Noise emission prediction of small scale twin pushing propellers through surrogate models

The global interest in small multi-rotor drones, urban air-taxis and unmanned aerial vehicles is growing rapidly. It is now widely accepted that noise is one of the main factors that could limit the public acceptance and adoption of this technology. The aim of this paper is a preliminary understanding of the noise impact of pusher propellers and developing surrogate models for its prediction. To this end, a series of laboratory scale experiments in an anechoic chamber were arranged involving two twin pusher propellers installed on a configurable frame. The test campaign was carried out varying the mutual configuration, the rotational speed and the blade pitch angle of the two propellers. Pressure fluctuations were acquired from eight microphones, two of them located in the nearfield to retrieve the rotational speeds, and six in the far field to study the directivity of the thickness and loading noise. The noise impact was evaluated through the calculation of the overall sound pressure level. The obtained database served as training set to build the surrogate model. Different techniques were tested and their performance evaluated in order to obtain a robust prediction of the noise emission with a small number of experimental test points.