Description of Invention
Optimization of a water turbine design is a non-linear problem with many trade-offs. This study aims to create a data-driven optimization workflow that learns aerodynamics of two-dimensional blade sections from experimental data. It involves constructing a rapid surrogate model using Support Vector Regression (SVR) and employing Particle Swarm Optimization (PSO) to maximize the Lift-to-Drag ratio (L/D), serving as a proxy for enhancing turbine power.