Hydroelectric unit commitment for power plants composed of distinct groups of generating units

Abstract

This paper proposes a new solution method for the unit commitment problem of hydro generation units for power systems in which there is prevalence of hydroelectric participation. Attention is focused on hydro power plants composed of distinct groups of generating units, each of which exhibiting different performance characteristics. Considering the forecasted load curves provided by short term operation planning studies, an optimization problem is formulated to determine how many and which generating units should be dispatched in each hour and at which power plant so as to meet load, operational and energy target constraints. The latter are established by medium term operation planning studies. The objective function to be minimized comprises distinct components related to loss of efficiency in the use of water resources, as well as unit startup and shutdown costs. The difficulties posed by the discontinuous nature of the loss of performance function, which actually comprises a collection of distinct curves, are dealt with by aggregating them into a continuous surface obtained via best fitting methods. This artifice allows the application of efficient mixed integer programming algorithms to solve the hydro unit commitment problem, and is seen as a contribution of this paper. The performance of such optimization model is assessed via its application to real power plants of the Brazilian interconnected power system. The results show that the proposed operation policy promotes the optimum use of water resources, leading to the minimum depletion of reservoirs as compared to all other unit combinations.