In the present report, experimental investigations on mechanical behavior of unsaturated subgrade soil

with fiber reinforcement and lime stabilization were conducted.

The soil samples were collected from the soil/aggregate laboratory at the Maryland State Highway

Administration (MD-SHA), Maryland Department of Transportation. The experiments were carried out to

investigate physical and mechanical properties of subgrade soil that was mixed with geofiber and lime.

The reinforced and stabilized soil with fiber and lime is considered as a composite material.

Investigations in this study included two phases. In the first phase, the investigation was for static

behavior of composite subgrade soil under the compressive shear loading. In the second phase, the

investigation emphasized dynamic behavior of reinforced and stabilized soil under cyclic shear loading. In

this research, three aspects of investigations are presented.

First, new constitutive models for static and dynamic loading were established for composite material.

In the present report, elastic constitutive relationship was assumed to describe both the linear and

nonlinear shear stress-strain relations of composite material. Static behavior was described with a

nonlinear elastic model, and dynamic behavior was expressed with a linear elastic model. For the

nonlinear model, elastic shear modulus was assumed to be a function of multiple variables such as shear

strain, contents of fiber and lime, confining pressure, and the curing period of samples. In contrast, for the

linear model, elastic modulus was not only defined as a function of confining pressure, contents of fiber

and lime, and the aging period of samples but also repetitions of cyclic loading. For convenience of

experimental investigation, the shear stress-strain relation in a three-dimensional stress space reduced to

that in a quasi-triaxial stress space in which the conventional triaxial shear tests were conducted.

Second, experimental investigations and calibration of constitutive models were conducted. Experimental data from laboratory tests were utilized to verify and justify the linear or nonlinear elastic

model suggested in this report. Constitutive parameters of linear and nonlinear models were investigated

and calibrated using experimental results from both static and dynamic triaxial shear tests. The linear

regression method was adopted to find constitutive parameters. The constitutive relationship of the

composite material made of soil, fiber and lime was established once constitutive parameters for the linear

and nonlinear models were determined. The elastic shear moduli were investigated, for example, the

initial shear and tangential moduli in the nonlinear elastic model under static loading, and the shear or

resilient modulus in the linear model under cyclic loading. Moreover, for static loading, the Coulomb –

Mohr’s failure criterion was applied. The strength indices c and φ were studied for the composite soil with

fiber and lime. A linear relation was introduced to describe parameters c and φ as a function of fiber and

lime content, and the sample-curing period. The coefficients of the linear relation for parameters c and φ

versus fiber and lime content, and sample aging time were found using the experimental data.

Finally, impacts of fiber and lime as well as other factors that affect mechanical behavior of the

composite material were discussed. Impact factors on shear moduli were introduced for both the linear

and nonlinear models. The impact factors for the nonlinear model under static loading were defined to

exhibit effect of variables such as cell pressure, fiber and lime contents, and the sample-curing period on

the initial modulus and soil strength, namely parameters 1/A and 1/B related to the shear modulus G. In

contrast, the impact factors for the linear model under cyclic loading were introduced to show effect of the

same variables (σ0, mF, mL and t) plus the repetition of cyclic loading on resilient modulus and dynamic

behavior of composite soils.