Adaptive-Sampling Physics-Informed Neural Network for Viscoacoustic Wavefield Simulation
- 24 June 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
Abstract
Seismic wave forward simulation is the basis of seismic inversion and imaging. In recent years, deep learning methods represented by physics-informed neural networks (PINN) have been widely applied in the field of scientific computing, and have obtained significant attention as mesh-free methods in seismic inversion. Currently, the training of PINN is usually constructed by fixed points in seismic wavefield simulation, and the large number of training points can result in high computational costs. This paper proposes an adaptive strategy for PINN training points selection based on the velocity distribution and residuals of viscoacoustic wave equations. This strategy initializes the positions and quantities of training points based on the velocity distribution of the model, and dynamically adjusts the training points based on the values of loss function during the calculation process. The proposed approach reduces the number of training points, thereby improving the training efficiency of PINN. Numerical examples including the homogeneous model, the layered model with irregular topography, and the Marmousi2 model show that the proposed sampling strategy effectively reduces the calculation time by 54.37% while ensuring accuracy.Keywords
Funding Information
- National Key Research and Development Program of China (2021YFA0716901)
- National Natural Science Foundation of China (41974132)
- Fundamental Research Funds for the Central Universities (xzy012023050)
This publication has 17 references indexed in Scilit:
- Physically Driven Self-Supervised Learning and its Applications in Geophysical InversionIEEE Transactions on Geoscience and Remote Sensing, 2024
- Helmholtz-equation solution in nonsmooth media by a physics-informed neural network incorporating quadratic terms and a perfectly matching layer conditionGeophysics, 2023
- Simulating seismic multifrequency wavefields with the Fourier feature physics-informed neural networkGeophysical Journal International, 2022
- Deep Learning Optimized Dictionary Learning and Its Application in Eliminating Strong Magnetotelluric NoiseMinerals, 2022
- A versatile framework to solve the Helmholtz equation using physics-informed neural networksGeophysical Journal International, 2021
- Solving the frequency-domain acoustic VTI wave equation using physics-informed neural networksGeophysical Journal International, 2021
- Domain Decomposition for Large-Scale Viscoacoustic Wave Simulation Using Localized Pseudo-Spectral MethodIEEE Transactions on Geoscience and Remote Sensing, 2020
- Seismic stratigraphy interpretation by deep convolutional neural networks: A semisupervised workflowGeophysics, 2020
- Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equationsJournal of Computational Physics, 2019
- Fast generation of 2-D node distributions for mesh-free PDE discretizationsComputers & Mathematics with Applications, 2015