MBR Design Choices: How MLSS, SRT and Membrane Flux Set Plant Cost
Why the settling tank was always the weak point
Activated sludge carries one structural dependency: the biomass has to be recovered from the treated water by gravity. That requirement caps how much sludge a plant can hold. Once the blanket rises too close to the surface, fragments cross the weir and the effluent fails, and no amount of biological tuning repairs it. A membrane bioreactor removes the dependency by placing a physical barrier where the settling tank used to be.
Modules are immersed in the aeration tank or mounted in an external loop. A pump draws water through a membrane whose pores retain flocs, free bacteria and dissolved macromolecules. Nothing has to settle, so poorly flocculating sludge stays in the system and keeps working. Everything else about the process follows from that one substitution.
Commercial uptake began in the 1980s, when modules became reliable enough for continuous duty. By the 1990s the process was being installed at municipal scale, and the installed range now runs from a few cubic metres per day to more than 13,000 m3/d. That spread is itself a warning: the same hardware covers duties so different that the design rules barely transfer.
Solids inventory resets the design basis
Once retention stops being a settling problem, the mixed liquor can be thickened far beyond conventional practice. In an MBR system, the activated sludge (MLSS) concentration can be raised to 8,000-10,000 mg/L or even higher, and the sludge retention time (SRT) can be extended to more than 30 days. Submerged designs generally run at MLSS of 8-20 g/L, which is 2.5 to 5 times the inventory of a conventional biological train.
Three benefits arrive together: more organisms per cubic metre, a longer window for slow growers, and a much smaller tank volume per kilogram of COD removed. Two costs arrive with them. Above a certain concentration the mixed liquor stops behaving like water, oxygen transfer efficiency falls, and the membrane has to filter through a thicker boundary layer. The practical ceiling is set by blower capacity and by viscosity, not by the biology.
The decoupling matters more than the concentration itself. Because the membrane holds the biomass back, the hydraulic retention time (HRT) and the sludge retention time (SRT) can be controlled independently, a flexibility a conventional train does not have. A plant can run a short HRT for a high flow and still keep a sludge age long enough for nitrifiers to establish themselves.
Nitrification is where the long sludge age pays
Nitrifying bacteria grow slowly, which is why they are the first population to wash out of an overloaded conventional plant in cold weather. In a membrane system they are retained regardless of flow, so nitrification becomes stable and NH3-N can be driven to a low residual. The same retention keeps specialised organisms that attack recalcitrant organics, extending their residence time until the compounds are broken down rather than merely adsorbed.
Reported removal rates reflect that: COD removal of 93% or better and essentially complete suspended solids capture. Permeate carries almost no suspended solids and very little turbidity, which is why MBR effluent is often reused directly instead of discharged.
What a duty actually looks like in numbers
A typical municipal design basis and the effluent it is expected to deliver:
- Influent: BOD5 below 300 mg/L, CODcr below 500 mg/L, SS below 300 mg/L, T-N below 45 mg/L.
- Effluent: BOD5 below 10 mg/L, NH4+-N below 10.0 mg/L, CODcr below 30 mg/L, turbidity below 5 NTU, total coliforms below 3 CFU/L.
Those figures are not a marketing claim; they are the reason a municipal plant can hand its permeate to a landscape irrigation system or an industrial user. Both BOD and COD fall by roughly an order of magnitude, and the coliform count collapses because the membrane is an absolute barrier rather than a settling device.
Membrane side: cut-off, fouling and conditioning
Most full-scale units use organic membranes, with polyethylene and polypropylene among the common materials. Side-stream systems typically run ultrafiltration modules with a molecular weight cut-off between 200,000 and 300,000. A larger cut-off gives a higher initial flux, though not necessarily a higher flux after a few months, because the initial advantage is often consumed by fouling.
Once the membrane is selected its surface chemistry is fixed, so operating mode becomes the dominant fouling variable. Sludge concentration and mixed liquor viscosity both act on flux, and so does the filtration behaviour of the liquor itself: floc size, floc strength, and the amount of colloidal and soluble material. Dosing powdered activated carbon (PAC) or a flocculant helps by building larger, less viscous flocs and reducing the chance of pore clogging. Overdosing is a real risk, since excess polymer or coagulant suppresses sludge activity and the reactor loses treating capacity.
Physical cleaning is the other half of the story. Submerged modules rely on air scouring to generate cross-flow across the surface, and the suction pump usually runs intermittently so the cake can relax and lift. Relaxation and backwash intervals, not chemical cleaning frequency, are what most operators actually tune.
Configuration choices that decide the power bill
Side-stream MBR was the first type studied and remains the easiest to maintain: modules sit outside the reactor, so cleaning, replacement and addition need no tank entry. High-velocity cross-flow keeps solids off the surface, and the price is pumping energy. It suits difficult, high-strength or toxic streams where biological reliability matters more than kilowatt-hours.
Submerged, or integrated, MBR puts modules straight into the reactor and pulls permeate by suction, with aeration providing the scouring. Energy demand and footprint both fall sharply, which is why nearly all large municipal plants use this arrangement. Adding carriers turns it into a hybrid, with attached growth supplementing the suspended biomass and damping shock loads.
For anaerobic duty the membrane solves a different problem. Granular sludge in a UASB reactor washes out when loading fluctuates, and a membrane retains it without a three-phase separator. Two-phase configurations benefit as well: retention keeps acidogens concentrated in the acidification stage and holds macromolecules there long enough to hydrolyse, which sustains a high acidification rate. Because there is no aeration to provide scouring, anaerobic MBRs are built in side-stream form to keep the sludge in suspension.
The trade-off is honest. Sludge production falls to roughly 30% of a conventional process, which simplifies dewatering, but the energy saved on sludge handling is partly spent on membrane aeration and pumping. Whether the balance closes depends on discharge limits, land cost and the value of the reuse water.
Where MBR stops being the right answer
If a plant has space, cheap power and a forgiving effluent limit, conventional activated sludge with a well-run clarifier will usually cost less over twenty years. MBR earns its capital premium in three situations: land is not available, the discharge limit is tight enough that a polishing step would be needed anyway, or the water is going to be reused. In each case the membrane is doing work that would otherwise require a separate unit.
What it cannot do is fix an influent problem. Oil and grease, abrasive grit, hair and fibres still have to come out upstream; a membrane asked to handle them will foul faster. Pretreatment carries more weight in an MBR than in a conventional plant, not less.