Why Starch Wastewater Is a Different SBR Design Problem
Raw starch liquor from corn, tapioca, wheat, and cassava lines carries 3,000–15,000 mg/L COD, 1,500–7,000 mg/L BOD5, 100–400 mg/L TN, and 50–250 mg/L NH3-N, with BOD5/TN ratios routinely below 4 (Zhongsheng field data, 2026). That envelope is the root cause of every nitrification failure in a starch plant: the residual dissolved BOD out-competes slow-growing autotrophic nitrifiers for dissolved oxygen, and once the liquor enters the aerobic stage the electron donor for denitrification is already gone. Field data show that when BOD5/TN drops below 4, total nitrogen removal plateaus at 40–60% no matter how much aeration is added (Zhongsheng field data, 2026).
The hydraulic pattern compounds the carbon problem. Starch lines run batch wash-downs and CIP cycles that push NH3-N spikes of 2–4× the daily mean through the equalization basin, and a 2× shock can push nitrifier SRT out of its working window unless the basin is sized for 8–12 h HRT and pH is held below 8.0. Per EPA's SBR factsheet (EPA 932-F-99-073), the SBR itself acts as an equalization basin during fill, which is part of why a single-basin time-sequenced system absorbs these spikes more gracefully than a continuous-flow train — but only if the cycle is long enough to let nitrifiers recover.
The design driver is therefore not the SBR hardware; it is the C/N correction step. Whether the engineer recovers carbon upstream (hydrolysis-acidification recycle, VSW sidestream), doses methanol downstream, or accepts a relaxed TN target, that decision must be locked before the cycle is timed and the MLSS setpoint is selected.
How the SBR Cycle Phases Map to Starch Organics and Ammonia
Per EPA's definition, an SBR is a fill-and-draw activated sludge system that operates in time rather than in space (EPA 932-F-99-073). The unit processes — equalization, biological treatment, secondary clarification — all happen in a single vessel under a timed control sequence. For starch wastewater, the six phases run as follows:
- Fill (1.0–1.5 h): Raw or equalized liquor enters the biomass already present in the reactor. Static fill (no aeration, no mixing) is used when the carbon needs to be absorbed onto the floc before nitrification starts; mixed fill runs when the operator wants to pre-denitrify.
- Anoxic react (1.5–2.0 h): Mixing only, DO held below 0.5 mg/L. This is the denitrification window and it must come before the aerobic phase when BOD5/TN is below 4 — the residual dissolved BOD is the electron donor, and once aeration starts, nitrifiers consume the DO and the carbon is gone.
- Aerobic react (4.0–6.0 h): Aeration on, DO 2–3 mg/L. Nitrification of the remaining NH3-N, plus residual BOD polishing. This is the longest phase and the one that protects the 20–30 day SRT window.
- Settle (1.0–1.5 h): Aeration and mixing off, sludge blanket forms. No influent or effluent currents interfere with settling, which is one of the SBR's structural advantages (EPA 932-F-99-073).
- Decant (0.5–1.0 h): A floating or fixed decanter removes the clarified supernatant. Decanter selection is the primary distinguishing factor between SBR manufacturers (EPA 932-F-99-073), and for starch service a weir-type decanter with scum baffle is the working choice because starch carryover floats.
- Idle / sludge wasting (0.5–1.0 h): Wasting window for SRT control. Wasted sludge goes to a thickener or directly to dewatering.
Total cycle time on starch liquor is 8–12 hours for a two-cycle-per-day SBR or 6–8 hours for a three-cycle-per-day configuration. Plants that cannot buffer influent into a discrete batch should consider the Intermittent Cycle Extended Aeration System (ICEAS), a continuous-feed modified SBR variant (EPA 932-F-99-073). The phase allocation below is a working envelope; operators re-phase on the fly to chase effluent quality or absorb a CIP peak.
| Phase | Duration (h) | DO target (mg/L) | Process function on starch liquor |
|---|---|---|---|
| Fill | 1.0–1.5 | 0–0.5 | Carbon absorption onto floc; volume make-up |
| Anoxic react | 1.5–2.0 | <0.5 | Denitrification using residual BOD as electron donor |
| Aerobic react | 4.0–6.0 | 2.0–3.0 | Nitrification of NH3-N; residual BOD oxidation |
| Settle | 1.0–1.5 | 0 | Sludge blanket formation, no currents |
| Decant | 0.5–1.0 | 0 | Supernatant withdrawal via decanter |
| Idle / waste | 0.5–1.0 | 0 | Sludge wasting for SRT control |
SBR Design Parameters for Starch Wastewater

The 2026 SBR envelope for starch plants is HRT 24–36 h, MLSS 3,500–5,000 mg/L, SRT 20–30 days, cycle 6–12 h, achievable NH3-N removal 90–96% and TN removal 75–85% (Zhongsheng field data, 2026). SRT is the binding control for nitrification because autotrophic nitrifiers grow slowly; at SRT above 20 days a 2× NH3-N shock recovers within 2–3 SRTs, typically 40–60 days, provided pH is held below 8.0 to keep free NH3 below 5 mg/L (Zhongsheng field data, 2026).
EPA's broader SBR HRT range spans contact stabilization at 3.5–7 h on one end and extended aeration at 18–36 h on the other (EPA 932-F-99-073). Starch sits at the extended-aeration end of that range because the high residual BOD and the need to keep nitrifiers alive through CIP spikes both push the design toward longer HRTs. DO targets are phase-specific: anoxic below 0.5 mg/L, aerobic react 2–3 mg/L. Aerating above 3 mg/L in the aerobic phase does not improve nitrification rate and wastes blower power; dropping below 2 mg/L risks incomplete nitrification and floating sludge.
The wheat-starch RSM benchmark is the proof point that this envelope is achievable. A response-surface-optimized SBR treating synthetic wheat-starch liquor (cow-dung bio-sludge seed) delivered 92% COD removal, 38 mg/L effluent TSS, and SVI of 57 mL/g in the optimum condition — SVI 57 is a healthy, well-settling sludge and confirms the cycle can hold a dense blanket without bulking (Desalination and Water Treatment, ScienceDirect, 2016). For plants evaluating a biological train against an MBR system for plants that exceed the SBR's 15 mg/L TN ceiling, the wheat-starch data also shows the SBR's settle-and-decant architecture produces a clean enough supernatant for direct filtration polishing — but the SBR does not hit <15 mg/L TN on its own, and that gap is where MBR pulls ahead.
| Parameter | Starch SBR envelope (2026) | Source |
|---|---|---|
| HRT | 24–36 h | Zhongsheng field data, 2026 |
| MLSS | 3,500–5,000 mg/L | Zhongsheng field data, 2026 |
| SRT | 20–30 days | Zhongsheng field data, 2026 |
| Cycle time | 6–12 h (2–3 cycles/d) | EPA 932-F-99-073; Zhongsheng, 2026 |
| F/M | 0.08–0.15 kg BOD/kg MLSS·d | Extended-aeration design range |
| DO (anoxic) | <0.5 mg/L | Zhongsheng field data, 2026 |
| DO (aerobic) | 2.0–3.0 mg/L | Zhongsheng field data, 2026 |
| NH3-N removal | 90–96% | Zhongsheng field data, 2026 |
| TN removal | 75–85% | Zhongsheng field data, 2026 |
| Effluent TSS (achievable) | 38 mg/L | ScienceDirect 2016 (wheat-starch RSM) |
| SVI (achievable) | 57 mL/g | ScienceDirect 2016 (wheat-starch RSM) |
SBR vs A2/O vs MBR for Starch Plants
For a 200–1,000 m³/d starch line, the choice between SBR, A2/O, and MBR is driven by three questions: what is the TN limit, what is the available footprint, and is the influent batch or continuous. All three trains assume the pretreatment envelope of screening, grit removal, equalization at 8–12 h HRT, and DAF for FOG and TSS (Zhongsheng field data, 2026). None of them will fix a C/N below 4 without carbon dosing.
SBR runs HRT 24–36 h, MLSS 3,500–5,000 mg/L, NH3-N 90–96% and TN 75–85%, and is the lowest-capex retrofit for plants under 1,000 m³/d with batch discharge because the existing equalization basin can often be repurposed as the reaction vessel (Zhongsheng field data, 2026). A2/O runs HRT 18–28 h, MLSS 3,000–4,000 mg/L, NH3-N 90–95% and TN 70–80%, and is the default greenfield train for corn and cassava starch plants in China discharging to GB 8978 Class I or II — it also delivers biological phosphorus removal as a built-in. MBR runs MLSS 6,000–10,000 mg/L, NH3-N 95–98% and TN above 85% with effluent TN below 15 mg/L without tertiary polishing, at roughly 60% of the conventional activated sludge footprint (Zhongsheng field data, 2026). The 2026 discharge crosswalk: China GB 8978 sets NH3-N at 15 mg/L and TN at 40 mg/L; EU food and drink BAT-AEL ranges for direct discharge sit at TN 10–15 mg/L; Vietnam QCVN 40:2011/BTNMT column A holds TN at 20 mg/L for industrial discharge. MBR meets all three without polishing; A2/O and SBR need a tertiary step for the EU and Vietnam limits (Zhongsheng field data, 2026).
Selection anchor by scale: SBR wins below 1,000 m³/d with batch discharge; A2/O wins at 1,000–5,000 m³/d with continuous flow and a phosphorus limit; MBR wins when TN limit is below 15 mg/L or footprint is constrained. The polish step applies to any of the three when BOD5/TN drops below 4 — for a 500 m³/d starch line with 200 mg/L TN and an MBR producing 20 mg/L effluent NO3-N, the methanol demand is roughly 350–450 kg/d at 3.5 g methanol per g NO3-N, sized for 7-day peak demand (Zhongsheng field data, 2026). Where the project is a retrofit with civil works already in place, the MBBR for distillery wastewater design guide covers a related biological train that can be co-located with an existing equalization basin when footprint is the binding constraint.
| Parameter | SBR | A2/O | MBR |
|---|---|---|---|
| HRT | 24–36 h | 18–28 h | 18–30 h (basin) + membrane |
| MLSS | 3,500–5,000 mg/L | 3,000–4,000 mg/L | 6,000–10,000 mg/L |
| SRT | 20–30 days | 15–25 days | 30–60 days |
| NH3-N removal | 90–96% | 90–95% | 95–98% |
| TN removal | 75–85% | 70–80% | >85% (effluent <15 mg/L) |
| Footprint index (vs A2/O=1.0) | 1.1–1.3 | 1.0 | ~0.6 |
| Capex tier | Lowest | Medium | Highest |
| Best-fit plant size | <1,000 m³/d, batch | 1,000–5,000 m³/d, continuous | Any, TN <15 mg/L or footprint-constrained |
| Polishing needed for EU/Vietnam TN | Yes | Yes | No |
| Phosphorus removal | Chemical only | Biological (built-in) | Chemical only |
Pretreatment and Equalization the SBR Demands

Three unit operations are non-negotiable before the SBR: screening, grit removal, and equalization at 8–12 h HRT (Zhongsheng field data, 2026). A typical starch line running 200 m³/h of liquor flow needs roughly 1,800–2,400 m³ of equalization volume, sized for the longest inter-batch peak so that NH3-N shocks of 2–4× the daily mean do not push the nitrifier SRT out of its 20–30 day working window. The equalization basin also buffers pH excursions from CIP cycles, which is the second-most-common cause of nitrification failure after carbon starvation.
Per EPA's SBR factsheet, primary clarifiers are typically not required before an SBR unless TSS or BOD exceed 400–500 mg/L (EPA 932-F-99-073). Starch liquor will exceed this on both parameters, so the working pattern is screening plus DAF pre-treatment ahead of the SBR rather than a primary clarifier — the DAF removes the bulk of the suspended solids and emulsified FOG that would otherwise blind the SBR settle phase and float into the decant. The screening step at the headworks is a GX series rotary bar screen at the headworks with 3–6 mm aperture, sized for peak instantaneous flow including CIP. After DAF, the liquor typically lands at TSS 200–400 mg/L and BOD5 1,200–5,500 mg/L, which is the envelope the SBR cycle is actually designed to handle.
Where the project is a greenfield rather than a retrofit, the equalization basin can be designed as the SBR reaction vessel itself — the SBR fills during what would otherwise be the equalization window, and the cycle timing absorbs the influent variability. Where the project is a retrofit, the existing equalization basin is repurposed with mechanical mixers, aeration grid, and a decanter; the civil works capex drops by roughly 30–40% versus a new SBR tank.
When SBR Is the Wrong Choice for Starch Wastewater
SBR is not a universal answer, and four disqualification conditions should kill the option before it reaches the design basis. The first is a TN limit below 15 mg/L — only MBR clears that without tertiary polishing, and a polishing step on an SBR effluent (denitrifying sand filter, moving-bed biofilm reactor) erases most of the capex advantage the SBR had over MBR in the first place (Zhongsheng field data, 2026). The second is flow above roughly 5,000 m³/d or strictly continuous influent — the cycle and decanter capacity cap out, and A2/O becomes more economical to operate. EPA notes the largest SBR in the world is 10 MGD (~38,000 m³/d) in the UAE (EPA 932-F-99-073), but that envelope is municipal and assumes multiple basins in parallel; for a single-line starch plant, 5,000 m³/d is the realistic ceiling.
The third disqualification is BOD5/TN below 4 with no external carbon dosing allowed. The denitrification mass balance cannot close without an electron donor, and an SBR that can only remove 40–60% TN is not a compliant biological train. The fourth is a phosphorus limit with no chemical dosing budget — A2/O delivers biological P removal as a built-in anaerobic phase, while SBR and MBR both need chemical precipitation (Zhongsheng field data, 2026).
When one of these four conditions is hit, the next step is a sizing check against A2/O or MBR. For projects that need a polishing or sidestream carbon-recovery add-on, the MBR installation and commissioning guide and the broader food processing wastewater engineering guide cover the upstream and downstream integration points a designer needs before locking the basis of design.
Frequently Asked Questions
What cycle time should an SBR run on starch wastewater?
Total cycle time on starch liquor should be 8–12 hours for a two-cycle-per-day configuration or 6–8 hours for three cycles per day, with the aerobic react phase allocated 4–6 hours at DO 2–3 mg/L and the anoxic react phase allocated 1.5–2 hours at DO below 0.5 mg/L (Zhongsheng field data, 2026; EPA 932-F-99-073). The cycle must be phased so the anoxic window runs before the aerobic window when BOD5/TN is below 4.
When is SBR the right pick over MBR for a starch plant?
SBR is the right pick for plants under 1,000 m³/d with batch discharge and a TN limit at or above 20 mg/L, because the existing equalization basin can be repurposed as the reaction vessel and the capex runs 30–50% below MBR. MBR is the right pick when the TN limit is below 15 mg/L, the footprint is constrained, or the discharge goes to an EU BAT-AEL or Vietnam QCVN 40:2011 receiver that requires polishing on the SBR or A2/O effluent (Zhongsheng field data, 2026).
What is the standard methanol dose for denitrification on a starch line?
3–4 mg methanol per mg NO3-N removed, corresponding to a COD:N ratio of 6:1 to 8:1 (Zhongsheng field data, 2026). For a 500 m³/d starch line with 200 mg/L TN and an MBR producing 20 mg/L effluent NO3-N, the methanol demand is 350–450 kg/d, with storage and dosing sized for 7-day peak demand. This dose applies to any of the three trains (SBR, A2/O, MBR) when BOD5/TN drops below 4.
How long does an SBR take to recover from an ammonia shock?
At SRT above 20 days, a 2× NH3-N spike recovers within 2–3 SRTs, typically 40–60 days, provided pH is held below 8.0 to keep free NH3 below 5 mg/L (Zhongsheng field data, 2026). Recovery time scales linearly with shock magnitude: a 3× spike takes 60–90 days at the same SRT, which is why equalization at 8–12 h HRT is non-negotiable for any starch line running batch CIP cycles.