Cottonseed pilot: 1-3 t/d, verify yield, gossypol, cake protein before scaling
Pilot line reference: 1-3 t/d cottonseed feed, 1× 200 ton hydraulic press (50-80 kg/barrel, 30-45 min cycle), pilot cooker 100-200 L jacketed (110-130°C), small grinder/stirrer, 50-200 L settling tank, plate-frame filter. Capex $30-60k ex-works. Outputs to verify per batch: yield (target 14-18% whole-seed / 28-35% kernel), residual cake oil 6-9%, crude oil FFA 1-3%, free gossypol ~0.04%, cake protein 23-44% (depending on dehulling), color Lovibond Y >60. Run 5-10 batches per variety/source to characterize before scaling.
Pilot scale: 1-3 t/d
1× 200 ton press, 50-80 kg/barrel, 30-45 min cycle = 4-8 cycles/shift = 200-640 kg seed/shift. Multi-shift = 1-3 t/d. Cooker 100-200 L, grinder 100-300 kg/h, settling 50-200 L.
Verification targets
Yield 14-18% (whole seed) / 28-35% (dehulled kernel), residual cake oil 6-9%, crude oil FFA 1-3%, free gossypol ~0.04%, refined free gossypol <0.045%, cake protein 23-44%, hull 35-45%.
Scale-up path
Pilot results map to full-scale design: cycle time × press count = throughput; cooker temp/time → scale to industrial cooker; refining gossypol reduction → confirms NBD refining capacity for 50-100 t/d mill.
Pilot test plan
What a cottonseed pilot must measure per batch
- Pretreatment effectiveness: delint % (target >85%), dehull % (target 35-45%), cooker outlet moisture (target 6-8%).
- Press cycle: yield 14-18% / 28-35%, residual cake oil 6-9%, cake density, oil clarity at discharge.
- Crude oil quality: FFA 1-3%, free gossypol HPLC ~0.04%, color Lovibond Y >60, phosphatides 1.5-3%.
- Refining test (small lab): caustic + clay + deodorize at 240-260°C → confirm free gossypol <0.045%, color Y <30.
- Cake/meal: protein 23-44% (depending on dehulling), free gossypol 0.02-0.04% (acceptable for ruminants).
- Run 5-10 batches per seed source/variety to capture variability before scaling to 10-100 t/d.
Common errors
Where cottonseed pilots fail to inform scale-up
- Skipping pretreatment in pilot ('we'll only test pressing') — yield and crude-oil quality both depend on cooker temp + moisture; without pilot pretreatment, scale-up is blind.
- No refining bench-test → cannot confirm whether full-scale NBD will hit free gossypol <0.045%.
- Pilot batch size too small (<20 kg) → press cycle dynamics differ, results don't translate.
- Buying full-scale equipment ($300-600k) without pilot ($30-60k) → 5-10× capital risk on first batch.
- Using one variety in pilot, then assuming it predicts another — gossypol and oil content vary 30-50% across cotton varieties.
Keep the engineering path moving
These next topics sharpen process, layout, and utility scope
Ready to size a line for your oilseed?
Share feed condition, pretreatment depth, shift output, post-press destination, and utility limits. We use that to narrow the scope to the pressing section, clarification loop, and real factory boundary.