Steel’s Thirst Meets Its New Playbook: Membranes, Math, and a March to ZLD

In an industry where water use runs deep, the numbers are blunt: typical steel plants pull in 28–30 m³ of water per tonne of steel and push almost as much back out (www.mdpi.com). One analysis pegs steelmaking at roughly 20% of global industrial water use (www.mheavytechnology.com). For integrated mills, withdrawals sit around ≈28.6 m³/t with 25.3 m³/t discharged; for electric‑arc mills, ≈28.1 m³/t in and 26.5 m³/t out (www.mdpi.com). The difference is essentially the small amount “consumed” by evaporation or incorporated into products.

Even modest recycling shifts millions of cubic meters. Targets are hardening: Tata Steel is driving to <1.5 m³/tonne freshwater by 2030 with Zero Effluent Discharge (ZED) (www.tatasteel.com), implying >90% reuse. Indonesia is tightening discharge standards (Government Reg No.22/2021), and observers have flagged enforcement gaps (www.mdpi.com)—a cue for voluntary best practice.

Process water mapping and quality bands

Getting to scale requires matching water quality to use. Total dissolved solids (TDS, dissolved salts), hardness, and oil drive the allocations. Reverse osmosis (RO, a pressure‑driven membrane process) sits at the core, often preceded by ultrafiltration (UF, 0.01–0.1 µm pore size) and sometimes nanofiltration (NF, softening‑oriented membranes).

Cooling towers and heat exchangers can tolerate relatively high TDS and hardness if scaling is controlled. One hot‑strip mill found recirculating cooling water needed very low chloride and salinity makeup to avoid surface defects on steel (www.mdpi.com). In practice, towers accept TDS on the order of 1,000–2,000 mg/L (at 4–6× cycles of concentration) with periodic blowdown. Membrane‑treated permeate post‑RO often has <500 mg/L TDS and slots neatly into cooling makeup (www.sbqsteels.com). Plants commonly deploy brackish-water RO trains for this duty after stable pretreatment.

Boiler feedwater demands high purity—near‑zero hardness and silica, conductivity <1–2 µS/cm—to avoid scale/corrosion. This is produced by demineralization or RO plus continuous electrodeionization (EDI, a resin‑membrane polishing step). Recycled water can supply cleaner boiler stages after polishing, with fresh makeup covering the rest. The push to cut boiler freshwater and blowdown underpins Tata’s <1.5 m³/t goal (www.tatasteel.com). Plants typically pair RO permeate with EDI polishing or a demineralizer for final purity.

Process/descaling water for pickling and annealing must be clean to avoid product defects; pickling lines prefer low hardness and alkalinity so acids (HCl or H₂SO₄) don’t form insoluble salts, with typical tolerance around <50 mg/L as CaCO₃. RO‑polished water, sometimes blended with fresh, fits these lines. Non‑critical uses—washdown, plant cleaning, dust suppression—can run on lower‑grade recycled water if oil <10 mg/L and solids <10 mg/L to avoid fouling; greywater (e.g., cooling blowdown) often feeds washing/firefighting.

Graded reuse works. One steelworks executed a closed‑loop design where >98% of process/cooling water is recycled, with RO‑permeate backfilling makeup and RO‑brine sent to evaporation (www.sbqsteels.com).

Wastewater streams and contaminant profiles

Oily wastewater from workshops and lubrication steps carries free and emulsified oil and grease. Bulk oil removal via oil‑water separation or dissolved air flotation protects membranes; residual oil can still foul, so polishing helps. Primary removal often pairs a separator with a DAF unit or dedicated oil removal skids.

Metal‑bearing effluent from pickling and finishing (Fe, Cu, Zn, Cr, Ni, etc.) responds to alkaline pH adjustment—metals precipitate as hydroxides or sulfides. Raising pH above 9.5 precipitates most heavy metals, enabling physical removal (smartwatermagazine.com). Industry practice often targets <1 mg/L of key metals for reuse streams; coagulants and polymers aid floc growth when dosed via a dosing pump and clarified in a clarifier.

Cyanide and organics are coke‑oven signatures: cyanide (free and complexed), phenols, thiocyanate, ammonia, and BOD (biochemical oxygen demand) can be significant. Cyanide can range from a few to tens of mg/L in raw influent. Biological treatment—specialized microbes degrading cyanide/thiocyanate and phenol—often follows pH adjustments, with oxidants such as chloride or H₂O₂ to convert cyanide (smartwatermagazine.com).

General industrial waste includes particulates, suspended solids, and scale fragments with elevated COD (chemical oxygen demand) in spots. Screens and primary sedimentation lead, often through physical separation lines. Stream segregation improves outcomes: stormwater, cooling‑tower blowdown, and boiler blowdown can route to different pretreatment trains than plating rinses. One case study described splitting blowdown for UF/RO recovery with only a small brine to disposal (www.mdpi.com), a strategy that relies on stable UF pretreatment to protect downstream membranes.

Reuse targets and “fit for purpose” specs

For non‑critical reuse (e.g., cooling makeup), typical targets include <50 mg/L oil & grease and <10 mg/L settleable solids, with TDS matched to cycles. After RO, the final effluent should land at TDS <500 ppm, turbidity <1 NTU, very low organics (<10 mg/L COD), near‑zero metals, and low hardness—category B/C, non‑potable reuse quality. RO‑permeate at this level can be safely allocated to cooling, washwater, and many process rinses (www.sbqsteels.com). Indonesia’s discharge regime (Government Reg. 22/2021) and reuse‑friendly standards commonly reference non‑potable Class B/C criteria (e.g., BOD <30 mg/L and no toxic metals).

Regulatory drivers and finance context

Asia‑Pacific is leading ZLD (zero liquid discharge) adoption, driven by industrial growth and water scarcity; the ZLD equipment market is projected to reach ≈$2.06 billion by 2033 (watermanaustralia.com). China has moved to mandate ZLD in coal‑to‑chemicals, pharmaceuticals, textiles, and other heavy users (watermanaustralia.com). Case experience is catching up: a U.S. steel mill integrated advanced membranes and hit 98% recycling, with compliance and cost upsides (www.sbqsteels.com).

For Indonesia specifically, adhering to pollutant limits under Government Reg No.22/2021 while aiming beyond discharge toward reuse reduces regulatory risk; noted enforcement gaps (www.mdpi.com) underscore the case for voluntary ZLD‑aligned programs.

Treatment design: staged and membrane‑centered

Pre‑treatment focuses on oil and solids. Primary step: oil‑water separation and flotation to pull free oil to <10 mg/L and capture large solids, followed by sedimentation in compact settlers. Plants standardize with automated intake screens (e.g., an automatic screen) and lamella clarifiers; a lamella settler minimizes footprint in tight plots.

pH adjustment and precipitation follow. Raising pH precipitates metals; pH >12 with NaOH/H₂O₂ (or adding Fe(II)/chlorine) oxidizes cyanide to cyanate and drops metals as hydroxides (smartwatermagazine.com). Coagulants (e.g., ferrous sulfate) and polymers build robust flocs for removal in a clarifier, with >90% of dissolved metals and organics commonly achieved in this stage. Plants often meter reagents with coagulants and trim performance with flocculants.

Biological oxidation comes next when coke‑oven or high‑organic streams are present. Aerobic reactors—activated sludge or sequencing‑batch systems—knock back BOD/COD and nitrify ammonia; cyanide/thiocyanate and phenols are biodegraded after pre‑oxidation and pH setpoints (smartwatermagazine.com). Plants commonly specify activated sludge trains or a flexible SBR unit, followed by secondary clarification.

Tertiary polishing uses membranes. UF eliminates turbidity, colloids, bacteria, and residual emulsions, defending RO from fouling; NF is applied selectively for partial softening. In practice, UF pretreatment and NF softening stabilize feed to the RO array.

RO is the workhorse. Multistage arrays strip >97% of ionic load; from a few‑thousand‑mg/L TDS feed, RO can reduce roughly three‑quarters of the dissolved load, producing permeate <500 mg/L (often <300) suitable for cooling, descaling, and washwater, with permeate conductivity around 100–300 µS/cm (www.sbqsteels.com). Plants standardize on high‑recovery RO skids and proven elements such as FilmTec RO membranes.

Brine concentration and ZLD finishing treat the RO concentrate. Mechanical vapor recompression (MVR) evaporators or thermal crystallizers boil off water for recovery and recycle, cutting brine volume by ~75–80% (watermanaustralia.com). The solid salts/sludge are disposed or potentially valorized. Hybrid designs—high‑recovery RO plus MVR—balance capex and energy.

A typical train: oil/grease trap → pH neutralization (FeSO₄ or NaOH) → primary clarifier → biological reactor → secondary clarifier → UF → 2‑pass RO → MVR/crystallizer. The configuration suits hot climates with higher natural evaporation and should be verified with a pilot or simulation (as used in Italy; www.mdpi.com). Where needed, supporting ancillaries integrate monitoring and control.

Implementation roadmap and KPIs

Months 0–6: audit and quick wins. Conduct a flow/quality audit, fix leaks, optimize cooling cycles, and capture “easy reuse” loops (e.g., reroute clarified water to dust suppression). Launch a 4R (Reduce, Reuse, Recycle, Recover) effort as advocated by Tata Steel (www.tatasteel.com).

Months 6–18: pretreatment upgrades. Install/upgrade oil‑water separators and chemical precipitation basins; stabilize pH across streams; start low‑grade reuse for washdown/toilet flushing. Trial nutrients or simple coagulants to improve removal. Expected outcome: 10–20% immediate cut in freshwater intake with a marked drop in COD/solids; tools range from coagulants to biological nutrients.

Year 1–2: pilot membranes. A pilot UF+RO skid validates flux, recovery, fouling, and CIP cycles on real effluent. UF typically recovers 95%+ into RO feed; RO yields ~70–80% permeate. Aim for permeate TDS <500 mg/L to fit cooling; demonstrate >50% of a chosen stream (e.g., cooling blowdown) can be reused (www.sbqsteels.com). Modular membrane systems ease piloting and scale‑out.

Year 2–3: full installation. Build continuous UF/RO trains sized to full flow and integrate a brine evaporator for ZLD ambitions. Tie RO‑permeate to cooling/process makeup; route brine to evaporation. Include real‑time SCADA for quality/flux monitoring. By ramp‑up end, aim for no fresh intake to cooling or medium‑quality lines; target ~80% reduction in freshwater withdrawal at this stage, consistent with designs that cut waste volumes by ~80% (watermanaustralia.com).

Year 3–5: optimization and ZLD finishing. Increment RO recovery, tune chemical regimes, and upgrade the evaporator (multi‑effect or hybrid compression) to improve energy. Harvest salts where viable. The end‑state is near‑zero external discharge; liquid effluent volumes can fall by ~80% in well‑designed ZLD (watermanaustralia.com). Track KPIs throughout: m³ of fresh water per tonne, % water reuse, and pollutant loads. One hot‑strip mill simulation of UF/RO reuse reported negligible change in intermediate water quality—a sign of a balanced system (www.mdpi.com).

Expected outcomes and economics

Freshwater use can fall from ~28 m³/t to 10–5 m³/t or less, depending on recycled fraction. Tata’s 1.5 m³/t target illustrates what aggressive programs can reach (www.tatasteel.com). An 80% cut equates to ~5.5 m³/t—material by any standard.

Reuse rates >90% are realistic, with projects citing 98% under advanced membrane regimes (www.sbqsteels.com). Final recycled water can meet stringent reuse criteria: BOD/COD <30–50 mg/L, oil <1 mg/L, heavy metals nearly nondetectable, and cyanide eliminated—adequate for cooling, irrigation, and non‑potable industrial uses.

Liquid discharge volumes plummet under ZLD; designs commonly cut wastewater by ~80% (watermanaustralia.com). While RO/thermal steps raise energy use slightly, reductions in raw‑water pumping and opportunities for heat recovery can offset some of the footprint. Utility bills fall with lower purchase and discharge fees; ZLD serves as regulatory hedging. Industry estimates put ZLD payback in the 3–5 year range in water‑scarce, high‑tariff settings, and high reuse rates have correlated with better environmental ratings and investor interest (watermanaustralia.com) (www.sbqsteels.com).

The approach aligns with broader trends: Asia‑Pacific’s ZLD market is growing fastest (watermanaustralia.com) and policy moves such as China’s five‑year plans encourage near‑zero liquid effluent strategies (watermanaustralia.com). Plants adopting this roadmap position alongside leading sustainable producers.

Source references

• Typical water use in steel production (www.mdpi.com); process recycling case studies (www.mdpi.com) (www.sbqsteels.com).
• Technology reviews: trending ZLD market (watermanaustralia.com) (watermanaustralia.com), membrane treatment in steel (www.sbqsteels.com).
• Industry goals and regulations (www.tatasteel.com) (www.mdpi.com).
• Environmental and design benchmarks (watermanaustralia.com) (smartwatermagazine.com).