BEHAVIOUR OF STRIP FOOTINGS ADJACENT TO REINFORCED EARTH SLOPES

Abstract

Ground improvement in weak soils and in difficult site conditions becomes necessary in view of heavy loads imposed by industrial structures; high rise buildings etc. A number of ground improvement techniques are available for improving the load carrying capacity and engineering properties of soils and sites which otherwise are not suitable for normal construction. CIVILENGINEERING structures are often forced to be constructed on sloping ground. This trend is more marked in areas where mountains dominate, and in some special situations. Examples of such structures are buildings or pads constructed in hilly regions, foundations for bridge abutments on granular fill slopes, transmission towers etc. Highway overpass bridges frequently connect with approach embankments that terminate in a sloping face dropping down to the underpass level. The bearing capacity of a foundation located near the edge of a slope assumes its importance in view of the fact that performance of such structures depends on the stability of the slope and the allowable soil pressure. These foundations when subjected to axial loading result in a reduction of ultimate bearing capacity as compared to that on a flat ground surface. The edge distance and the slope angle further affect the stability of a foundation constructed on top of a slope. The estimation of ultimate bearing capacity and the load settlement behaviour of foundations on or near the slope are important to the safe and efficient design of such structures. Several procedures have been proposed to enhance the performance of foundations located on or near the crest edge of slopes. These range from the incorporation of deep deposits of granular fill material to the incorporation of piles and high tensile strength reinforcing materials to supplement the load carrying capacity of the foundation. An understanding of the behaviour of a surface footing on reinforced slopes is of practical importance to geotechnical engineers. To investigate and propose a method for bearing capacity improvement of strip footing resting on reinforced cohesionless soil slope on a competent foundation soil and 11 identify the factors affecting the performance of the footing. The problem taken up was bifurcated into two parts: (i) estimation of improvement in bearing capacity of strip footings resting adjacent to a stable cohesionless soil slope by the inclusion of geogrid reinforcement layers, and (ii) estimation of bearing capacity of strip footing resting adjacent to a geogrid reinforced cohesionless steep slope which is unstable (FOS<l) when unreinforced. (i) The analysis proposed for estimating the improvement in bearing capacity of a strip footing resting adjacent to reinforced cohesionless stable slope is based on the approach developed by Binquet and Lee (1975b), and modified by Kumar (1997). Prior to developing the analysis, it is required to have pressure - settlement curve for the strip footing adjacent to unreinforced slope. A methodology has been developed to predict pressure-settlement characteristics of strip footings placed adjacent to the slopes and subjected to vertical loads using constitutive law of soils. Sharan (1977) studied the behaviour of strip footing resting on flat ground consisting of both clayey and sandy soils using non-linear constitutive laws. Initially the footing is analyzed considering as a flexible footing, and later the results are interpreted as that of rigid footing. For a given pressure intensity, the maximum difference between the average settlement of a flexible footing and the settlement of a rigid footing was found to be less than 3.5%. Advantage of this concept has been used in the present analysis. The method makes use of elasticity theory only for stresses calculation; the stress-strain relation required for the subsequent calculations is obtained from the experimental results (triaxial test results) of the actual soil. This concept of obtaining the stress components in soil mass beneath the strip footing resting on top of slope was verified by comparing predicted values with the experimental results of vertical stress observed along the central line of footing in a fullscale unreinforced test embankment by a strip footing load at various depths beneath the footing, as reported by Bathurst et al. (2003). The procedure used to compute the stress components below the footing in developing the pressure - settlement characteristics has been used to obtain the non-dimensional force and length coefficients, which are used in the Binquet and Lee (1975b) method to determine the (BCR) bearing capacity ratio i.e., ratio of bearing capacity of reinforced slope to the bearing capacity of unreinforced slope in at the same settlement. The non-dimensional coefficients have been presented in the form of charts for ready use. These coefficients are sensitive to number of factors, like slope angle, edge distance of footing from crest, width of footing, depth of reinforcement layer, height of slope, and density of fill soil unlike flat ground. An attempt has been made to derive the equations for the dimensionless coefficients by regression analysis. The factor of safety (FOS) of the reinforced slope with different footing pressure intensities and at different locations obtained after inclusion of reinforcement layers were checked using commercial software SLOPE/W. The procedure developed has been illustrated by an example. To validate the proposed approach, plain strain small scale tests were performed on the model footing in the laboratory. The model slopes were constructed of locally available Ranipur sand with angle of internal friction as 35°. The slopes were constructed in a steel tank with one lateral side provided with Perspex sheet, so as to observe the pattern of failure of soil below the footing on loading. Model slopes were 900mm high, 607mm wide and 3000mm long. Model tests on three unreinforced slopes (34°, 24°, and 21° with the horizontal) and one reinforced slope (34°), with footing located at different edge distances and different number of reinforcement layers were conducted. The model footing was made of 25mm thick steel plate 150mm wide and 607mm long. The footing was glued with sand paper cloth at the bottom so as to ensure rough base simulating the field conditions. All the tests were performed under plain strain condition. The reinforcement used in the experimental investigation was stratagrid (SGI50) supplied by M/S STRATA Systems India Pvt. Ltd. The load was applied by screw jack. For establishing the constitutive laws for both unreinforced and reinforced sand used in the investigation, triaxial tests were performed on oven dry sand samples of size100 mm diameter and 200 mm height, with varying number of reinforcement layers. The tests were conducted under confining pressures of 150, 250, 350 and 500kN/m . It was observed that the influence of reinforcement was on increase of soil stiffness. Large size direct shear tests were performed in shear box of dimensions 300mm x 300mm x 200mm, underthe normal stress of 50, 100, and 150kN/m2, with reinforcement placed on the shear plane. Tests were conducted at 55% relative density of sand. Friction IV coefficient computed from the results obtained from the tests was 0.612 and the interface friction angle was obtained as 31.5°' Coefficient of friction between reinforcement and sand was also determined by conducting pull-out tests in specially designed tank of size 980 mm x 360 mm with 320 mm height. In these tests effect of normal pressure and length of reinforcement on the friction coefficient was studied. The results indicated a friction coefficient of 0.604. Constitutive law parameters obtained from triaxial tests on unreinforced samples of sand used in the experimental study were used to predict the pressure-settlement characteristics of footing on unreinforced model slopes by the proposed procedure. Similarly data from pullout tests and pressure-settlement values from tests on unreinforced slope were used in the procedure developed to predict the improvement in bearing capacity with the inclusion of different number of geogrid layers in model slope. Pressure-settlement plots were obtained from the model tests conducted on reinforced and unreinforced model slopes. The results were then synthesized and compared with the results obtained from the analytical procedure proposed. Good agreement was observed between the predicted and observed values. (ii) In the second part of research, numerical parameter investigation was carried out on the behaviour of strip footing adjacent to a steep geogrid reinforced slope resting on competent foundation soil. The load -settlement response of strip footing resting adjacent to such a slope was obtained using finite element software PLAXIS 8.2. Extensive twodimensional FEM parameter study was carried out on a sand slope of height 10m, to better understand the behaviour of strip footing on steep reinforced slope. The computations were performed for plain strain conditions. The slope domain extending laterally two and half times the height of slope from the toe is discretized into 15-noded triangular elements. PLAXIS uses the Gaussian integration within the triangular elements. For 15-node elements, the integration is based on 12 sample points. The hardening soil model (HSM) was used to represent the behaviour of the sand. The geogrid was modeled as elastic line element represented by a single axial stiffness value. The interaction between the geosynthetic and the surrounding soil is simulated by interface elements located between the reinforcement and the soil surfaces. The interface v friction angle and adhesion between the contact surfaces are modelled by assigning a suitable value for the strength reduction factor in the interface compared with the corresponding soil strengths which for most of geogrids is 0.9. The facing was modelled by providing a competent soil layer with cohesion of 200kN/m2 and an internal friction angle of 35° and contribution of facing to load-settlement response was not taken into account. Updated Lagrangian method is employed in the analysis to update the stiffness matrix at every iteration. In this method, the coordinates of nodes are updated by adding the corresponding displacements of nodes at every load step. Without this scheme it is not possible to model important effects such as the additional tensile resistance due to the membrane action of the reinforcement. The parameters investigated in the numerical analysis thought to influence the load - settlement response of footing adjacent to reinforced steep slope are: (i) width of footing, (ii) slope angle, (iii) edge distance of footing, (iv) axial stiffness of reinforcement, (v) vertical spacing of reinforcement layers, and (vi) length of reinforcement. The results are presented in the form of charts. vi