A neural networks approach to aerofoil noise prediction

Abstract

A neural network noise prediction model for a turbulent boundary layer noise mechanism has been created using a feed forward multilayer perceptron and a noise spectrum database collected from a family of NACA 0012 areofoils. The results of the neural network model were compared against the Brooks model and it was found that the quality of the prediction was improved was improved over the entire range of the data. The model was also validated against experimental data not utilized the training of the neural, with positive results.