Aerodynamic Design Optimization of a Rotor Blade-Sleeve Section using a Surrogate-Based Approach

The ever increasing air traffc demands for a reduction of aircraft's emissions: this could be achieved by reducing the fuel consumption, which is highly affected by the drag produced by the aircraft. In helicopter aerodynamics, one of the main drag contributors are rotor heads. As part of the Clean Sky 2 programme, TUM's FURADO Project aims at optimizing the aerodynamic design of a full fairing rotor-head via CFD simulations. The optimization task is performed using a genetic algorithm, a very effective but expensive method. In this Thesis, a Surrogate-Based approach is proposed so as to perfom the aerodynamic design optimization of a rotor blade-sleeve section belonging to the Airbus' RACER demonstrator. A Bayesian surrogate model, speci fically a Gaussian Process, is used to partially replace the CFD solver, thus increasing the computational effciency of the whole optimization process. The procedure is directly compared to a standard fully CFD-based approach. Results reveal that the proposed method, although showing some drawback in designs constraints handling, has a good potential to obtain a substantial reduction of the computational effort, nontheless being able to provide promising designs within a few number of optimization's iterations.