Batch-to-batch variability in dry pulse protein fractionation arises from raw seed differences, milling drift, classifier cut-point shift, powder moisture changes, unstable hopper discharge and filter/equipment fouling. Consistent protein content and particle size require a closed-loop control strategy covering incoming raw material qualification, pre-processing, milling, air classification, real-time monitoring, sampling, QC testing and standardized changeover protocols.
1. Raw material incoming qualification and homogenization
Raw pulse seed quality is the root source of most batch variation.
- Incoming lot testing: Test each new seed batch for protein content, moisture, seed size, hardness, hull fraction and inherent microstructure before releasing to production. Reject or segregate material outside specification.
- Blending / homogenization: Mix multiple harvest lots together to average out natural variation in seed protein and hardness. Create a composite feedstock with tight target ranges for bulk protein and moisture.
- Controlled dehulling: Dehull under fixed rotor speed, gap and aspiration settings. Residual hull fibre increases PSD spread and interferes with protein liberation. Monitor hull content of dehulled material; adjust dehuller when hull residue drifts outside limits.
2. Pre-milling and milling process control (critical for particle size)
The mill determines liberation of protein bodies from starch granules, which directly controls both PSD and maximum achievable protein enrichment.
- Fixed milling setpoints: maintain constant feed rate, mill rotor speed, grinding gap and air assist.
- Real-time PSD monitoring after milling: Continuous on-stream particle size measurement tracks D50, D90 and span. If PSD drifts, automatically adjust mill parameters before flour enters the air classifier.
- Keep mill temperature stable: excessive heat changes powder moisture and particle friability. Monitor bearing temperature and inlet air humidity.
- Regular mill maintenance: replace worn grinding elements. Worn mill liners or rotors gradually coarsen the product PSD batch over batch.
3. Air classifier operating parameter locking and closed-loop tuning
Classifier wheel frequency, primary airflow and feed rate are the main levers controlling cut point, particle split and protein enrichment.
- Use recipe management: store saved parameter recipes per pulse species. Load the same recipe for each batch of the same product grade.
- Stabilize inlet air: tightly control inlet air temperature and relative humidity. Humidity variation causes particle agglomeration, which changes effective aerodynamic particle size and protein partitioning.
- Continuous monitoring: track classifier wheel speed, airflow, feed mass flow, filter differential pressure. Rising baghouse backpressure reduces airflow and shifts cut point.
- Real-time PSD sampling at fine and coarse outlets: any shift in particle distribution triggers small corrections to rotor frequency or airflow, rather than waiting for offline Kjeldahl protein results.
4. Consistent powder feeding and material handling
Unsteady hopper feed creates fluctuating mass loading into the classifier, causing batch drift.
- Bin activators, low-friction hopper liners and aeration controls prevent bridging and ratholing, delivering constant mass flow.
- Calibrate loss-in-weight feeders before each production run. Verify volumetric or gravimetric feed rate at startup and periodically during batch runs.
- Maintain consistent powder moisture: hygroscopic protein fines change flow behaviour with minor moisture swings. Control process air dehumidification across all transfer and collection systems.
5. Representative sampling and laboratory QC validation
Online sensors provide process feedback, while offline chemical analysis calibrates protein results.
- Standardized sampling protocol: grab representative samples from feed, fine fraction and coarse fraction at steady-state production, not only at batch start or end. Avoid segregated material from hopper walls.
- Test each sample for: moisture, protein (using fixed nitrogen conversion factor), full PSD by laser diffraction.
- Mass balance check: verify closure of total mass and protein mass for every batch. Material imbalance signals sampling error, leaks or product loss.
- Trend logging: store PSD, protein content and process parameters in a batch database. Identify slow drift trends over many batches, such as gradual classifier wheel fouling.
6. Batch changeover, cleaning and equipment maintenance
Cross-contamination and residual powder buildup are major sources of between-batch variation.
- Complete line cleaning between pulse species or product grades. Remove residual powder from cyclones, baghouses, ducts and classifier chamber.
- Inspection of classifier wheel blades, filter bags and rotor surfaces. Fouling changes airflow and particle separation performance.
- Scheduled calibration: recalibrate feeders, load cells, PSD sensors, airflow transmitters and analytical lab equipment (Kjeldahl/Dumas, laser particle analyser) on fixed intervals.
7. Operational rules and limit alarms
Set hard operating limits and automated alarms:
- Alarms for out-of-spec PSD, feed rate deviation, high filter delta-P, out-of-range air RH/temperature.
- Hold / divert logic: automatically divert off-spec material when parameters breach limits, preventing non-conforming product from being blended into good finished batches.
- Standard operating procedures for startup and shutdown. Startup transition material has unstable PSD and protein; isolate transition material and do not include it in finished product.
8. Post-collection blending of finished fractions
Even after consistent classification, minor within-batch variation can exist.
- Controlled blending of collected fine protein concentrate from multiple collection hoppers to homogenize PSD and protein content before packaging or silo storage.
- Avoid segregation during storage and transfer: design silos and transfer lines to minimize particle size segregation of fine protein and residual starch particles.
Batch-to-batch consistency of protein content and particle size is achieved through a complete control chain: homogenized raw feed, reproducible dehulling and milling, stable conditioned process air, gravimetric feeding, closed-loop classifier tuning, real-time PSD monitoring, standardized sampling, regular equipment calibration and controlled batch changeover. The core principle is to stabilize protein liberation in milling and maintain a constant classifier cut point; these two variables are the primary drivers of both particle size distribution and final protein enrichment across batches.
