Summary
Following the Brumadinho dam disaster, a Brazilian iron ore producer urgently required a credible, independent monitoring system for its 45 Mm³ tailings storage facility. Hypotenuse implemented a comprehensive instrumentation and AI platform that achieved Certifiable Stability Status within 6 months and now generates continuous Factor of Safety estimates validated by independent geotechnical engineers.
Background & Context
Tailings storage facilities (TSFs) represent the single largest source of catastrophic risk in the mining industry. The host company faced regulatory pressure to demonstrate enhanced monitoring capability or face operational suspension. Existing instrumentation was largely manual read-point piezometers with monthly download cycles — wholly inadequate for a structure of this consequence.
Geotechnical Challenge
The TSF is constructed using the upstream raise method, which is the highest-risk embankment type. Liquefaction of the tailings beach during seismic events or rapid drawdown is the primary failure mode. Phreatic surface control is critical — the free-draining zone must extend at least 100 m from the crest.
The structure spans 3.2 km and its heterogeneous internal structure, deposited over 22 years, creates highly variable hydraulic conductivity zones that are difficult to characterise.
Sensor Deployment
Digital Twin & Risk Modelling
A dynamic seepage model (SEEP/W) is updated nightly with the latest piezometer readings, generating a continuously revised phreatic surface. The AI layer predicts phreatic surface position 72 hours ahead based on rainfall forecasts and historical hydrological responses. When the predicted surface rises within 40 m of the crest, an automated advisory is issued.
Slope stability analysis (SLOPE/W) is run automatically against the updated seepage model, generating FoS estimates for three critical cross-sections. Results are reported to the Responsible Engineer and the regulatory authority via a shared dashboard.
Key Outcomes & Results
Regulatory Status
Achieved Certifiable Stability Status in 6 months — avoided $85M production suspension
FoS Reporting
Continuous automated FoS reporting replacing 4-weekly manual reports
Alert Lead Time
72-hour advance warning on phreatic surface rise events
Piezometer Data Availability
>99.2% uptime across 64 automated instruments
Audit Readiness
Full data trail available for regulatory audit at all times
64
Auto piezometers
72h
Seepage prediction horizon
99.2%
Data uptime
$85M
Suspension avoided
Deployment Snapshot
Location
Minas Gerais, Brazil
TSF Volume
45 million m³
Crest Length
3.2 km
Embankment Type
Upstream raise
Sensor Count
149 instruments