Table of Contents
10. OEE Calculation Reference
This chapter documents exactly how CedarInsight computes its figures, so every displayed value can be audited. All analysis views, reports and the AI Assistant use the same calculation functions; figures are consistent everywhere.
10.1 Data captured per production run
| Quantity | Source |
|---|---|
| Actual quantity (units in) | Input-stage counter × units-per-count [PBX] |
| Good quantity (units out) | Output-stage counter × units-per-count [PBX] |
| Defect quantity | Defect/Reject-stage counter; can be corrected by the operator at run end (5.2.4) |
| Target quantity | Entered at changeover |
| Running time | Total time in states Running and Running Slow (from state events) |
| Unplanned downtime | Total time in state Down |
| Planned production time | Wall-clock run duration (start → end) |
| Ideal cycle time | From the product (or program default) configuration |
10.2 Availability
Availability = Running time / (Running time + Unplanned downtime)
Availability relates run time to operating time. Time in Planned Stop, Changeover and Draining is not an availability loss, only unplanned Down time reduces Availability. This follows the standard OEE definition in which planned stops are excluded from the loss calculation.
10.3 Performance
Performance = (Ideal cycle time × Actual quantity) / Running time
Performance compares the time the production should have taken at ideal speed with the time it actually ran. Slow cycles and minor stops (shorter than the stage timeout) reduce Performance. The value is capped at 100%.
Performance is only as accurate as the configured ideal cycle time (chapter 7.2).
10.4 Quality
Quality = Good quantity / Actual quantity
Good quantity is the output-stage count; actual quantity is the input-stage count. Operator corrections of the defect quantity at run end are reflected here.
10.5 OEE
OEE = Availability × Performance × Quality
10.6 Aggregation across runs
When several runs are combined (period views, grouping by product/program, reports), the factors are weighted, not simply averaged:
| Factor | Aggregation |
|---|---|
| Quality | Volume-weighted: Σ good / Σ actual over all runs |
| Availability | Time-weighted: Σ running / Σ (running + unplanned downtime) |
| Performance | Running-time-weighted mean of the per-run values, mathematically equal to Σ (ideal cycle × count) / Σ running time |
| OEE | Recomputed from the three aggregated factors: A × P × Q |
This means a long run influences the aggregate more than a short one; a simple arithmetic mean of run OEEs would distort the result.
10.7 OEE colour scale
OEE values are colour-coded throughout the application. The default bands:
| Band | Default threshold | Colour |
|---|---|---|
| Excellent | ≥ 85 % | dark green |
| Good | ≥ 75 % | light green |
| Fair | ≥ 60 % | yellow |
| Poor | ≥ 40 % | orange |
| Critical | < 40 % | red |
The thresholds are configurable under System → System → OEE Thresholds (chapter 9.1).