Optimisation of Propellers with Noise-Based Constraints Including a Deep Learning Method for Aerofoil Prediction

The increased interest in Urban Air Mobility in recent years has created a new class of aircraft design concepts that will be subject to strict noise emission limits due to the desired proximity to population centers. As such, it is necessary to be able to predict the noise produced by these vehicles at the earliest stages of the design process. Propeller design is an important aspect in which vehicle noise can be reduced. In this work a BEMT based propeller simulation and optimisation framework is presented, capable of predicting the aerodynamic and aeroacoustic performance and optimising the blade geometry subject to acoustic constraints. As part of this tool aerofoil aerodynamics are obtained using a machine learning approach, allowing for optimisation of the blade profile as well as an increase in computational efficiency. The application of this tool to a baseline experimental propeller is included, increasing the aerodynamic efficiency by over 10%. Including the aerofoil shape in this optimisation process allowed for a maximum additional efficiency increase of 2%. The effect of noise constraints are also studied and were found to reduce the optimal efficiency. The inclusion of aerofoil shape in the optimisation was found to mitigate the deterioration of optimal efficiency as more stringent noise constraints are imposed.