Multipoint Discrete-Adjoint Optimization of Two-Dimensional Axial Turbine Blade Profiles Using CAD-Based Curve Parameterization

Modern turbomachinery is required to maintain high aerodynamic performance over a wide operating range. To address this requirement, a multipoint discrete-adjoint optimization framework is developed based on the open-source SU2 platform and blade-profile parameterization based on computer-aided design (CAD). Two representative operating points, corresponding to transonic and subsonic flow conditions, are considered in the multipoint optimization, while two corresponding single-point optimizations are also performed for comparison. The results show that the multipoint optimized blade achieves a 17.3% reduction in weighted total pressure loss relative to the baseline blade, with reductions of 20.4% and 12.0% at the two operating points, respectively. Although the single-point optimized designs show better performance at their respective design points, the subsonic single-point optimized blade suffers suction-side trailing-edge separation under the transonic condition, leading to an enlarged vortex region, intensified wake mixing, and degraded aerodynamic performance. Geometrically, the optimized blades exhibit a reduced geometric chord length through a decrease in stagger angle under the fixed axial-chord constraint, and the multipoint optimized profile lies between the transonic and subsonic single-point design tendencies near the suction-side leading edge. The corresponding flow-field changes include an upstream shift of the high-Mach-number region, smoother postpeak deceleration, weakened shock intensity, and reduced flow loss. In addition, the multipoint optimized blade shows lower turbulent kinetic energy levels near the trailing edge, increased trailing-edge base pressure, and reduced wake-vortex size, indicating weakened wake mixing and lower local energy dissipation. These results demonstrate the effectiveness of the proposed framework for multipoint aerodynamic optimization of turbine cascades.