Characterisation of Resilient Modulus for Unbound Granular Materials Using Quick Shear Strength Test

Abstract

<title>Abstract</title> This study investigates the correlation between the resilient modulus (M _R ), a critical parameter in mechanistic-empirical pavement design, and the Quick Shear Strength (QSS) at 1% strain, using remoulded Unbound Granular Materials (UGMs) sourced from various regions across Tanzania. M _R characterises the elastic (recoverable) stiffness of UGMs under repeated loading, typically measured using the Repeated Load Triaxial (RLT) test, a procedure that is both complex and time-consuming. In contrast, the QSS test provides a measure of the shear strength of UGMs under confining pressure and offers a simpler, more cost-effective alternative. While previous research has explored correlations between M _R and the California Bearing Ratio (CBR), the predictive reliability of CBR remains questionable. This is primarily due to the fundamental differences between the two parameters: CBR is stress-independent, whereas M _R is stress-dependent. Additionally, significant differences exist in specimen size and the application of confining pressure between the two test methods. In this study, a strong correlation was established between M _R and QSS results, tested under three moisture conditions 0.75 OMC, OMC, and 1.15 OMC to simulate field moisture and humidity variations. The analysis was based on 20 UGM samples commonly used in road construction across Tanzania. The developed M _R -QSS model demonstrates a high level of reliability, with a coefficient of determination (R²) of 0.99. The proposed model offers a practical and consistent alternative for estimating M _R , significantly reducing testing time while maintaining consistency in specimen size and confining pressure. It provides a more accessible and reliable substitute for both conventional M _R testing and traditional M _R -CBR correlations. Based on the research findings, critical observations are also provided regarding the comparative suitability of the M _R -QSS model versus existing M _R -CBR models.