A compact, multi-stage plant design cools, skims, digests, and polishes palm oil mill effluent to hit Indonesia’s toughest limits—and spins off biogas along the way.
Industry: Palm_Oil | Process: Palm_Oil_Mill_Effluent_(POME)_Treatment
Palm Oil Mill Effluent (POME) is as abundant as it is challenging: 2.5–3.5 m³ per tonne of crude palm oil (CPO) with biochemical oxygen demand (BOD) often exceeding 20,000 mg/L and chemical oxygen demand (COD) ranging from ~16,000–100,000 mg/L (www.mdpi.com; www.mdpi.com; www.mdpi.com). In Indonesia alone, 42.87 million tonnes of CPO in 2018 translates into roughly 100–150 million m³ of POME every year (www.mdpi.com; www.mdpi.com).
The stakes are high. Updated Indonesian standards demand final BOD₅ ≤100 mg/L, COD ≤350 mg/L, total suspended solids (TSS) ≤250 mg/L, oil & grease ≤25 mg/L, total nitrogen ≤50 mg/L, and pH 6–9 (earlier BOD/COD limits were 250/500 mg/L) (www.mdpi.com). This guide lays out a comprehensive, multi-stage design—pre‑treatment, high‑rate anaerobic digestion, then aerobic polishing—targeted at senior environmental engineers and mill managers.
Regulatory limits and load profile
Untreated POME typically presents BOD ≈10,000–44,000 mg/L and COD ≈16,000–100,000 mg/L, with ~0.6–0.7% oil by volume (v/v) (www.mdpi.com; www.mdpi.com). POME is >95% water with a few percent solids/oil, but the organic load is extreme and pushes conventional pond systems to their limits (www.mdpi.com).
The discharge targets—BOD₅ ≤100 mg/L, COD ≤350 mg/L, TSS ≤250 mg/L, oil & grease ≤25 mg/L, total nitrogen ≤50 mg/L at pH 6–9—set the specification for the downstream polishing train (www.mdpi.com).
Pretreatment: cooling, screening, oil removal
Raw POME exits clarification at ~80–90 °C and must be cooled to protect the biology downstream. A “cooling pond” with 7–10 days residence is conventional, while compact plants rely on a heat exchanger or a short pond of 1–2 days to bring the stream to ~35–38 °C for mesophilic operation (www.mdpi.com; www.mdpi.com). Coarse solids are removed via rough screening (~1–2 mm) and short settling, which can strip >20–30% of suspended solids (www.mdpi.com).
In practice, screens are deployed at the headworks; compact plants often adopt an automatic screen, with coarse debris sometimes handled by a manual screen upstream.
Residual oil—commonly ~0.6–0.7% v/v—requires an oil–water separator or fat‑trap tank. Typical designs provide ~8–10 hours retention or a de‑oiling tank of ~0.3–0.5 days; a “fat trap” of ~0.4–1 day hydraulic retention time (HRT) is reported to remove >80% of free oil, with skimmed oil recycled or sold (palm oil sludge) (www.mdpi.com; www.mdpi.com). Industrial skimming is commonly delivered through dedicated oil removal units.
Well‑configured pretreatment removes coarse TSS and oil & grease (O&G), reducing influent BOD/COD by roughly 10–20% before biological treatment; after pretreatment, the near‑neutral pH stream feeds the anaerobic stage (www.mdpi.com). Upfront physical works are often bundled as primary separation packages.
High‑rate anaerobic digestion (UASB/hybrids)
The anaerobic reactor removes the bulk of the organic load and generates biogas. Mesophilic operation near 35 °C stabilizes rates; pH is maintained around 6.8–7.2 with mixing via mechanical means or gas recirculation (www.sciencedirect.com). Typical HRT is 15–25 days, with 22 days reported as optimal in a large‑scale POME digester (www.sciencedirect.com).
Organic loading rate (OLR) targets sit around 5–15 kg COD/m³·d, with reported ranges from ~2 to 15 kg COD/m³·d in POME UASBs (www.mdpi.com). For a 100 t/d CPO mill (~300 m³/d POME), a 20‑day HRT implies ~6,000 m³ of digester volume—typically delivered in parallel tanks or stages for sludge retention and control.
Performance is robust: well‑operated high‑rate systems typically remove 80–95% of influent COD, with reported values up to ~96–97% (Table 3) in UASB‑based systems; a conservative design basis is ~90% COD removal (www.mdpi.com). Overall, the high‑rate digester should achieve roughly 80–90% BOD removal (well above 95% for COD), leaving an effluent around ~2,000–3,000 mg/L BOD if the influent is ~20,000 mg/L.
Biogas is ~60–65% methane (CH₄), with yields around 0.23–0.30 m³ CH₄ per kg COD removed. An optimized plant running at 22 days HRT with a 148% recirculation ratio generated ~859 m³/h total biogas and 0.259 m³ CH₄ per kg COD removed; at 0.259 m³ CH₄, yields translate to ~2.5–3.0 kWh of electrical power per kg COD removed (www.sciencedirect.com). For a 100 t/d CPO plant (~300 m³/d POME), order‑of‑magnitude expectations are ~200–300 m³/d methane, or roughly 8–12 MWh/d of energy.
Reactor configurations proven on POME include UASB, UASBF (fixed film), two‑stage acidogenic/methanogenic trains, and hybrids such as UASFF or UASB‑HCPB, which retain sludge and handle higher loads (www.mdpi.com). Suppliers typically deliver these within integrated anaerobic digestion systems, with start‑up inoculation often aided by starter bacteria.
Biogas capture and utilization
All biogas is captured and used—none is openly vented. At ~60–65% CH₄, options include combined‑heat‑and‑power to supply mill electricity and steam, or supplemental heat in dryers/boilers. In one case, ~859 m³/h biogas (~520 m³/h CH₄) supported a 3–5 MW engine (www.sciencedirect.com).
The climate implication is material. Roughly ~9–18% of a mill’s CO₂‑equivalent emissions can stem from POME methane—biogas recovery directly knocks down that share (www.researchgate.net).
Aerobic polishing to compliance
The anaerobic effluent—dark, colloidal, near‑neutral pH (~7), with residual soluble organics—feeds an aerobic reactor to meet final limits. Activated sludge (continuous) or sequencing batch reactor (SBR) configurations are common; SBRs offer control under variable loads (www.mdpi.com). Many mills standardize on activated sludge basins, while others prefer SBR trains for batch control or MBBR biofilm reactors.
Design anchors include a 20–24 hour HRT, F/M of ~0.2–0.5 kg BOD/kg MLSS·d, mixed liquor suspended solids (MLSS) of ~3,000–5,000 mg/L, and dissolved oxygen ≥2 mg/L. In polishing duty, aerobic removal is powerful: 96–99% BOD removal is typical in ~1 day, with one study reporting >99% BOD removal in a 22‑hour SBR (www.mdpi.com).
Effluents then settle in a clarifier before discharge or reuse. With BOD driven to ~20–50 mg/L and COD, for example, to ~160 mg/L from a 4,000 mg/L influent at 96% removal, the stream meets Indonesian COD ≤350 mg/L and BOD ≤100 mg/L specifications (www.mdpi.com). Settling is typically handled by a final clarifier.
Nutrients, nitrification, and TN control
Nitrogen emerges from anaerobic breakdown as ammonia (typically 500–1,000 mg/L N). Given sufficient sludge retention time (SRT ≥10–15 days) and oxygen, the aerobic stage nitrifies ammonia to nitrate; if total nitrogen (TN) ≤50 mg/L is enforced—as in Indonesia—additional denitrification or land application may be required (www.mdpi.com). Plants often manage nutrients with standard nitrification/denitrification controls and biological nutrient aids such as nutrient adjuncts and engineered nutrient removal stages.
Performance and footprint metrics
Stacking ~90% anaerobic with ~98% aerobic removal yields >99% overall BOD reduction; a nominal 20,000 mg/L BOD influent falls to <20 mg/L. COD removal exceeds ~98%, taking POME from ~40,000 mg/L to ≈200–300 mg/L, well under the 350 mg/L limit (www.mdpi.com; www.mdpi.com).
Operators often target final BOD <50 mg/L and COD <300 mg/L, with oil & grease <5 mg/L (vs 25 mg/L allowed) and TSS <30 mg/L to provide margin under the 250 mg/L limit. Total hydraulic residence for the compact line—cooling + anaerobic + aerobic—runs ~25–30 days, versus 80–100 days for open pond trains (www.mdpi.com; www.mdpi.com).
For the same 100 t/d CPO reference (~300 m³/d POME), a ~6,000 m³ anaerobic stage pairs with an ~300 m³ aerobic volume—orders of magnitude smaller than extended pond footprints—while still delivering clarifier underflow as manageable sludge.
Energy recovery and emissions cuts
Using literature yields of ~0.26 m³ CH₄/kg COD removed at ~60% CH₄, one example plant (≈10,000 kg COD/d) would produce ~2,600 m³ CH₄/d, equivalent to ~26,000 kWh (95 MJ) of energy; the setup can drive 2–3 MW of power continuously. A published full‑scale case reached 859 m³/h biogas on POME with 0.259 m³ CH₄/kg COD removed—consistent with these design targets (www.sciencedirect.com).
Implementation and operations
Capital costs exceed earthen ponds, but the engineered system slashes land take and runtime. Core spends cover digester tanks (concrete or steel), gas handling (flare or engine), aeration blowers, and clarifiers; many mills assemble these with wastewater ancillaries such as control valves and instrumentation.
Operations hinge on careful anaerobic start‑up (granule development), pH control, and aeration tuning; online monitoring (BOD sensors, ORP, gas flow) tightens performance. Primary and secondary sludge are handled via dewatering; anaerobic sludge can be co‑digested or dried and used as a soil conditioner after pathogen kill, while excess activated sludge may be land‑applied or further treated.
Energy recovery is central. At 0.259 m³ CH₄ per kg COD removed, methane production is predictable for generator sizing and long‑term ROI (www.sciencedirect.com).
Data‑driven rationale and cited sources
Pretreatment imperatives—cooling to 35–38 °C, oil trapping with ~0.5‑day retention, and coarse solids screening—are drawn from conventional pond practice and modern compact analogs (www.mdpi.com; www.mdpi.com). UASB and hybrid reactors consistently report ~90–97% COD removal on POME (bench to field), with 20–22 days HRT frequently optimal; the observed methane yield of 0.26 m³/kg COD removed aligns with mechanistic expectations (~0.35 kg COD → 0.25 m³ CH₄, ~75% of theoretical) (www.mdpi.com; www.sciencedirect.com).
For aerobic polishing, SBR trials demonstrate 99% BOD removal over ~22 hours, consistent with the 96–99% range cited for polishing duty (www.mdpi.com). The regulatory end‑points—BOD₅ ≤100 mg/L, COD ≤350 mg/L, TSS ≤250 mg/L, oil & grease ≤25 mg/L, TN ≤50 mg/L, pH 6–9—are compiled for Indonesia in Table 2 of Saputera et al. (2021) (www.mdpi.com).
Notably, with 98% COD removal on a 40,000 mg/L influent, final COD is ~800 mg/L—still above the 350 mg/L limit; the additional 99% by aerobes drives it down to ~10 mg/L. Even allowing conservative 90+98% totals, final COD falls below 350 mg/L; a cautious design (e.g., 95% & 98%) ensures margins (e.g., 40,000 → 2,000 → 40 mg/L) (www.mdpi.com; www.mdpi.com).
