Situation
A packaging site calculated OEE in spreadsheets from shift logs. Supervisors reclassified stops at end of shift, so engineering could not trust pareto charts. Leadership wanted automatic downtime capture aligned with PLC alarm classes and machine states already used for operations.
Problems encountered
- Micro-stops under thirty seconds were ignored in logs but dominated availability loss.
- Alarm floods during jams produced duplicate downtime events.
- Planned changeover was not distinguished from unplanned stops in the historian.
- Corporate KPI definitions did not match PLC state names on the floor.
Engineering response
TECHNLOGIX PTY LTD mapped PLC states and priority-one alarms to a finite set of downtime reason codes agreed with production and quality. Debounce and suppression rules collapsed alarm storms into single events with duration. Planned stops required operator acknowledgment on HMI to enter changeover mode, preventing misclassification.
SCADA reports exposed availability, performance, and quality with shift boundaries matching payroll clocks.
Outcomes
Top-three loss categories were more consistent after the first month of automatic capture; improvement work could focus on specific jam causes rather than generic “mechanical” labels. OEE meetings spent less time reconciling spreadsheets and more time reviewing timed event lists.
Lessons for similar sites
OEE projects succeed when reason codes are few and tied to automation truth, not a free-text field. Involve operators in naming states they already use verbally.