Combined finite-discrete element modelling of key instabilities which characterise deep-seated landslides from massive rock slope failure

The expression “landslide from massive rock slope failure” (MRSF) is used to indicate large-scale landslides characterised by a variety of complex initial failure processes and unpredictable postfailure behaviour. In this context, deep-seated landslides are classified as “landslides from massive rock slope failure”. Typically, deep-seated landslides are slow mountain deformations which may involve movement along discrete shear surfaces and deep seated mass creep. The long-term development of deep-seated slope deformations creates suitable conditions for the subsequent occurrence of other slope deformations. Deep-seated landslides in mountain areas can be spatially interconnected with other types of slope deformations such as debris flows, debris slides, rock avalanches, topple, translational, rotational and compound sliding and complex type of mass movements. It is to be recognized that many aspects of large-scale landslides need be investigated in order to gain the necessary confidence in the understanding and prediction of their behaviour and in the associated risk assessment. The present thesis is to contribute to such understanding with specific reference to a number of mass movements which characterize large-scale landslides. An advanced numerical technique (FDEM) which combines finite elements with discrete elements has been applied in this thesis for improving such understanding. The open source research code, called Y2D, developed at the Queen Mary, University of London by Prof. Munjiza has been used. Considering that this code has not yet been applied to slope stability problems, a series of numerical tests have been carried out to assess its suitability to properly and efficiently simulate geomechanical problems. To this purpose standard rock failure mechanisms as well as laboratory tests have been modelled first and the results obtained have been compared with available analytical and numerical solutions. The advantages of the application of FDEM has been outlined by showing that both the simulation of failure initiation and progressive development to fragmentation of the rock mass is possible as this is deposited at the slope toe. The case study of interest for this thesis is the Beauregard massive landslide located in the Aosta Valley (Northwestern Italy). At this site the presence of an extensive deep-landslide insisting on the left abutment of an arch-gravity dam is well recognised. Based on detailed studies, the investigated area has been subdivided into zones which are characterised by different geomorphologic and geostructural features. Different landslide mechanics as well as different landslide activities upstream of the dam site have been identified and studied in detail. Such an area is thought to be at an intermediate stage of development of the deep seated landslide compared with the sector which insists on the dam. The observed failure mechanism has been ascribed to a large sliding on a compound surface. Some other failure mechanisms have been recognized, such as large flexural toppling and local block toppling instability. The final part of the thesis has been devoted to the FDEM numerical modelling of a large scale failure mechanism based on brittle behaviour of the rock mass. The aim is to apply the “total slope failure” approach through the application of FDEM. Such a technique has demonstrated the significant potential in predicting the development of possible slope instability phenomena.