Abstract

Capturing the spatial variability in soil is crucial for ground response analyses in the context of seismic hazard mitigation. The lateral variability in thickness and properties of the different soil layers is one of the main factors that determines the variability of the ground motion spectrum from one location to another. The absence of such lateral variability information in the subsoil in between the locations of Cone Penetration Tests (CPTs) may be compensated by the use of more densely sampled seismic data. In this research we aim to derive a shear-wave velocity field through seismic full-waveform inversion that yields a model resolution approaching that of high-resolution seismic CPT surveys. Following this, a datadriven correlation between geophysical and geotechnical information is attempted through the application of new machine-learning-based approaches.

Abstract

Non-invasive site characterization techniques have the potential to rapidly evaluate large subsurface volumes to guide subsequent invasive geotechnical site investigation. Among these methods, seismic full waveform inversion (FWI) stands out for its potential to recover detailed two-dimensional (2D) images of the subsurface. However, FWI’s need for substantial computational resources and sensitivity to the initial starting model has limited its utilization as a generalpurpose geotechnical site characterization tool. Addressing this, prior studies have shown data-driven methods can predict 2D subsurface structures composed of soil over rock. In the present study, we aim to generalize these findings to all nearsurface conditions. We propose a novel model generation framework that utilizes techniques from geostatistics to generate complex 2D subsurface models. The generated models include dipping soil and rock layers, soil lenses, boulders, and underground utilities; none of which have been considered previously. We use our model generation framework to simulate 100,000 2D subsurface models. We simulate field data acquisition along these 100,000 synthetic models, by numerically solving the elastic wave equation using an impulse source at the model’s center surrounded by 24 receivers (12 on either side). The data-driven predictive model, trained on 90% of the simulated data, achieved a mean absolute percent error on the testing set of 19%. Furthermore, these predictions are made within fractions of a second circumventing the computational and starting-model-related challenges associated with traditional 2D FWI. These results demonstrate that data-driven methods can predict complex images of the subsurface to enable rapid subsurface imaging for geotechnical applications.

Abstract

Non-invasive site characterization techniques have the potential to rapidly evaluate large subsurface volumes to guide subsequent invasive geotechnical site investigation. Among these methods, seismic full waveform inversion (FWI) stands out for its potential to recover detailed two-dimensional (2D) images of the subsurface. However, FWI’s need for substantial computational resources and sensitivity to the initial starting model has limited its utilization as a generalpurpose geotechnical site characterization tool. Addressing this, prior studies have shown data-driven methods can predict 2D subsurface structures composed of soil over rock. In the present study, we aim to generalize these findings to all nearsurface conditions. We propose a novel model generation framework that utilizes techniques from geostatistics to generate complex 2D subsurface models. The generated models include dipping soil and rock layers, soil lenses, boulders, and underground utilities; none of which have been considered previously. We use our model generation framework to simulate 100,000 2D subsurface models. We simulate field data acquisition along these 100,000 synthetic models, by numerically solving the elastic wave equation using an impulse source at the model’s center surrounded by 24 receivers (12 on either side). The data-driven predictive model, trained on 90% of the simulated data, achieved a mean absolute percent error on the testing set of 19%. Furthermore, these predictions are made within fractions of a second circumventing the computational and starting-model-related challenges associated with traditional 2D FWI. These results demonstrate that data-driven methods can predict complex images of the subsurface to enable rapid subsurface imaging for geotechnical applications.

Abstract

The multi-channel analysis of surface waves (MASW) is a widely employed surface wave method. The fundamental mode inversion is the usual scheme in MASW, because separating the fundamental mode from other high modes is feasible with an adequately long array. However, the extracted fundamental mode represents only part of the dynamic characteristics, it may deviate from the theoretical fundamental mode due to truncation effect, and part of it may correspond to another higher mode or leaky mode, resulting in mode misidentification. Conversely, the spectral analysis of surface waves (SASW) method features a more rigorous inversion scheme by matching the effective mode, but it may suffer from the tricky phase-unwrapping in the dispersion data reduction. This study introduces an improved dynamic response solution for elastic-layered media subjected to vertical loads. The proposed dynamic response solution is fast and accurate, facilitating the full wavefield inversion in terms of the frequency-velocity spectrum. The MASW frequencyvelocity spectrum inversion considers testing configuration and comprehensively models all wave phenomena, including near-field effect, truncation effect, and leaky waves. The MASW frequency-velocity spectrum inversion merges the convenience of MASW dispersion analysis with the rigorous inversion scheme adopted by SASW. The new MASW frequency-velocity spectrum inversion is compared with the SASW effective mode inversion and MASW fundamental mode inversion. The results show that both SASW effective mode inversion and MASW frequency-velocity spectrum inversion produce better inverted results than the MASW fundamental mode inversion, while the process of MASW frequency-velocity spectrum inversion is more convenient and robust. Finally, a field example is used to demonstrate the applicability of frequency-velocity spectrum inversion.

Abstract

Understanding subsurface conditions is critical to creating and maintaining resilient infrastructure systems, such as dams and levees. Seismic geophysical tools can be very effective for site characterization of these structures as they directly measure the elastic moduli and can provide insight into both the soil properties and groundwater conditions. Full waveform inversion (FWI) is one processing option for seismic geophysics that seeks to overcome some of the limitations in the traditional approaches by using the full time-domain recording of the wavefield to develop 2D or 3D profiles of shear wave velocity. In addition to providing characterization data, FWI can also potentially be used as a monitoring tool for dams and levees to assess how elastic moduli are changing with time and to infer how these changes might relate to changes in the hydromechanical properties of the soil. This study seeks to explore the use of seismic FWI as both a characterization and monitoring tool through numerical simulations of seismic surveys on a hypothetical levee with a low velocity anomaly in the foundation. The simulations are used to assess both the spatial resolution and the ability of the simulations to detect changes in properties that might be related to softening of the foundation or development of internal erosion failure modes. The findings from the study will be used to highlight potential benefits and challenges to using seismic FWI for characterization and monitoring of dams and levees.