Full waveform inversion to evaluate liquefaction triggering in soils exhibiting natural spatial stiffness variability

Abstract

Liquefaction triggering procedures based on the standard penetration test (SPT) and cone penetration test (CPT) each have seen a significant history of development. More recently, liquefaction triggering methodologies based on shear wave velocity (VS ) from geophysical measurements have been developed. Typically, estimates for VS in these procedures have been acquired using downhole geophysical methods. SPT, CPT, and downhole geophysical techniques are point sources of information that only provide data in the localized region surrounding the measurement location. Therefore, these methods may not provide sufficient information regarding liquefaction triggering in soils that exhibit appreciable natural spatial variability with respect to stiffness. Surface wave methods can cover a larger spatial area and address this concern. However, surface wave methods suffer from uncertainty and spatial averaging due to the wavefield transformations used to evaluate dispersion characteristics of the underlying soils and the inherent one-dimensional assumption built into typical inversion algorithms. This study numerically modeled the propagation of surface waves in a spatially correlated Gaussian random field to simulate the effects of natural soil variability on data acquired using the multichannel analysis of surface waves (MASW). The goal was to study the capabilities of a full waveform inversion (FWI) approach when used to evaluate liquefaction triggering in spatially variable soil conditions. The results demonstrated that a FWI approach outperforms the typical dispersion-based MASW approach when implementing VS -based liquefaction triggering procedures in spatially variable soil conditions.