AI-Based Generative Geometrical Design of Concentrated Solar Thermal Tower Receivers

Abstract

An artificial intelligence (AI) aided generative design workflow for the optimization of cavity receivers for concentrated solar thermal (CST) energy systems is presented. The workflow integrates the Non-dominated Sorting Genetic Algorithm III (NSGA-III) with generative design methodologies and optical evaluation through Monte-Carlo ray-tracing in an interoperable way, to optically optimize the geometry of cavity receivers according to a set of objective functions for a given heliostat field. As a demonstrator test case, the workflow is used to provide an optimal geometrical design of a cavity receiver given the Cyprus Institute’s PROTEAS heliostat field. It is shown that the workflow is able to generate unconventional, non-intuitive and efficient receiver designs in an automated manner, which are often not conceived by traditional design approaches.