Wind-Turbine Blade Optimization Using MATLAB and CFD + AI

A student-laptop project that optimises the blades of a 10 kW, three-bladed, 7.2 m small wind turbine for annual energy at a Rayleigh site with 6 m/s mean wind. Aerofoil polars come from a panel method with a boundary layer (100 NACA 4-digit polars); blade-element-momentum (BEM) theory gives the power curve, annual energy (AEP) and loads; the spar caps are sized for strength and a 10 % tip-deflection limit. 1,400 Latin-hypercube blades train four AI surrogate families (neural network mean test R2 0.988), which drive a constrained optimisation. The optimum raises AEP from 26.89 to 28.98 MWh per year (+7.8 %) with the root flap moment kept within 110 % of the baseline; a Pareto-knee design gains +7.2 % with -6.0 % blade mass. An axisymmetric Navier-Stokes actuator-disk CFD solver checks the momentum part of BEM. A web application ("Blade Studio") lets engineers try designs with BEM or the AI surrogate. The workflow is also written in MATLAB (port, not run - no MATLAB on the build computer); all numbers come from the Python implementation and are computed, not measured.

Category: Mechanical Engineering / Wind Energy / Aerodynamics and CFD / Machine Learning Surrogates / Optimisation / MATLAB and Python

Mechanical Engineering Project M06 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.