With total nitrogen caps tightening to 30 mg/L and even 15 mg/L in places, refinery wastewater can’t rely on hydrocarbon cleanup alone. The new playbook: aerobic nitrification plus anoxic denitrification, tuned by hard parameters and real-time control.
Petroleum refineries increasingly face discharge targets that treat nitrogen as seriously as oil. China’s GB 31570-2015 sets total nitrogen (TN) below 30 mg/L, and some local limits push TN below 15 mg/L (pubs.rsc.org). Typical raw refinery wastewater—after preliminary oil/fats‑oils‑grease removal via primary steps like waste-water physical separation and purpose‑built oil removal—can still carry about 20–30 mg/L TN, largely as ammonium and nitrate.
One real‑world snapshot: a Liaohe refinery “CAST” intermediate measured ~0.82 mg/L NH4–N and 22.3 mg/L NO3–N (TN≈24.2 mg/L) (pubs.rsc.org). And while a refinery activated‑sludge plant logged >93% chemical oxygen demand (COD) removal, it struggled with nitrogen—spurring a move to simultaneous nitrification/denitrification (SND) (researchgate.net). The countermeasure is firmly biological: nitrification (NH4+→NO3−) followed by denitrification (NO3−→N2 gas).
Two‑zone reactor architecture (MLE)
The workhorse is a Modified Ludzack–Ettinger (MLE) layout inside standard activated‑sludge basins: influent first enters an anoxic tank to mix with recycled nitrified liquor (nitrate‑rich) and its own readily biodegradable COD, then flows to an aerobic tank for nitrification (suezwaterhandbook.com). Plants return the nitrified mixed liquor back to anoxic at roughly 100–350% of influent flow to ensure nitrate is available where carbon is (suezwaterhandbook.com).
Carbon‑to‑nitrogen balance is pivotal. An influent BOD/N near ~4:1 generally underwrites complete denitrification; at 4:1, post‑denitrification NO3–N around ~5–7 mg/L can hit a 10 mg/L TN target from 40 mg/L TN influent (suezwaterhandbook.com). If organics are low or TN is high, that same nitrate cut demands added carbon or more reactor volume.
Variants: step‑feed, three‑zone, and cyclic
For tighter targets, a three‑zone process adds a third, often endogenous, anoxic stage after aeration to mop up residual NOx (suezwaterhandbook.com). Step‑feed distributes influent across multiple anoxic cells to better match nitrate with carbon. Oxidation ditches and sequencing‑batch‑reactor (SBR) cycles achieve the same aerobic/anoxic choreography in one tank by switching aeration on/off using real‑time ammonia and nitrate signals—air on only after anoxic denitrification consumes NOx, off when sufficient NH4+ has accumulated (suezwaterhandbook.com).
Nitrifiers grow slowly (roughly 1–2‑day doubling times), so sludge age matters: conventional designs hold solids retention time (SRT) around 10–15 days to avoid washout. Many plants raise retention with biofilm carriers—think moving bed bioreactors (MBBR), IFAS, or granular sludge—to decouple nitrifiers from daily wasting.
DO, pH, and alkalinity control
Dissolved oxygen (DO) is the master knob. Nitrification thrives near ~2 mg/L DO, while denitrification happens only when DO is below ~0.5 mg/L; anoxic tanks must be mixed without aeration so microbes use nitrate, not oxygen, as the electron acceptor. Too much oxygen in anoxic zones undercuts denitrification; over‑aeration in aerobic zones burns energy and can strip organics that would have fueled nitrate reduction.
Nitrification consumes alkalinity and generates acidity. About 7.14 mg/L as CaCO3 alkalinity disappears per 1 mg/L NH4–N oxidized (cwea.org). Without enough buffer, pH can sag (often stalling nitrification below ~pH 6.8) (cwea.org). Operators typically target pH 7–8 and ensure at least ~70 mg/L CaCO3 remains after nitrification (cwea.org), dosing lime or bicarbonate as needed via a precise dosing pump. Denitrification restores some alkalinity, but only partially offsets the overall demand; plants track pH and alkalinity (or nitrate profiles) to tune lime feed and keep pH ~7.0–8.0.
Temperature sensitivity and seasonal effects
Temperature swings hit nitrifiers hard. Growth rates roughly halve for every 10°C drop below ~30°C; nitrification becomes extremely slow below about 10°C. In tropical settings like Indonesia, ambient ~25–30°C is favorable, though rainy‑season cooling can slow rates; denitrifiers are less temperature‑sensitive and still perform to ~10°C given enough carbon, but total N removal still eases off.
Carbon supply and DAF sidestreams
Denitrification needs readily biodegradable carbon. If COD/TKN is below ~3–5 after primary and secondary treatment, external carbon (e.g., methanol or acetic acid) or an internal sidestream is required. A heavy‑oil refinery study found acetate COD/NO3–N ≈3.75 was optimal (pubs.rsc.org).
There’s also a low‑chemical tactic: use secondary dissolved air flotation (DAF) effluent with COD/NO3 ≈5.4 as the carbon source. In that study, adding 25% of this effluent volumetrically sustained stable denitrification without pure chemicals (pubs.rsc.org). This dovetails with how many refineries already run DAF for solids and oil separation upstream. As the SUEZ design rule notes, keeping BOD/N around ~4:1 typically delivers effluent TN ~5–7 mg/L and meets a 10 mg/L TN standard from 40 mg/L influent TN (suezwaterhandbook.com); when that ratio dips, operators either add carbon or adjust internal recycle.
Sludge age, MLSS, and mixing
Ammonia‑oxidizing (AOB) and nitrite‑oxidizing bacteria (NOB) grow slowly and yield little new biomass, so many designs keep mixed liquor suspended solids (MLSS) around ~3–6 g/L and SRT ≈10–15 days. In one refinery wastewater study, raising MLSS to ~10 g/L under SND significantly improved TN removal (researchgate.net).
Mixing quality matters: anoxic tanks need robust mixing without aeration to prevent dead zones and keep sludge suspended. In a Tehran refinery experiment, vigorous mixing (300 rpm in a 5 L reactor) helped complete SND (researchgate.net). For attached growth, immobilized media and biofilm stages align with fixed‑bed bio‑reactors that stabilize nitrifiers and denitrifiers.
Instrumentation and automated control
Reliable removal is a control problem as much as a biology problem. Plants increasingly automate DO, pH, and oxidation‑reduction potential (ORP) with online probes and ancillary gear sourced as wastewater ancillaries. In oxidation ditches, advanced “Green Bass” sequencing toggles aeration based on ammonia and nitrate: after anoxic denitrification depletes NOx, air is pulsed to nitrify until ammonia drops, then aeration halts again (suezwaterhandbook.com). Chemical additions—lime for alkalinity, acetate for carbon—are fine‑tuned with metered feeds through a dosing pump.
Performance benchmarks and case data
With correct design and control, studies report >80–90% total nitrogen removal and single‑digit mg/L effluents. A lab‑scale system treating simulated refinery wastewater achieved ~95% TN removal on a synthetic feed and ~87% on real effluent; nitrification wrapped in ~4 hours and denitrification in ~3 hours (researchgate.net).
In a heavy‑oil refinery pilot using an immobilized denitrification reactor followed by an aerobic biofilm reactor, researchers reported a volumetric N removal rate of 0.82 kg N/m³·day and cut TN from 24 mg/L to ~1.5 mg/L—well below a 15 mg/L limit (pubs.rsc.org, pubs.rsc.org). Even at hydraulic retention times of 6–8 hours, effluent NH4–N was ~0 mg/L and NO3–N ~<2 mg/L (pubs.rsc.org). A separate SBR‑based approach in Iran likewise delivered efficient nitrification and denitrification under tight DO control; at MLSS 10 g/L the team reported “very efficient” N removal by SND (researchgate.net, researchgate.net).
Full‑scale expectations track the pilots: activated‑sludge nitrification routinely trims ammonia to below ~0.5 mg/L. With adequate carbon, downstream denitrification frequently secures >90% ammonia removal and >80% total N removal. If influent TN is ~30 mg/L, sound biological nutrient removal (BNR) design can deliver ~<5–10 mg/L TN, depending on carbon supply. Design rules of thumb back this up: to meet TN<10 mg/L from 40 mg/L, hold BOD/N ≥4 and internal recycle ≈150% (suezwaterhandbook.com).
The process picture in one line
Conceptually, the flow is simple: an anoxic denitrification zone plus an internal nitrate recycle feeding an aerobic nitrification zone—and back again. Configurations vary, but every scheme partitions oxic and anoxic work, preserves nitrifier biomass, and controls carbon, DO, pH, and temperature to keep the biology on‑song.
Sources and technical notes
This report draws on peer‑reviewed refinery wastewater studies (pubs.rsc.org, researchgate.net, pubs.rsc.org), engineering manuals (suezwaterhandbook.com, suezwaterhandbook.com), nitrogen‑removal process reviews (sciencedirect.com, sciencedirect.com), and alkalinity/pH guidance (cwea.org). Regulatory context is detailed here (pubs.rsc.org). Quantitative outputs cited include >0.8 kg N/m³·day removal, ~90–95% N removal, and effluent NH4–N and NO3–N in the low mg/L range.
