A two-level decomposition algorithm for robust unit commitment in large-scale power systems with variable wind generation

As wind generation integration in power systems has become more common, the two-stage robust unit commitment (TSR-UC) problem, which accounts for wind power uncertainty, has emerged as a critical research area. For large-scale power systems, the TSR-UC has a large computational burden and difficulty obtaining an optimal solution in an acceptable amount of time. To address this problem, we propose a two-level decomposition algorithm for TSR-UC to reduce the complexity of the model from the time dimension. In the first level, the column and constraint generation (C&CG) algorithm is used to split the original problem into a master problem and a subproblem. In the second level, the entire scheduling horizon is partitioned into several consecutive intervals, and the C&CG master problem is further decomposed into smaller-scale problems by replicating the variables in the last period of each interval and introducing on/off time counters. Similarly, the C&CG subproblem is decomposed into smaller-scale subproblems by introducing an uncertainty subset budget. Moreover, the analytical-target-cascading algorithm, coupled with an effective initialization strategy, is proposed to solve the decomposed C&CG master problems/C&CG subproblems in parallel. Finally, numerical experiments across three different scales of power systems demonstrated that the proposed algorithm significantly enhanced solution speed while maintaining solution quality compared with the C&CG algorithm.