Influence of the enclosure on the performance of radial fans

Due to restrictions in the space that is available for mounting fans, a fan sometimes has to work in a narrowed enclosure. This leads to reductions in the fan performance and an increase in fan noise. A common arrangement is a radial impeller confined in a square housing. The flow patterns in the enclosure and their influence on the fan performance are investigated. For this investigation the complex three dimensional transient flow is described using a simplified two dimensional stationary model, reducing the complexity of the problem to the main design parameters. The analysis is performed using Computational Fluid Dynamics (CFD). 
Different sizes of the enclosure and different impeller locations are investigated as well as different types of impeller (slow runner and fast runner). As the enclosure narrows, the performance mostly decreases. For certain configurations the enclosure leads to a pressure increase at the the outlet. If the enclosure is too big, the influence on the fan performance is negligible. If the enclosure is too small, the performance is strongly diminished and the influence of other design parameters is negligible. The pressure at the outlet and thus the fan performance is strongly influenced by the flow patterns in the enclosure. The fan performance is mainly affected by the maldistribution of the air at the outlet, the flow and losses in the confinement and the influence of the non-uniform pressure distribution at the impeller outlet. These effects depend on the size and dimensions of the enclosure and the impeller diameter, as well as the location of the impeller and flow patterns at the impeller outlet. 
A correlation of the performance reduction with the geometric parameters (dimensions of the enclosure and fan location relative to impeller diameter) and flow pattern at the impeller outlet is attempted. Using the results some indications for an optimization of the performance of radial impellers in square enclosures are derived.