Aprendizado de máquina aplicado a simulações de radiação solar em edificações

Abstract

This research aimed to investigate the application of a supervised machine learning technique to reduce updates and modifications to the parametric design model focused on solar radiation analysis in buildings. Initially, a complex geometry was created through parametric programming, aiming for the model's versatility in different situations. The analysis of incident solar radiation was conducted to understand the energy performance of the architectural design, verifying the input and output data of the parametric model to ensure the accuracy of the analyses. Subsequently, a supervised machine learning technique was proposed to improve and accelerate the solar radiation analysis, automating part of the process and reducing the time required to obtain the desired results. Finally, execution times between the proposed methods and the results were compared using new datasets to evaluate the efficiency of the machine learning technique compared to the traditional parametric programming method. The results obtained contribute to the advancement of architectural design practices with a focus on energy efficiency, providing valuable insights into the application of machine learning in optimizing parametric workflows for solar radiation analysis in buildings.