-Scale Dispatchable Loads

Risk-cognizant power dispatch techniques are urgently needed towards achieving the goal of high-penetration renewables in the future power smart grids. In this paper, day-ahead stochastic market clearing based on the DC optimal power flow (OPF) model is pursued accounting for the stochastic availability of renewables. The objective is to minimize the grid-wide total cost which consists of the conventional generation cost, the end-users' utility, as well as the energy transaction cost utilizing the conditional value-at-risk (CVaR). The proposed CVaR-based transaction cost serves as a smart regularizer to mitigate the potentially high risk of inadequate wind power. The sample average approximation method is introduced to bypass the prohibitive high-dimensional integral in the resulting optimization problem. Furthermore, to address the challenges of respecting end-users' privacy and the computational complexity incurred by large-scale dispatchable loads, a fast ADMM-based solver is developed with guaranteed convergence. Numerical results are reported to corroborate the merits of the novel framework and the proposed approaches.