Programação diária da operação de sistemas termoelétricos de geração utilizando otimização bio-inspirada em colônia de formigas

Abstract

The daily schedule of thermoelectric systems consists of determining the strategy to set the generation units to be put in operation to meet the load, meeting also the operational and functional constraints of the respective power system. This problem can be split into two subproblems: (i) schedule of units that must operate in accordance with a given load, or Thermal Unit Commitment and (ii) set the power generation for each committed unit, or Economical Schedule. Due to load variations the schedule involves hourly generation decisions, in a horizon that varies from one day to two weeks. Researches related to bio-inspired optimization strategies applied to the daily thermal system operation show that the combination between bio-inspired computing techniques and other optimization methods has an important role in order to obtain better solutions in a shorter computing time. Following this, the present work makes use of a methodology based on Ant Colony Optimization to minimize the costs of the thermal system daily scheduling. This proposed method uses a Sensitivity Matrix (SM) based on information from Lagrange Multipliers related to the problem in order to improve the bio-inspired process. In this way, a percentage of the individuals make use of the provided information in the colony evolution process. The results obtained through those simulations indicate that the use of the SM presents better quality solutions with a reduced number of individuals.