RESPONSE OF WELL FOUNDATIONS UNDER HORIZONTAL LOADS

Abstract

Well foundations of bridges are subjected to vertical forces due to dead load of structure, live load of vehicles, buoyancy pressures and vertical inertial forces due to earth quakes. They are subjected to horizontal forces caused by action of vehicles, contraction or expansion effects of super structures and forces caused by water, wind, soils and earth quakes. Due to the combined action of these loads a well undergoes rotation relative to the surrounding soil. The action is resisted by normal and frictional reactive forces of the soil acting on the side faces and the base of the well. The design problem in well foundations subjected to the combined loads and moments is to work out its safe depth of embedment in soil so as to limit the horizontal and vertical movement at the bearing level within safe range and provide adequate factor of safety against failure. A review of literature shows that the current method of solving the well problem are based on Tcrzaghi's (1943) elastic and plastic approaches to the analysis of rigid bulkheads and on inference drawn from some model studies. None of the solutions, however, takes into account the non-linear behaviour of the soil. Frictional forces on sides and base are also not accounted for in an adequate manner. In this thesis, a theory has been developed for analy sing the lateral resistance of well foundations which takes into account, the non-linear pressure versus displacement (ii) characteristics of the soil on the sides as well as the base. The partial or full mobilization of frictional force on the soilwell contact is also considered. The soil pressure at any depth zx below the general ground (scour) level is taken as follows: p • % zi° yT where p is the soil pressure, mh anon-linear coefficient of horizontal subgrade reaction and y the lateral displacement. The indict n and r are to be determined by tests. For noncohesive soils n = 1. To substantiate the theory, static lateral load tests were performed on small scale models of square wells embedded in dense sand. The following variables have been studied in the model tests, i] Size of well model, 15 cm and 20 cm side ii) Vertical load from zero to avalue sufficient to over come frictional resistance on side facts iii) Depth of embedment iv) Position and magnitude of lateral load v) Friction coefficient on faces vi) Stiffness of subgrade at base vii) Sloping surcharge due to scour pit around the well Pressure distribution and frictional force on faces and base were obtained with specially designed earth pressure cells and friction cells respectively. The extent of soil disturbed at ground level under failure condition of well has also been observed. (iii) The tests have given the average values of rrv as .0674 to .0624 in kg. cm units and values of r as 0.55 to 0.65 for the dry sand having a density of 1.658 g/cm3 . The observed results have oeen compared with the theore tical values using these values of m^ and r and are found to have very good agreement with each other. Dynamic behaviour of wells has also been studied by free vibration tests and cyclic lateral load tests on these models. Besides the small scale laboratory models tested in prepared beds stated above, afield model of reinforced concrete, 1.5m x 1.5m in plan and 2.25m depth, was sunk in natural soil deposit and tested under the combined action of vertical and horizontal loads. Larger size cells were fitted on the sides and base for observing earth pressures and friction of cells for observing frictional resistance. Free vibration tests were also performed. Its behaviour hrs been found to be similar to that of laboratory models. The following are the main conclusions from the study: 1. Lateral load resistance of a well increases with the increase in its size and depth of embedment, with the increase in vertical load or stiffness of subgrade at base and due to inclined surcharge above de-pest scour level. This resistance decreases if the coefficient of friction is reduced. 2. Under increasing lateral load, the instantaneous point of rotation of well starts at the base level at a distance more than 0.5B from the axis of well and goes on shifting upwards (iv) and towards the axis. At sufficiently large tilts its position comes between 0.05D to .25D above the base and at a distance between O.lB to 0.2B from the axis of well where B is the width of well and D the depth of well. 3. The non-linear theory developed for analysing the lateral load resistance is very well coroborated by the model tests and could be used for design of well foundations in noncohesive soil deposits* provided mh and r are determined by tests for the soil.