Colloquia

Samenvatting

This research investigates the accuracy of a semi-empirical broadband noise prediction method for UAV (Unmanned Aerial Vehicle) and UAM (Urban Air Mobility) applications. An accurate broadband noise prediction method for the design and operational optimization of UAVs and UAM blades can help to reduce noise pollution, particularly in urban environments.

XFOIL has been used to gain insight into boundary layer properties, airfoil characteristics, and flow separation effects on a DJI Matrice 600 pro drone blade geometry.

A Brooks, Pope, Marcolini (BPM) model is used to predict broadband noise from a single rotor blade under hover conditions. The boundary layer parameters which are needed as an input for the BPM model are determined from the XFOIL results, or from semi-empirical expressions derived by Brooks, Pope and Marcolini. Further input parameters were gained from measurements at relative flight conditions for UAV and UAM applications. FLIGHTLAB was used to account for 3D effects (induced velocity and angle of attack) in the 2D flow results from XFOIL. Acoustic Wind Tunnel Testing was conducted and used to compare empirical data with model predictions.

It was found that the BPM model provides a practical broadband noise characteristic prediction which is sufficiently accurate for early-stage design and analysis. However, it was also found that the BPM method underpredicts Sound Pressure Levels (SPL) and overpredicts peak frequencies. It was concluded that semi-empirical prediction models like BPM should only be used to model sound characteristics and their sensitivity to a change in (design) parameters. The absolute values gained from predictions are imprecise and should not be used (for example to predict if a rotor blade would comply with regulatory requirements).

It was furthermore found that adapting the BPM model such that computed blade-specific boundary layer parameters from XFOIL simulations can be given as an input into the BPM model, instead of using the semi-empirical relations from the original BPM model, does not improve agreement of SPL or peak frequency predictions with wind tunnel test results for a single rotor hover case of a DJI Matrice 600 pro drone blade.

This work contributes to the ongoing development of predictive noise modelling tools, guiding the design of quieter UAV and UAM systems that align with society. It points towards further work for better engineering models and future research into high-fidelity modelling and experimental validation, paving the way for sustainable and publicly acceptable urban air mobility solutions.