Reducing statistical uncertainty in geotechnical engineering design relying on targeted field investigation: A random field approach

Abstract

This doctoral research aims to investigate numerically the effect of targeted field investigation on the reliability of different geotechnical engineering structures, i.e. pile foundations, earth-retaining structures (studying both the active and passive states) and shallow foundations (studying both the elastic settlement and bearing capacity analysis problems). The targeted field investigation refers to sampling from a specific point or a set of points of the ground semi-space (i.e., adopting a sampling strategy) so that the statistical uncertainty in the design is minimized. The latter, generally, is attributed in the international literature to limited material soil testing. The sampling strategy leading to the minimum statistical error is called optimal. The framework of the present thesis is based on the random finite element method (RFEM), properly considering soil sampling in the analysis. The RFEM method combines finite element method with the random field theory. The random fields are generated using the Local Average Subdivision Method and mapped onto the finite element mesh, taking full account of element size in the local averaging process. Contrary to the common belief that statistical uncertainty decreases with increasing number of samples, this doctoral research clearly shows that the statistical error for the problems examined can only be minimized relying on targeted field investigation. Additionally, it shows that the benefit from a targeted field investigation is much greater as compared to the benefit gained from the use of characteristic values in a limit state design framework.