A New Multi-Stage Turbine Stator Design for Improved Performance Retention

Abstract

Abstract An experimental and computational investigation has been undertaken into the effects of deterioration of the first stage rotor shroud knife-edge seal clearance in a two-stage turbine which has engine representative cavity geometries. Four values of deterioration were investigated which cover the new-condition to old-condition knife-edge seal clearance. Measurements within the first stage rotor shroud cavity show that whilst the leakage mass flow rate increases with deterioration, the angle at which the leakage flow approaches the downstream stator is essentially fixed and independent of the flow coefficient. This is in agreement with a simple over-tip leakage model. Because of the engine representative cavity geometry, the over-shroud leakage flow undergoes little mixing when it re-enters the mainstream and approaches the downstream stator at more than 60° negative incidence. Detailed measurements at the exit of the second stage stator identified two large positive vortices which were not consistent with the horseshoe vortex model for secondary flow. A computational investigation revealed that one vortex originates from the rolling-up within the stator passage of the streamwise vorticity sheet associated with the first stage rotor over-shroud leakage. This roll-up vortex cannot be eliminated. The second vortex is generated within the stator passage by the separation of the over-shroud leakage flow at the leading-edge due to the large negative incidence. It was hypothesised that this separation vortex might be eliminated by locally redesigning the stator. A new stator was designed, manufactured and tested. As predicted, the roll-up vortex was still present but the separation vortex was eliminated. For all the values of deterioration tested the entropy loss coefficient of the new stator and the unchanged second stage rotor were reduced. It is estimated that the new stator would improve the lifetime average efficiency by 0.5% compared to the original.