The short-term operation planning problem aims to define the generation units should operate at minimum cost to fully supply the demand while meeting the system constraints over the study horizon. The problem is of a non-linear, integer-mixed, large scale optimization type, whose solution is a rather challenging task. This paper uses the Lagrangian Relaxation to decompose the original problem into simpler subproblems to be solved in sequence. In this decomposition, the subproblems are constructed taking advantage of their specific characteristics. One of the resulting subproblems of the relaxation scheme refers to the commitment of hydroelectric units. In order to solve this problem, an algorithm for exhaustive enumeration is proposed. This algorithm makes use of the Augmented Lagrangean method, which solves the resulting constrained nonlinear subproblems for each examined state. A Quasi-Newton method is used for the solution of the unconstrained subproblems resulting from the application of the Augmented Langrangean method. The proposed model is applied to two hydroelectric units of the Brazilian power system, demonstrating its consistency and practical feasibility.
Power System Operation Programming; Unit Dispatch; Lagrangian Relaxation; Augmented Lagrangian
