Multi-objective optimization and simulation for multiple models of walls to estimate heat gain using Artificial Neural Network model

Abstract

In homes all around the world and other hot climates, different types of walls are frequently used to provide small heat gains from the outside. Climate-related factors have been the subject of several studies, but there has been relatively little quantitative study on the best wall types and thicknesses. In this study, an effort was made to determine how much heat would be gained by a different form of wall. The HAP program allows us to change the parameters flexibly and easily calculate the heat gain. We used this advantage to collect data for our analysis, 432 models have been produced for three distinct wall types—light, medium, and heavy—made of various materials and thicknesses using the HAP software, where the direction of the walls and the color of their outer surface were considered. Using a function from a neural network that was built later, the optimal wall model was determined by the multi-objective Genetic algorithm method. The results showed that for the optimal wall, which were 9.548, 9.598, and 9.62 W/m², heat gain for ANN, HAP, and transient thermal analysis by Ansys, respectively, showed a close agreement among the methods.