Abstract:
Burgers' equation is used to predict the effect of nonlinearity on the power spectral density of plane broadband noise traveling in a nondispersive thermoviscous fluid. The source signal is assumed to be stationary Gaussian noise, which, because of nonlinear propagation distortion, becomes non-Gaussian as it travels. As opposed to time-domain methods, the method presented here is based directly on the power spectral density of the signal, not the signal itself. The Burgers equation is transformed into an unclosed set of linear equations that describe the evolution of the joint moments of the signal. A method for solving the system of equations is presented. Only the evolution of appropriately selected joint moments needs to be calculated in order to predict the evolution of the power spectral density of the signal. The results are in good agreement with a time-domain code [Cleveland et al., J. Acoust. Soc. Am. 98, 2865(A) (1995)]. The method can be also applied when the source condition is a stationary, ergodic, and Gaussian stochastic process. [Work supported by NASA.]