Abstract
Today numerical models are widely used within the underwater acoustics community for solving complex wave‐propagation problems. While it is easy for anybody to produce “interesting” field solutions with a model, it is extremely difficult even for experts to generate numerical results that can be considered an accurate solution to a stated acoustical problem. This difficulty derives from insufficient knowledge both of the approximations introduced to formulate a solvable set of equations and of the accuracy and convergence problems associated with the numerical implementation itself. When experimental data are not available for checking a numerical solution, there are, in principle, only two ways to gain confidence in the numerical result (it is here assumed that straightforward checks of reciprocity and energy conservation have been performed): (1) Use a different model to confirm the validity of the original solution; or (2) compare the numerical result to an accepted reference solution to a similar propagation problem. Currently, the intermodel comparison is the only viable approach, since few reference solutions are available, none of which is for general range‐dependent problems. Illustrative examples of both easily generated incorrect field solutions and artfully constructed correct field solutions will be given for range‐independent environments using computer codes based on normal‐mode and fast‐field theory, and for range‐dependent environments using codes based on adiabatic modes, coupled modes, and the parabolic equation.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call