Modelling the field behaviour of a granular expansive barrier Eduardo Alonso * , Christian Hoffmann Department of Geotechnical Engineering and Geosciences, Universitat Polite `cnica de Catalunya, c/ Jordi Girona 1-3, Campus Norte, Edificio D-2, 08034 Barcelona, Spain Received 25 April 2005; received in revised form 10 April 2006; accepted 26 April 2006 Available online 23 October 2006 Abstract The large scale ‘‘Engineered Barrier’’ (EB) experiment, performed at the Mont Terri Underground Laboratory is described. A coupled hydromechanical model is then used to simulate the test performance. Constitutive parameters for the bentonite granular backfill are based on experimental work described in a companion paper. An elastoplastic model describes the granular fill, while the host rock is simulated by a damage model. Predictions of EDZ development around the tunnel are compared with some indirect measurements. Calculated evolutions of relative humidity and stresses within the buffer are compared with sensor records. Good agreement was found for the bentonite blocks supporting the canister. The granular expansive fill exhibit a more irregular behavior. Calculated displacements of the canister agree in absolute terms with actual measurements. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Case history; Hydro-mechanical model; Waste disposal; Bentonite; Pellets; Elastoplastic constitutive model; Excavation disturbed zone 1. Introduction The engineered barrier (EB) experiment (AITEMIN, 2001) was designed to demonstrate the feasibility of granu- lar expansive barriers. The proposed concept envisages placing the nuclear canister in a horizontal drift and using a combination of two bentonite materials for filling the empty space surrounding the canister. A lower bed made of compacted bentonite blocks is first constructed to emplace the canister and then a bentonite pelletized mate- rial fills the remaining space. Both sealing materials, blocks and pellets, are made of the same FEbEX bentonite (ENRESA, 2000). The aim of crushed pellets in the upper part of the bar- rier is to avoid the difficulties of barrier emplacement of alternative concepts previously tested based on block assemblies, as in FEBEX. Since pellets can be projected into the drift from some distance by mechanical means, a simpler emplacement operation is possible. Some aspects, which appear to be solved with this type of emplacement technique, are that the emplacement of the pellets can be automated, the upper remaining gap usually found when working with full bentonite block barriers can be mini- mized, a better contact between the canister and the back- fill material is achieved and the emplacement of the canister becomes more accurate. The full scale demonstration test was performed in the underground laboratory of Mont Terri. The rock hosting the experiment, Opalinus clay, is an impervious clay shale and therefore natural hydration was incapable of providing significant barrier saturation in a reasonable time (say two years). As a result, a forced hydration system was designed and built in order to keep test duration within reasonable times. Field instruments were emplaced to monitor the evo- lution of the different variables of the experiment (buffer and host rock). The laboratory experimental program was performed in order to gain insight into the hydro-mechanical response of the material and to derive the material parameters for mod- elling purposes. 1474-7065/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.pce.2006.04.039 * Corresponding author. Fax: +34 934017251. E-mail address: eduardo.alonso@upc.edu (E. Alonso). www.elsevier.com/locate/pce Physics and Chemistry of the Earth 32 (2007) 850–865