Direct shear strength testing of unsaturated soils
Abstract
The more rigorous approach on the behavior of unsaturated soils using two independent stress state variables is increas_ingly gaining acceptance. The proposed shear strength equation based on this approach (i.e., τ = c' + (ua - u w tan φ b + (σ - ua) tan φ' ), requires that the stress circle corresponding to the failure conditions be plotted on a three dimensional diagram. In this diagram, the axes in the horizontal plane are the stress state variables,(σ - ua) and (ua - uw), and the ordinate axis is the shear strength, τ. The failure conditions define a surface. If one stress variable is maintained constant, say the variable (σ - ua), then the relationship between τ and (u a - uw can be portrayed on a two dimensional plot. The resulting envelope in this plane of τ versus (ua - uw will be called the suction envelope.
The validity of the form of the proposed shear strength equation has been verified by results from laboratory testing of unsaturated soils where the stress state variables at failure are known. So far, these results indicated that the failure surface is planar, i.e, the suction envelope is linear. Results also showed that the friction angle,φ b, associated with the suction envelope is always smaller than the angle of shearing resistance, φ'.
A completely constant φb (i.e., a linear suction envelope) , however, is found to be inconsistent with pieces of experimental evidence obtained from tests on saturated soils under suction. The linear suction envelope also frequently results in an abrupt break in the failure surface when crossing from positive pore-water pressure over to negative pore-water pressure. The result is that there is not a smooth transition from soil with positive pore-water pressure to soil with negative pore-water pressure.
A theoretical investigation of unsaturated soils from a phenomenological standpoint predicts a curved suction envelope comprising of two linear sections connected by a transition curve. The angle of the slope of the initial linear section is equal to Ø' while the angle of the slope of the latter linear section, Ø b, is smaller than Ø'. An experimental program of multistage direct shear tests on a Saskatchewan till was conducted. Theses tests were carried out over a suction range of 0 kPA to 500 kPa. Results confirmed that the suction envelope is nonlinear and that the slope of this envelope varies from Øb = Ø = 25.5 degrees, to Øb = 7 degrees constant.
For the investigation in this thesis, a direct shear box was modified, utilizing the axis translation technique, to render the equipment suitable for unsaturated soil testing. There are two reasons for wanting a direct shear
apparatus capable of testing unsaturated soils. First, unsaturated soil testing using the triaxial apparatus is complicated and takes a long time. Second, larger strain is possible in the direct shear which makes is more suitable for multistage testing. Multistage testing is desirable to reduce the number of variables involved.