Prediction of nonlinear jet noise propagation

by Gee, Kent L.

Abstract (Summary)
The role of nonlinearity in the propagation of noise radiated from high-performance jet aircraft has not been a well-understood phenomenon in the past. To address the problem of finite-amplitude noise propagation, a hybrid time-frequency domain model has been developed to numerically solve the generalized Mendousse-Burgers equation, which is a parabolic model equation that includes effects of quadratic nonlinearity, atmospheric absorption and dispersion, and geometrical spreading. The algorithm has been compared against analytical theory and numerical issues have been discussed. Three sets of experimental data have been used to evaluate the model: model-scale laboratory jet data, field data using a large loudspeaker, and static engine run-up measurements of the F/A-22 Raptor. Comparison of linearly- and nonlinearly-predicted spectra demonstrates that nonlinearity does, in fact, impact the noise propagation in all three sets of data. Additionally, the extensive comparison with the Raptor data shows that the model is successful in predicting the measured spectrum over multiple angles and engine conditions, demonstrating that the model captures much of the physics of the propagation, despite its current neglect of multipath interference and atmospheric refraction and turbulence effects. Two additional studies have been carried out in order to address fundamental questions relevant to the nonlinear propagation of jet noise: “What is the impact of nonlinearity on perceived levels?” and “At what point does the propagation become linear?” An investigation of the perceived impact of nonlinearity shows that there are only minor differences between nonlinear and linear predictions in calculations of power-based, single-number metrics, such as A-weighted overall sound pressure level. On the other hand, the actual perceived differences between nonlinear and linear waveforms are substantially greater and consequently do not correlate well with calculated metrics. This investigation suggests the need for further research in order to better quantify the perceptual differences between nonlinear and linear propagation. Study of the additional generation of nonlinearity in the long-range propagation iii of a noise waveform suggests that the high-frequency energy decays very slowly and the spatial rate of change of the difference between nonlinearly- and linearlypredicted sound pressure levels does not go to zero, which would indicate linear propagation, but appears to asymptotically approach nonzero constant behavior. This implies that a finite-amplitude noise waveform may never truly decay as linear theory would indicate, which could be important from a human-perception standpoint. iv
Bibliographical Information:


School:Pennsylvania State University

School Location:USA - Pennsylvania

Source Type:Master's Thesis



Date of Publication:

© 2009 All Rights Reserved.