Assessing the Performance of Liquefaction Settlement Models Following Blast Testing in Silty Sands

Soil liquefaction can significantly impact the settlement and load-bearing capacity of soil, which in turn affects structural stability and safety. The increase in pore pressure from seismic events and its subsequent dissipation can lead to settlement, impacting the stability of existing structures. However, quantifying these effects remains challenging due to the unpredictable nature of earthquake loading, the reliance on laboratory-based models that may not fully capture in situ behavior, and the difficulties associated with establishing correlations between pre-earthquake conditions and post-seismic deformations. To address these challenges, this study utilizes settlement data from blast-induced liquefaction testing at a field site in Christchurch, New Zealand, offering a unique opportunity to observe and evaluate liquefaction effects under controlled conditions. Multichannel analysis of surface waves, dynamic cone penetrometer (DCP), and terrestrial light detection and ranging coupled with a profilometer were deployed to measure shear-wave velocity (VS), dynamic cone penetration resistance (qd), and ground settlement (S), respectively, before and after blasting. The findings revealed immediate significant reductions in VS and DCP resistance post-blast, suggesting strain accumulation and soil softening in the affected layers. Comparisons of observed settlement with predictions from multiple settlement models were conducted. The cone penetration test-based model predictions closely matched the measured settlement, with a predicted settlement of 4.18 cm compared to a measured value of 4.66 cm. In contrast, the DCP-based approach, which used correlations between cone tip resistance (qc) and dynamic cone resistance (qd), tended to underpredict settlements, yielding estimated settlements around 1.8 cm. The VS-based approach significantly overpredicted settlement, with estimated settlements reaching ∼21 cm. These comparisons highlight the variability in predictive performance when applying different in situ measurements.