Applications of decomposition techniques to unit commitment with transmission flow and voltage constraints

This thesis presents an efficient algorithm for unit commitment with transmission security and voltage constraints. The objective for incorporating these constraints in unit commitment is to ensure the feasibility of network-constrained generation schedule. The problem is a large mixed-integer programming and has a two-level hierarchical structure suited for Benders decomposition. Using decomposition, a master problem solves unit commitment with all prevailing constraints, except transmission security and voltage constraints, by augmented Lagrangian relaxation method. Given the unit commitment schedule, the subproblem minimizes transmission violations by adjusting unit generation and phase shifters. Likewise, the subproblem minimizes voltage violations by adjusting reactive power and tap-changing transformers. The reactive power subproblem is a linear program with block-angular structure which is suited for Dantzig-Wolfe decomposition. In the case of infeasibility in the subproblem, a proper constraint (Benders cut) is introduced and placed in the master problem for rescheduling the generating units. The iterative process between the master problem and subproblem provides a minimum cost solution for generation scheduling while satisfying transmission and voltage constraints. A wide range of tests on the modified IEEE-30 bus system with 9 units and a 118 bus network with 36 units are presented to demonstrate the efficiency of the proposed method. The battery storage is incorporated in the unit commitment for peak load shaving with promising results.