Efficacy of deep replacement (cement-stabilised sand column) length in mitigating liquefaction: shaking table tests under earthquake loading

Abstract

This research examines the impact of deep mixing column length on its effectiveness in mitigating soil liquefaction, a significant concern in geotechnical engineering, particularly during earthquakes. The study used a series of shaking table tests to simulate earthquake conditions, specifically referencing the 1999 Kocaeli earthquake. Cement-stabilised sand columns of three different lengths (15 cm, 25 cm, and 35 cm) were installed in a triangular arrangement within medium-dense Firouzkoh sand, and scaled models of 10- and 15-story buildings were placed on top. The experiment followed Iai’s scaling law (λ=20) to ensure realistic simulations. Results clearly showed that longer columns perform much better at reducing excess pore water pressure and limiting soil liquefaction. Shorter columns only provided minor improvements compared to untreated soil, while longer columns prevented full liquefaction and kept ground settlement below 1 cm; by contrast, shorter columns allowed over 2.5 cm of settlement. The stiffer, longer columns also increased the composite shear modulus and reduced shear strains. However, post-test analysis revealed bending cracks in the longest columns, indicating significant bending moments. In summary, the study highlights that columns should reach at least 80% of the liquefiable layer and must be designed to handle seismic-induced bending for optimal performance.