PERFORMANCE ANALYSIS OF A THERMAL SYSTEM OPTIMIZATION BY USE STOCHASTIC METHODS
Leonardo Rodrigues de Araujo1; João Luiz Marcon Donatelli2; José Joaquim Conceição Soares Santos2; Edmar Alino3
1 IFES; 2 UFES; 3 ITA
doi:10.20906/CPS/COB-2015-0925
Resumo
Thermal systems are essential in facilities such as thermoelectric plants, cogeneration plants, refrigeration systems and air conditioning, among others, in which much of the energy consumed by humanity is processed. In a world with finite natural sources of fuels and growing energy demand, issues related with thermal systems design, such as cost estimative, design complexity, environmental protection and optimization are becoming increasingly important. Therefore the needs to understand the mechanisms that degrade the energy, improve the energy sources use, reduce environmental impacts and also reduce project, operation and maintenance costs. In recent years have occurred consistent developments of procedures and techniques for computational design of thermal systems. And in this context, the fundamental objective of this study is a performance comparative analysis of structural and parametric optimization of a cogeneration system by uses of three stochastic methods (genetic algorithm, simulated annealing and particle swarm). This research work uses a superstructure, modeled in a process simulator (IPSEpro of SimTech), in which the appropriate alternatives options to design are included. Accordingly, the cogeneration system optimal configuration is determined as a consequence of the optimization process, restricted within the configuration options included in the superstructure. The optimization routines are written in the "MSExcel - Visual Basic" to work perfectly coupled with the simulator process. At the end of the optimization process, the system optimal configuration, given the characteristics of each specific problem, should be defined.
Palavras-chave: Stochastic Methods; Genetic Algorithm; Simulated Annealing; Particle Swarm; Thermal System Optimization