Session: 32-07 Endwall and Secondary Flows

Paper Number: 154131

Secondary Flow Losses Mitigation for Low-Pressure Turbine Cascade Using Response Surface Optimization 

End wall secondary flows are one of the major contributors to turbine loss generation. The effect is crucial in compactly placed low-pressure turbines. The interaction of the incoming flow boundary layer with the blade leading edge forms the horseshoe vortex. Later, this horseshoe vortex contributes to the counterproductive formation of passage vortex within blade rows. The multi-vortex structure reduces effective power generation, thereby degrading turbine performance. In recent times, methods like non-axisymmetric end wall contouring have shown the potential to suppress the detrimental effects of secondary flow streams. The method involves the systematic redistribution of the flow, thereby reducing end wall non-uniformities and improving the end wall static pressure distribution. In the present paper, a baseline experimental validation study is performed with an underlying secondary flow field in the first place for a low-speed, low-pressure linear turbine cascade vane. Thenceforward, the non-axisymmetric end wall design strategy, optimization method, and effect of end wall contouring on the secondary flow field are studied.

This study integrates a computational fluid dynamics solver with an optimization algorithm based on the parameterization method. Each design of the experiment is solved using the steady Reynolds-Averaged Navier-Stokes equations, closed by the shear stress transport-based γ-Re_θ turbulence model for low entry turbulence intensity and chord exit Reynolds number 1.6 × 10⁵. Numerical simulations are performed for each generated configuration, and the responses are collected. Finally, the single-objective optimization is carried out using a response surface algorithm to obtain the optimal non-axisymmetric end wall position parameters. The baseline configuration features a non-contoured end wall, and the numerical results show good predictive accuracy against experimental data. Baseline configuration has revealed detailed secondary flow field originates from the leading-edge horseshoe vortex and propagates towards the suction side and pressure side. The major secondary loss is observed from the passage vortex and the pressure-side stream of the horseshoe vortex, the end-wall cross-flow, and the incoming boundary layer primarily influences the generation of the passage vortex within the blade row. The non-axisymmetric end wall parameterization involves creating a hill near the pressure side and a valley near the suction side, which is achieved by performing pitch-wise parameterization of the control points within the design space. The study revealed that extending the design space of the non-axisymmetric end wall from 20% Cax upstream to the leading edge to 70% Cax aids in the breakdown of the inlet boundary layer. Reduction in the end wall cross-flow mitigation and span-wise extension of the passage vortex is identified for an optimized profile. The optimal range for amplitude variation within the design space (20% Cax upstream to the leading edge to 70% Cax) is typically 2 - 4% of the span. However, extending the design space beyond 70%Cax results in the formation of corner vortices, leading to an increase in the loss coefficient, and is therefore not advisable.

Presenting Author: Anand Darji Sardar Vallabhbhai National Institute of Technology - Surat

Presenting Author Biography: Anand is a PhD Research Scholar at the Department of Mechanical Engineering, Sardar Vallabhbhai National Institute of Technology, Surat, Gujarat, INDIA. His area of research is CFD in Axial and Radial Turbomachines, Optimization, and linear cascade Experiments for Low-Pressure Turbines.

Authors:

Anand Darji Caterpillar India Pvt Ltd.
Beena Baloni Sardar Vallabhbhai National Institute of Technology - Surat
Chetan Mistry Indian Institute of Technology - Kharagpur