AIAA 96 - 1706 Numerical Algorithms for Acoustic Integrals – the Devil is in the Details

The accurate prediction of the aeroacoustic eld generated by aerospace vehicles or nonaerospace machinery is necessary for designers to control and reduce source noise. Powerful computational aeroacoustic methods, based on various acoustic analogies (primarily the Lighthill acoustic analogy) and Kirchho methods, have been developed for prediction of noise from complicated sources, such as rotating blades. Both methods ultimately predict the noise through a numerical evaluation of an integral formulation. In this paper, we consider three generic acoustic formulations and several numerical algorithms that have been used to compute the solutions to these formulations. Algorithms for retarded-time formulations are the most e cient and robust, but they are di cult to implement for supersonic-source motion. Collapsing-sphere and emission-surface formulations are good alternatives when supersonicsource motion is present, but the numerical implementations of these formulations are more computationally demanding. New algorithms|which utilize solution adaptation to provide a speci ed error level|are needed.