Plant roots provide mechanical reinforcement to soils upon shearing and seismic loading. The effects of roots distribution and their orientations on any changes in soil anisotropy in terms of shear strength, critical state and maximum shear modulus (G_max) are crucial to engineering design and analysis but have not been thoroughly studied. Existing failure criteria of rooted soils, that are predominantly derived based on the test results of direct shear, could not capture the strength anisotropy of rooted soils under general loading conditions. Moreover, whether, and how, roots provide resistance to soil liquefaction upon cyclic loading and the mechanisms of dissipated energy involved at liquefaction state have rarely been studied. Two series of drained and undrained monotonic triaxial t...[ Read more ]

Plant roots provide mechanical reinforcement to soils upon shearing and seismic loading. The effects of roots distribution and their orientations on any changes in soil anisotropy in terms of shear strength, critical state and maximum shear modulus (G_max) are crucial to engineering design and analysis but have not been thoroughly studied. Existing failure criteria of rooted soils, that are predominantly derived based on the test results of direct shear, could not capture the strength anisotropy of rooted soils under general loading conditions. Moreover, whether, and how, roots provide resistance to soil liquefaction upon cyclic loading and the mechanisms of dissipated energy involved at liquefaction state have rarely been studied. Two series of drained and undrained monotonic triaxial tests upon compression and extension stress path were conducted on soils reinforced by the roots of a deep-rooted species (vetiver grass, (Chrysopogon zizanioides L.) to investigate the shear strength of soil and to aid the development of a new generalised 3-D anisotropic failure criterion for rooted soils. The underlying mechanism of liquefaction of rooted soil was investigated by the data of a series of undrained cyclic triaxial tests on rooted soils. A series of bender element tests on rooted soils were carried out following the isotropic loading and unloading paths, aiming to explore the effects of roots on (G_max) and to develop new semi-empirical equations of (G_max) for rooted soils.

It was discovered that the root reinforcement effect was anisotropic and path dependent. Roots with predominant orientation aligning in the tensile strain direction contribute the most to soil strength. In the case of vetiver grass, which has a taproot system, their roots show the strongest reinforcement effect in conventional triaxial extension path, in which the maximum portion of roots are subjected to tension. Comparing to bare soil, rooted soils were less compressible, and its recoverable elastic deformation were also smaller. Interestingly, the critical-state lines (CSLs) of bare and rooted soils in the υ-Lnp' space for the compression and extension paths converged. However, the gradient of the CSL in the q-p' space is dependent upon the stress path due to the fabric anisotropy and anisotropic distribution of roots in the soil.

A new generalised 3-D anisotropic failure criterion was derived for rooted soils. The projection of the microstructure fabric tensors of soil and root network on stress tensors to address the anisotropic effects of root network and soil fabric on the shear strength parameters of rooted soils upon various effective stress paths. This model addressed anisotropies of both cohesion and friction angle and explains why most of studies by direct shear reported that roots affect the cohesion but not the friction angle. Indeed, this stress paths are within section I of the deviatoric plane, where the effects of soil anisotropy on friction angle are not noticeable.

The presence of roots in cultivated samples increased the (G_max) , whereas that in the artificially rooted samples reduced its (G_max) . A new semi-empirical model was derived to address the effects of roots on (G_max) of high compressible cultivated rooted soil by incorporating elastoplastic constitutive relationships, which took account the effects of mean effective stress and plastic volume change hardening. Additionally, A new equation was devised for sandy soil since the cultivated (G_max) model was ineffective for artificially rooted soil due to difficulties in locating the isotropic consolidation line and negligible impact of over consolidation on (G_max) .

The liquefaction resistance of artificial samples was improved with an increase in root volume, and this improvement was more remarkable at higher cyclic stress ratios. The grass roots, beyond certain volume, prevented the soil from experiencing limited flow failure that occurred in unreinforced sand, and switched the failure mode to cyclic mobility. Normalised cumulative dissipated energy (ΣΔW/σ_c’) of rooted soil depended on cyclic stress ratios (CSRs) and RVRs and this energy is uniquely correlated with the cyclic resistance ratio at the cycle number of 15 (CRR15). It was discovered that roots that were predominantly orientated in the direction perpendicular to the major principal stress of extensive path reduced soil anisotropy upon cyclic loading. Besides, the ΣΔW/σ_c’ was linearly correlated with the normalized cumulative strain energy (Σ4W/σ_c’) with a gradient of approximately 2.