betaFeedback

WorksMathematics

A probabilistic analysis of the Brumadinho tailings dam-breach using uncertainty quantification and global sensitivity analysis

Abstract

On 25th January 2019, Dam B1 in Brumadinho, Brazil, suddenly collapsed and produced a catastrophic tailings flow with severe human, environmental and infrastructural consequences. Previous numerical back-analyses of the tailings flow have relied on deterministic approaches, which provide useful reference results but do not explicitly incorporate uncertainty associated to rupture hydrograph and tailings' physical properties. This work presents a probabilistic approach to study the tailings dam-breach flow caused by the rupture of Dam B1, focusing on the quantification and interpretation of the effects of input uncertainty on hazard-related hydrodynamical outputs. Deterministic data was obtained through HEC-RAS v6.6 by considering tailings as an one-phase Bingham fluid. The probabilistic framework combines dimensionality reduction, metamodeling and flooding classification to perform fast prediction of unexplored input scenarios. It allows to obtain data distributions that reflect output uncertainty and to estimate variance-based sensitivity indices to quantify the influence of input uncertainty over output variability. Results show that uncertainty is strongly dependent on time and spatial location. Hydrograph-related parameters mainly control the timing and magnitude of the early flow, whereas yield stress becomes more influential in regions associated with flow attenuation, spreading and possible arrest. Overall, the framework provides probabilistic hazard estimates and insights into the physical mechanisms governing the dynamics of tailings flow.

Paper details

vlnc:W_dpk2mzwg
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…