Development of a realistic soil-structure interaction system

Abstract

In this study attempts have been made to construct an analytical model that closely resembles the real field behaviour of Dhaka soil. In developmg the model, the critical state program - CRISP - has been used in its core and the formulation of interface element has been incorporated into itS computer code. Whereas the ensuing model is expected to work well with various types of soil-structure interaction systems; the model has been developed and tested against deep (pile) foundations and shallow (footing) foundations. Availability of reliable data of pile load tests performed in- and interaction of piles with both clayey and sandy layers of Dhaka soil prompted a detailed study of pile-soil systems. Here, apart from proposing a methodology for fixing various mesh parameters, a study concerning the sensitivity of various input parameters of pile-soil system has been conducted. A design formula has been developed, after an extensive parametric study, for predicting the load at the onset of nonlinearity in the pile-soil system. The interaction system has been applied to square footings resting on Dhaka soil at various depths and the load-displacement relationship of footings of various sizes has been studied. The guidelines proposed and implemented for mimicking various structure-soil systems have been found to be very effective. It has been understood that special care should be taken in specifying in-situ stresses in soil prior to the installation of the structural member, in order to simulate field behaviour faithfully. While studying the interaction of pile-soil and footing-soil systems, it has been revealed that the horizontal and vertical extent of soil to be included in the .finite element idealization has a pronounced effect on the satisfactory prognosis of the system. Although the performance of the finite element model is affected by the thickness of the interface dement, for a width-to-breadth ratio of O. 1 for the interface element, such an effect has been found to be minimal. Prior to the frnal analysis of any soil-structure system, the loading rate has to be determined individually for the case concerned. In case of interaction analysis involving consolidation, it has been observed that excess pore water pressure does not dissipate much during the time span considered in case of pile load testing in the field. The onset of nonlinearity of pile-soil system has been found to be sensitive to the variation of parameters like .the unit weight of soil, depth of clay layer, the angle of friction of soil and, of course, the pile size. On the other hand, the responses have been found not to be very sensitive to the variation of cohesion, critical void ratio and the slopes of the virgin compression and swelling lines. Although the displacement predictions were affected by the variation in the value of the initial tangent modulus of structural- and soil-elements, the failure load of deep (pile) foundations remained independent of such variations. The design rationale suggested in this study for designing pile foundations has been found to match the finite element predictions satisfactorily. Although some deviations from the results obtained from a traditional design method were detected, such divergence could be explained. The load-displacement relationship of square footings has been found to be related by a hyperbolic function; the ensuing loaddisplacement equation traced the finite element predictions faithfully. The resulting loaddisplacement relationship of square footings may be conveniently used for calculating expected settlements of such footings of a superstructure. Apart from assessing differential settlements, footing sizes and depths may be chosen, albeit approximately, using the equation developed, via settlement equalization of footings.