How to Size MBR for E-Coat UF Reject: 2026 Engineering Guide

Why E-Coat UF Reject Is Harder Than Municipal Sewage

E-coat UF reject routinely cuts membrane life in half when engineers apply municipal MBR defaults to a paint stream, because the reject carries 1,500–5,000 mg/L COD, 2,000–8,000 µS/cm conductivity, and a pH window of 5–9 that drifts with every CIP dump (Zhongsheng field data, 2026). The reject is the 5–15% concentrate split from the ultrafiltration loop that returns the other 85–95% of permeate to the paint tank as rinse water; the MBR is treating only the bleed, not the whole paint bath. That concentrate is hot, low in suspended solids, and high in dissolved resin and pigment precursors — the precise combination that fouls 0.1 µm PVDF flat-sheet membranes within days when flux is set at municipal numbers. A correctly specified integrated MBR system treats this stream as a high-TDS, high-conductivity biological problem first and a filtration problem second.

Municipal MBRs are typically designed around 250–500 mg/L COD and 25–30 LMH sustainable flux, with MLSS of 6,000–10,000 mg/L and coarse-bubble scouring at 0.2–0.3 m³/m²·h (Judd, 2010, as cited in the 2025 Water Air & Soil Pollution review). None of those defaults survive contact with e-coat reject: dissolved resin fragments block 0.1 µm pores, pigment drag-out spikes the organic load, and aeration has to do triple duty — supply oxygen, keep MLSS suspended, and scour the membrane. The 2025 review states explicitly that in submerged MBRs, "aeration supplies oxygen, sustains the activated sludge in suspension, scours the membrane surface and restricts membrane fouling" — which is why the paint-stream numbers in the steps below run hotter on aeration, lower on flux, and higher on MLSS than any municipal design basis.

Step 1 — Build the Flow and Load Balance

The first design input is the daily reject volume, calculated from the paint tank size, the drag-out rate at the workpieces, and the UF recovery. For a 100 m³ e-coat tank with 0.5% drag-out and 90% UF recovery, the reject flow is 100 × 0.005 × 0.10 = 0.05 m³ per workpiece cycle, or roughly 5–15 m³/day for a single-shift line and 30–80 m³/day for a two- or three-shift automotive plant (Zhongsheng field data, 2026). Apply a 1.2–1.5× safety factor on top of the average to absorb rinse-cycle spikes and CIP dumps, so a "designed for 50 m³/day" stream should actually be sized to 60–75 m³/day.

The organic load drives everything downstream. At 50 m³/day and 3,000 mg/L influent COD, the MBR has to process 150 kg COD/day, with biological oxygen demand roughly 1.5× that load once the resin is oxidized (per the 2025 Water Air & Soil Pollution review). The review also notes that aerobic MBR "has already found widespread application in low to medium strength municipal wastewater treatment" — e-coat reject sits at the upper end of "medium strength" and bleeds into "high strength," which is exactly why every derating factor in the steps below pulls harder than a municipal design basis would suggest. Confirm the flow with a one-week on-site audit before locking the design; published tank-turnover numbers undercount CIP contributions by 20–40%.

Step 2 — Characterize the Influent and Set the MLSS Target

Lock MLSS at 8,000–12,000 mg/L for paint reject — higher than the 6,000–10,000 mg/L municipal band — because the higher biomass dilutes the incoming resin load per unit MLSS and reduces the foulant flux reaching the membrane surface. The 2025 Water Air & Soil Pollution review identifies a critical aerobic granular sludge size of 1–1.2 mm as the point where membrane fouling drops sharply, and well-suspended activated sludge at 8,000–12,000 mg/L flocs into that range under proper aeration. Below 8,000 mg/L the mixed liquor carries too much unbound resin; above 12,000 mg/L oxygen transfer efficiency collapses and the MLSS itself becomes the fouling layer.

The food-to-microorganism ratio (F/M) for paint streams should fall in the 0.05–0.15 kg COD/kg MLSS·day band — tighter than municipal because the resin fraction is slowly biodegradable. At 50 m³/day × 3,000 mg/L COD = 150 kg COD/day, an 8,000 mg/L MLSS reactor needs a wet-sludge mass of 150 ÷ 0.10 = 1,500 kg MLSS in the tank, which at typical mixed-liquor density implies a working reactor volume near 18–20 m³. Run HRT at 12–36 hours to absorb diurnal swings, and SRT at 30–60 days so the slow-growing biomass that tolerates paint chemistry has time to establish itself (Zhongsheng field data, 2026). Shorter SRTs wash out the resin-degrading population; longer SRTs push the system toward endogenous decay and pin-floc problems.

Step 3 — Apply a Derated Design Flux

Municipal design flux of 25–30 LMH will foul a 0.1 µm PVDF submerged flat-sheet module on paint reject within 5–10 operating days, because dissolved resin fragments and pigment precursors form a gel layer on the membrane surface that compresses under suction and resists backwash. The 2025 Water Air & Soil Pollution review confirms that "active backwashing process is not valid for flat-plate membranes," which removes the most common fouling recovery lever and pushes all of the fouling control onto operating margin. Set net flux at 15–18 LMH as the design band for e-coat reject — that is the number the membrane can sustain at 8,000–12,000 mg/L MLSS with continuous coarse-bubble scouring.

The tighter 0.04 µm nominal pore size MBRs documented in the 2025 review deliver strong virus rejection because "the membrane with smaller pore size is preferable to reject virus by sieving impact," but they also drop sustainable flux another 20–30% and are reserved for reuse applications where pathogen barriers matter more than throughput. For a discharge-to-drain or RO-polish reuse loop, 0.1 µm PVDF flat-sheet at 15–18 LMH is the right trade. The DF series 0.1 µm PVDF flat-sheet modules are built around exactly this operating envelope, which is why the rest of this article uses DF-series numbers in the worked example.

Step 4 — Calculate Membrane Area and Number of Modules

The headline sizing formula is straightforward once the derated flux is locked: Membrane area (m²) = Daily flow (m³/day) × 1,000 ÷ (Flux LMH × 24) × safety factor. For the 50 m³/day worked example at 16 LMH with a 1.3× safety factor, the math is 50 × 1,000 ÷ (16 × 24) × 1.3 = 50,000 ÷ 384 × 1.3 = 169 m². Round up to the next available cassette geometry: two 80 m² modules plus one 50 m² module (210 m² installed) gives 41 m² of headroom, or three standard 80 m² modules (240 m² installed) gives 71 m² of true headroom and the option to drop to a lower MLVSS setpoint during a particularly dirty production week.

The DF series flat-sheet range covers 80–225 m² per cassette, with throughput windows of 32–135 m³/day per cassette at the 15–18 LMH design band. The 169 m² target maps cleanly to a two-cassette skid (one 100 m² + one 80 m², or two 90 m² units) plus a third smaller cassette held in standby. Effluent quality from this configuration is <1 µm filtered, typically <50 mg/L COD and <5 mg/L TSS, which is suitable for RO polishing or direct discharge to a municipal sewer at most paint-shop sites (Zhongsheng field data, 2026). Plan a dedicated permeate manifold with sample ports so the operator can confirm the DF series performance monthly without shutting the line down.

Step 5 — Size the Aeration System for Both Biology and Scouring

Aeration has to cover two independent loads, and most vendor proposals under-spec one of them. The biological oxygen demand for 150 kg COD/day at the Step 1 numbers is roughly 1.5 × 150 = 225 kg O₂/day, which at 1.2 kg O₂/kWh standard oxygen transfer efficiency implies about 7.5–8 kW of blower duty for the biology alone. The membrane-scouring load is independent of biology and runs at 0.3–0.5 m³ air per m² membrane per hour per the 2025 review — for 169 m² of membrane, that is 51–85 m³/h of air dedicated purely to keeping the membrane surface clean.

Run the two loads on the same coarse-bubble manifold directly under the membrane cassettes, which is how the DF series modules are built — aeration in the mixed liquor alone is not enough because fine-bubble diffusers do not generate the crossflow velocity the membrane needs. The 2025 review states that in submerged MBRs, "aeration supplies oxygen, sustains the activated sludge in suspension, scours the membrane surface and restricts membrane fouling" — that sentence is the engineering justification for sizing aeration above the biological minimum by 30–50%. Continuous duty, not intermittent, is the operating mode; intermittent scouring lets the gel layer re-form between cycles and shortens cleaning intervals.

Step 6 — Pretreatment That Protects the Membrane

Upstream protection is what separates a five-year membrane life from a one-year membrane life on paint reject. The headwork workhorse is a rotary bar screen at 2–3 mm aperture to strip oversize paint chips, rag fragments, and phosphatization-stage carryover before they reach the bioreactor. A 24-hour equalization tank with mechanical mixing dampens both flow and load swings from rinse cycles and CIP dumps — without EQ, the reactor sees pH and COD swings that knock MLSS off the 8,000–12,000 mg/L target band and drive the system into fouling excursions.

Install a PLC-controlled chemical dosing skid to hold bioreactor pH at 6.5–8.0; e-coat drag-out can swing pH from 4 to 10 across a single CIP, and biomass activity collapses outside the 6.5–8.0 window. If the upstream pre-treatment stages (degreasing, phosphating) co-mingle oil and grease with the reject, add a dissolved air flotation stage ahead of the bioreactor — the DAF removes free oil before it coats the membrane and is sized to surface loadings of 15–25 m³/m²·h per the ZSQ-series DAF specifications. Skipping any of these steps shows up as a 30–50% reduction in membrane life within the first 12 months.

Design Output Summary

The table below consolidates every parameter calculated across the six steps, with the 50 m³/day worked example as the reference row. Use it as the design basis document when you compare against an incoming vendor proposal, and as a starting point for your own scaling calculations.

Cross-reference the parameter set against the integrated MBR system scope to confirm the vendor is including pretreatment, EQ, blower skid, and membrane cassettes in a single equipment list. For plants that already send CIP wastewater to a similar train, the MBR configuration for cleanroom CIP wastewater guide walks through reuse-versus-discharge trade-offs. If the e-coat line includes copper-bearing rinses upstream, fold in chemical precipitation for heavy metals ahead of the bioreactor to protect MLSS activity.

Frequently Asked Questions

What flux should I design a paint wastewater MBR for?

Use 15–18 LMH net flux on 0.1 µm PVDF submerged flat-sheet membranes for e-coat UF reject. Municipal defaults of 25–30 LMH foul a paint-loaded membrane within 5–10 days because dissolved resin and pigment form a gel layer that active backwashing cannot remove on flat-plate geometries (per the 2025 Water Air & Soil Pollution review).

What MLSS should a paint wastewater MBR run?

Target 8,000–12,000 mg/L, which is higher than the 6,000–10,000 mg/L municipal band. The higher biomass dilutes resin foulant per unit MLSS and flocculates into the 1–1.2 mm granule size that the 2025 review identifies as the fouling-reduction threshold. Hold F/M between 0.05 and 0.15 kg COD/kg MLSS·day.

Why choose submerged MBR over sidestream (cross-flow) for paint streams?

Submerged flat-sheet modules use 10–20× less energy than external cross-flow loops because the membrane sits inside the reactor and the aeration system does the crossflow for free. The 2025 review confirms submerged is the dominant MBR configuration globally, and on a hot, spiky paint stream the absence of a high-flow recirculation pump also reduces shear-related floc break-up.

How often does a paint MBR membrane need chemical cleaning?

Operate continuous aeration with a relaxation cycle (5–10 minutes of permeate-off every 30–60 minutes), and plan a maintenance chemical CIP every 30–90 days using alkaline surfactant followed by acid wash. Membrane life is 3–5 years when MLSS, flux, and aeration stay inside the design bands, versus 12–18 months when any one of those drifts (Zhongsheng field data, 2026).

Can MBR effluent from e-coat reject be reused?

Yes. MBR effluent at <50 mg/L COD and <5 mg/L TSS is suitable feed for an RO polishing stage or for direct reuse in non-critical rinses after breakpoint chlorination. Typical reuse rates of 60–80% of the reject volume are achievable when the RO is sized at 18–22 LMH on a brackish-water element, recovering about 75% of the MBR permeate as deionized-quality reuse water.