Loss Mechanisms of Interplatform Steps in a 1.5-Stage Axial Flow Turbine

Abstract

In this paper, the detailed steady and unsteady numerical investigations of a 1.5-stage axial flow turbine are conducted to determine the specific influence of interplatform steps in the first stator—as caused by deviations in manufacturing or assembly. A basic first stator design and a design consisting of a bow and endwall contours are compared. Apart from step height, the position and geometry of the interplatform border are varied for the basic design. To create the steps, every third stator vane was elevated, together with its platforms at hub and shroud, such that the flow capacity is only little affected. The results show that the effects of steps on the platform borders in front and aft of the first stator can be decoupled from those occurring on the interplatform steps. For the latter, being the main contributor to the additional loss, the intensity of recirculation zones and losses increase substantially when the platform border is located close to the suction side. Using a relative step height of 1.82% span, the entropy production doubles when compared to a position close to the pressure side, which can be explained by differences in local flow velocity level. Regarding a circular-arc-shaped platform, the losses can be more than halved—mainly due to lower included angles between step and endwall flow streamlines. The findings can be explained by a nondimensional relation of the local entropy production using local values for step height and characteristic flow quantities. Furthermore, a reduction in step height leads to an attenuation of the otherwise linear relationship between step height and entropy production, which is mainly due to lower local ratio of step height and boundary layer thickness. In the case of laminar or transitional flow regions on the endwall, typical for turbine rigs with low inlet turbulence and low-pressure turbines under cruise conditions, the steps lead to immediate local flow transition and thus substantially different results.