Improvement of fan airflow characteristic by modification to blade geometry (CFD) / Azmir Mohd Isa

This research focused on the aerodynamic efficiency of automotive cooling fan via two-dimensional CFD simulation. The subject of this research was to study blade geometry design as a function of flow separation behavior and vortex formation. By doing blade design modifications in conjunction with CFD forecast results, improvement in boundary layer characteristic was achieved within the CFD realm. Irregular or unequal spacing of blades in subsonic turbomachinery has been observed in the field as well as in the literature to possess superior noise quality and this irregular spacing was modeled into the CFD domain. Flow vectors were studied from baseline blade geometry and improvement in flow vectors were achieved by modifying the blade geometry. For accurate and reliable flow vector behavior, the two-dimensional cascade model was pre-conditioned with realistic boundary conditions based on cited references and observed data from vehicle measurements in the industry. A more expensive turbulence model and very refined grid size near the wall was employed in order to tackle complex flow phenomena such as flow separation and vortex formation. Choice of turbulence model was based on cited references. Grid independence and convergence criteria were based on stability of monitored residuals, which necessitated very low residual values. A technique for obtaining CFD fan characteristic was established by extracting sufficient CFD data points from base fan model as a function of ram air and correlating with measured data of actual base fan. Blade airflow behavior near the surface revealed an airflow improvement near the trailing edge, instead of near the leading edge as in the case of a patent by GM (Razinsky et al, 1987). Roughness near trailing edge had reduced the magnitude of the reversing flow from trailing edge. This CFD modeling technique has a potential for robust usage in vehicle architecture of an automotive design industry.