Estimates of run-out distance for tailings impoundments are mandatory and are required by Authorities in many countries around the world. There are widely accepted differences between water dams and tailings dams regarding their failure modes and discharge processes. Empirical methods based on statistical analysis of past events are available to estimate the discharge volume, peak discharge and aid engineers to predict potential consequences and assess hazard levels downstream of the dam. On the other hand, simulation tools are available to include rheological parameters and to generate more sophisticated models if necessary.
Figure 1 Ajka failure, Hungary: predicted (yellow zone) versus real flooded area (red line area)
Despite the above, this kind of analysis still possesses a strong hypothetical approach and presents uncertainties. Questions arise such as: What are the most common uncertainties of such a simulation? Is rheological data so important for tailings modelling, especially when tailings slurry and significant volumes of water are considered? What effect can nearby water bodies have on the run-out distance? What about the accuracy of our prediction models? What about the confidence level of alert times for contingency plans?
This paper presents the Ausenco view on tailings-related flood hazard mapping. The approach includes a comparison between prediction models, real occurrences and available post-event models. Well documented tailings disasters showing Newtonian behaviour are highlighted and analysed for their use from an estimation of consequences viewpoint. Understanding the particularities of these events could help general thinking about prediction of run-out distance and wave transit periods.
Based on the above, Ausenco is applying its engineering judgment following the Canadian Dam Association (CDA) guidelines in order to perform Dam Break Analysis for Dam Classification based on the proposed CDA failure consequences particularly for Peruvian projects.
To learn more about tailings-related flood hazard mapping, read the full paper here.
For more information, please contact Roberto Jesús Flores Jimenez.