Computational design and experimental evaluation of using a leading edge fillet on a gas turbine vane

With the desire for increased power output for a gas turbine engine comes the continual push to achieve higher turbine inlet temperatures. Higher temperatures result in large thermal and mechanical stresses particularly along the no.-le guide vane. One critical region along a vane is the leading edge-endwall juncture. Based on the assumption that the approaching flow to this juncture is similar to a two-dimensional boundary layer previous studies have shown that a horseshoe vortex forms. This vortex forms because of a radial total pressure gradient from the approaching boundary layer. This paper documents the computational design and experimental validation of a fillet placed at the leading edge-endwall juncture of a guide vane to eliminate the horseshoe vortex. The fillet design effectively accelerated the incoming boundary layer thereby mitigating the effect of the total pressure gradient. To verify the CFD studies used to design the leading edge fillet, flowfield measurements were performed in a large-scale, linear vane cascade. The flowfield measurements were performed with a laser Doppler velocimeter in four planes orientated orthogonal to the vane. Good agreement between the CFD predictions and the experimental measurements verified the effectiveness of the leading edge fillet at eliminating the horseshoe vortex. The flow-field results showed that the turbulent kinetic energy levels were significantly reduced in the endwall region because of the absence of the unsteady horseshoe vortex.