Solving the hydrothermal scheduling problem considering network constraints

This paper addresses the Short-Term Hydrothermal Scheduling – STHS problem of power systems. In this problem, besides the constraints associated to hydro and thermal plants, network constraints are included due to the significant number of hydro plants located far from the load. In the STHS problem, hydro and thermal plants must be coordinated in order to supply the demand at a minimum cost and comply with a set of constraints, over a short-term horizon. Aiming to solve this problem considering the presence of nonlinearities, large number of decision variables and constraints coupled in time periods, we propose a strategy based on Lagrangian Relaxation – LR and Augmented Lagrangian – AL techniques. By using LR scheme, based on a variable splitting technique, the resulting separable dual problem is solved by a Bundle method. A primal feasible solution is obtained using the AL technique, based on the Auxiliary Problem Principle – APP approach, starting from the primal and dual solutions supplied by the LR phase. A computational model that makes use of the proposed strategy is tested on a hydrothermal configuration, whose data were extracted from the Brazilian Hydrothermal power system.

► The high accuracy in the modeling generation and transmission systems.
► A coordination with the medium problem by means of a Future Cost Function.
► The original STHS problem is split into five smaller and easy-to-solve subproblems.
► The proposed strategy to solve the resulting STHS problem.
► A pseudo-primal point is also used as a warm starting to find a feasible solution.