Evaluation of CFD accuracy for the ventilation study of a naturally ventilated broiler house

Using the CFD model, a new ventilation system design will be found later taking into consideration the ventilation efficiency such as uniformity, stability, and suitability of environmental factors in a naturally ventilated broiler house. Because conducting a field experiment for the ventilation study presented so many difficulties, a reliable 3-dimentional computational fluid dynamics (CFD) model had to be developed to investigate the natural ventilation. Before investigating its accuracy, a wind tunnel and particle image velocimetry (PIV) test was initially conducted to find their best experimental conditions and improve the PIV accuracy(13,15). A 1/20 scale model of a naturally ventilated broiler house was used to get qualitative and quantitative airflow distribution in the broiler house using the PIV and CFD. To improve the CFD accuracy, the PIV and CFD computed airflows in the broiler house were compared, particularly on the distribution, local air velocity, and turbulent intensity in the house. The quality of the mesh density and the design of the boundary condition, especially the wind velocity and turbulence profiles, were found to be very important for getting accurate results. Assuming the PIV results were accurate, the most accurate CFD results were obtained when using a RNG k-epsilon turbulence numerical model. The average error of the CFD computed air velocity when using the RNG k-epsilon models was -6.2%.