Integrating Physics-Informed Neural Networks for Earthquake Modeling: 2D Verification, Applications

Written by seismology | Published 2024/08/02
Tech Story Tags: neural-networks | earthquake-modeling | seismic-hazard | fault-friction-parameters | inverse-problems | rate-and-state-friction | computational-seismology | deep-learning-in-geophysics

TLDRThis paper is available on arxiv.org/abs/2312.09403 under CC BY 4.0 DEED license. The physics-informed deep learning framework is able to solve both forward and inverse problems to reasonable accuracy. Verification is an essential first step to ensure credible results 20, 12.via the TL;DR App

Authors:

(1) Cody Rucker, Department of Computer Science, University of Oregon and Corresponding author;

(2) Brittany A. Erickson, Department of Computer Science, University of Oregon and Department of Earth Sciences, University of Oregon.

Table of Links

Abstract and 1. Context and Motivation

  1. Physics-Informed Deep Learning Framework
  2. Learning Problems for Earthquakes on Rate-and-State Faults
  3. 2D Verification, Validation and Applications
  4. Summary and Future Work and References

4. 2D Verification, Validation and Applications

When computational methods for physical problems are used to address science questions, verification is an essential first step to ensure credible results [20, 12]. While validation with observational data is the focus of future work, we must first verify that our physics-informed deep learning framework is able to solve both forward and inverse problems to reasonable accuracy.

4.1. Verification with the Method of Manufactured Solutions

This paper is available on arxiv under CC BY 4.0 DEED license.


Written by seismology | Uncovering the secrets of the Earth's core, Seismology shapes resilient foundations for a sustainable tomorrow.
Published by HackerNoon on 2024/08/02