Train-induced vibrations provide useful information for passive seismic exploration. Such signals are repeatable and environment-friendly, and hence can provide a cost-effective way to analyze subsurface structure and estimate the medium parameters. We propose a workflow that use recurrent neural networks (RNN) and seismic interferometry to monitor the railway by producing real-time 1-D shear-wave velocities. In this paper, we first analyze the time-dependent character of the railway source and verify it by comparing the frequency spectrum of real data and the synthetic one. We then introduce the RNN-based surface-wave dispersion inversion method and validate the trained network using the overthrust model. Finally, seismic interferometry and RNN-based surface-wave inversion are applied to a synthetic record. The synthetic tests verify that train-induced vibrations can be a useful tool for real-time monitoring of the subsurface along the railways.
Presentation Date: Monday, October 12, 2020
Session Start Time: 1:50 PM
Presentation Time: 3:05 PM
Location: 351D
Presentation Type: Oral
