home / case studies / Cement producer with remote plants in Africa and South Asia (composite)
IT/OT Convergence

Edge Analytics and Predictive Maintenance for Rotating Equipment

We deployed edge computing nodes that analyse vibration and process data close to the equipment and raise early-warning alerts. Maintenance teams now plan interventions before failures stop the kiln line.

Cement producer with remote plants in Africa and South Asia (composite) · IT/OT Convergence

Edge Analytics and Predictive Maintenance for Rotating Equipment

Challenge

Limited bandwidth at remote sites and scarce local IT support. Edge-first processing, store-and-forward, and zero-touch device provisioning made the solution operable with minimal on-site expertise.

Approach

Vibration sensors and existing SCADA tags feed industrial edge nodes running containerised analytics (Python, ONNX models) managed remotely through a fleet management tool. Models detect bearing wear, misalignment and imbalance signatures on fans, mills and conveyors. Alerts are pushed to the CMMS as work-order requests with supporting trend data. The edge nodes buffer data during connectivity outages and sync when links recover.

Outcome

Earlier detection of developing equipment faults Illustratively ~20–35% reduction in unplanned downtime on monitored assets Maintenance planned into scheduled stops rather than emergency repairs Works reliably with intermittent connectivity Alerts integrated into existing CMMS workflows

✓Earlier detection of developing equipment faults
~20–35%Illustratively reduction in unplanned downtime on monitored assets
✓Maintenance planned into scheduled stops rather than emergency repairs

Problem solved

Unplanned stops on critical rotating equipment caused costly production losses, and route-based manual vibration checks missed fast-developing faults.

What we built

Vibration sensors and existing SCADA tags feed industrial edge nodes running containerised analytics (Python, ONNX models) managed remotely through a fleet management tool. Models detect bearing wear, misalignment and imbalance signatures on fans, mills and conveyors. Alerts are pushed to the CMMS as work-order requests with supporting trend data. The edge nodes buffer data during connectivity outages and sync when links recover.

Benefits

Earlier detection of developing equipment faults Illustratively ~20–35% reduction in unplanned downtime on monitored assets Maintenance planned into scheduled stops rather than emergency repairs Works reliably with intermittent connectivity Alerts integrated into existing CMMS workflows

Industrial edgePythonONNXSCADACMMSStore-and-forward

Talk to us