Loss mechanisms in labyrinth seals of shrouded axial turbines

Abstract

The interaction flows associated to open cavities in shrouded, high pressure tur- bines were experimentally investigated in this dissertation. For this purpose a two-stage, shrouded, axial turbine was built and commissioned. The measure- ment campaign focused on the rotor tip labyrinth seal, comprising two seal gaps, 0.3% and 0.8% blade height. The labyrinth seal consists of an open inlet cavity, closed labyrinth cavities and an open exit cavity. The size of those cav- ities is small in comparison to the main flow channel height (15% of blade height). Therefore, a new probe measurement technology of minimum block- age effect was developed. The new virtual four sensor, fast response aerody- namic probe (FRAP) resolves the deterministic, unsteady flow field up to 25kHz in terms of flow angles, velocity and total and static pressure. The inlet cavity to the labyrinth seal is subject to strong in and out flows, which involve up to four times the leakage mass with a superimposed unsteady fluc- tuation of ±70% of the leakage mass flow. The in and out flow happens in spe- cific flow regions set up by the downstream flow field of the stator. This results in unsteady fluctuations on the downstream rotor inlet flow field. A control vol- ume analysis revealed the forces which act on the fluid close to the interaction zone. The radial equilibrium of forces is applied to the inflow streamline, ex- plaining the inviscid exchange mechanism. Due to the sucking of the leakage the flow loses circumferential and axial momentum across the interaction zone, which results in a negative incidence at the rotor blade tip. A flow model is de- rived which predicts negative incidence angles of -9° for gap widths of the or- der of 1.5% blade height. The cavity flow is dominated by two toroidal vortices, which swirl around the annulus at a high tangential velocity of 82% blade tip speed. Both vortices are stretched in space and time, due to the fluctuating ve- locity gradients setup by the moving pressure fields. A small portion of losses was found to be caused by the inlet cavity flow itself, which are mainly due to wall friction effects (2.7% of stage loss). Additional losses may be induced within the rotor blade row due to the negative incidence in combination with fluctuating flow angles and velocities at the rotor tip inlet. This region is of importance to the secondary flow formation within the rotor passage. The inlet boundary conditions to the rotor tip region found in this con- figuration is discussed in detail. The exit cavity flow comprises three flow features, which interact among each other: the cavity fluid flow being a toroidal vortex, the leakage jet and the main flow. The cavity flow is mainly driven by the jet due to viscous shear. The tan- gential velocity of the cavity flow was found to depend on the strength of the