How to Size MBR for Paper Machine Seal Water: 2026 Specs

What Paper Machine Seal Water Is and Why It Needs a Dedicated MBR Sizing

Paper machine seal water (PMSW) is a low-pressure, continuous leak and shower stream collected from vacuum pumps, felt and roll showers, hydraulic power-unit seals, and bearing-cooling circuits on a paper machine. It is not the same as white water or total mill effluent: PMSW typically carries 200–600 mg/L COD, 50–150 mg/L TSS, and runs hot — 35–60 °C year-round — because it absorbs heat from the dryer section, vacuum pumps, and steam-heated showers. Flow usually lands between 5–15% of total mill discharge and is essentially continuous, which is why a dedicated side-stream treatment is more defensible than blending it into the main biotank (Zhongsheng field data, 2026).

Generic MBR sizing articles under-design hot PMSW because they default to a 25 °C feed and ignore two stream-specific penalties: (1) membrane flux must be derated as temperature rises — roughly 2% per °C above 25 °C — so a 45–50 °C feed costs 40–50% of nominal membrane area, and (2) intermittent starch and defoamer carryover perturbs the MLVSS/MLSS ratio, which throws off F/M and SRT calculations. The five-step workflow below (characterize → biological volume → membrane area → module count → derate) is built around those penalties, not around municipal sewage defaults. Submerged MBR architecture — microfiltration or loose ultrafiltration integrated with an aerated activated-sludge tank — is the right configuration here, as documented in the MDPI Water review on pathogen removal in MBRs (Water, 2014-11, doi:10.3390/w6123603). For an integrated MBR membrane bioreactor system sized to PMSW service, the steps below produce a tender-ready design basis in a single afternoon.

Step 1 — Characterize the Seal Water Flow and Load

The first sizing error to avoid is locking the design on a single grab sample. PMSW temperature swings 20–25 °C between winter shower water and summer vacuum-pump discharge, and starch/defoamer slugs coincide with grade changes. Collect a 24-hour composite sample and repeat across at least three operating seasons before finalizing the design basis. Minimum inputs: average and peak flow (Qavg, Qpeak, with a peak-to-average ratio typically 1.5–2.0), influent COD and BOD5, TSS, temperature min/max, pH, and residual H2O2 if the mill runs peroxide-bleached furnish. BOD/COD for PMSW sits near 0.35 because the load is largely dissolved starch and short-chain organic additives rather than fibrous solids (Zhongsheng field data, 2026).

When the engineer is sizing before characterization data exists, the table below provides the default ranges used in Zhongsheng tender submissions for a single paper machine. Override every value with site data the moment it is available.

Equalization upstream of the MBR is non-negotiable. A side-stream equalization tank with 8–12 h HRT (100–150 m³ per 1,200 m³/d in the worked example) smooths starch and defoamer slugs and brings temperature swings inside a ±5 °C envelope, which keeps the biological stage and the membrane flux on a single rating. The submerged MBR configuration remains the relevant architecture (Water, 2014-11, doi:10.3390/w6123603) once the feed has been equalized and pre-screened.

Step 2 — Size the Biological Tank Volume (F/M and HRT)

The aeration tank volume is governed by two independent checks that must agree to within 20%; if they disagree, the larger of the two governs. The food-to-microorganism ratio for PMSW runs lower than municipal design because the COD is largely dissolved and readily biodegradable:

F/M = (Q × COD) / (V × MLVSS), with a target of 0.08–0.15 kg COD/kg MLVSS·d. The parallel hydraulic check is HRT = V / Q, with a target of 4–6 h at MLSS 8,000–10,000 mg/L. The MLVSS/MLSS ratio for paper-mill sludge is typically 0.7–0.8; drop the assumed ratio to 0.6 if the mill runs heavy defoamer (silicone or fatty-acid based) because the inert fraction rises sharply (Zhongsheng field data, 2026).

Worked check at Q = 1,200 m³/d, COD = 450 mg/L, target F/M = 0.10, MLSS = 9,000 mg/L (so MLVSS ≈ 6,750 mg/L using a 0.75 ratio):

• F/M volume: V = (1,200 × 0.450) / (0.10 × 6.75) ≈ 800 m³
• HRT cross-check: V = (1,200 / 24) × 5 = 250 m³, far below the F/M requirement — confirming that F/M governs at this load and that headroom exists for peak flows and slugs.
• Stabilized design volume (rounded, with 10% safety margin): ≈ 720 m³ (per the original design assumption) to 900 m³, depending on how aggressively the engineer wants to size for peak flow.

A common F/M-based undersizing trap: an engineer who assumes MLSS = 12,000 mg/L without checking the MLVSS/MLSS ratio will land a tank that is 20–30% too small once defoamer carryover drives the volatile fraction down. Lock the ratio to 0.75 unless you have a recent TSS/VSS dataset that says otherwise.

Step 3 — Size the Membrane Area and Module Count

Membrane area converts hydraulic load into installed equipment, and on PMSW it is the step where temperature derating does most of the damage to a generic design. The flux equation is:

A = Q / (Jnet × toperation), where Jnet = Jclean × TCF × FF. TCF (temperature correction factor) is 1.0 at 25 °C and derates at roughly 2% per °C above 25 °C; FF (fouling factor) is the 1.2–1.4× multiplier the engineer adds for irreversible fouling between cleans, which is the dominant design constraint in MBR service (Water, 2018-06, doi:10.3390/w10070847). Design flux for submerged PVDF flat-sheet modules is 12–18 L/m²·h at 25 °C clean water, applied to PMSW as 8–14 L/m²·h net after TCF and FF.

Using the DF series PVDF flat-sheet membrane module as the reference (0.1 μm PVDF, 80–225 m² per cassette, 32–135 m³/d per cassette per product spec), the worked calculation at Q = 1,200 m³/d, design T = 45 °C, target Jclean = 14 L/m²·h is:

• TCF at 45 °C: 1.0 − (20 × 0.02) = 0.60 (conservative) to 0.80 (typical vendor curve).
• Jnet: 14 × 0.80 = 11.2 L/m²·h, then × 0.77 fouling factor (1/1.3) = 8.6 L/m²·h design net flux.
• Required area: 1,200,000 L/d ÷ (8.6 × 24) ≈ 5,800 m² installed.
• Module count: 5,800 / 225 ≈ 26 cassettes (round up to 28 with one redundancy cassette per train).

At 50 °C, the same calculation lands closer to 7,200 m² — a 25% jump for a 5 °C feed rise. This is the line item that routinely blows the capex envelope on PMSW jobs, and the reason equalization is best justified by membrane cost rather than by biology.

Step 4 — Aeration, Sludge Wasting, and Pretreatment Sizing

Three ancillaries decide whether the MBR actually runs at the rated flux. Aeration has two duties: BOD oxidation at roughly 1.5 kg O2/kg BOD removed, and membrane scouring at 0.3–0.5 m³ air per m² membrane area per hour for flat-sheet cassettes. For 5,800 m² of membrane at 0.4 Nm³/m²·h, the membrane-scour blower alone is sized at ~2,300 Nm³/h; add biological oxygen demand of (1,200 × 0.158 × 1.5) ≈ 285 kg O2/d ≈ 12 Nm³/h, and total aeration for the DF-series aeration box lands near 1,300–1,500 Nm³/h on a duty/standby basis (Zhongsheng field data, 2026).

Waste activated sludge (WAS) is sized from SRT and observed yield. At SRT 20–30 d and Y ≈ 0.3 kg MLVSS/kg COD removed, daily waste from a 1,200 m³/d, 450 mg/L COD feed is (0.3 × 1,200 × 0.450 × 0.90) ≈ 145 kg MLVSS/d, or roughly 2–3% of the tank inventory per day. Confirm the downstream dewatering unit — a plate and frame filter press rated for the equivalent dry-solids load — can take that throughput before locking the SRT.

Pretreatment is where paper-mill MBRs fail. A GX series rotary mechanical bar screen with ≤ 2 mm openings ahead of the MBR removes fibers, stickies, and plastic shavings that would otherwise blind flat-sheet channels. Pair the screen with an automatic chemical dosing system sized for anti-foam (silicon-emulsion breaker) and in-line CIP recipes (typically 0.5–1.0% NaOCl + 0.5–1.0% citric acid, monthly per train); treat these as recurring opex rather than a dedicated capex line in the 2026 submission.

Step 5 — Worked Example and Cost Sanity Check

The deliverable below consolidates every step above into a single row a procurement reviewer can price. Inputs and outputs are tied 1:1 so any change to Q, COD, or T propagates through the same chain.

2026 capex sanity check: integrated MBR skids for PMSW service land at $400–$700 per m² of installed membrane area, with the swing driven by tank material (carbon steel vs. 304/316 stainless), automation scope, and whether the equalization tank is in scope. For 5,800 m² that is a $2.3–4.1 M envelope before civil works. The two largest cost drivers are the membrane modules and the stainless tankage; the two most common oversights are ignoring the temperature derate (which alone can add 20–30% membrane area) and skipping the fiber pre-screen (which shortens cassette life by 30–50% in service). Hand procurement a P&ID, equipment list, I/O list, and CIP recipe in the same package, and the design basis is tender-ready. For a related sizing workflow on the cooler, starchier sister stream, see this MBR sizing for paper mill white water walkthrough; for a side-by-side approach where dissolved air flotation makes more sense than membranes, the DAF sizing for white water discharges guide is a useful counterpoint.

Frequently Asked Questions

What flux should I design a paper machine seal water MBR at?

12–18 L/m²·h at 25 °C clean water, applied as 8–14 L/m²·h net after a temperature correction factor of ~2% derate per °C above 25 °C and a fouling factor of 1.2–1.4×. At 45 °C PMSW, the net design flux typically lands near 8–10 L/m²·h.

Do I need a flow equalization tank before the MBR for PMSW?

Yes. An 8–12 h HRT equalization tank smooths starch and defoamer slug loads, narrows the temperature envelope to ±5 °C, and is the single most cost-effective way to reduce membrane over-sizing on a hot stream.

What MLSS should I target for PMSW in an MBR?

8,000–10,000 mg/L is the operating window that balances treatability against membrane air-scour demand. Run higher MLSS only if the diffuser layout and scour blower have been verified — flat-sheet cassettes lose air distribution uniformity above ~10,000 mg/L.

Is hollow-fiber or flat-sheet membrane better for hot PMSW?

Flat-sheet PVDF tolerates higher solids, is easier to clean in place, and survives the starch and stickies load characteristic of PMSW better than hollow fiber. Hollow fiber has a lower capex per m² but higher irreversible-fouling risk on this stream.

What is the typical effluent quality from a properly sized PMSW MBR?

COD < 50 mg/L, BOD5 < 5 mg/L, TSS < 5 mg/L, and turbidity typically < 1 NTU — suitable for cooling-tower makeup or in-mill reuse after polishing, and well below typical direct-discharge limits for pulp and paper effluent.

Further Reading

• MBBR sizing for factory white water