Simulation-Based Optimization of Smart Building Energy Using Artificial Neural Network

Abstract

The present study aims to investigate the influence of the window size, the glazing properties and the shading overhang specifications on the energy consumption in smart buildings located in Alexandria, Egypt taking in account thermal and visual comfort. In this study, single objective and multi-objective optimizations are carried out on four objective functions, namely annual cooling, annual heating, annual lighting and annual total energy consumption using genetic algorithm. The simulations are performed using EnergyPlus through Openstudio to generate database that used to train an artificial neural network for the four objective functions. Results indicate the most significant factors are window wall ratio, glazing solar transmittance and glazing visible transmittance. Furthermore the increasing of window wall ratio and glazing solar transmittance produces an increase in cooling energy consumption and a decrease in heating energy consumption in addition window wall ratio and glazing visible transmittance have a high positive impact on lighting consumption in both cities. The multi objective optimization study is a powerful and useful tool that can save time while searching for the optimal solutions with conflicting objective functions.