Mini-Symposium Topic: Aerodynamics II

Organizer: Robert W. Derksen (Univ. of Manitoba) and Jeffrey W. Yokota (Univ. of Alberta)

Aerodynamic Optimization of Axial Flow Fan Hub Nose Fairings

R. W. Derksen and J. R. Bender

University of Manitoba

Winnipeg, Manitoba, Canada

The purpose of this work was to improve the efficiency of medium and large scale existing industrial axial flow fans by producing a more uniform velocity distribution over the fan blades. Existing fans have been designed with the assumption that the flow at the blade inlet plane is uniform, which is clearly not true for most axial flow fan designs. The improved velocity distribution was obtained from the development of an aerodynamically optimized hub nose fairing.

The procedure for optimizing the nose fairing uses existing aerodynamic optimization methods programmed into a completely self contained FORTRAN program. The type of optimization algorithm used here is the use of design optimization to solve an inverse design problem. The different modules of the program include a surface vorticity panel method flow solver, a Bezier curve surface definition routine and a minimization method. Three different minimization methods were tested to determine the most appropriate one, this being the downhill simplex method in multidimensions. Many different sized fans and nose fairings were tested, with two different types of optimized fairings being discovered. Short inlet fairings typically make use of a very blunt inlet hub with a slight hump or rise above the hub radius. Longer inlet fairings make use of a more curved inlet cone with no hump. It was also shown that the hub-to- tip ratio of the fan has a significant impact on the velocity distribution and therefore the efficiency of the fan unit. As well, it was shown that only a small change in the length of the hub nose faring could affect the velocity distribution. It was also found from this work that the relatively simple methods used can provide an adequate modeling of the problem and a reasonable solution.


Saturday, 11:50 a.m. - 12:10 p.m. Room 2270