Sensibilidade a parâmetros reológicos na modelagem de fluxo hiperconcentrado proveniente de rupturas de barragens de rejeitos de mineração

• Ano de defesa: 2022
• Autor(a) principal: Malena Lúcia Sousa Freire de Melo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA SANITÁRIA E AMBIENTAL; Programa de Pós-Graduação em Saneamento, Meio Ambiente e Recursos Hídricos; UFMG
• Palavras-chave em Português: Fluido não-newtoniano; Barragem de rejeito; Propriedades reológicas; HEC-RAS; Análise de sensibilidade; Engenharia sanitária; Recursos hídricos - Desenvolvimento; Reologia; Fluídos não-newtonianos; Modelagem de dados; Avaliação de riscos; Barragens de rejeitos - Brumadinho (MG)
• Link de acesso: http://hdl.handle.net/1843/55006

The risk of failure of tailings dams is relatively small, but historically it has been shown to be considerably higher than the risk of failure of water storage dams. Flood modeling allows the mapping of the most vulnerable locations and populations to guide the alert and evacuation of potentially affected communities. The evaluation of the sensitivity of hydrodynamic models to rheological parameters aims to identify uncertainties in the simulated results caused by the high variability of these parameters, with the purpose of obtaining more reliable mappings. The new non-Newtonian module of HEC RAS allows the simulation of hyperconcentrated flow and has four rheological models: Bingham, Herschel-Bulkley, Quadratic and Clastic-Grain Flow. In this study, we seek to evaluate the influence of hyperconcentrated fluid rheology parameters on the results of modeling floods resulting from tailings dam failures with the HEC-RAS. A probabilistic approach was used and the automation of simulations in HEC-RAS was developed through programs in Python language. The rheological parameters of the Bingham and Herschel-Bulkley models were defined as test parameters and the ranges of values for these parameters were obtained based on a literature review. The parameters were sampled using the Latin Hypercube Sampling (LHS) method. Two case studies were developed, the first of the fictitious dam proposed by Zenz and Goldgruber (2013) in a study promoted by ICOLD, where the impact on the results was analyzed. And in the second, the calibration of rheological parameters of hydrodynamic modeling was sought from real failure data. To this end, a modeling of the rupture of the old B1 dam of Mina do Córrego do Feijão, in Brumadinho-MG, which ruptured in 2019 was carried out. The results showed that in the scenario of the Bingham model of the ICOLD case study, the increase in values of volumetric concentration of solids (Cv) and the b parameter of the yield stress curve caused a reduction in the flooded area and an increase in the maximum depths. In the Herschel-Bulkley scenario, higher values of yield stress also caused a reduction in area and an increase in depths. Regarding the arrival time results for a depth of 0.3 m and 0.61 m, the values varied more significantly in the analysis section further downstream (15 km) of the dam. The variation of the flooded area in the ICOLD scenarios occurred in the range of the flood wave, with little variation in the width of the flood boundary. In the scenario of the Brumadinho case study, the calibration of the simulated flood area from the area observed in the real rupture was equal to 85.4% according to Fit metric F. The simulated arrival times differed between 1 and 3 minutes from the observed times at points of interest. Based on the analyses, it is inferred that although other aspects of the modeling may cause more significant influences on the modeling, the sensitivity of the hydrodynamic modeling to the rheological parameters cannot be neglected, since significant variations were observed in some results.