The Challenge of Real-Time Oilseed Extraction Control
Soybean crush facilities operate in a high-volume, low-margin environment. Extraction efficiency for crude soybean oil and crude protein (CP) content in the remaining soybean meal dictates daily profitability. Traditionally, operators adjust de-solventizer toasters (DT) and solvent extraction parameters based on composite samples analyzed in the lab every 4 hours. This lag leads to significant protein give-away or out-of-spec oil leftovers.
Real-Time FT-NIR Composition Monitoring
By installing the USTECH ProLine2550 inline FT-NIR system at the discharge of the cooling line before shipment silo packaging, a commercial crushing plant monitored moisture, crude protein, and residual oil levels continuously. The real-time measurement feedback loop allows for precise temperature adjustments on the DT to prevent protein denaturation and optimize solvent levels.
| Parameter | Target Level | NIR Lab Correlation (R²) | Standard Error of Prediction (SEP) |
|---|---|---|---|
| Crude Protein | 47.50% | 0.96 | —0.15% |
| Moisture | 12.00% | 0.98 | —0.11% |
| Residual Oil | 0.80% | 0.94 | —0.06% |
Operational Return on Investment
Optimizing the dryer and solvent loops based on instant 5-second scans allowed the plant to increase meal moisture levels safely closer to the 12.0% specification limit and reduce average protein give-away by 0.45%. This resulted in an extra 450 tons of product yield per year, translating to $145,000 in raw material savings.
References
- AOCS Official Method Am 1-92 - Determination of Oil, Moisture, and Protein by Near-Infrared Reflectance.
- "Process Control of Solvent Extraction Loops in Large Scale Soybean Crushing Plants," Journal of American Oil Chemists' Society, 2024.