| This thesis proposes a new methodology to define the risk zone associated with the surge wave following a dam-break. It incorporates the conventionally used maximum water levels and also the sediment movement in the river bed and the possibility of bank failure. The current definition of the risk zone is defined as the inundated area; however, significant lateral (bank) erosion leading to landslides due to bank instabilities could accompany the inundation. Clearly therefore, although the flow hydraulics is generally considered the primary component in evaluating the zone of risk, this study indicates that the riverbank stability, analyzed on the basis of geotechnical considerations, may influence the delineation of the risk area.; This work clearly shows that it is possible to use the combined disciplines of hydraulics and geotechnics to precisely follow the evolution of the riverbed and riverbanks during a flood event.; A structured methodology to arrive at the final result is followed. First, the St. Venant shallow water wave equations are used to determine the area likely to be flooded without taking into account any possible sediment transport. Next, some notions on the theory of sediment transport are presented. The fundamental physical processes responsible for bank retreat are fluvial erosion or entrainment of bank material into the flow, followed by possible lateral bank failure for example due to channel incision. For the former process, the minimum energy dissipation rate theory is used; and for the latter, i.e. evaluating the possibility of bank failure, Bishop's modified method for the analysis of slope stability. On the basis of a comprehensive overview of numerical modeling of flows over movable beds, complemented by a review of some of the available numerical models, the GSTARS Version 2.1 numerical model was chosen. Two models: BISHOP for the automatic determination of the safety factor and its corresponding slip circle, and REDISSED for the distribution of the eventually eroded bank material in the transverse section were developed as part of this project. After validation of BISHOP, the three models, GSTARS 2.1, BISHOP and REDISSED were combined to produce the model RIVIERE, used in the third phase of the new methodology which consists of four steps: (1) A diagnostic phase that provides the necessary information to globally qualify the extent of the damage after a flood event, as well as the necessary elements for the choice of the high impact areas of potential instability (HIAPI), (2) The identification of the HIAPI phase, (3) Detailed local predictions on HIAPI, where the model RIVIERE is used, and finally, (4) The determination of the risk area.; The validation of the model RIVIERE was realized on a reach of the Ha! Ha! River which was affected in the 1966 Saguenay "deluge". Finally, the new methodology is applied to the Outaouais River at Notre Dame du Nord as a predictive tool. For this case, the new methodology produces a risk area of much larger extent than that obtained when only the inundated area is considered. |
