Assessment of Second Moment Closure Models for Modelling of Natural Convection in a Tall Cavity and Displacement Ventilation

ABSTRACT This study examines turbulence models for predicting natural convection in two scenarios: a differentially heated cavity (AR = 28.68) and a displacement‐ventilated room. For the tall cavity, the elliptic blending Reynolds stress model (EBRSM), the standard Reynolds stress model (RSM), the model, and the model are investigated. In contrast, in the displacement‐ventilated room scenario, EBRSM and are examined. For the EBRSM, the elliptic blending differential flux model (EBDFM) computes the turbulent heat flux. The RSM employs a differential flux model (DFM), while the and models use the Simple Gradient Diffusion Hypothesis (SGDH). Transient simulation is conducted for the tall cavity, whereas a steady‐state simulation is performed for displacement ventilation. The simulation results indicate that the EBRSM delivers improved predictions of both the turbulent and mean quantities near the wall in the 2D case. Both EBRSM and RSM provide a better representation of the secondary motions in the cavity, particularly at the corners. Except for the model, all models yield good predictions of the wall heat flux. In the displacement ventilation scenario, the EBRSM and EBDFM combination provides better predictions of the temperature and velocity fields, capturing sharper gradients than the model. Additionally, it predicts a stronger, more pronounced plume structure compared to the model.