Abstract:Abstract Physics-informed neural networks (PINNs) have been used by researchers to solve partial differential equation (PDE)-constrained problems. We evaluate PINNs to solve for frequency-domain acoustic wavefields. PINNs can solely use PDEs to define the loss function for optimization without the need for labels. Partial derivatives of PDEs are calculated by mesh-free automatic differentiations. Thus, PINNs are free of numerical dispersion artifacts. It has been applied to the scattered acoustic wave equation, which relied on boundary conditions (BCs) provided by the background analytical wavefield. For a more direct implementation, we solve the nonscattered acoustic wave equation, avoiding limitations related to relying on the background homogeneous medium for BCs. Experiments support our following insights. Although solving time-domain wave equations using PINNs does not require absorbing boundary conditions (ABCs), ABCs are required to ensure a unique solution for PINNs that solve frequency-domain wave equations, because the single-frequency wavefield is not localized and contains wavefield information over the full domain. However, it is not trivial to include the ABC in the PINN implementation, so we develop an adaptive amplitude-scaled and phase-shifted sine activation function, which performs better than the previous implementations. Because there are only two outputs for the fully connected neural network (FCNN), we validate a linearly shrinking FCNN that can achieve a comparable and even better accuracy with a cheaper computational cost. However, there is a spectral bias problem, that is, PINNs learn low-frequency wavefields far more easily than higher frequencies, and the accuracy of higher frequency wavefields is often poor. Because the shapes of multifrequency wavefields are similar, we initialize the FCNN for higher frequency wavefields by that of the lower frequencies, partly mitigating the spectral bias problem. We further incorporate multiscale positional encoding to alleviate the spectral bias problem. We share our codes, data, and results via a public repository.

Effects of Sampling Strategies and Optimization Methods on Solving Wave Equation with Physics-Informed Neural Network

An Efficient Approximation Method Based on Enhanced Physics-Informed Neural Networks for Solving Localized Wave Solutions of PDEs

Adaptive-Sampling Physics-Informed Neural Network for Viscoacoustic Wavefield Simulation

Pre-Training Physics-Informed Neural Network with Mixed Sampling and Its Application in High-Dimensional Systems

A comprehensive study of non-adaptive and residual-based adaptive sampling for physics-informed neural networks

TCAS-PINN: Physics-informed neural networks with a novel temporal causality-based adaptive sampling method

Diminishing Spectral Bias in Physics-Informed Neural Networks Using Spatially-Adaptive Fourier Feature Encoding

Physics-informed neural networks with residual/gradient-based adaptive sampling methods for solving PDEs with sharp solutions

An Efficient Framework for Solving Forward and Inverse Problems of Nonlinear Partial Differential Equations Via Enhanced Physics-Informed Neural Network Based on Adaptive Learning

A Gradient-Enhanced Physics-Informed Neural Networks Method for the Wave Equation

Accelerating the Convergence of Physics Informed Neural Networks for Seismic Wave Simulation

A two-stage physics-informed neural network method based on conserved quantities and applications in localized wave solutions

Practical Aspects of Physics-Informed Neural Networks Applied to Solve Frequency-Domain Acoustic Wave Forward Problem

Multiple Wavefield Solutions in Physics-Informed Neural Networks Using Latent Representation

Multi-Net Strategy: Accelerating Physics-Informed Neural Networks for Solving Partial Differential Equations.

An Adaptive Sampling Algorithm with Dynamic Iterative Probability Adjustment Incorporating Positional Information

Pre-training strategy for solving evolution equations based on physics-informed neural networks

Physics-Informed neural network solver for numerical analysis in geoengineering

Wave simulation in non-smooth media by PINN with quadratic neural network and PML condition

Data-Guided Physics-Informed Neural Networks for Solving Inverse Problems in Partial Differential Equations

Causality-guided adaptive sampling method for physics-informed neural networks