PLANEJAMENTO DE MÉDIO PRAZO DA OPERAÇÃO HIDROTÉRMICA FORMULADO COMO UM PROBLEMA DE OTIMIZAÇÃO ESTOCÁSTICA

Abstract

A central aspect in the medium-term hydrothermal system operation planning is the handling of uncertainties associated with stochastic water inflows at the system reservoirs. In this paper, a two-stage stochastic optimization approach is adopted for handling such uncertainties. Instead of generating a set of synthetic time series, which must be able to represent both time and space complex coupling relations of the plants, the approach proposed works directly on the historical records of water inflows, known as scenario fans, which inherently represent such relations. The first stage of the model proposed is concerned with the calculation of control actions in the first period of the problem (i.e. here-and-now decisions), while the control actions in the remaining periods are calculated for each scenario realization (i.e. wait-and-see decisions). The hydraulic production function of each individual unit in a hydro plant is formulated considering a linear function. The generation of thermal units are integrated into an equivalent plant having a quadratic cost curve by means of an optimization model also proposed. The multi-stage characteristic of the problem may be explored in the proposed approach if a rolling window solution structure is adopted. The results show that the model proposed is able to represent the uncertainties in water inflows, which are adequately captured by the scenario fans technique.