A statistical approach combined with numerical simulations for ventilation problems

We propose a method for optimizing analyses of ventilation solutions using CFD. The method combines the CFD with the Taguchi method in an attempt to reduce calculation costs and analytical computation limitations when performing CFD posed by multiple factors. It thereby reduces the need for a large number of Computational Fluid Dynamics (CFD) simulations when various factors at various levels are considered in the model. Moreover, the underlying mathematical relations between the investigated parameters and desired objects do not need to be known. The outcome of the method is an optimal solution. We examined this method by exploring two specific scenarios. One scenario aimed at identifying the best ventilation solution for the given set of input parameters. The other one focused on reducing exposure to infectious agents in an environment with displacement ventilation. The first scenario presents a technique that combines the Taguchi method, ANOVA, and Grey Relational Analysis (GRA) to achieve the optimum solution according to the investigated solution space. Through a two-dimensional benchmark test case, we modeled draft risk and the mean age of air with five independent variables. Using an L25 orthogonal array, we reduced the number of cases from 3,125 to 25, then applied CFD, Taguchi, and ANOVA analyses to identify the best solutions for draft risk and the age of air. GRA identified the best joint solution for both draft and age of air. The CFD simulation of the best solution confirmed its superiority, demonstrating the credibility of the proposed method. In the second scenario, we extended the application of the proposed method to study pathogen concentration when a displacement ventilation system is used. Analyzing factors such as room dimensions, inlet and outlet locations, velocities, and temperatures, we used the Taguchi method to manage complexity and reduce the cases from 19,683 to 27 combinations. Inlet velocity emerged as the most influential factor in minimizing pathogen concentration; room volume had a very small influence on the concentration. The optimal solution was modeled using CFD and validated. Results were subsequently compared with the results of the Wells-Riley model, reaffirming the importance of inlet velocity.