Optimizing CFD Configuration for Accurate Prediction of Wind Pressure Coefficients on Circular Arched Roof Structures

Abstract

This study evaluates the influence of Computational Fluid Dynamics (CFD) configurations, including domain size, mesh resolution, and turbulence model, on the prediction of  (mean wind pressure coefficients) on circular arched roof structures, and determines the optimal configuration for accurate and efficient simulations. A parametric study was conducted using RANS-based CFD simulations for six different arched roof geometries. The downstream, lateral and upper sides of the computational domain were varied as geometry-based distances in order to assess flow stabilization and blockage effects. Four mesh densities proportional to the structure dimensions were assessed to obtain high accuracy with low-cost computational demand. Finally, five turbulence models were compared for their ability to predict wind pressure coefficients on six circular-arched roof geometries using statistical metrics. The results indicate that domain dimensions and mesh resolution have a slight impact on the accuracy of CFD predictions for arched roof geometries. A downstream domain length of12S (where S is the greater of the total height and half the span) and a 3.45% blockage ratio provide a good compromise between accuracy and computational efficiency. A mesh resolution of 0.0125W (W is the average of the total height and half the span) based on the structure dimensions was found sufficient to meet the requirements of precision and low computational cost. Among the turbulence models, the Realizable  model produced the most accurate results across all geometries, obtaining an average of 0.9 for R2 and 0.25 for RSR, followed by the RNG  model, whereas theStandard and SST  models provide a good distribution but underestimate suction magnitude. Overall, the findings provide a guideline for efficient and accurate CFD simulations for arched roof structures, offering recommendations for domain size, mesh resolution, and turbulence model selection. These guidelines can improve the efficiency of future aerodynamic analyses of similar structural configurations.