Data-driven prediction of flow parameters in a ventilated cavity using high-fidelity CFD simulations

Abstract

In this study, we develop a machine-learning-based data-driven model, which predicts comfort-related flow parameters in a ventilated room. The model is based on the results of high-fidelity computational fluid dynamics (CFD) simulations with different geometrical configurations and boundary conditions. The developed model could be used as a cheaper alternative to CFD for applications where rapid predictions of complex flow configurations are required, such as model predictive control. Even though the developed model provides acceptable accuracy for most of the tested configurations, more input data is required to improve the model performance.