Submerged Aerated Filters and Biological Aerated Filters: Volumetric Loadings, Media Choice and the Hydraulics That Decide Performance

SAF and BAF: the same biology, different hydraulics

Both are submerged fixed-film reactors. What separates them is what the media does with the solids:

• A SAF uses structured or random plastic media of large voidage (90 percent or more). Solids sloughed from the biofilm pass through and are removed in a downstream settlement stage; there is no backwash. Such a reactor is a biological contactor only.
• A BAF uses a granular bed (expanded clay, shale or similar, 2 to 6 mm, specific gravity about 1.5 to 1.8, or floating polystyrene beads) that acts as a biological reactor and a depth filter at once. Suspended solids are captured in the bed, headloss builds, and the bed is backwashed with air and water once or twice a day, producing a backwash liquor that is settled or returned to the inlet. No final settlement is needed.

A SAF is simpler and tolerant; a BAF is more compact and produces a better effluent but needs backwash pumps, blowers, storage and control, and its performance depends on media size: 3 to 6 mm for carbonaceous duty, 2.5 to 5 mm where nitrification is required, with smaller media giving more surface and better filtration at the cost of more frequent washing. Flow configurations vary (upflow and downflow, co-current and counter-current air) and each vendor has a preferred one; the loading relationships below apply across them.

Volumetric loading limits

Fixed-film capacity is expressed as mass of pollutant per cubic metre of media per day. Operating data from UK BAFs give the working envelope:

Hydraulic loading is often the real limit in a BAF, since the bed must filter as well as treat: 2 to 6 m/h is the usual range, higher for tertiary duty. SAF loadings are lower per cubic metre of reactor because the media surface area is lower and the solids are not retained; typical design values are 0.5 to 1.5 kg BOD/m³/d and 0.1 to 0.3 kg NH₄-N/m³/d, with temperature correction in the same way as any nitrifying process (see the temperature guide). All these figures assume adequate oxygen: a fixed film at 2 kg BOD/m³/d demands about 2 to 2.5 kg O₂/m³/d, which with coarse-bubble transfer efficiencies of 1 to 2 percent per metre of depth sets an air rate of the order of 20 to 40 m³ air per m³ of media per hour. Air also scours the biofilm and keeps it thin, which is why under-aerated SAFs blind and fail.

Worked sizing: a 2,000 PE works with a 5 mg/L ammonia permit

Load: 2,000 × 60 g = 120 kg BOD/d and 2,000 × 8 g = 16 kg NH₄-N/d after primary settlement removes about 30 percent of the BOD, leaving 84 kg/d to the SAF.

• Carbonaceous SAF stage at 1.2 kg BOD/m³/d: 84/1.2 = 70 m³ of media, say two cells of 36 m³.
• Nitrifying SAF stage at 0.2 kg NH₄-N/m³/d at 10 °C: 16/0.2 = 80 m³, two cells of 40 m³. The nitrifying stage is larger than the carbonaceous one, which is normal for a tight ammonia permit.
• Air: about 2 kg O₂/m³/d across the first stage and 4.6 kg O₂ per kg of ammonia oxidised in the second, 74 kg O₂/d, which at 3 m depth and 5 percent field transfer efficiency (air carries about 0.28 kg O₂/m³) is roughly 220 m³/h of air for the second stage alone. Scour at 20 to 40 m³/m³/h on 80 m³ of media is far larger, so scour governs the blower duty.
• Settlement: a humus tank sized at 1 to 1.5 m/h surface loading at peak flow for the sloughed solids.

A BAF alternative at 2.5 kg BOD/m³/d combined duty would need about 34 m³ of granular media for the carbonaceous load but would not nitrify reliably at that loading, so a two-stage BAF of similar total volume to the SAF, without the humus tank but with backwash plant, is the comparison. The choice between them is usually made on operator skill and site constraints rather than on treatment capacity.

What the hydraulics inside the reactor do to performance

Design is normally done on inlet and outlet concentrations, but tracer studies on a 7.7 m³ pilot SAF with internal recirculation, run at media fill ratios from 0 to 100 percent with and without biofilm, show that the internal flow pattern decides what the loadings above actually deliver. With the media and aeration in place the pilot behaved as an efficient plug-flow reactor, and the optimum fill ratio differed by pollutant: a 50 percent fill gave the best BOD removal (70 percent, a mass removal of 1.6 kg/m³/d) because the free volume allowed mixing and solids transport, while a 100 percent fill gave the best ammonia removal (60 percent at 0.14 kg/m³/d) because the nitrifiers need the surface area and the quieter, more plug-flow conditions. Back-mixing between cells improved as flow passed to the second cell.

Practical lessons: a carbonaceous SAF benefits from free volume and vigorous mixing, a nitrifying SAF from full media and plug flow, which supports staging them separately rather than asking one reactor to do both; forward velocity through the media, not just hydraulic retention time, governs mass transfer to the film; and a tracer test at commissioning, which takes a day, reveals short-circuiting and dead zones that no amount of sampling the effluent will. Tracer methodology is covered in the residence time distribution guide.

Operation and the common failure modes

BAF problems, from operating experience at UK works, cluster around five causes: influent load or hydraulic peaks beyond the design envelope, including toxicity from trade effluent; poor process air distribution and control, which lets parts of the bed go anaerobic; backwash rate, frequency and the handling of the backwash liquor, which if returned unsettled to the inlet simply recirculates the solids; media loss or attrition; and headloss build-up from biological growth that the wash does not remove, the "mud-balling" familiar from sand filters. SAF failures are simpler: under-aeration leading to a thick, sloughing film and solids carry-over; media blinding by fats or fibrous material where screening is inadequate; and ammonia breakthrough in winter where the nitrifying stage was sized for summer.

Both reactors suit packaging. Steel or GRP cells with media, aeration grid and blowers delivered as a tested unit are the basis of many packaged treatment plants, and the enclosure helps the winter temperature case. The related attached-growth options, moving-bed reactors and trickling filters, are compared in the MBBR guide and the trickling filter guide.

How to size a two-stage SAF for a small works

1. Set the loads at the design temperature. BOD and ammonia mass per day after primary settlement, at the winter wastewater temperature and the peak-week load.
2. Size the carbonaceous stage. Media volume from the BOD load at 0.5 to 1.5 kg BOD/m3/d, in at least two cells; allow free volume for mixing.
3. Size the nitrifying stage. Media volume from the ammonia load at 0.1 to 0.3 kg NH4-N/m3/d corrected to winter temperature, with full media fill and plug-flow configuration.
4. Size the aeration. Oxygen demand from BOD removed plus 4.6 kg per kg ammonia oxidised, divided by the field transfer efficiency at the media depth; include scour.
5. Size the settlement. Humus tank at 1 to 1.5 m/h at peak flow; return or treat the humus sludge.
6. Commission with a tracer test. Confirm plug flow and the absence of short-circuiting before loading the reactor, and log temperature and ammonia through the first winter.