Engineering topics
Publications and source records attributed to Karel, S..
Results of applications of algorithm suggest that it has potential application to variety of related fluid flow problems, such as presently intractable separation problem of aerodynamics. Details of mathematical development, as well as computation of explicit error estimates, are available.
Results are described of an analytical study of the accuracy and limitations of a technique that permits the mathematical extrapolation of near-field noise data to far-field conditions. The effects of the following variables on predictive accuracy of the far-field pressure were examined: (1) number of near-field microphones; (2) length of source distribution; (3) complexity of near-field and far-field distributions; (4) source-to-microphone distance; and (5) uncertainties in microphone data and imprecision in the location of the near-field microphones. It is shown that the most important parameters describing predictive accuracy are the number of microphones, the ratio of source length to acoustic wavelength, (L/wavelength), and the error in location of near-field microphones. If microphone measurement and location errors are not included, then far-field pressures can be accurately predicted up to L/wavelength values of 15 using approximately 50 microphones. For maximum microphone location errors of + or - 1 cm, only an accuracy of + or - 2-1/2 db can be attained with approximately 40 microphones for the highest L/wavelength of 10.
An algorithm for solving the nonlinear stationary Navier-Stokes problem is developed. Explicit error estimates are given. This mathematical technique is potentially adaptable to the separation problem.
Results are presented for an analytical study of the accuracy and limitations of a technique that permits the mathematical extrapolation of near-field noise data to far-field conditions. The effects of the following variables on predictive accuracy of the far-field pressure were examined: (1) number of near-field microphones; (2) length of source distribution; (3) complexity of near-field and far-field distributions; (4) source-to-microphone distance; and (5) uncertainties in microphone data and imprecision in the location of the near-field microphones. It is shown that the most important parameters describing predictive accuracy are the number of microphones, the ratio of source length to acoustic wavelength (L/lambda), and the error in location of near-field microphones. For maximum microphone location errors of plus or minus 1 cm, only an accuracy of plus or minus 2.5 dB can be attained with approximately 40 microphones for the highest L/lambda of 10.