Relative Importance of Rock Mass Geometrical Parameters on Erosion in a Dam Spillway Tailrace using a Small-Scale Physical Model

ABSTRACT: Dams spillway tailrace are at risk of erosion, given the excavation of the rock mass, as well as due to the high flow rates applied to the unlined rock. Studies have already been carried out on the subject, but several aspects, including the relationship between the erosion parameters, remain uncertain. To study this problem, a small-scale model of a spillway, which reproduced the flow rate and the real model dimensions, was built in a laboratory of the University of Quebec at Chicoutimi (UQAC). This model allows the simulation of the rock mass in the flow channel with concrete blocks of variable size, layout and density. Different parameters of the rock mass are simulated using these blocks. The objective of these tests is to assess the relative effect of the geometric parameters of the rock mass on erosion, such as joint opening, block protrusion, joint orientation, joint spacing and block volume. 1 INTRODUCTION Spillways are used to evacuate excess water in dam reservoirs. They consist of channels excavated in the bed rock. The excavation process as well as the natural discontinuities of the rock mass, coupled with hydraulic characteristics of the flow in the channel as high flow velocity and turbulence can make the rock mass vulnerable to erosion. Erosion of rock mass in high velocity flow channels is due to pressure variations on the top and bottom of blocks. When upward forces due to high flow velocity in the joints becomes higher than downward forces due to the weight of the block, the friction and the water pressure above the block, lifting of the block occur (Bollaert and Schleiss 2002; Pan et al. 2014). In empirical methods, erosion is defined as the limit where the flow power becomes greater than the resistance of rock mass to erosion. The resistance of rock mass to erosion is defined by different authors who use different rock mass parameters. Kirsten (1982) and Pells (2016) developed rock mass erodibility indices. Kirsten's Index (N) was used by different authors (Annandale 1995; Kirsten et al. 2000; Van Schalkwyk et al. 1994) to define an erodiblity threshold. The results obtained varied for each study. The Kirsten Index was first developed to assess rock mass excavatability rather than erodiblity, which could lead to some discrepancies in the evaluation of rock mass erodibility. For example, the excavation force was considered a bulldozer rather than flow power. The pressure applied by a bulldozer is on a very limited surface, counter to flow power, where water pressure is applied on all submerged surfaces of rock blocks. Pells (2016) developed the erodibility Geological Strength Index (eGSI) and the Rock Mass Erodibility Index (RMEI). The former adds an erodiblity parameter (Edoa) based on joint orientation to the Geological Strength Index GSI, while the latter is based solely on visual observation of rock mass erosion in unlined spillways. For these two indices, classification of erosion degrees errors for spillway erosion varies between 15% and 75% (Boumaiza et al. 2021). The effects of rock mass characteristics on its erodibility are still less documented. More specifically, relative effect of joint orientation, joint opening, protrusion, block size and shape as well as joint shear strength are parameters known to have an effect on rock mass erosion in spillway channels (Boumaiza et al. 2019). Because of a lack of a comprehensive database on the erosion process combined to measurement difficulties of the effect of several parameters at in-situ scale, a physical small-scale model could be a very good tool for studying the effects of these parameters. Previous physical flow model channels were built to study erodibility of rock mass in unlined spillway, as Reinius' model (1986), Annandale's model (1998), George

RÉSUMÉ: Les évacuateurs de crues des barrages sont à risque d’érosion dû à l’excavation du massif rocheux ainsi qu’aux forts débits appliqués à la roche non recouverte. Des études portant sur ce sujet ont déjà été réalisées, mais certains aspects comme la relation entre les paramètres d’érosion du massif rocheux demeurent incertains. Pour étudier cela, un modèle à échelle réduit d’un évacuateur de crues, qui reproduit le débit d’écoulement et les dimensions du modèle réel, a été construit dans un laboratoire de l’Université du Québec à Chicoutimi. Ce modèle permet de simuler le massif rocheux au sein d’un canal d’écoulement avec des blocs de béton de taille, disposition et densité variables. Les objectifs des essais sont de déterminer l’effet relatif des paramètres géométriques du massif rocheux sur son érodabilité : l’ouverture des joints, la saillie des blocs, l’orientation des joints, l’espacement des joints et le volume des blocs.