Cluster article
Wheat Blending Strategy: How to Hit a Protein Target Every Day
Batch versus continuous blending, proportioning equipment, the arithmetic of hitting a specification, and why most mills lose money by blending badly rather than buying well.
Key takeaways
- Blending lets you hit a protein target within ±0.2%; without it you hit ±0.6% and lose contracts.
- Continuous proportioning with weigh-belt feeders is the only method that holds specification at 60 t/d and above.
- Blend on cost, not just protein — two wheats at the same protein can differ 20% in price.
Why blend at all
Because no single wheat does everything, and because the cheapest way to hit a specification is usually a mixture.
| Situation | Blending answer |
|---|---|
| Local wheat 10.5%, market wants 12.5% | Blend with imported 14% spring wheat |
| Harvest varies 11.8–12.9% | Blend high and low lots to hold 12.3% |
| Two grades required simultaneously | Blend differently for each product |
| Cheap wheat available but weak gluten | Blend with strong wheat for structure |
The arithmetic
Two-component blend:
protein_target = x · protein_A + (1 − x) · protein_B
Target 12.0% from 13.5% (A) and 10.0% (B):
12.0 = 13.5x + 10.0 − 10.0x 2.0 = 3.5x x = 0.571 → 57.1% wheat A
Three components
Target 12.0% from 14.0%, 11.0% and 9.5%:
Choose proportions, then check. With A 40%, B 40%, C 20%: (0.4 × 14.0) + (0.4 × 11.0) + (0.2 × 9.5) = 5.6 + 4.4 + 1.9 = 11.9%
Adjust: A 45%, B 40%, C 15% → 6.3 + 4.4 + 1.43 = 12.13%
Three components give you a cost-optimisation problem, not just an arithmetic one — see below.
Blending on cost, not just protein
Two wheats at identical protein can differ enormously in price.
| Wheat | Protein | Price | Cost per % protein |
|---|---|---|---|
| Local hard | 12.0% | USD 290/t | USD 24.2 |
| Imported spring | 14.2% | USD 355/t | USD 25.0 |
| Local soft | 9.8% | USD 250/t | USD 25.5 |
Here the local hard wheat is the cheapest protein source, so blend toward it. In another season the ranking reverses. The blend should be recalculated every time a purchase is made, not fixed once.
A simple linear program — minimise cost subject to protein ≥ target, falling number ≥ target, moisture ≤ limit — is worth 2–5% on the wheat bill at any scale above 100 t/d. Spreadsheets do this adequately; dedicated mill software does it continuously.
Batch versus continuous
| Method | Accuracy | Scale | Cost |
|---|---|---|---|
| Manual batch weighing | ±1.0% | Under 10 t/d | Negligible |
| Batch weigh hopper + mixer | ±0.5% | 10–30 t/d | USD 15,000–40,000 |
| Continuous weigh-belt proportioning | ±0.2% | 20 t/d+ | USD 40,000–150,000 |
| Continuous with NIR feedback | ±0.1% | 100 t/d+ | USD 120,000–300,000 |
Why continuous wins at scale
At 60 t/d you blend 78 t of wheat per day. A batch system requires a weigh hopper, a mixer and a buffer bin — the buffer alone is 2–4 hours of capacity, and every transition creates a slug of off-specification wheat.
Continuous proportioning draws from each silo at a controlled rate onto a common conveyor. There is no buffer, no transition slug, and the correction is instantaneous when a new silo is switched in.
Equipment
| Component | Function | Notes |
|---|---|---|
| Weigh-belt feeder | Meters each wheat at a set rate | One per component silo |
| Collecting belt | Combines streams | Must handle total rate |
| Belt scale | Verifies total | Cross-check on feeders |
| Controller | Holds ratios, logs data | PLC with recipe storage |
| Moisture sensor (optional) | Corrects for water content | Dry-matter blending |
| NIR (optional) | Measures protein in real time | Feedback control |
Dry matter correction
Blending on wet weight is wrong when moisture differs. Two wheats at 12.0% and 14.0% moisture, blended 50/50 by wet weight, give 13.0% moisture — but the protein on a dry basis is not what the arithmetic predicted.
Correct every component to dry matter before calculating:
dry mass = wet mass × (1 − moisture)
For precise work, the controller should accept moisture per silo and convert automatically.
Verification
| Check | Frequency |
|---|---|
| Feeder calibration | Weekly |
| Blend sample protein | Every shift |
| Falling number on blend | Daily |
| Moisture on blend | Every shift |
| Full analysis | Weekly |
A blend that is calculated correctly but metered incorrectly is worse than no blend at all, because it produces confident-looking paperwork attached to out-of-specification flour.
The failure modes
Blending to hide a defect. Blending a 180 s falling number wheat into sound wheat does not fix the enzyme — it dilutes it, and at 20% inclusion the flour still fails. Falling number is the one parameter where blending is a weak tool.
Blending without silo segregation. If you cannot draw from a specific silo, you cannot blend. This is a storage design decision made before the mill is built — see the silo article.
Blending on paper only. Purchase contracts specify protein at 14% moisture basis; the mill measures at whatever moisture the wheat actually is. Without correction, the blend is wrong by 0.1–0.3% routinely.
Never re-optimising. Wheat prices move weekly. A blend ratio fixed at the start of the season is almost certainly not the cheapest blend by month three.