Ottimizzazione aerodinamica e aeroacustica di propeller di droni = Aerodynamic and aeroacoustic optimization of drone propeller blades

Recently, interest in battery-powered Unmanned Aircraft Vehicles has grown as potential sustainable aerial transport. Reference is made to several Sustainable Development Goals of the 2030 Agenda thanks to the wide field of application of UAVs, covering health, food security, infrastructure, cities, and human settlements. Among the challenges that the integration of UAVs in the existing aviation airspace has to overcome, noise regulation and public acceptance appear, as the noise signature of propellers could critically penalize their perception in the community. In propeller design optimization, noise reduction typically entails a trade-off with aerodynamic efficiency or thrust generation. Moreover, small size and low flight speed collocate drones in the range of Low Reynolds Number, 10^4 < Re_c <10^5, where conventional airfoils exhibit a critical performance level. In this thesis, a multi-objective optimization is conducted to find an efficient and silent propeller, starting from a two-blade propeller similar to an APC 9x6, with a diameter of 30 cm, reshaped with NACA 4412 airfoil sections, operating at 4000 rpm and at low advance ratio. Specifically, the condition at J=0.1 is studied, and the influence of modifications in blade design concerning chord length and twist angle on the figure of merit and tonal noise sound pressure level is investigated. Second-degree chord distributions are parameterized, identified by the chord value at the root and tip, along with third-degree twist angle distributions, identified by the twist angle value at the root of each blade. Blade Element Momentum Theory is employed for calculating the aerodynamic load distribution across the blade, combining data from Xfoil at Re=50000, while the frequency-domain Helicoidal Surface Theory (Hanson 1980) is implemented for predicting noise. A study of the several tonal noise components at low frequencies is performed. Using a genetic algorithm 'gamultiobj', multiple compromise solutions were found, and it was possible to achieve a noise reduction of about -2 dB and an increase in the figure of merit by 5% on the same blade.