ENStarting point
OEE describes losses in productive equipment utilisation within a defined period. Material yield and specific consumption answer different questions. Management needs to keep the measurement bases separate and analyse the indicators together, using data by campaign and material family.
Choose the decision before the indicator
Start with the operational question. Does the line lose production because it is idle, runs below its possible rate, generates out-of-specification output or receives material with little recoverable content? Each hypothesis requires a different measurement. A single “efficiency” percentage can hide these differences and encourage actions that improve the dashboard's appearance without improving results.
Define the equipment or line section, period, product family and those responsible for records. If the assessment covers the whole plant, identify how intermediate stocks affect the relationship between inputs and outputs. An extruder may produce from stock prepared the previous week; comparing its output with waste received today without reconciliation creates a yield that belongs to no actual campaign.
Use OEE with stable definitions
The usual formula combines availability, performance and quality. Availability relates operating time to planned production time; performance compares output during operation with the ideal reference rate; quality considers the share accepted on the first pass. Vorne's methodological documentation presents these factors and their multiplication. [1]
In continuous recycling, adapting the calculation to mass requires consistency: use the same output and material family for the reference rate and measured production. Do not replace quality with the separation yield of heterogeneous waste. Necessary removal of dirt or distinct components is not automatically a product defect. Declare the adaptation used to avoid comparing your figure with indicators calculated across different boundaries.
Fix planned time and the reference rate
Determine in advance which periods are planned for production and how changeovers, maintenance and lack of demand will be treated. Do not remove a stoppage from the denominator after observing that it worsened the result. When the schedule legitimately changes, record the change and retain the history needed to understand the difference.
The reference rate must correspond to technically substantiated ideal capacity for that product and equipment condition. Do not use a reduced target simply to raise the percentage. If performance exceeds 100%, investigate the reference, unit, time and records. Materials with very different bulk densities or behaviour may require separate comparison families. Classification must be established for a technical reason, not chosen afterwards to justify a result.
Review an example applied to a homogeneous stage
Hypothetical example of a granulation stage: during 7 planned hours, the equipment operates for 6 hours. The ideal output rate is 1 t/h for the material considered. Total production is 5.40 t, of which 5.13 t is accepted on the first pass. Availability = 6 ÷ 7 = 85.71%; performance = 5.40 ÷ (6 × 1) = 90%; quality = 5.13 ÷ 5.40 = 95%.
The product of the factors is approximately 73.29%. The same result is obtained by dividing the 5.13 t accepted by the 7 t that the ideal rate would allow during the planned time. The calculation assumes consistent weighing, no double counting through rework and a defined output boundary. On its own, it does not show how much original waste was needed to produce the granulate.
Visual explanation · hypothetical example
Three factors, one product
Availability, performance and quality are multiplied. The result is not the average of the three factors.
- Availability
- 85.71%
- Performance
- 90.00%
- Quality
- 95.00%
Common scale for the factors: 0 to 100%.
OEE = availability × performance × quality. Yield from the original waste requires a separate balance.
View the example's data and calculations
| Factor | Calculation | Result |
|---|---|---|
| Availability | 6 h / 7 h | 85.71% |
| Performance | 5.40 t / 6.00 t | 90.00% |
| Quality | 5.13 t / 5.40 t | 95.00% |
| OEE | Availability × performance × quality | 73.29% |
Keep yield and recovery in separate calculations
Material yield relates a defined output to the corresponding input. If additives, water, different products or changes in stock are involved, the formula should make their treatment explicit. A recovery rate for a particular polymer may use the estimated content of that polymer in the waste as its denominator, while overall yield uses total feed mass. These indicators should not share the same name.
Record accepted outputs, other recovered fractions, rejects and work in progress. For a separation line, purity and recovery must be examined together: increasing capture may bring more contaminants. The article on validating WEEE plastics separation explores this relationship in greater depth. A change should only be considered an improvement after checking the destination and quality of the fractions obtained.
Relate energy to production that creates value
Measure energy within the declared boundary: equipment alone, a cell with auxiliaries or the entire plant. Use the same boundary in comparisons. In the preceding hypothetical example, consumption of 3,240 kWh corresponds to 600 kWh per tonne of total output, since 3,240 ÷ 5.40 = 600. Based on accepted output, the value is approximately 631.58 kWh/t, since 3,240 ÷ 5.13 = 631.58.
Both results are correct for different bases. The second highlights the effect of quality on the energy embodied in accepted material. Separating consumption during operation, waiting and start-up can help select actions. A fall in kWh/t does not, by itself, prove a reduction in all environmental impacts; that conclusion depends on the relevant assessment and the other streams involved.
Prioritise losses by effect and cause
Keep a simple event log with start, end, equipment, material and reason. Distinguish confirmed causes from the initial symptom description. “No feed” may result from planning, stock, transport failure or a quality hold. If all these cases receive the same treatment, the chosen solution may address the wrong area.
Compare frequency, duration and effect on accepted production. A long, rare stoppage may warrant a different action from frequent short interruptions. Check whether a local improvement transfers the constraint to another section. Monitoring aims to decide on an action, owner and deadline, then verify whether the effect persists under comparable conditions. A dashboard without that feedback merely records losses that have already occurred.
Publish a measurement dictionary and review the data
For each indicator, record its name, purpose, formula, unit, source, frequency and exclusions. Keep a history of definition changes and avoid comparing incompatible series without explanation. Check weighing, clocks, manual entries and the treatment of reprocessed material. High apparent precision does not correct an incomplete dataset.
Present OEE, yield, quality, energy and stock together at operational meetings. For industrial closed-loop systemsprogrammes, add delivery to the destination and demonstrated reincorporation. The desired outcome is to identify where capacity or material is lost and verify whether the action taken increased useful production, with repeatable quality and known resource consumption.
Sources and references
References accessed on 5 October 2026. The numbering in the text indicates the concepts supported by each source.
Circuvia corporate content to support project discussions. Methods, acceptance criteria and applicable requirements are defined within the scope of each operation.
