| 摘要 |
Propagation of sound in an ocean is accurately predicted at all ranges from a source by a computational method that uses a newly created Decomposition Method to make the predictions at long range from the source. The predictions are done in a limited region of the ocean without the need to predict propagation of sound outside the region. The vertical boundaries of the region are parallel to a given direction of propagation of sound from the source. These boundaries are transparent, and sound penetrates them without reflection. The Decomposition Method is based on a special splitting of the sound field into primary and secondary components. The primary components are related to the sound field that would exist in the region if the vertical boundaries of the region were acoustically hard. They are found by integrating a system of coupled parabolic equations that are derived from the coupled mode equations for the sound field in the region. The primary components contain spurious reverberations that originate at the vertical boundaries of the region. The secondary components identify and cancel the spurious reverberations in the primary components. The secondary components are found by integrating standard parabolic equations subject to transparent boundary conditions at the vertical boundaries of the region. These boundary conditions also include source terms that are obtained from the primary components. When the primary and secondary components are summed, the spurious reverberations are cancelled in the sum. The actual sound field in the region is found from this sum. A modification of the Decomposition Method incorporates an extrapolation procedure to ensure accuracy of the predictions at long range from the source.
|
