The following is an article by Marcelo Llano, Red Earth Engineering, a Geosyntec Company, MJ Seyedan, Geosyntec Consultants, Jim Hazzard, Itasca Consulting Group which was published in the Mine Closure 2024 supplement.
A paradigm shift in geotechnical risk management: tailings dam breach analysis using material point method techniques for inactive and closed tailings storage facilities
Risk is defined as a function of two key factors: likelihood and consequence. The likelihood aspect of risk refers to the probability that an event can occur. The consequence component of risk refers to the potential impact of the event. The consequences of a potential dam failure are multifactorial and can include environmental damage, loss of life, economic costs and reputational harm to mining.
The potential consequence of a tailings dam breach is assessed through tailings dam breach analysis (TDBA). The Canadian Dam Association (CDA) technical bulletin on TDBAs, released in 2021, is one of the more recent guidelines published to guide dam safety professionals. CDA constituted a step forward in a discipline that inherited most of the methodologies developed for the water dam industry several decades ago —the CDA guideline highlighted, among other key factors, the complex non-Newtonian flow behaviour exhibited by tailings (CDA 2021). A common misconception in the industry is that when tailings are assumed to behave like water, it results in conservative TDBA consequence estimates (Llano-Serna 2023). The assumption can be partially correct since behaviour like water results in over-predicted inundation areas. However, it under-predicts mudflow depths, particularly near the dam toe, where mining and processing plant infrastructure is usually located. This misinterpretation often results in inadequate risk mitigation measures and mischaracterisation of risk profiles.
Numerous factors and failure modes can cause tailings dam failure. For example, tailings liquefaction has been demonstrated to have played a significant role in recent dam failures like Cadia in Australia, or Fundao and Brumadinho in Brazil. The International Commission on Large Dams (ICOLD) Bulletin 194, recommends following National Academies of Sciences, Engineering, and Medicine (NASEM) for guidance on state-of-practice guidelines to assess earthquake-induced soil liquefaction and its consequences (ICOLD 2023). NASEM highlights the emergence of techniques such as the material point method (MPM) to predict liquefaction and its consequences (NASEM 2021). The United States Society on Dams (USSD), in its most recent update about the analysis of seismic deformation of embankment dams in 2022, highlighted MPM as one of the promising methods for modelling the deformation of embankment dams (USSD 2022).
MPM has been applied in multiple projects in Australia. For example, Figure 1 shows the cross-section of an upstream tailings storage facility (TSF) with a structural zone built using mud farming methods.
Figure 2 shows the hypothetical flow slumping modelled using MPM.
Selecting which technologies to assess potential dam failure is a critical step and the decision depends on multiple factors. CDA classifies the types of breach outflows based on the tailings solids contents:
a) water flood
b) mud flood
c) mudflow
d) flows slide slumping.
Changes in solids concentration affect the flow types. Figure 3 illustrates the relationship between solids concentration and the types of tailings breach outflow as classified by the CDA.
Assuming tailings behave like water, could be an appropriate assumption when solids concentrations are low. Different non-Newtonian rheological models could be appropriate for modelling mud floods and mudflow. Recent experience in Australian and overseas projects has shown that MPM techniques apply to model tailings dam breaches that are expected to behave like mudflows, with high solids content and flow slide slumping – see Seyedan et al. (2024).
The stars in Figure 3 show the location within the CDA plot of some past successful projects using MPM techniques. The pink shadow show the recommended region of tailings solids content that could be modelled with MPM.
The CDA has established four conceptual cases for tailings dam breach analysis:
- Case 1A: liquefiable tailings with a supernatant pond
- Case 2A: liquefiable tailings without a supernatant pond
- Case 1B: non-liquefiable tailings with a supernatant pond
- Case 2B: non-liquefiable tailings without a supernatant pond.
MPM has been particularly effective in modelling sunny day failure scenarios such as Case 2A and Case 2B, where tailings are liquefiable or non-liquefiable and no supernatant pond exists. MPM has also demonstrated its effectiveness in more complex cases involving liquefiable tailings. Seyedan et al. (2024), in their study of an upstream TSF in Australia, showed that MPM could accurately simulate the effects of a flow slide slump, provided pond containment is maintained at safe distances — typically 60 m from the dam crest. This example shows how MPM can inform dam breach, uncontrolled release of reservoir (URR), and rate of breach/URR for a Case 1A type of facility. Inactive and closed TSFs can often be classified as Case 2A or 2B. For example Figure 4 shows the potential worst-case scenario of TDBA using 3D MPM modelling for a closed anonymous TSF in Australia. The figure shows potentially impacted regions, including roads and infrastructure.
An exciting new initiative
A challenge for the wide adoption of MPM in the mining industry has been the lack of a commercial software that embraces this industry need and is accessible. Most of the MPM modelling done currently is undertaken using academic codes (or software addons). These academic codes offer great flexibility; however, quality control and workflows are time-consuming. The industry is about to change with the soon-to-be-released Material Point Analysis of Continua (MPAC). MPAC is an MPM code developed by ITASCA Consulting Group (ITASCA) (Purvance & Coetzee 2024). ITASCA is a trusted software developer for geotechnical applications and its flagship products are widely used in the mining industry including FLAC and PFC. Figure 4 shows the progressive failure of a 3D box cut modelled using MPAC.
The release of MPAC is expected to improve risk management perceptions in the tailings dam and geotechnical engineering industries. Conventional geotechnical tools have not yet made it easy for engineers to estimate the probability of failure needed for effective risk management, let alone consequence estimation.
The wider use of MPM techniques unlocks the potential for geotechnical engineers to assess risk holistically. For example, it allows practitioners to understand the potential consequences of failure scenarios, see Figure 6. The seamless integration of FLAC, utilising the finite volume methods and the MPM modelling technique used in MPAC, captures a wide range of scenarios to model large-strain problems (such as the TDBA explored here) while also allowing the stresses to be set-up with ease using FLAC. The program intelligently switches from continuum to material points as failure progresses. Furthermore, the numerous constitutive models already available in FLAC will be the same as those available in MPAC.
Red Earth Engineering, a Geosyntec Company (REE), has partnered with ITASCA to beta-test the code MPAC in the FLAC environment. This represents an exciting opportunity to solve real-world problems with innovative solutions. It is anticipated that MPAC will make addressing the complex failure problems with TDBA more accessible to the wider community, enhance the accuracy of predictions and provide valuable insights. Ultimately assisting the industry to make more informed decisions in managing the risks associated with tailings dams; elevating the standard for safety and risk mitigation. REE is excited to be part of this important initiative.
References
CDA 2021, Technical Bulletin: Tailings Dam Breach Analysis, Canadian Dam Association, Ontario.
ICOLD 2023, ICOLD Bulletin 194 Tailings Dam Safety, ICOLD, Chatou.
Llano-Serna, MA 2023, On the consequences of getting dam breach analysis wrong, Friction, Brisbane, https://www.friction.news/news/on-the-consequences-of-getting-dam-breach-analysis-wrong
NASAEM 2021, State of the Art and Practice in the Assessment of Earthquake-Induced Soil Liquefaction and Its Consequences, The National Academies Press, Washington, DC, https://doi.org/10.17226/23474
Purvance, M & Coetzee C 2024, ‘MPAC – material point analysis of continua’, in J Hazzard, M Nelson, T Katsaga & J Sanftenberg (eds), Proceedings of the 6th International ITASCA Symposium on Applied Numerical Modeling in Geomechanics, ITASCA International Inc, Minneapolis.
Seyedan, S, Arenas, A & Llano-Serna, M 2024, ‘Advances in dam breach analysis appropriate for dewatered tailings storage facilities’, in AB Fourie & D Reid (eds), Paste 2024: Proceedings of the 26th International Conference on Paste, Thickened and Filtered Tailings, Australian Centre for Geomechanics, Perth, pp. 247–256, https://doi.org/10.36487/ACG_repo/2455_20
USSD 2022, Analysis of Seismic Deformations of Embankment Dams, United States Society on Dams, Aurora.