Abstract:
A comprehensive literature survey and review of all interface elements so far
developed and implemented within their limited scope of applications have been carried out
in this study. From the literature survey, three major types of interface elements such as,
Goodman (1968) Type, Katona (1983) Type and Desai (1984) Type have been selected for.
further study depending on their successful application to physical problems. These three
most important interface elements are judged in an existing nonlinear incremental-itrative
generalized elasto-plastic Finite Element Method (FEM) program.
'* The parameters controlling the main features of each interface elements have been , ,
identified. These parameters are examined.critically in a typical problem setting of directshear
box simulation. In order t~ investigate the evolution of different internal variables like
deformation, strain or stresses, various types of figures have been drawn. Each case of
analysis resulted in outputs of: (a) deformed mesh; (b) force-deformation and stress-strain
relationship inside the interface; (c) shear strain contours; and (d) displacement fields. As
most of the figures were drawn at different load steps in order to trace the progress of the
variables, it was rather easy to identify the trouble spots for a particular type of interface.
The contrast between an analysis in direct-shear box without an interface element
.1" and with an interface element is very clear. Goodman (1968) type of interface performed
well in direct shear simulation. The performance of this interface depends on the relative
stiffness of the interface compared to the surrounding soil. Its behavior also depends on the
thickness of the interface and load eccentricity from the interface. It shows distinct si,gnof
mesh penetration at higher shear deformation. Katona (1983) type interface .element is the
simplest interface in terms of number of parameters required. It does not have any
penetration or separation problem. The performance in slip mode is ideal although not
realistic. The performance of Desai (1984) type interface element seems to be the most
realistic although it showed tendency of mesh penetration. It also showed better convergence
ii
properties. Reduction in the thickness of the element improves the performance of the Desai .
,{ type element.
Finally, it can be concluded that among the three interfaces studied in this research,
Katona and Desai type of interface element are much better in accuracy and although,
practically Goodman type of interface is easy to formulate and to implement.
