Comparative study of the analysis technique of mat foundation

Abstract

Mat foundations are becoming increasingly popular in the construction of highrise buildings as they are usually found to be economical in cases where the base soil has a low bearing capacity and/or the column loads are very large. An alternative to mat is pile foundation which is not only expensive but also causes problem to rail tunnel construction. There are several methods available for analysis of mat foundation; all such techniques i.e. conventional method, ACI approximate flexible method, and . discrete element methods developed so far are based on several assumptions and idealisations of mat on soil. In this study, a detailed comparative study of the performance of these analysis techniques has been carried out. Also the effects of several parameters such as modulus of subgrade reaction of soil, mat thickness, column number and spacing, column loads and their combination, Poisson's ratio and strength of concrete on the behaviour of mat foundation have been studied. Conventional method of analysis has been found to be inadequate to give realistic results. It has been revealed that ACI approximate flexible method yields moments and deflections comparable to discrete element methods at points away from edges, but in the vicinity of edges, the method has been found to be inadequate. The two discrete element methods i.e. finite difference and fInite grid method yield moments and deflections which are comparable to each other. It has been found that maximum moments at column points are due to fInite grid method with soil spring zoning. Finite grid method without zoning, fmite difference method and ACI approximate flexible method yield next higher values. Column point moments obtained by conventional method are very small compared to other methods. As for moments between column points, the maximum among comparable methods is due to fInite grid method without soil spring zoning. The fInite grid method with zoning, fmite difference method and ACI approximate flexible method yield next higher moments respectively. Finite grid method without zoning and fmite difference method have been found to give similar deflections at all the points. With soil spring zoning, the VI deflections towards the centre have been found to be much higher. Soil spring zoning has been proven to simulate mat response to load more realistically. Mat deflections have been found to decrease almost exponentially with an increase in modulus of subgrade reaction. A slight increase in moment at column points has been noticed with an increase in modulus of subgrade reaction. With an increase in mat thickness, both moments at column points and deflections have been found to be reduced. Moments in both directions at column points have shown an increase with an increase in Poisson's ratio but it has demonstrated no effect on deflections. Both moments at column points and deflections are reduced with an increase of concrete strength. it appeared that keeping stiffness of the mat-soil system unchanged, change in external loads results in a corresponding change in deflections and moments accordingly, as it should in a linear analysis. it has been revealed that if spans remain unchanged in a direction of mat, the summation of moments at column points in that direction remains unchanged no matter how many parallel column lines are present. Apart from identifying the merits and demerits of different analysis techniques of mat, efforts have been made to put forward rational explanations of the findings of the present study.