Accuracy of CFD Simulations on Indoor Air Ventilation: Application of Grid Convergence Index on Underfloor Air Distribution (UFAD) System Design

Abstract

Underfloor air distribution system (UFAD) mesh flow velocity was simulated using Computational Fluid Dynamics (CFD). Three mesh sizes were used to explore the domain's core x-y plane velocity contour and profiles. Compared to medium and fine, the coarse mesh underestimated the velocity significantly. A slight discrepancy occurred where the shear flow was dominant. The symmetrical flow velocity for both sides of the room length was shown in the xy-plane at the centre of the inlet. The mean error for coarse and medium mesh was larger than for the medium and fine mesh. It shows that the difference between the medium mesh and the fine was accepted. The computational time for medium mesh was acceptable for simulation, and it will not vary substantially even if the grid is refined further. The normalised mean square error (NMSE), the factor of two observations (FAC2), the factor of 1.3 observations (FAC1.3), and the fractional bias (FB) are used to measure the performance of the models and the value of the outcomes was exceptional. As a result, the accuracy of the finding can be improved by conducting additional research with manikins and in a fully occupied room under real-world conditions. In addition, this study could analyse and anticipate the optimal scenario regarding ventilation performance, etc.