A dimensional description of the unconfined compressive strength of artificially cemented fine-grained soils

Abstract

This study aims at establishing a universal predictive model for the unconfined compressive strength (UCS) of artificially cemented fine-grained soils. Model development, its validation and calibration were carried out using a comprehensive database gathered from the research literature. The dimensional analysis concept was successfully extended to the soil–cement UCS problem, thereby leading to a practical dimensional model capable of simulating the UCS as a function of the blend’s index properties — that is, cement content, specific surface area, curing time, and the compaction state parameters (including water content and dry density). The predictive capability of the proposed model was examined and further validated using routine statistical tests, as well as conventional fit-measure indices which resulted in R2 > 0.95 and NRMSE < 5%. A sensitivity analysis was also carried out to quantify the relative impacts of cement content, curing time and soil plasticity on the UCS. The higher the soil plasticity, the higher the positive sensitivity to cement content, implying that soils of higher plasticity would require higher cement contents for stabilization. On the contrary, the higher the soil plasticity, the lower the positive sensitivity to curing time, indicating a more effective cement hydration in soils of lower plasticity. Finally, an explicit calibration procedure, involving a total of three UCS measurements for three recommended soil–cement mix designs, was proposed and validated, thus allowing for the proposed model to be implemented with confidence for predictive purposes, preliminary design assessments and/or soil–cement optimization studies.