Seismic traveltime simulation for variable velocity models using physics-informed Fourier neural operator

2023/11/07 by Chao Song, Tianshuo Zhao, Song, Chao +6 · 5 citations
Computer Science · Earth and Planetary Sciences · #FOS: Physical sciences #Geophysics (physics.geo-ph) #Seismic Imaging and Inversion Techniques #Seismic Waves and Analysis #Seismology and Earthquake Studies

paper · pdf · doi:10.48550/arxiv.2311.03751

openalex publication_date 2023/11/07 · openalex created_date 2023/11/09 · openalex updated_date 2026/07/28

Abstract

Seismic traveltime is critical information conveyed by seismic waves, widely utilized in various geophysical applications. Conventionally, the simulation of seismic traveltime involves solving the eikonal equation. However, the efficiency of traditional numerical solvers is hindered, as they are typically capable of simulating seismic traveltime for only a single source at a time. Recently, deep learning tools, particularly physics-informed neural networks (PINNs), have proven effective in simulating seismic traveltimes for multiple sources. Nonetheless, PINNs face challenges such as limited generalization capabilities across different models and difficulties in training convergence. To address these issues, we have developed a method for simulating multi-source seismic traveltimes in variable velocity models using a deep-learning technique, known as the physics-informed Fourier neural operator (PIFNO). The PIFNO-based method for seismic traveltime generation takes both velocity and background traveltime as inputs, generating the perturbation traveltime as the output. This method incorporates a factorized eikonal equation as the loss function and relies solely on physical laws, eliminating the need for labeled training data. We demonstrate that our proposed method is not only effective in calculating seismic traveltimes for velocity models used during training but also shows promising prediction capabilities for test velocity models. We validate these features using velocity models from the OpenFWI dataset.

Cited by

• Diffusion priors enhanced velocity model building from time-lag images using a neural operator
• Microseismic event classification with a lightweight Fourier Neural Operator model
• Learning from Imperfect Labels: A Physics-Aware Neural Operator with Application to DAS Data Denoising
• An effective physics-informed neural operator framework for predicting wavefields
• Equivariant Eikonal Neural Networks: Grid-Free, Scalable Travel-Time Prediction on Homogeneous Spaces

Related

• Seismic first-arrival traveltime simulation based on reciprocity-constrained PINN
• Solving Seismic Wave Equations on Variable Velocity Models with Fourier Neural Operator
• PINNtomo: Seismic tomography using physics-informed neural networks
• PINNPStomo: Simultaneous P- and S-wave seismic traveltime tomography using physics-informed neural networks with a new factored eikonal equation
• Deep-learning inversion: a next generation seismic velocity-model building method