The paper presents an automatic generator of approximate nonreflecting boundary conditions, analytical and numerical, for scalar wave equations. This generator has two main ingredients. The first one is a set of local Trefftz functions -- outgoing waves approximating the solution in the vicinity of a given point of the exterior boundary of the computational domain. The second ingredient is a set of linear test functionals (degrees of freedom). One example of such functionals is the nodal values of the solution at a set of grid points; in that case, one obtains a numerical condition -- a finite difference scheme at the boundary. Alternatively, the functionals may involve derivatives or integrals of the solution, in which case the proposed "Trefftz machine" yields analytical nonreflecting conditions. Corners and edges are treated algorithmically the same way as straight boundaries. With specific choices of bases and degrees of freedom, the machine produces classical conditions such as Engquist-Majda and Bayliss-Turkel. For other choices, one obtains a variety of analytical and numerical conditions, a few of which are presented as illustrative examples.
A Trefftz Discontinuous Galerkin Method for Time Harmonic Waves with Generalized Impedance Boundary Conditions
Institute for Mathematics and its Applications, College of Science and Engineering(数学及其应用研究所,科学与工程学院);Department of Mathematical Sciences, University of Delaware(特拉华大学数学科学系);Departamento de Matemáticas, Universidad de Oviedo(奥维耶多大学数学系)
Stability and convergence analysis of high-order numerical schemes with DtN-type absorbing boundary conditions for nonlocal wave equations
研究了无界空间区域上一维和二维非局部波动方程高阶数值逼近的稳定性和收敛性。首先用基于正交的有限差分格式离散空间非局部算子,然后用显式差分格式逼近时间导数,得到一个完全离散的无穷系统.在此基础上,构造了Dirichlet-to-Neumann(DtN)型吸收边界条件(ABCs),将无限大离散系统化为有限大离散系统。为此,我们首先采用[Du,Zhang and Zheng,\emph{Commun. Comput. Phys.},24(4):1049--1072,2018 and Du,Han,Zhang and Zheng,\emph{SIAM J. Sci. Comp.},40(3):A1430--A1445,2018]中的思想,分别导出一维和二维情况下的Dirichlet到Dirichlet(DtD)型映射。然后利用离散非局部Green第一恒等式从DtD型映射得到离散DtN型映射。由此产生的DtN型映射使得有可能进行减少的问题的稳定性和收敛性分析。数值实验证明了该方法的准确性和有效性。