Barbara Lipska
Bibliographic info:
The International Journal of Ventilation, Vol. 7 N°3, December 2008

CFD modelling can be used in the design of ventilation systems. However, simplifications in room airflow models may lead to errors and discrepancies between predictions and reality. The aim of this paper is to present the problems and errors connected with the prediction of the air distribution based on CFD codes. The sources of error are pointed out, as well as possibilities for eliminating or reducing them based on program options and experimental identification of the predicted flows. As an example, a real complex room was considered. The investigations consisted of a test room with heat and contaminant sources and a local exhaust. Mixing ventilation using a ceiling square cone diffuser and a displacement system with laminar diffusers were applied. Numerical calculations were carried out using Flovent 6.1. The correctness of the predictions of the airflow pattern and contaminant propagation was assessed by comparison with visualization of actual flow. The predicted profiles of the parameters of air and tracer gas were compared with the results of measurements. Discrepancies were found and attempts were made to reduce them. It was found that some of these discrepancies could be eliminated by applying respectively the options which are available in the program, including an adequately selected discretization grid. Unfortunately, however, there were also problems which were difficult to overcome. These include the impossibility of the simulation to exactly represent the complicated construction of the diffusers and the characteristics of the jet leaving them.