foratvise
Get Started
Process defects (scrap and rework)

Selection/sorting error in separation process

Process defects (scrap and rework)QualitySeparation (centrifugation, filtration, membranes)

In short

It appears in about 1 of every 100 lines studied.

When it appears, it takes on average about 48 % of the line's lost time. Reported by 0 of the 37 studies that could observe it (sparse evidence).

What happens

During separation, the sorting or classification mechanism (mechanical screens, optical or sensor-based sorters, centrifugal or gravity separators) fails to correctly discriminate between acceptable and non-acceptable product fractions. This mis-selection lets off-spec material pass as good product, or conversely rejects good product, generating first-pass defects without any mechanical breakdown.

What is done about it

Recalibrate and periodically verify the sorting sensors or screening elements against reference samples, tighten detection thresholds and maintain cleanliness of optical or mechanical sensing surfaces, and implement statistical monitoring of reject/accept ratios to detect drift before it causes sustained defect generation.

Practical guidance written from engineering knowledge; it is not a statistical estimate.

1Probability and evidence

Pooled over all segments. A wide interval is a result: it states how much the evidence allows us to know.
Probability of presence
0.6% [0.1–1.8%]
Expected share of lost time
0.3% [0.0–0.9%]
Share of lost time when present
48% [35–59%]
Studies / units
37 / 44
Evidence grade
D

Grade D: 0 or 1 study; the estimate is driven by partial pooling and priors, and the wide interval is the result.

2Probability by segment

Partially pooled estimates for each industry segment.
  • Dairy products
    0.7% [0.1–2.6%]5 studies
  • Soft drinks, bottled water and juices
    0.7% [0.1–2.8%]1 study
  • Brewing, wine and spirits
    0.7% [0.1–2.7%]1 study
  • Bakery and pasta
    0.7% [0.1–2.9%]0 studies
  • Meat, poultry and seafood
    0.7% [0.1–2.6%]2 studies
  • Fruit and vegetable processing
    0.7% [0.1–2.8%]3 studies
  • Grain and oilseed milling
    0.7% [0.0–2.8%]6 studies
  • Sugar and confectionery
    0.7% [0.1–2.8%]3 studies
  • Snacks, coffee, tea and other foods
    0.7% [0.1–2.7%]4 studies
  • Animal feed and pet food
    0.7% [0.0–2.7%]5 studies
  • Food manufacturing (unspecified or mixed)
    0.7% [0.0–2.8%]7 studies

3Mechanism, SCADA signature and countermeasure

Engineering guidance that helps recognise and address the cause; it is not part of the statistical estimate.

Mechanism

Engineering guidance

During separation, the sorting or classification mechanism (mechanical screens, optical or sensor-based sorters, centrifugal or gravity separators) fails to correctly discriminate between acceptable and non-acceptable product fractions. This mis-selection lets off-spec material pass as good product, or conversely rejects good product, generating first-pass defects without any mechanical breakdown.

How it looks in SCADA data

Engineering guidance

SCADA/MES shows the separator running in normal operating state with no fault or downtime alarms, while downstream quality or reject counters record an elevated proportion of misclassified units; sensor or detector signal traces may show drift, noise or threshold crossings inconsistent with the setpoint without triggering a stop.

Countermeasure

Engineering guidance

Recalibrate and periodically verify the sorting sensors or screening elements against reference samples, tighten detection thresholds and maintain cleanliness of optical or mechanical sensing surfaces, and implement statistical monitoring of reject/accept ratios to detect drift before it causes sustained defect generation.

4Sources

Every extracted datum with its study, value and verbatim quote.
  1. Defect rate: 4.25 % · as reported: “error rate in selection process”
    Show verbatim quote
    The results of the Poka Yoke template showed reduced error rates: 4.25% in selection process, 2.86% in toasting and 2.73% in the semi-toasting.