Laboratory and field evaluation of modulus-suction-moisture relationship for a silty sand subgrade

Abstract

A change in soil modulus in response to suction-moisture variation after compaction plays a vital role on modulus-based compaction control during earthwork construction. This study investigated the small-strain modulus-suction-moisture relationship for a silty sand subgrade using the laboratory free-free resonant frequency (FFR) and the in-situ spectral analysis of surface waves (SASW) in a trial section of highway construction project in Thailand. The post-compaction soil water retention curves (SWRCs) at different compaction states were investigated over the entire range of suction. Relationships between the SWRC parameters, i.e., air-entry suction, water-entry suction, residual suction etc., and as-compacted dry density and water content were obtained using the multiple-linear regression analysis. Such relationships are useful for the prediction of post-compaction suction in the field. The variation of SASW and FFR shear moduli with the suction stress, inferred from SWRCs, was found to be linear. A modified model was proposed, considering the combined effect of void ratio and the suction stress on the shear modulus, represented as Parameter A which was found to decrease as suction increased. The inclusion of suction stress and void ratio within the model variables yields better accuracy of prediction as compared to those models that considered either only suction or water content.

Introduction

Long-term performance of compacted earthfill is highly related to its suction-moisture regime which can significantly affect the corresponding mechanical responses, especially the stiffness and deformation characteristics. Stiffness measurements in pavement engineering can be performed based on a variety of techniques namely resilient modulus (M_R) tests using cyclic triaxial apparatus, in-situ automated plate load test [1], [2], [3], and non-destructive tests of small-strain modulus such as falling weight deflectometer (FWD), portable falling weight deflectometer (PFWD), light weight deflectometer (LWD), soil stiffness gauge (GeoGauge), spectral analysis of surface waves (SASW), intelligent vibratory roller compactors, etc [4], [5], [6], [7], [8]. More commonly adopted in pavement design and analysis, the resilient modulus (M_R) tests normally involve up to some hundred cycles of loading, typically in the strain range from 10⁻² to 10⁻¹% [1], [2], [3]. However, in-situ resilient modulus determination involved relatively complicated test set-up and time-consuming to carry out in the field.

In contrast, wave propagation methods used for determining small-strain modulus of geomaterials in a much smaller strain range of less than 10⁻³%, were mainly employed for construction quality control/assurance of compacted earthwork [4], [5], [6], [7]. In general, the tests can be performed instantly during the earthwork construction, resulting in an increasing number of inspection points and a better control of compaction uniformity as well as a comparison with the mechanistic design parameters on the basis of strain-dependent modulus degradation curve [9], [10]. Nevertheless, a limitation of modulus-based compaction control is that the modulus can be affected by suction-moisture conditions [11], [12], [13], [14], [15], [16], [17]. It has been generally recognized that without a proper understanding of suction and dry density influence on modulus, the field modulus measurements could be misleading and thus the target modulus may not be achieved.

The soil-water retention curve (SWRC), a.k.a. soil-water characteristic curve (SWCC), representing the relationship between soil water content and suction, is generally regarded as one of the important fundamental properties of unsaturated soils in practice [18], [19]. The SWRC can be used, directly or indirectly, to estimate the suction stress which represents the contributions of capillarity and adsorption water force on mechanical response of unsaturated soils such as shear strength [20], [21], [22], [23], [24], [25], and modulus [11], [12], [13], [14], [15], [16], [17]. The SWRC is fundamentally related to pore-size distribution of the soil, which is in turn influenced by various factors, such as soil structure and fabric [26], [27], stress level and stress history [28], [29], [30], [31], [32], [33], [34], initial moisture content, initial dry density and gradation [35], [36], [37], [38], [39], [40], [41], [42], [43], deformability of the soil [44], [45], [46] etc. A number of SWRCs for different soil types have been reported in the literature along with some predictive models [47], [48], [49], [50], [51], [52]. However, the test procedure adopted in most studies involved initial saturation of the compacted soil sample before gradually increasing suction from zero to the desired values, which is a conventional approach in the soil science discipline. Such test procedure does not mimic the field compaction condition, where the soil is unsaturated in its initial as-compacted state. It is very likely that the field condition follows the first wetting or first drying paths from the as-compacted state. This paper therefore investigates the effects of as-compacted moisture content and density on SWRCs of a silty sand subgrade from a highway construction project in Thailand [53], following a first wetting path and a first drying path over the whole range of suction (zero to 1,000,000 kPa). The small-strain shear moduli of the material having different suctions were investigated in the laboratory and in the field using the free-free resonant frequency (FFR) and the spectral analysis of surface waves (SASW) tests respectively, in order to develop the modulus-suction-moisture relationship for the compacted subgrade. Results from this study were used to improve the prediction model for shear moduli over the entire range of suction.