Problem
At HMB, my client was ABB, a multinational industrial automation firm. Maintenance ran on a fixed schedule, so a machine could fail between visits, and parts were sourced only after something broke.
Approach
- Built features for managing ingest from thousands of IoT devices monitoring vibration.
- Consolidated data and built dashboards across enterprise customers with multiple sites.
- Tested and validated the data.
- [TBD: the stack, for example ingestion, storage, and model type.]
Outcome
- The system flagged machines expected to need maintenance within 3 to 6 months, so downtime could be avoided or scheduled instead of arriving unannounced.
- Parts and inventory could be routed to a site ahead of time, so what was needed was there before the technician showed up.
- Inventory could be managed across several sites, rather than stocked everywhere just in case.
- [TBD: measured results, if available.]