Optimization of Axial Fans with Highly Swept Blades with Respect to Losses and Noise Reduction

D1 - Computational Aeroacoustics

The main objectives in axial fan design are high efficiency and low sound emission over a wide operating range. A successful design strategy with respect to noise is blade sweep. How-ever, blade sweep also increases losses due to secondary flow effects and thus decreases effi-ciency. The work presented deals with a redesign of a pre-existing highly swept axial fan con-sidering both, sound emission and efficiency. The old design is exclusively based on a quasi 2D blade element method whereas the new design strategy addresses 3D flow effects in more detail. Secondary flows were brought into special focus and examined by means of steady state CFD simulations and an optimization method.
Firstly, the magnitude and position of maximum camber of the utilized NACA profile were varied along the blade span. Secondly, a simple kind of axisymmetric hub contouring was applied to address losses linked with the near-wall circumferential pressure gradient. Geome-trical parameters were defined and the aerodynamic optimum was approached using numerous runs of RANS simulations and the Nelder-Mead optimization method. As target function the total-to-total efficiency at the design point has been selected. An important constraint in the optimization cycle was constant chord length in spanwise direction ensuring equal sweep at leading and trailing edge whereas in the old design the superposition of sweep angles speci-fied at the blade stacking line and variable chord length lead to sweep reductions either at the leading or trailing edge.
It becomes evident that both, a wave shaped hub contour along the blade channel and a new distribution of magnitude and position of maximum camber can manipulate unfavorable pres-sure gradients and hence reduce near wall secondary flow losses. The latter design feature means turning away from predefined spanwise load distributions (e.g. free vortex) and shifts more load towards the outer blade part. The angle of attack was not optimized but is still se-lected according to the 2D airfoil polars. Eventually the characteristic curves of the optimized design were predicted via RANS and experimentally validated. The test results showed good agreement with the simulations. The aerodynamic benefit of the new design was a considera-ble extension of operating range and a moderate increase in efficiency near design point. Moreover, acoustic investigations in a semi-anechoic chamber showed a considerable reduc-tion in sound power over the complete operating range.