Comparison and Validation Study of Prediction Methods for Jet Noise Shielding

This study compares methods for predicting the jet noise shielding effect and examines their uncertainties and computational characteristics. Jet shielding is a promising approach to reducing landing and takeoff noise in future supersonic aircraft. However, modeling and analyzing noise sources remain challenging due to the distributed nature of jet noise sources. Experimental and numerical analyses were performed on two geometries of varying complexity to assess differences in uncertainty and computational cost across three jet noise shielding prediction methods: a ray-tracing method employing multiple directional point sources, Fast Multipole Boundary Element Method (FMBEM) with a distributed point source model, and FMBEM with a wave packet model. All three prediction methods captured the qualitative trends of the shielding effect. Although the average discrepancy between predicted and measured values varied by method, it remained within 2.5 dB across all cases. Even when using the same FMBEM framework and computational mesh, the wave packet model exhibited lower average error and standard deviation than the distributed point source model. This difference was particularly significant for the more complex geometry. Among the three methods, the ray-tracing approach required the least computational time, whereas the FMBEM with the wave packet model was the most expensive. The computational cost of the ray-tracing method remained nearly constant across frequency variations, making it advantageous for broadband predictions. Regardless of the prediction method, incorporating the observed axial sound pressure distribution is essential for accurate prediction. Therefore, precise measurement of the isolated jet’s source distribution as an input is critical for the successful estimation of the shielding effect.