Why Ultrafiltration Fits Edible Oil Wastewater
Vegetable-oil processing wastewater carries a contaminant profile that defeats gravity separators and most biological systems: free oil that skims, 1–20 µm emulsified oil droplets that pass dissolved air flotation, colloidal phospholipids and proteins, and a COD envelope of 3,000–25,000 mg/L typical of the food-processing industry (per broader food-plant UF engineering guide, 2026). An ultrafiltration system for edible oil wastewater operates in the 0.01–0.1 µm pressure-driven window — large enough to pass water and low-MW salts, tight enough to retain the emulsified phase that defines this duty. UF sits between DAF (free FOG) and RO or biological polishing rather than replacing either: DAF is ineffective on 5 µm emulsified droplets, and RO cannot tolerate the TSS and oil loading UF is built to remove. The same logic that made UF the standard for dairy whey and brewery trub recovery applies directly to degumming water, caustic refining wash water and deodorizer condensate once the free-oil layer is lifted upstream.
Edible Oil Wastewater Characteristics and Design Basis
Micellar-enhanced UF with LAS surfactant cuts turbidity 98%, oil and grease 95.7%, COD 84.7%, electrical conductivity 90.6% and TDS 90.7% on real edible-oil processing wastewater (per micellar-enhanced UF study, 2024). Without surfactant dosing, plain UF on emulsified oily feeds demonstrates >99% oil rejection at 61 GFD steady-state flux, but the COD and conductivity reductions are lower because the free organics remain in true solution (per Synder Filtration case study, 2024). The surfactant choice matters: LAS outperformed SDS on every indicator in the 2024 study, so a process engineer specifying a micellar-enhanced stage should default to LAS and run bench confirmation on the actual refinery stream. The temperature anchor is 40 °C from the food-plant literature — every 10 °C above 40 °C roughly halves PVDF membrane life expectancy, so the design loop should treat 40 °C as a ceiling (per broader food-plant UF engineering guide, 2026). The table below carries the influent and UF-permeate envelope for a process datasheet.
| Parameter | Influent range (edible oil) | UF permeate (with LAS) | Removal efficiency |
|---|---|---|---|
| Oil & grease | 200–1,500 mg/L | <30 mg/L (with surfactant) | 95.7% |
| COD | 3,000–25,000 mg/L | ~15–20% of influent | 84.7% |
| Turbidity | 500–2,000 NTU | <40 NTU | 98% |
| Electrical conductivity | 2,000–6,000 µS/cm | ~10% of influent | 90.6% |
| TDS | 1,500–4,500 mg/L | ~10% of influent | 90.7% |
Membrane Selection: MWCO, Material and Module Geometry
Dairy whey protein fractionation specifies 5–10 kDa to retain β-lactoglobulin and casein, while brewery spent liquor and edible-oil degumming/wash water fit the 10–30 kDa band (per broader food-plant UF engineering guide, 2026). For vegetable-oil refining specifically, 10–30 kDa is the defensible starting point because it retains emulsified oil droplets and phospholipid micelles while passing salts and low-MW organics — using a tighter MWCO offers little extra oil rejection and triples the fouling rate. PVDF hollow-fibre dominates food duty because outside-in flow tolerates suspended solids and supports air-scour cleaning, keeping CIP intervals measured in shifts rather than hours. Spiral-wound elements provide higher packing density but foul faster on protein- and fat-rich feeds, and tubular ceramics handle the highest TSS and viscosity at a CAPEX penalty that rarely pays back on an edible-oil stream. These membrane options dictate the operating window of 20–80 LMH flux, 0.5–2 bar TMP, and 1–3 m/s cross-flow velocity (per broader food-plant UF engineering guide, 2026). The table below maps module geometry against edible-oil stream characteristics.
| Module type | TSS tolerance | Cleaning support | Best fit on edible-oil duty | Trade-off |
|---|---|---|---|---|
| PVDF hollow-fibre | High (outside-in) | Air-scour backwash | Degumming water, wash water, deodorizer condensate | Lower packing density |
| Spiral-wound (PES/PVDF) | Low–medium | Chemical CIP only | Pre-screened, low-fat polishing | Fouls fast on protein/fat feeds |
| Tubular ceramic | Very high | Aggressive CIP, backflush | High-viscosity, high-TSS refinery waste | High CAPEX, niche duty |
The DAF → UF → RO Process Train
The full vegetable-oil flowsheet sequences a rotary screen or drum filter for coarse solids, a ZSQ series DAF for free-oil removal, pH adjustment to 6.5–7.5, UF, and finally an optional RO step for reuse or ZLD targets. DAF sits upstream of UF in edible-oil and meat plants to drop free FOG before it blinds the membrane — running UF on raw refinery wastewater collapses flux within hours because free oil coats the membrane surface faster than cross-flow can remove it (per broader food-plant UF engineering guide, 2026). RO sits downstream of UF whenever reuse or ZLD is the project driver, as UF permeate (TDS <500 mg/L) is the appropriate feed for a high-pressure membrane. Recovery above 80–85% is rarely economic because flux decline turns exponential as the cross-flow concentration approaches the protein/fat gel point; chasing the last 5% of recovery roughly doubles membrane area, which is an inefficient use of CAPEX. The full train logic, with MBR and ZLD alternatives, is detailed in the broader food-plant UF engineering guide.
Fouling Control and CIP for Edible-Oil Duty
Four fouling modes occur on edible-oil duty, requiring specific chemical responses: cake formation from protein and phospholipid concentrates, pore blocking from emulsified fat droplets, adsorption of surfactants and pigments onto PVDF surfaces, and biological fouling from oil- and protein-fed biofilms. A defensible CIP regime runs in three steps. Alkaline wash at pH 11–12 (typically NaOH at 50–60 °C) hydrolyses proteins and saponifies residual fats to perform the bulk of recovery. Acid wash at pH 2 (typically citric or nitric) removes mineral scale and breaks protein-mineral complexes. Enzymatic cleaners (lipases and proteases) handle residual biofilm, particularly in plants running below the design temperature. Air-scour backwash at 0.5–1.0 Nm³/m²·h between CIP cycles keeps flux recovery above 90% on PVDF hollow-fibre (per broader food-plant UF engineering guide, 2026). Realistic membrane life is 3–5 years with proper pretreatment and CIP, dropping to 1–2 years when DAF and screening are inadequate. A PLC-controlled chemical dosing skid tied to UF run-hours maintains CIP consistency.
CAPEX, OPEX and Reuse Targets
A typical food-plant UF CAPEX split runs 35–45% on membranes and modules, 20–25% on skids and piping, 10–15% on controls and instrumentation, and 15–20% on installation and commissioning (per broader food-plant UF engineering guide, 2026). OPEX includes energy at 1–3 kWh/m³ permeate, CIP chemicals, membrane replacement on a 3–5 year cycle, and labor minimized by automated CIP and dosing skids. UF permeate typically lands at TDS <500 mg/L, suitable for reuse as cooling-tower make-up or boiler feed after RO polishing. Standard UF does not reject short-chain PFAS such as PFOA and PFOS; however, it remains essential in a PFAS train by stripping the emulsified oils, proteins and surfactants that would otherwise foul high-pressure RO membranes. Design for 2026 should specify UF → RO (or UF → NF → RO) rather than UF alone. Since Koch and Dow control over 51% of UF membrane unit volume (per Al Aani et al., 2020), qualifying a second supplier for spare UF membrane elements is recommended. For downstream sizing, the nanofiltration sizing guide covers the RO/NF step that typically follows UF in a reuse or ZLD train.
Frequently Asked Questions
What pore size, TMP and flux should I specify for an edible-oil UF?
Specify a 0.01–0.1 µm pore size (10–30 kDa MWCO on PVDF hollow-fibre), a transmembrane pressure of 0.5–2 bar, and a flux operating window of 20–80 LMH with cross-flow velocity of 1–3 m/s (per broader food-plant UF engineering guide, 2026).
What removal efficiencies can I expect from UF on edible-oil wastewater?
On real edible-oil processing wastewater, micellar-enhanced UF with LAS surfactant has demonstrated 98% turbidity reduction, 95.7% oil and grease reduction, 84.7% COD reduction, 90.6% electrical conductivity reduction and 90.7% TDS reduction (per micellar-enhanced UF study, 2024). Without surfactant dosing, plain UF on emulsified oil still delivers >99% oil rejection at 61 GFD steady-state flux but lower COD and conductivity reductions (per Synder Filtration case study, 2024).
How do I clean the membrane and how long will it last?
Run a three-step CIP: alkaline at pH 11–12 and 50–60 °C, acid at pH 2 (citric or nitric), then enzymatic cleaners (lipases and proteases) for residual biofilm, with air-scour backwash at 0.5–1.0 Nm³/m²·h between cycles (per broader food-plant UF engineering guide, 2026). Properly pretreated and CIP'd PVDF hollow-fibre life is 3–5 years, dropping to 1–2 years when DAF and screening are inadequate or when the loop runs above 50 °C.
What is the energy and cost case for UF versus RO?
UF runs 1–3 kWh/m³ permeate versus 4–6 kWh/m³ for RO, which is why UF is the energy-correct step to place ahead of an RO polishing unit (per broader food-plant UF engineering guide, 2026). UF CAPEX is dominated by membranes and modules (35–45% of total) and OPEX by energy, CIP chemicals and a 3–5 year membrane replacement cycle.
Do I need DAF upstream of UF, and where does UF permeate go for reuse?
DAF is required upstream of UF in edible-oil and meat plants to remove free FOG before it blinds the membrane, and RO sits downstream whenever reuse or ZLD is the design target (per broader food-plant UF engineering guide, 2026). UF permeate at TDS <500 mg/L is suitable as RO feed for cooling-tower make-up, boiler feed or any PFAS-aware reuse loop; for 2026 PFAS expectations, specify UF → RO (or UF → NF → RO) rather than UF alone.
