Direct answer
Calculate MTBF from operating time divided by defined failure events and MTTR from repair time divided by those events. Keep the event definition stable, separate waiting from hands-on repair, and analyse by failure mode so the averages lead to specific spares, maintenance, training or design actions.
Define a failure before calculating
Decide which unplanned events count: stops requiring maintenance, stops over a duration threshold, loss of a safety function or any intervention that prevents normal output. Exclude planned changeover and planned maintenance consistently.
Changing the definition each month makes the trend meaningless.
Separate uptime, detection, waiting and repair
MTBF uses operating time between defined failures. MTTR is often improved by separating diagnosis, waiting for access, waiting for parts, hands-on repair, test and return to production.
A long restoration time may be caused by parts availability or escalation rather than difficult physical repair.
Analyse failure modes, not only one average
Group events by component, symptom, machine, product and cause. A high overall MTBF can hide one repetitive nuisance fault, while one rare major failure can dominate MTTR.
Use Pareto analysis and repeat-event review to select the first reliability action.
Turn the metric into action
Use MTBF to assess recurrence and maintenance strategy. Use MTTR to improve diagnostics, access, spares, procedures and training. Confirm that actions improve good output rather than only the metric definition.
Review safety-critical failures separately even when they are infrequent.
Information to prepare
- Stable failure-event definition
- Operating time source
- Failure start and restore timestamps
- Waiting, diagnosis and repair breakdown
- Failure mode and component coding
- Repeat-event and Pareto review
- Action owner and post-change trend
Packserv
Packaging-machinery service, breakdown support and planned maintenance routes.