Predictive Maintenance

Have you ranked your equipment by how much impact a failure has on production?

Level 1 → 5World class at level 4

1. 1All equipment treated the same
2. 2Informal understanding of which machines are most critical
3. 3Written list of critical machines but no formal ranking
4. 4Formal criticality matrix (impact × likelihood) for all assetsWorld class
5. 5Criticality ranking drives sensor investment and maintenance priority

How do you currently monitor the physical condition of your critical equipment?

Level 1 → 5World class at level 4

1. 1No condition monitoring — we rely on operators noticing changes
2. 2Manual checks (temperature by hand, listening for unusual sounds)
3. 3Handheld measurement tools used periodically
4. 4Fixed sensors on some critical assets (vibration, temperature)World class
5. 5Full sensor coverage on all critical assets with continuous data feed

When a condition monitoring alert fires, what happens?

Level 1 → 5World class at level 4

1. 1No alerts exist
2. 2Alerts go to an email inbox that may not be monitored promptly
3. 3Alerts go to a specific person who decides what to do
4. 4Defined response protocol: who responds, within what timeframe, what actionWorld class
5. 5Alert automatically creates a CMMS work order and notifies the right person

How do you use condition data to predict future failures?

Level 1 → 5World class at level 4

1. 1We don't use historical data to predict failures
2. 2We occasionally review graphs when a machine seems to be deteriorating
3. 3Regular manual review of trends; experienced techs make judgment calls
4. 4Statistical baselines set; deviation triggers review and planningWorld class
5. 5ML model trained on failure history predicts failures with lead time