The challenge
Mining productivity in South Africa is squeezed by ageing infrastructure and unplanned downtime — and on a mine or plant, a conveyor is a single long machine whose failure propagates: a seized pulley bearing, a belt tracking off and shredding, or a drive failure can take out a shift, destroy an expensive belt, or start a fire. Conveyors are also long, dusty and hard to inspect by hand, so problems develop in the runs nobody walks often. Maintenance was reactive — fix it when it breaks — which on a critical conveyor is the most expensive option of all.
The principle: don't instrument every idler. Instrument the failure modes that actually stop the conveyor — the drive, the key pulleys, belt tracking — and let the data say when to act. Scoping beats sensor count.
Our approach
- Rank by cost of failure. Start with the conveyor whose stoppage hurts most and whose runs are hardest to inspect (the approach in our conveyor monitoring guide).
- Watch the right signatures. Vibration, temperature and current on the drive motor and gearbox; temperature and vibration on head, tail and drive pulley bearings; belt-drift detection at key points.
- Baseline each machine. Capture normal behaviour and set thresholds per drive and pulley — not a generic table — so alerts mean "go look," never "ignore me".
- Close the loop. Every actionable alert routes to a maintenance owner and becomes a planned work order — the difference between data and downtime avoided.
- Built for the environment. Sensors rated for dust and heat, edge-buffered through load shedding, on the addaNet platform alongside the rest of the plant's assets.