Enhancing building sustainability through aerodynamic shading devices: an integrated design methodology using finite element analysis and optimized neural networks

In the quest for sustainable building solutions, attention has increasingly turned to innovative structural designs and technologies that minimize environmental impact. Within this context, this study explores the integration of advanced sun-breakers and aerodynamic features into building design. Recent developments in shading devices are analysed to optimize the use of airflow in aerodynamic performance. Due to the computational intensity of Finite Element Analysis (FEA), an Artificial Neural Network (ANN) based on Stochastic Gradient Descent (ANN-SGD) is employed as a substitute for such limitations. A predictive model was developed to estimate fluid flow fluent full iteration range variables which included wind velocity, static pressure, air density, and turbulent kinetic energy. These predictions are initially validated using a first short run of workbench simulations then the ANN-SGD model is applied to expedite the computational process. The adopted ANN-SGD model exhibited high regression accuracy that was evidenced by a Root Mean Square Error (RMSE) of 4.66% while the Coefficient of Determination (R²) had scored a highest score of 0.995 in addition to the Coefficient of Variation of the RMSE (CVRMSE) scoring 8.33%. The findings of this work offer future development perspectives and strategies in building sustainability and aerodynamical shading devices’ design.