Search for a command to run...
Development of infrastructure in areas with poor subsurface soil conditions is often encountered in civil engineering. In these situations, the engineering characteristics of problematic soils can be improved by ground improvement techniques tailored to site-specific constraints. Jet grouting is among the most widely used methods to stabilize poor subsurface soils, particularly for underpinning of existing structures and cutoff walls for excavation support. During jet grouting, a stabilizing fluid (cementitious or chemical grout) is injected into the surrounding soil through small nozzles under very high pressure and velocity. The rotary motion of the injection nozzle creates columnar elements of soilcrete as the nozzle is withdrawn from the subsurface. The columns are often overlapped to create a wall or prevent horizontal movement of unstable soils and/or groundwater. However, local variations in the subsurface conditions can lead to variability in the column diameter. This in turn can cause untreated zones where the columns cease to overlap. Typical quality assurance efforts rely on real-time recording of jet grout parameters such as fluid pressure, flow rate, rotational speed, and rate of withdrawal/insertion. This is combined with a comprehensive test program prior to production efforts that establishes the jet grout column spacing, overlap, and overall depth geometry, typically using destructive methods such as excavations and corings of the soilcrete columns. However, these efforts do not allow for a comprehensive site-wide assessment of the as-constructed column geometry given the significant spatial variability present in the subsurface of many sites. Consequently, there exists demand for a high-resolution non-destructive testing (NDT) method that can reliably identify untreated zones caused by lack of column overlap after jet grouting. This paper numerically examines the use of full waveform inversion (FWI) of seismic data as a potential NDT approach for quality assurance of jet grouting efforts. FWI seeks to optimize the misfit between synthetic waveforms and observed seismic data. An initial model is subsequently updated by a linearized inversion procedure for optimizing the wave propagation properties of the subsurface soils. The current study adopts the spectral element method as a robust forward modeling technique to solve the wave propagation equation and updates the models using a quasi-newton approximation of the inverse Hessian to lower the computational cost of the process. The results of the current study highlighted the potential of the FWI technique to reliably image anomalous conditions in the overlapping jet grout column and detect any untreated zones.