Convolutional perfectly matched layer for elastic second-order wave equation.

Area of Science:

• Computational physics
• Wave propagation modeling
• Numerical methods

Background:

• Simulating wave propagation in unbounded elastic media requires effective absorbing boundary conditions (ABC).
• Traditional perfectly matched layers (PML) can exhibit spurious reflections, particularly for surface waves at oblique incidence.
• Existing methods struggle with numerical instabilities in anisotropic media.

Purpose of the Study:

• To extend the convolutional perfectly matched layer (C-PML) for simulating acoustic wave propagation in elastic media.
• To enhance the attenuation of outgoing surface waves compared to classical PML.
• To address and stabilize absorbing layers in anisotropic solids.

Main Methods:

• Developed a C-PML method for second-order acoustic wave equations.
• Implemented C-PML within finite element method (frequency domain) and pseudo-spectral (time domain) algorithms.
• Investigated C-PML performance for oblique incidence and anisotropic media.

Main Results:

• The C-PML ABC demonstrates superior attenuation of outgoing surface waves compared to classical PML.
• The proposed method is more effective than classical PML at oblique incidence, reducing spurious reflections.
• A modified C-PML formulation enhances stability in anisotropic solids.

Conclusions:

• The C-PML is a more effective ABC for simulating acoustic wave propagation in elastic media.
• This method offers improved accuracy and stability, especially for surface waves and anisotropic materials.
• The C-PML provides a robust solution for truncating unbounded domains in wave propagation simulations.