Most large scale cascade experiments and the corresponding computational fluid dynamics are conducted on geometries without fillet radii, that is with sharp edges. Despite that the fact that all real machines have a fillet radii there has not so far been a detailed examination of the influence of fillet radii on the secondary flow structures in turbines.

This paper presents detailed measurements on a large scale, low speed cascade with significant secondary flows. A series of detailed traverses inside the blade row were obtained allowing the influence of fillet radii on the formation of secondary flows to be examined. The experimental campaign is backed up by RANS computations.

The fillets were found to increase loss by around 10% compared to the base case. This loss could not be accounted for by a simple model of boundary layer behaviour and an accounting of the complex secondary flow interactions is required to explain the observed changes in performance. The secondary flow interactions include a larger radial migration of secondary flows from the endwall and the elimination of the corner vortex. The potential to improve performance by varying the fillet radii around the blade was identified.

This content is only available via PDF.
You do not currently have access to this content.