Regression Model for Soil Compaction Assessment via Shear Wave Velocity in Kaolin-sand Mixtures

Abstract Ensuring uniform soil compaction is critical for the long-term performance of transportation infrastructure. Conventional density-based methods remain destructive, time-consuming, and provide limited spatial representation. This study aims to establish a fines-sensitive regression model to predict shear wave velocity (Vs) as a non-destructive indicator of compaction quality, addressing the research question: Can Vs reliably predict compaction parameters by incorporating both dry density and fines content in kaolin–sand mixtures? A laboratory investigation was conducted on kaolin–sand mixtures (K100, K70, K50, K30) prepared under Standard Proctor conditions. Compaction characteristics, including Maximum Dry Density (MDD) and Optimum Moisture Content (OMC), were obtained. Bender Element (BE) testing, enhanced with a claydough–foam coupling interface to improve signal clarity, was performed to measure Vs. Multiple linear regression was applied to model Vs as a function of fines content (FC) and dry density (ρ d ). Strong correlations were observed between Vs and compaction parameters, with R² values of 0.9978, 0.7723, 0.7139, and 0.4071 for K70, K50, K100, and K30, respectively. The proposed model, Vs = 160.11 − 0.992FC + 0.1107ρ d , achieved high predictive accuracy (R² = 0.9207), demonstrating fines-sensitive stiffness prediction capability. The findings confirm that Vs can serve as a reliable non-destructive indicator for evaluating soil compaction, particularly in fine–coarse blends. The developed model lays essential groundwork for integrating Vs-based assessment into intelligent compaction and field quality control practices.