The steel mill that drinks its own water: inside the API‑DAF‑filter loop turning oily scale into supply

Hot and cold rolling mills push vast volumes of water through descaling and cooling—then take it back. Integrated steel plants draw ≈28.6 m³ per tonne and discharge ≈25.3 m³ per tonne, for a net ≈3.3 m³ per tonne consumed, according to Colla et al. (mdpi.com). Electric‑arc furnaces are even leaner: ≈1.6 m³ per tonne net (mdpi.com). In practice, ~88–94% of intake is recycled (mdpi.com), and a survey of modern mills shows reuse rates routinely above 70%—25 of 31 Chinese steelworks (~81%) reported >70% reuse and about 32% exceeded 90% (iwaponline.com).

Designing for those numbers starts with a scale pit to drop heavy iron oxides, adds oil‑water separation to strip lubricants, and finishes with filtration to polish out fine suspended solids. The prize is reuse—cooling towers, quench circuits, and other non‑potable loops—at the quality and reliability rolling lines demand.

Wastewater loads and reuse targets

Rolling trains shed mill scale (iron oxides) and oil/grease from coolants and lubricants, plus trace metals. Integrated mills recycle ~88–94% of intake water in practice (mdpi.com), and many plants now target 80–90% reuse. Surveys back it up: ~81% of Chinese steelworks reported >70% reuse, ~32% exceeded 90% (iwaponline.com).

Case in point: in Indonesia, Krakatau Posco treats 17,000 m³/d of process water—including rolling mill effluent—and reuses about 700 m³/d (lestari.kompas.com), meeting national effluent (“baku mutu”) standards for discharge or reuse (lestari.kompas.com).

Scale pit sedimentation design

The first barrier is a deep scale pit—a sedimentation basin that lets heavy ferrous oxides settle by gravity. Detention is typically ~30–60 minutes (nepis.epa.gov) to drop 50–100 μm particles. A typical layout is long and rectangular (length:width ≥5:1) with depth ~8–10 m or more, surface oil skimmers, and bottom sludge pumps; heavy sludge is scraped or dredged periodically (often weekly) for recycle to sinter plants.

Real‑world loads are not trivial. A US EPA study found a trio of hot‑strip scale pits removed ≈4,085 t of iron scale in 11 months (≈124 t/day), equivalent to ~1,300 mg/L of solids in the influent (nepis.epa.gov). In that case the pits discharged “low net quantities of TSS” (total suspended solids) (nepis.epa.gov), but they do not remove oil or fine emulsions; the cited study found no effect on oil/grease levels (nepis.epa.gov). For primary hardware integration, many plants match this gravity stage with compact oil devices from waste‑water physical separation portfolios.

Gravity and flotation oil removal

Once the scale is gone, oil is next. An API‑type gravity separator (a horizontal‑flow tank that uses density differences) removes free oil and any remaining settleable solids—oil (density ~0.8) floats, solids sink, and clarified water passes through (parsianfarab.com). Design rules of thumb are L:W ≈5:1 and velocity <3 ft/min, with 80–90% of free oil skimmed off (parsianfarab.com). Free oil recovery can be paired with packaged units like oil removal skimmers to keep top layers clean.

Emulsified droplets need flotation. A DAF (dissolved‑air flotation) injects air‑saturated water to form microbubbles that attach to oil and fine solids, floating them into a skimmable sludge blanket (parsianfarab.com). In oily steel mill service, DAF units handle influent oil/TSS up to ~500 mg/L and remove on the order of 95% of oil and TSS (parsianfarab.com); many facilities deploy purpose‑built packages such as DAF systems. It’s common to dose a coagulant or flocculant just ahead of DAF to agglomerate fines—managed with dosing pumps and chemicals like coagulants—because an API on its own will not reduce emulsified oil below ~50–100 mg/L.

Tertiary filtration and membrane polishing

With oil removed, tertiary filters sweep up residual fines. Multimedia beds with sand and anthracite are standard; dual‑media designs use media such as sand/silica filtration topped with anthracite media to reach low turbidity and a few mg/L TSS. For tight cuts or duty protection upstream, mills layer in cartridge filters (1–100 μm).

Examples are industrial “high‑rate filters” packed with slag/rock media; in one case the filter effluent was clear enough to feed the cooling towers (givemechallenge.com). Where trace organics linger, plants may add activated carbon. And when closing loops aggressively, membranes come into play: Colla et al. note that coupling ultrafiltration with reverse osmosis can yield “very high quality” water suitable for reuse (mdpi.com). Typical building blocks include ultrafiltration for pretreatment and brackish‑water RO for conductivity control within a broader membrane system architecture.

Reuse pathways and water balance

The polished effluent—essentially mill‑scale‑free and oil‑free—heads back into cooling circuits. In one hot‑strip mill, filtered effluent was pumped directly to cooling towers for reuse (givemechallenge.com), echoing Colla et al., who describe hot strip mill water that after sedimentation and decantation “goes to the cooling tower” (mdpi.com).

Targets are ambitious but proven. Modern integrated mills reuse ~88–94% of process water (mdpi.com), and the Chinese survey shows ~81% of firms above 70% reuse and ~32% above 90% (iwaponline.com). The economic impact is large: at 1 tonne steel, fresh water demand falls from ~28 m³ to just ≈3 m³ (mdpi.com). In the Indonesian example, the treated water (<80 μS after full RO) is explicitly used in making steel (lestari.kompas.com), underscoring why UF+RO loops are favored for high recovery and quality (mdpi.com).

Compliance and performance benchmarks

Regulatory endpoints are tight. US EPA guidance assumes 30 mg/L TSS (30‑day average) is attainable with clarifiers and flotation (nepis.epa.gov). Indonesian rules (PermenLHK) set strict TSS and oil limits for industrial discharge; the Krakatau Posco system treats to “baku mutu” for sea discharge or reuse (lestari.kompas.com), and process trains often target ≤30–50 mg/L TSS and near‑zero oil based on achievable norms (nepis.epa.gov).

Performance data points are consistent with those targets. One Indian hot‑strip mill reports polishing water to ~2–3 mg/L solids so it can be recycled (givemechallenge.com). A Chinese project showed UF+RO giving “very high quality” water with 80 μS conductivity and favored reuse over discharge (mdpi.com) (lestari.kompas.com). By meeting or exceeding such standards, treated water can safely return to the plant or local watershed.

System outcomes and circular gains

A properly designed train—scale pit + API/DAF + filters—typically removes >95% of TSS and oil, meeting regulatory limits and enabling high reuse. Effective treatment of scale, oil, and TSS can cut fresh intake by ~85–90% and sharply reduce effluent, aligning steel production with circular‑economy goals (mdpi.com) (iwaponline.com). As the Indonesian case demonstrates, treated water—down to <80 μS after full RO—can be explicitly used in making steel (lestari.kompas.com), while API and DAF ensure the front‑end knocks out free and emulsified oils at the rates expected in industry (parsianfarab.com) (parsianfarab.com).

References: Colla et al. (2017) for water balances and reuse (℞28.6 m³/ton intake, 3.3 m³/ton net) (mdpi.com); Liang et al. (2023) for surveyed reuse rates (≈85%–98%) (iwaponline.com); US EPA for achievable TSS standards (30 mg/L) (nepis.epa.gov) and real‑world pit performance (≈1,300 mg/L influent solids; ≈4,085 t removed in 11 months; no oil/grease effect) (nepis.epa.gov) (nepis.epa.gov); technology summaries for API/DAF (parsianfarab.com) (parsianfarab.com); Indonesian mill results and “baku mutu” compliance (lestari.kompas.com) (lestari.kompas.com) (lestari.kompas.com). All figures and claims reflect those sources.