Throughout the project, we expect to develop innovative algorithms for 2D/3D adaptive feature-based parametrizations in 3D point clouds, including dynamic geometry partitioning and parameter reduction for shape optimization. We will implement software procedures for dynamically adaptive parametric representations, enabling faster and more precise optimization of complex shapes. In the experimental phase, we will complete modifications to the wind tunnel for precise flow and torque measurements and perform 3D scanning of Savonius turbine blades. Through iterative CFD-FEM simulations, we will optimize blade geometry to reduce vortices and achieve the target power coefficient (cP > 0.40). We will test three VAWT prototypes with different configurations, measuring power output (4–12 m/s), pressure distribution, velocity profiles, and dynamic loads. Results will be published in at least one Q1/Q2 journal, alongside developing test cases using FEM/CFD simulations and our own geometric modeling methods.