CFD Analysis of Turbine Cascade Unsteady Aerodynamics Using a Hybrid POD Technique

This study presents a computational investigation into the unsteady aerodynamics of a low-pressure turbine cascade, utilizing computational fluid dynamics (CFD) with a primary focus on enhancing efficiency. The proposed approach combines a classical proper orthogonal decomposition with a modern machine learning technique. This hybrid methodology demonstrates its effectiveness by accurately predicting the unsteady flow over the turbine blade. Crucially, the solution retains the essential features of the original physics-based computational model. This study represents a potential significant advancement in improving the efficiency of CFD solutions, enabling future resource-conscious scale-resolving simulations of complex aerodynamic flows without sacrificing solution accuracy.