Keywords: back analysis, coupled hydro-mechanical, expansive clay, nuclear waste, sensitivity analysis, optimization.

Compacted bentonite has attracted considerable attention as a candidate for buffer and backfill material by its high swelling capacity, low hydraulic conductivity, and high radionuclide adsorption capacity. The swelling pressure is an important aspect of all high-level radioactive waste disposal projects. The mechanical properties of compacted bentonite e.g. stiffness and hydraulic conductivity also needs to be analyzed in detail.

This paper presents numerical simulation and back analysis based on the constant volume swelling pressure test. The laboratory tests have been carried out on Calcigel bentonite with various initial dry density. The finite element code, Code-Bright (UPC)[1] is used for the numerical simulation of the tests considering a coupled hydro-mechanical problem for multi-phase processes in unsaturated soil. A modified Barcelona Basic Model (TEP model) is employed to simulate the swelling behaviour of the material during the wetting period. A direct inverse approach, which consists of an automated iterative procedure correcting the trial values of the unknown parameters by minimizing the error function [2], is applied for the identification of the model parameters. Particle swarm optimization (PSO)[3], a method to optimize a problem by iteratively trying to improve a candidate solution with regard to a given measure of quality, is selected as an optimization algorithm.

In addition, a sensitivity analysis was carried out in order to estimate the rate of change in the model output with respect to the change in the model parameters. The result of the sensitivity analysis shows that intrinsic permeability plays the most important role in the model response in terms of the development of the swelling pressure.

The general result shows that the swelling pressure increases with an increase of the dry density. The swelling, hardening and softening phenomenon during the wetting process is reproduced using particular mathematical equations. The results of optimization for different tests show that, the different densities of the samples result in different optimized sets of parameters. There are two groups of parameters: the initial density-dependent parameters, and the initial density-independent parameters. A density-independent set of parameters might be obtained by taking the mean value of the three sets of parameters. Finally, a good agreement between the numerical simulation and experimental data measured from the tests on Calcigel bentonite in constant volume conditions is obtained after back analysis.

1

DIT-UPC, "CODE_BRIGHT user's guide", Universitat Politècnica de Catalunya, Barcelona, Spain, 2009.

2

M.M. Zimmerer, "VARO²PT. User Manual", 2010.

3

J. Kennedy, R. Eberhart, "Particle Swarm Optimization", in Proceedings of IEEE International Conference on Neural Networks, IV, 1942-1948, 1995. doi:10.1109/ICNN.1995.488968

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £130 +P&P)