Assembling the drive…
Conveyor drive · CV-03
Predictive maintenance twin
Motor → reducer 1:16.0 → head drum · 17:70 · 18:70
- Load
- 70%
- Drum
- 92.4 rpm
- Shaft freq.
- 24.7 Hz
Gear train 16.01 : 1
GMF 419 Hz · stage 1
Fleet model · all bearings
Healthy
All four bearings inside their learned baseline. Nothing to do.
Vibration · B2 0.71 mm/s RMS · zone A
Health trend t = 0.0 d
Alerts
- d 0.0Baseline learned from 30 days of healthy running
Camera
Cutaway
Inject a fault
Pick a fault and watch the model catch it: spectrum, health, RUL and alerts all read the same simulation.
Simulated data · illustrative model · drag to orbit, tap a part
How the twin thinks
Catch the failure weeks before it stops the line.
- SenseVibration and temperature on four bearings, streamed every second.
- LearnA healthy baseline per machine; spectral bands (1×, 2×, BPFO, gear mesh) scored against it.
- PredictA degradation model turns the anomaly trend into remaining useful life — with an honest uncertainty band.
- ActOne sentence for the maintenance planner: what to replace, and by when.
This is the problem my PhD in Computing at Curtin works on — machine learning for predictive maintenance of industrial assets — and the kind of system I built for around 20 mining assets. The data on this page is simulated so the model can be shown failing on demand.