Core crack-filling by upstream gap-graded soils in zoned dams R. Correia dos Santos & L. Caldeira Laborat´ orio Nacional de Engenharia Civil (LNEC), Lisboa, Portugal E. Maranha das Neves Instituto Superior T´ ecnico (IST), Lisboa, Portugal ABSTRACT: The crack-filling action in zoned dams, by a granular upstream zone located upstream of a damaged core, was investigated experimentally using the Crack-Filling Erosion Test (CFET). The CFET allows testing specimens with three distinct zones: the upstream zone, the core and the downstream filter. The results of a series of laboratory tests are presented. A total of 34 tests were conducted combining 6 coarse-grained (gap-graded) upstream materials, 2 core soils, and 2 granular filters. The results of the CFETs showed that the crack-filling action is mainly controlled by some of the properties of the upstream zone and of the filter. Core soils with moderately slow erosion, or less erodible, should not have an influence on the crack-filling action. This is so mainly because the filling mechanism should occur over a very short period. The factors influencing the crack-filling by an upstream material are addressed, and some rules which give dam engineers a tool for decision-making about the potential of a upstream material to limit progression of erosion in concentrated leaks are indicated. 1 INTRODUCTION Internal erosion in cracks is one of the main causes of earth dam failures all over the world (Foster et al. 2000, ICOLD 2013). In some case studies of earth dams with a core, for example the Balderhead Dam (Vaughan and Soares 1982) or the Matahina Dam (Sherard 1973, Gillon 2007), despite the occurrence of concentrated leaks similar to those indicating de- velopment of imminent failure in the embankment, the flow has stopped or stabilised, allowing sufficient time for remedial actions to be effective. The most rel- evant issue distinguishing these incidents from those that lead to breach formation appears to be related to the presence of some types of materials upstream of the cracked core (Fell et al. 2008). These materials may induce the occurrence of two mechanisms, here named the flow-limiting action and the crack-filling action. This manuscript is focused on the latter. A de- tailed explanation about the flow-limiting action can be found in Correia dos Santos (2014) and Correia dos Santos et al. (2014). The crack-filling action involves soil particles of an upstream material being washed into core cracks. These particles are transported by the concentrated flow from the interface with the cracked core, up to the downstream filter. This process fills the crack in the core, self-heals the concentrated leak in the core and stops the excessive concentrated leakage, limit- ing the progression of the internal erosion. A more detailed explanation of the crack-filling action can be found in Correia dos Santos et al. (2015a, 2015b). There are almost no previous laboratory experi- ments focused on this particular topic. Previous lab- oratory tests are related only with the effectiveness of a particular fine sand (here termed as soil A0) to act as upstream filler (Maranha das Neves 1987, 1989, 1991). In the interest of reducing costs while providing a safe design, it is valuable to investigate if other types of materials from borrow pits explored during construction can be used as effective upstream crack fillers. In particular, in this paper we investigate ex- perimentally whether naturally occurring gap-graded materials with no (or few) fines can be trusted to pro- vide the crack-filling function, and to what extent. When a crack forms in the core, high gradients may arise in the upstream zone. If the upstream zone is gap-graded, the finer fraction near the upstream soil- core interface susceptible to suffusion may then be transported into the damaged core, due to seepage forces through the space formed by the stable coarser fraction. The material transported from the upstream zone may fill in the flaw in the core, if retained effec- tively at the filter face adjacent to the core. In this study, the Crack-Filling Erosion Test