Comparing of CFD contours using image analysing method: a study on velocity distributions

Abstract

Contour plotting, a widely utilized graphical technique for visualizing CFD (Computational Fluid Dynamics) outcomes, is highly valuable. It provides an effective and practical approach to analysing distributions of magnitudes belonging to fluid domains such as; velocity, temperature, pressure, volume fraction, etc. Nevertheless, when analysing multiple contours, especially showing similar distribution, identifying the ideal contour can be difficult and open to speculation. In this research, the issue was addressed by employing the Image Analysis Method for the classification of velocity distribution contours. This led to determining which picture has the best distribution among a few of the contour’s pictures. Firstly, velocity distribution contours downstream of the diffuser located in Air Handling Unit (AHU) unit were obtained by using CFD. The contour pictures were then transferred to MATLAB environment. With pixel analysis in MATLAB, the pictures were able to be classified based on which parameters had an effect on the velocity distribution. Variable parameters are the length of the fan channel (x) and the ratio of cross-sectional areas of the AHU (A/Ao). The results showed that x=250 mm and A/Ao=0.5 improved velocity distributions by 6% and 20%, respectively.