Reliability Analysis of Tailings Storage Facility Embankments: Investigating the Effect of Zonation on Slope Stability

The stratified nature of tailings storage facilities (TSFs) remains a major stratigraphic or transformation uncertainty in slope stability assessment. Unlike natural soils, tailings are hydraulically deposited and develop alternating layers of coarse, dilative sands and fine, contractive slimes that exhibit contrasting hydraulic and mechanical behaviours. These inherent differences govern the stress–strain response of the deposit, yet most design analyses still treat TSFs as single mass of soil. Such simplifications overlook the interaction between contractive and dilative responses, especially under varying phreatic conditions, and may lead to misleading assessments of safety and reliability. This study examines how stratigraphic zoning, defined as the subdivision of the embankment into zones of distinct hydraulic and mechanical characteristics, influences probabilistic evaluations of TSF stability. Three conceptual embankment models were developed to represent different phreatic conditions: Model A with a moderate phreatic surface, Model B with a low phreatic surface, and Model C with an elevated phreatic surface. Within each model, three zoning configurations were analysed to represent increasing stratigraphic complexity. The vertical layering of contractive and dilative materials was maintained within each zone to reflect hydraulic deposition processes. The single-zone configuration represents a single mass deposit with no horizontal stratigraphic differentiation. The two-zone configuration divides the slope into a lower (coarse–dilative) and upper (fine–contractive) section, while the four-zone configuration introduces alternating horizontal zones of contractive and dilative materials to model enhanced stratigraphic variability across the slope. Variability in shear strength within the saturated zone was represented through Beta-distributed friction angles (φ). The Beta distribution was selected because it is bounded between user-defined minimum and maximum limits and can flexibly reproduce a wide range of strength conditions. The distribution parameters were adjusted to reflect different proportions of contractive and dilative tailings, with higher mean φ values corresponding to increasingly dilative behaviour. Randomised Beta-distribution parameters statistics were imported into Slide2 for Monte Carlo simulations to propagate uncertainty in shear strength and generate distributions and statistics of Factor of Safety (FoS) which was used to compute Reliability Index (β), and Probability of Failure (Pf). Slope stability was evaluated using Spencer’s limit-equilibrium method, and statistical hypothesis testing (Student’s t-test) was performed to assess the significance of differences across zoning configurations. The results indicate that stratification has a significant influence on reliability outcomes. While zoning had minimal impact on the calculated Factors of Safety typically less than 4% variation between the one-zone and four-zone configurations for the same simulation type, the Reliability Index (β) increased by more than 100% between these configurations. This demonstrates that β is more sensitive to stratification and inherent uncertainty than FoS. The influence of zoning was also strongly dependent on phreatic conditions, under low and moderate phreatic surfaces (Models A and B), zoning effects on both FoS and β were pronounced, whereas elevated phreatic conditions (Model C) suppressed effective stress and reduced reliability. Statistical hypothesis testing confirmed that differences in β and FoS across zoning cases were statistically significant for Models A and B, but insignificant for Model C due to the elevated phreatic surface. In general, the results show that using stratification in a probabilistic framework gives a more realistic and justifiable way to judge TSF stability. This method makes design forecasts more reliable and helps manage tailings storage facilities in a way that is safer and more informed about risks.

Grantor dc:publisher

Stellenbosch : Stellenbosch University

Year dc:date.issued

2026

Author dc:creator

• Bamisaye, Rapheal Temitayo

Advisor dc:contributor.advisor

• MacRobert, Charles John

Language dc:language.iso

en

Repository record dc:identifier.uri

https://scholar.sun.ac.za/handle/10019.1/135596

OAI identifier oai:identifier

oai:scholar.sun.ac.za:10019.1/135596