Large-Scale 3D Ground-Motion Synthesis with Physics-Inspired Latent Operator Flow Matching

2026-03-18

AI Summary

This work addresses the high computational cost of traditional physics-based ground-motion simulations, which hinders large-scale infrastructure uncertainty analyses requiring vast ensembles of spatially and temporally consistent seismic time histories. To overcome this limitation, the authors propose Ground-Motion Flow (GMFlow)—a latent-space implicit operator flow-matching framework that integrates physical priors to efficiently generate regional three-dimensional ground-motion fields conditioned on source and site parameters. By uniquely combining physics-informed latent operators with flow matching, GMFlow enables mesh-independent generation of large-scale spatiotemporal physical fields. In a case study of the San Francisco Bay Area, GMFlow produces high-fidelity ground-motion simulations across more than 9 million grid points in seconds, achieving a speedup of approximately 10⁴ compared to conventional methods, thereby substantially enhancing the efficiency and scalability of seismic risk assessment.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Computer Vision: Low Level & Physics-based Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Earthquake hazard analysis and design of spatially distributed infrastructure, such as power grids and energy pipeline networks, require scenario-specific ground-motion time histories with realistic frequency content and spatiotemporal coherence. However, producing the large ensembles needed for uncertainty quantification with physics-based simulations is computationally intensive and impractical for engineering workflows. To address this challenge, we introduce Ground-Motion Flow (GMFlow), a physics-inspired latent operator flow matching framework that generates realistic, large-scale regional ground-motion time-histories conditioned on physical parameters. Validated on simulated earthquake scenarios in the San Francisco Bay Area, GMFlow generates spatially coherent ground motion across more than 9 million grid points in seconds, achieving a 10,000-fold speedup over the simulation workflow, which opens a path toward rapid and uncertainty-aware hazard assessment for distributed infrastructure. More broadly, GMFlow advances mesh-agnostic functional generative modeling and could potentially be extended to the synthesis of large-scale spatiotemporal physical fields in diverse scientific domains.

Problem

• ground-motion synthesis
• spatiotemporal coherence
• uncertainty quantification
• large-scale simulation
• earthquake hazard analysis

Innovation

• physics-inspired generative modeling
• latent operator learning
• flow matching
• spatiotemporal coherence
• large-scale ground-motion synthesis