Impellers are critical components in hydraulic turbomachinery, as they directly influence energy efficiency, structural reliability, and cavitation resistance, factors which are essential to the sustainability of water systems. This study introduces an innovative hybrid optimization framework that combines high-fidelity Computational Fluid Dynamics (CFD) and Finite Element Method (FEM) simulations with AI-based predictive modeling, complemented by evolutionary algorithms such as NSGA-II and advanced statistical techniques including RSM and BBD. A systematic review of over 100 high-impact studies reveals that hybrid approaches can enhance hydraulic efficiency by up to 25%, reduce rotor mass by over 30%, and significantly mitigate cavitation.

The proposed modular framework enables component-specific optimization (of the blades, hub, and casing), effectively balancing simulation accuracy and computational cost. Predictive models based on Artificial Neural Networks (ANNs), XGBoost, and Deep Graph Neural Networks (DGNNs) accelerate the optimization process and support decision-making under uncertainty. Notable strengths include adaptability to various flow regimes (steady and unsteady), integration with open-source tools such as OpenFOAM, and experimental validation through additive manufacturing.

By bridging traditional CFD-driven design and data-centric optimization, the proposed methodology offers a reproducible path for improving rotor performance in water-based turbomachinery. This work not only consolidates the current state of the art but also delivers a scalable and transferable process for applications in hydropower, water treatment, and fluid transport systems.