Distributed and asynchronous unit commitment and economic dispatch

The drastic increase of uncertainties associated with intermittent renewables create major challenges in power system operations. Current centralized unit commitment and economic dispatch (UCED) used by Independent System Operators (ISOs) may not efficiently handle uncertainties in the future as penetration levels of renewables increase. In this paper, to overcome this difficulty, a distributed and asynchronous framework is presented whereby the process of solving unit subproblems is delegated to unit solvers. To investigate the crucial coordination aspect, simplified deterministic UCED is formulated and price-based coordination is performed by asynchronously updating prices. The difficulty of the approach is that multipliers are updated using subproblem solutions obtained using multipliers of different vantages and divergence may occur. To overcome this difficulty, a novel algorithm is developed under an assumption that each subproblem is solved at least once within a finite number of coordinator iterations. This assumption together with diminishing stepsizes ensures that multipliers get closer and closer to the optimum with a certain periodicity and the overall coordination approach converges to the optimum. To test the coordination aspect, the deterministic UCED problem with 1000 units without transmission capacity constraints is considered, and results indicate that the method is robust and fast. To further explore the scalability of asynchronous coordination, a simplified UCED problem with 10,000 units each with generation capacity constraints only indicates that the method converges, is robust and more efficient as compared to the alternate direction method of multipliers (ADMM).