Enhancing aerodynamic and aeroelastic performance of axial compressor through multi-degree-of-freedom parameterization and data-driven multidisciplinary optimization

Multidisciplinary design optimization (MDO) facilitates a comprehensive consideration of inter-disciplinary coupling, overcoming the lengthy processes of conventional serial design and charting a progressive trajectory for compressor optimization. However, the presence of multiple disciplines poses challenges concerning computational efforts and system organizational integration. This paper introduces a novel MDO system by integrating a directly manipulated free-form deformation (DFFD) method and a data-driven model with pre-screening strategy. The DFFD enables precise deformation of compressor surfaces with fewer control points to reduce variables, while data-driven models can significantly accelerate the convergence of complex multi-objective optimization problems. It is applied to an axial compressor to enhance its aerodynamic and aeroelastic performance. A fluid-domain mesh-based method is employed to construct the blade finite element model, addressing the fluid–structure interface interpolation. By progressively increasing the outlet pressure to search the compressor near surge point, the prediction accuracy of surge margin during the optimization process is effectively improved. Compared to the prototype, the isentropic efficiency of optimized compressor at the design point and peak increases by 1.22 and 0.8 percentage points, respectively, and the surge margin improves by 3.23%. Besides aerodynamic performance, the mechanical properties of the rotor blades also show substantial improvement, with the maximum equivalent stress at the blade root reduces by 35.7%, and the low-order resonance margin is also enhanced. The optimization results demonstrate that the developed MDO system can effectively handle the aerodynamic and aeroelastic collaborative optimization of compressor and improve its comprehensive performance.