Algoritmo genético de codificação real aplicado à otimização termoeconômica

Abstract

This work proposed the use of a Real-Coded Genetic Algorithm (RCGA) in classical exer- goeconomic analysis substitution on optimization of a thermal power generation system by solid waste incineration. The optimization is performed by using payback as objective fun- cion. The RCGA is based on natural selection theory and looks for optimal solutions from randomly generated initial solutions that are improved along iterations called generations, by operators that simulates crossover, mutation and selection as observed in nature. The system is modelled based on 1st and 2nd laws of thermodynamics with implementation in Python. From a standard case with municipal solid waste (MSW) with no costs and 3.89 years of paybak, optimization by RCGA leads to a 32,90% reduction, reaching 2.61 years. In this scenario, when compared with the optimal case by classical exergy analysis, with 3.06 years of payback, the reduction is of 14.71%. The otpimization with RCGA results show that the exergetic efficiency of some equipments is decreased. Some heat exchangers had their thermal exchange areas reduced in the optimal case. Optimization takes place by the 57,60% cash flow increase (BRL/h) and 5,57% increase in total capital investment cost (BRL). The solid waste flow (m˙MSW ) increase from 525 kg/h to 700kg/h in optimal case.