A rejected fish meal consignment
Your 5 t/day fish meal plant processes trimmings, frames and market waste from a harbour. Raw material arrives warm (28–33 °C) in open bins and is sometimes held for up to 30 h before cooking; it is not preserved in any way. Cooking, pressing and rotary drying follow, and the drier is fired to bring the meal to 8–10 % moisture. The latest 20 t consignment analysed as follows: crude protein 58 %, fat 11 %, moisture 10.8 %, salt 3.1 %, sand 1.4 % (the buyer's limit is 1 %), FFA 6.4 % of the fat (limit 5 %), TVB-N 135 mg/100 g, peroxide value 22 meq O₂/kg fat, and the pressed oil has FFA 8.5 %. The poultry feed customer has rejected the load and asked for a corrective action report. Your plant has no chilling, no antioxidant dosing point, and the drier outlet temperature is not recorded.
1. Identify the root causes of the rejection and link each to a specific analytical result.
Read the worked answer
Cause 1 - raw material held warm for up to 30 h with no preservation.
- Bacterial and enzymatic lipolysis releases free fatty acids → FFA 6.4 % in meal, 8.5 % in pressed oil.
- Bacterial action on protein produces amines and ammonia → TVB-N 135 mg/100 g.
- This is the dominant cause. Everything downstream starts from spoiled raw material, and no drying regime can reverse it.
Cause 2 - uncontrolled drying and no antioxidant.
- The drier outlet temperature is not recorded, so the meal may be overheated at the surface while the centre is still wet. Heat plus air plus 11 % fat gives oxidation → peroxide value 22 meq O₂/kg.
- No antioxidant is dosed at the press or in the meal, so there is no brake on oxidation during storage and transport.
Cause 3 - handling hygiene and salt load.
- Open bins on the floor take in grit → sand 1.4 % against a 1 % limit.
- Salt 3.1 % indicates salted waste or seawater contamination entering the raw material mix; this also matters to poultry feed formulators.
Note what is not a cause: crude protein at 58 % and moisture at 10.8 % are acceptable, so the plant is capable of making a good product when the raw material is sound.
2. Propose a prioritised corrective action plan with the expected effect on each failing parameter.
Read the worked answer
Priority 1 - control the raw material (biggest effect, lowest cost).
- Set a maximum holding time of 12 h in the season and 8 h in hot weather; log arrival time and cooking time for every batch.
- Introduce preservation: ice or chilled water where volumes allow, or acid preservation with formic acid (about 2–3 % of the raw material weight, pH brought to about 4) for material that cannot be processed within the day. Acid-preserved material must be declared to the buyer if the specification requires it.
- Expected effect: FFA below 4 %, TVB-N below 100 mg/100 g.
Priority 2 - control the drier.
- Fit a thermometer at the meal outlet and record it every batch; keep the meal temperature in the range that dries without scorching (target roughly 80–90 °C product temperature at the outlet) and control moisture to 8–10 %.
- Expected effect: lower and more consistent PV.
Priority 3 - antioxidant dosing.
- Add a dosing point at the press or on the meal (a permitted antioxidant such as BHT/BHA, or a natural tocopherol blend) at the label-declared level.
- Expected effect: PV below 10 meq O₂/kg and a longer shelf life in the buyer's store.
Priority 4 - handling and sand.
- Cover the bins, raise them off the floor, and sieve or screen the meal before packing.
- Expected effect: sand below 1 %.
Verification: analyse each lot for moisture, fat, FFA, PV, TVB-N and sand; send the customer a corrective action report with three consecutive compliant lots before resuming normal supply.
3. The plant owner says the cheapest fix is to blend the rejected meal with a good batch to bring the averages inside specification. Evaluate this proposal.
Read the worked answer
Reject it. Blending is a fraud issue, not a technical fix.
Technical objection:
- Some parameters average linearly, but damage does not. A blend of a 6.4 % FFA meal with a 3.6 % FFA meal gives 5.0 % on paper, sitting exactly on the limit with no safety margin. Any variation in the next batch puts the customer over the limit again.
- Oxidation is not stopped by dilution; the peroxide value continues to develop in the blended lot because the pro-oxidant metals and hydroperoxides are still present.
- Sand and salt dilute the same way, so the blend again sits at the limit rather than below it.
Commercial objection:
- The customer has asked for a corrective action report. Blending answers a different question, and if the practice is discovered the supply relationship ends.
- Feed customers audit suppliers; lot traceability and honest deviation records are part of the qualification.
Legal/ethical objection:
- Selling blended material without disclosure as compliant product misrepresents the goods. The correct routes for a non-conforming lot are rework with a documented protocol, sale to a lower-grade outlet with disclosure, or rejection.
Correct judgement: treat the 20 t as a deviation, keep it segregated and labelled, negotiate a discounted feed or fertiliser outlet with full disclosure, and spend the effort on holding time, acid preservation and drier control so the next lot is compliant by design.