Attenuation of Vane Distortion in a Transonic Turbine Using Optimization Strategies: Part II—Optimization

Abstract

Single-stage high-pressure turbines often operate in the transonic regime, resulting in strong shock interactions between the vanes and the rotor blades. With the present study, a Multi-Objective Optimization is applied to the redesign of a transonic vane. The objective of the steady-flow optimization is to attenuate the propagation of shock waves downstream of the vane, while reducing the vane losses. Computation of the resulting unsteady forcing on the rotor is performed to validate the reduction in high-cycle fatigue risk. The first part of this paper presents the optimization code developed at the von Karman Institute. An evolutionary optimization strategy is applied with a differential evolution algorithm to address multi-objective problems. Turbine vane design is performed using a parametrization of 2D sections and of the 3D stacking line. Aerodynamic performances are evaluated with the TRAF solver considering robust mesh generations. Metamodels are investigated to provide a fast approximation of the Navier-Stokes computation. A methodology is proposed to assess the accuracy and robustness of the metamodel prediction, i.e. to evaluate the metamodel training effectiveness. This second part of this paper targets at the application of the optimization technique to the transonic turbine vane design. The objective is to reduce the downstream pitchwise static pressure distortion, with no increase of vane losses. The operating condition are an isentropic exit Mach number of 1.2 and an outlet angle of −74 degree. The methodology includes 2D section steady-flow optimization design at mid-span. Optimal airfoil passages present a convergent-divergent contraction channel that reduces the trailing edge shock system propagation. Validation of the geometry with the computation of the unsteady force response on the rotor confirms a reduction in high cycle fatigue risk. A multipoint optimization highlights the conflict between reducing the stator/rotor interaction and limiting the losses at off-design. A 3D optimization is finally performed considering simultaneously the lean and 2D section adaptation and enables further improvements of stator/rotor interaction and losses.