Predictive maintenance on production lines
Turning unplanned downtime into scheduled intervention
Mid-size manufacturer, 3 lines on double shifts. Unplanned downtime accounts for 6% of operating hours. Modelling the gain from anticipating equipment failure.
- Modelled yearly gain
- €454k – €756k
- Payback
- 2.9 to 4.8 months
- Over 3 years
- €1.06M to €1.97M net over 3 years
modelled yearly gain
A mid-size manufacturer runs three production lines on double shifts, roughly 4,000 operating hours per line per year. Unplanned downtime accounts for 6% of those hours — a common level in process manufacturing. Every lost hour costs unproduced margin, idle labour and, at month end, late-delivery penalties.
- Instrumentation of critical equipment (vibration, temperature, power draw) with continuous signal capture
- Anomaly detection models trained on failure history and process data
- Shift from calendar-based to condition-based maintenance planning, arbitrated with production
- Feedback loop: every intervention feeds the model and sharpens alert thresholds
Everything on the table
- Operating hours
- 3 lines × 4,000 h = 12,000 h/year
- Unplanned downtime rate
- 6% → 720 h/year
- Share of downtime caused by addressable equipment failure
- 60% → 432 h/year (scope haircut)
- Cost of one hour of downtime (lost margin + idle labour)
- €3,500/h
- Year 1 investment (sensors, platform, integration, change management)
- €180,000
- Annual running cost in later years
- €60,000/year
Three scenarios, not one number
| Parameter | Conservative | Central | High |
|---|---|---|---|
| Reduction in addressable downtime applied | −30% | −40% | −50% |
| Production hours recovered | 130 h | 173 h | 216 h |
| Gross yearly gain | €454k | €605k | €756k |
| Payback period | 4.8 months | 3.6 months | 2.9 months |
| Net cumulative gain over 3 years | €1,062k | €1,515k | €1,968k |
- The gain is only monetisable if demand absorbs the recovered hours. On a line already running below capacity, predictive maintenance cuts maintenance cost without creating revenue — the model then has to be rebuilt on the maintenance line item alone.
- Without usable failure history (at least 12 to 24 months of time-stamped events), the learning phase pushes payback out by 6 to 12 months.
- The €3,500 hourly downtime cost is a working assumption. It varies by a factor of five depending on the line's value added — it is the first parameter to replace with your own.
- Critical equipment identified and ranked: you do not instrument everything, you instrument what stops the line.
- A CMMS that actually records interventions — otherwise the model has no ground truth.
- An explicit production/maintenance agreement on the right to stop a machine on alert, before failure.
- A pilot on a single line, measured before and after, ahead of any rollout.
What these scenarios are, and what they are not
These are models, built for illustration on public sector benchmarks. They are not results observed at Cardan-AI clients, and they constitute no commitment as to outcome. The improvement rates come from the studies cited; the choice of assumptions, the scope haircuts and the arithmetic are Cardan-AI's, and are shown in full so they can be challenged. Transposed to your organisation, these orders of magnitude can vary widely — in both directions.
Test these assumptions against yours (30 min, free)
We swap this scenario’s assumptions for your real figures, and you leave with a model you can defend internally.