Summary
A major copper mining operation in Zambia faced chronic slope instability challenges across a 420 m deep open-pit mine. OctaSense deployed a dense geotechnical sensor network combined with a multi-variate AI model to provide continuous slope stability assessment and early warning — reducing unplanned downtime and preventing two imminent slope failures.
Background & Context
The Nchanga open-pit mine, operated by a Tier-1 copper producer, processes over 6 million tonnes of ore annually. The pit walls — composed of alternating schist, quartzite, and shale — had experienced three significant bench failures in the preceding five years, each requiring month-long recovery operations costing an estimated $4–12M per event.\n\nThe existing manual slope monitoring regime relied on fortnightly surveying with a single robotic total station and periodic piezometer readings. This approach provided insufficient temporal resolution to detect the rapid onset of failure precursors, particularly during the wet season when pore pressure fluctuations are most pronounced.
Geotechnical Challenge:
The pit operates in a complex hydrogeological environment where pit-wall de-watering is impeded by inter-bedded low-permeability shale layers. Pore pressures can spike by 40–60 kPa within 24 hours of heavy rainfall events, which dramatically reduces effective stress on the failure planes.The primary failure mode is bi-planar wedge failure controlled by two conjugate joint sets dipping at 35° and 58°. Secondary rotational slumps are also possible in the weathered upper benches. The challenge was to differentiate normal elastic deformation from irreversible plastic creep — the latter being the reliable precursor to collapse.
Failure Mechanisms:
Three interconnected failure mechanisms were identified:\n\n1. Pore Pressure Surge: Rapid rainfall infiltration through tension cracks raises phreatic surface, reducing effective normal stress on joint planes.\n2. Progressive Creep: Sustained shear displacement accumulates along the critical joint set until residual friction angle is reached.\n3. Blast Vibration Fatigue: Repeated production blasting degrades intact rock bridges between discontinuities, reducing the Factor of Safety (FoS) over time.The OctaSense model integrates all three drivers into a unified risk score.
Sensor Deployment
Key Outcomes & Results
Failure Events Prevented
2 imminent slope failures detected and managed — est. $18M in avoided losses
Warning Lead Time
Average of 72 hours of advance warning before critical threshold breach
Downtime Reduction
Unplanned operational disruptions reduced by 64% year-on-year
False Positive Rate
< 3% — preventing unnecessary production halts
Safety
Zero LTIs associated with slope movement during the deployment period
72h
Avg warning lead time
64 %
Downtime reduction
$18M
Losses prevented
'<3%
False positive rate
Deployment Snapshot
Location
Zambia, Sub-Saharan Africa
Pit Depth
420 m
Sensor Count
61 Instruments
Monitoring Period
28 months ongoing
Alert Response
< 5 minutes