Geometric optimization of linear Fresnel reflector solar collector using a genetic algorithm

Abstract

In this work, the optimization of a Linear Fresnel Reflector solar concentrator through a parallelized Multi-offspring Genetic Algorithm is presented. An analytical description of the concentrator was developed and used to compute specific annual collected energy; this is the fitness function to be maximized. Selection, reproduction, and mutation operators were defined to produce diversity of individuals. Elitism was applied to the best subject of the population, so that it is kept in the following generations. The results show a tendency towards systems that privilege larger gaps between heliostats, as this reduces shading and blocking effects. Additionally, concentration was found to be at the lower bound of search, the receiver tends to be positioned as high as possible and mirrors to be as short as possible. So, for these constraints we found an optimal fitness of 1238.35 kWh/m2 for Évora.