Improved Empirical Methods for Rocket Noise Prediction Through CAA Computation of Elementary Source Fields

Abstract

Sound pressure levels contours and spectral sound distributions about the VEGA launch system are computed by means of a standard empirical model developed by NASA, and by means of a novel hybrid CAA-empirical approach. The empirical prediction is validated against available noise spectra measured on a scaled mock-up, showing a fairly good agreement when standard values for the model constants are used. The jet sources employed in the empirical prediction are then used to convolute a database of tailored Green’s function computed by means of the frequency-domain CAA code GFD. This second approach permits to take into account the real launch pad geometry and, in the limit of the grid resolution drawn by the acoustic wavelength, mean flow refraction effects in the jet stream. The mean flow is supplied by a Fluent simulation and projected onto the CAA grid. This is a Cartesian grid generated automatically and used in connection with an immersed boundary approach. It is shown that the hybris CAA-empirical prediction is able to explain some discrepancies observed between experimental data and the empirical predictions. In addition it generates a basis of elementary fields to be used in a source localization method under development in the framework of the research project CAST funded by the Italian Aerospace Agency.