Redes neurais artificiais recorrentes: uma aplicação à otimização da operação de sistemas de geração.

Abstract

This work dcals with the study of artificial neural networks(ANN) to solve optmization problems and their applications in the mid-tcrm operation planning of hydrothermal gencration systcms. The operation planning problem deals with economic power dispatches, that is, with the scheduling of hydro and thermal power plants that minimizes the overall production cost while satisfies the load demmand. The study of ANN as optmization tools for solving large scale problems was motivated by the necessity to being up to date with the state of the art of this new technology. This technique has a great potential for hardware VLSI implementation, in which could be more efficient then traditional optimization techniques. The operation planning of hydrothermal generation systems is a large scale problem, whose complexity increases as the planning horizon increases and the aceuracy of the system modeling increases. Hence, to solve such a large problem an efficient optimization technique is always necessary. This work considers recurrents ANN to solve linear and quadratic programming problems. These networks are based on the solution of a set of differential equations that are obtained from a transformation of a Lagrangian energy function. This network also provides the corresponding Lagrange multiplier associated with each constraint, which is the marginal price. The ANN was applied to solve the economic power dispatches of the CHESF/ELETRONORTE interconnected system to which was calculated the optimized solution, the marginal costs and the water values associated with each hydro plant