Shear Strength of Unfilled and Rough Rock Joints in Sliding Stability Analyses of Concrete Dams

Sammanfattning: The horizontal water load combined with uplift forces implies that concrete damsare sensitive for sliding. At the same time, the safety of concrete dams againstsliding in the rock foundation is associated with large uncertainties. One of themain uncertainties regards the shear strength of rock joints.Several failure criteria exist to express the shear strength for unfilled and roughrock joints. However, these criteria do not in general consider a possible scaleeffect which means that the shear strength could be lower at larger scales. Somesuggestions exist for how a possible scale effect could be considered, but theseare mainly based on empiric grounds. This means that there exists a need of amore detailed and conceptual understanding on the scale effect of rock joints.In an attempt to increase the understanding on the scale effect of unfilled andrough rock joints, a conceptual model was derived. The model is based on theassumptions that contact points occur at the steepest asperities facing the sheardirection and that their total area could be expressed with adhesion theory. Fractaltheory is used in order to idealize the surface roughness by superposition ofasperities at different scales. Based on changes in the size and number of contactpoints, the conceptual model suggests that the scale effect does not occur for alltypes of rock joints. Perfectly mated joints are suggested to not exhibit any scaleeffect while a considerable scale effect could be expected for unmated joints.The practical implications from this, for foundations with unfilled joints, is thatunmated joints with large aperture are most critical for the sliding stability ofconcrete dams, since these joints probably are the ones with longest persistenceand lowest shear strength.In order to study the scale effect of rock joints further, eighteen shear tests atdifferent scales were performed. All of the samples were taken from the rockfoundation at Långbjörn hydropower station. Possible scale effects could beobserved, but no firm conclusions could be made, mainly due to different surfacecharacteristics of the tested joints. Three of the samples were also used toinvestigate the accuracy of the conceptual model. This investigation revealed thatit may be necessary to distinguish between weathered and unweathered joints,since the distribution of contact points appears to become more randomlydistributed for a weathered joint which in turn results in lower friction angles.In Sweden, dam safety is governed by the Swedish power company’s guidelinesfor dam safety, RIDAS. When the Swedish guidelines were compared withguidelines and regulations in other countries, it was realized that there exists aneed for a more balanced evaluation of the sliding stability for concrete damsfounded on rock in Sweden. In a first step of development, it can be based onsafety factors and an increased use of investigations. However, in a second step, anatural way of the development of RIDAS would be in a direction towardsreliability based methods.