Non-reflecting boundary conditions

Numerical solution of problems on unbounded domains. a review

Computational fluid dynamics (CFD) has made tremendous progress especially in aerodynamics and aircraft design over the past 20 years. An obvious question to ask is "why not use CFD methods to solve aeroacoustics problems?" Most aerodynamics problems are time independent, whereas aeroacoustics problems are, by definition, time dependent. The nature, characteristics, and objectives of aeroacoustics problems are also quite different from the commonly encountered CFD problems. There are computational issues that are unique to aeroacoustics. For these reasons computational aeroacoustics requires somewhat independent thinking and development. The objectives of this paper are twofold. First, issues pertinent to aeroacoustics that may or may not be relevant to computational aerodynamics are discussed. The second objective is to review computational methods developed recently that are designed especially for computational aeroacoustics applications. Some of the computational methods to be reviewed are quite different from traditional CFD methods. They should be of interest to the CFD and fluid dynamics communities.

/pdf/computational-aeroacoustics-issues-and-methods-27gvcjx7jp.pdf

Computational aeroacoustics - Issues and methods

High-order finite-difference schemes are less dispersive and dissipative but, at the same time, more isotropic than low-order schemes. They are well suited for solving computational acoustics problems. High-order finite-difference equations, however, support extraneous wave solutions which bear no resemblance to the exact solution of the original partial differential equations. These extraneous wave solutions, which invariably degrade the quality of the numerical solutions, are usually generated when solid-wall boundary conditions are imposed. A set of numerical boundary conditions simulating the presence of a solid wall for high-order finite-difference schemes using a minimum number of ghost values is proposed. The effectiveness of the numerical boundary conditions in producing quality solutions is analyzed and demonstrated by comparing the results of direct numerical simulations and exact solutions.

Wall boundary conditions for high-order finite-difference schemes in computational aeroacoustics

I review the development of numerical evolution codes for general relativity based upon the characteristic initial-value problem. Progress in characteristic evolution is traced from the early stage of 1D feasibility studies to 2D-axisymmetric codes that accurately simulate the oscillations and gravitational collapse of relativistic stars and to current 3D codes that provide pieces of a binary black-hole spacetime. Cauchy codes have now been successful at simulating all aspects of the binary black-hole problem inside an artificially constructed outer boundary. A prime application of characteristic evolution is to extend such simulations to null infinity where the waveform from the binary inspiral and merger can be unambiguously computed. This has now been accomplished by Cauchy-characteristic extraction, where data for the characteristic evolution is supplied by Cauchy data on an extraction worldtube inside the artificial outer boundary. The ultimate application of characteristic evolution is to eliminate the role of this outer boundary by constructing a global solution via Cauchy-characteristic matching. Progress in this direction is discussed.

/pdf/characteristic-evolution-and-matching-50f8eok8zg.pdf

Characteristic Evolution and Matching

https://ntrs.nasa.gov/api/citations/19890013026/downloads/19890013026.pdf

Hong Jiang

Papers

Absorbing boundary conditions for second-order hyperbolic equations