Tailings Dam Collapse and Propagation Modelling: A Study Case in Brazil

Tailings Dams are structures constructed to store the waste of mining activities. Nonetheless, its collapse is a recurring phenomenon, due to the lack of control and monitoring of the material deposition and the stability conditions of the reservoir. In this context, the Brumadinho tailings dam failure occurred in January of 2019, in Brazil, causing around 270 casualties, destroying residential and administrative buildings, native vegetation, agricultural fields; and polluting water resources. The damage extent of the phenomenon can be back-calculated from numerical analyses, using rheology laws, which can describe the behavior of the flowing material. In this thesis, it was employed the Voellmy and Bingham methodologies for modeling the runout, and they are expressed as a function of some intrinsic parameters, such as the basal friction angle and the turbulence coefficient, for the former, and the yield stress and plastic viscosity, for the latter. It is important to perform a calibration of these values such that the real phenomenon is accurately reproduced. However, if a predictive analysis is performed to analyze the potential of failure, known values of similar events could be used and adapted to the case under study. Using this methodology, it is possible to reproduce the path followed by the flow, the velocity profile, and the flow depth; and to analyze who and what was affected by the runout. More importantly, even if this study was a post-failure analysis, the techniques discussed here can be used in a monitoring study to evaluate the hazard posed by the rupture of tailings dam and the consequences it would cause.