Using a physics-informed neural network and fault zone acoustic monitoring to predict lab earthquakes

• Published in: Nature communications Vol. 14; no. 1; p. 3693
• Main Authors: Borate, Prabhav, Rivière, Jacques, Marone, Chris, Mali, Ankur, Kifer, Daniel, Shokouhi, Parisa
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 21.06.2023; Nature Publishing Group; Nature Portfolio
• Subjects: 639/766; 704/2151/2809; Acoustic tracking; Datasets; Deep learning; Earthquake prediction; Earthquakes; Elastic properties; Failure; Geological faults; Humanities and Social Sciences; Laboratories; Learning algorithms; Machine learning; multidisciplinary; Neural networks; Physics; Science; Science (multidisciplinary); Seismic activity; Tectonics; Training; Transfer learning
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-023-39377-6

Predicting failure in solids has broad applications including earthquake prediction which remains an unattainable goal. However, recent machine learning work shows that laboratory earthquakes can be predicted using micro-failure events and temporal evolution of fault zone elastic properties. Remarkably, these results come from purely data-driven models trained with large datasets. Such data are equivalent to centuries of fault motion rendering application to tectonic faulting unclear. In addition, the underlying physics of such predictions is poorly understood. Here, we address scalability using a novel Physics-Informed Neural Network (PINN). Our model encodes fault physics in the deep learning loss function using time-lapse ultrasonic data. PINN models outperform data-driven models and significantly improve transfer learning for small training datasets and conditions outside those used in training. Our work suggests that PINN offers a promising path for machine learning-based failure prediction and, ultimately for improving our understanding of earthquake physics and prediction. When attempting to predict laboratory quakes with a small amount of training data, a Physics-Informed Neural Network (PINN) outperforms purely data-driven models. PINN models also improve transfer learning when applied to a similar, yet differing dataset.