Why Busan, Why Now: The Korean Water-Energy Squeeze on Hyperscale Builds
Seoul-region data center power demand is forecast to exceed 1,450 MW by 2028, with only about 1.9% of grid-connection applications approved as of 2025 — electricity is now the binding constraint on Korean hyperscale growth (Mint, 2025). Water is the second. The Ulsan underwater pilot, which targets roughly 100,000 servers at ~20 m depth using ~13.3°C seawater and claims a ~12% carbon reduction, signals that Korean planners are already willing to redesign the cooling envelope itself rather than keep paying the freshwater tariff. That same pressure shows up on shore: a 100 MW facility on conventional evaporative cooling draws on the order of 2 million L/day (IDE Tech, 2026), 75–90% of the world's data centers rely on water-based cooling, and 70–80% of that withdrawal is lost to evaporation while 20–30% leaves as blowdown (KETOS, 2026). Busan's particular reality is a coastal site with one of the country's tightest water balances, drawing source water from the Nakdong River basin during dry-season low flow, discharging to the Busan Metropolitan City sewer or the coast, and sitting next to the reclaimed-water network around Busan Eco Delta City. Any 2026 Busan build therefore has to be designed around cycles of concentration higher than the 3.5–4 baseline, treated blowdown that can be reused as make-up, and a discharge envelope that survives a sulfate <500 mg/L rule. For a comparable European reference architecture under the same Busan-style freshwater squeeze, the data center wastewater and cooling blowdown treatment train used in Warsaw, Poland (2026 guide) shows the same NF + precipitation logic applied to a softer feed.
What a Busan Data Center's Wastewater Streams Actually Look Like
Cooling tower blowdown (CTBD) is the dominant liquid stream at 20–30% of cooling water withdrawal, typically operating between 3.5 and 6 cycles of concentration. At a 3.5-cycle design basis, representative CTBD characteristics are TSS 42–55 mg/L, iron 0.5–1.3 mg/L, calcium ~80–110 mg/L, magnesium ~30–45 mg/L, and sulfate that concentrates proportionally from the Nakdong baseline (Brown & Caldwell, 2025 IWC). As cycles push toward 6, the same ions land in the 1,400 mg/L Ca, 370 mg/L Mg, 9,500 mg/L sulfate range inside an NF reject — a chemistry that conventional BWRO cannot reach. The secondary streams matter because they still need a permitted outlet: chiller and adiabatic cooler blowdown (low volume, high TDS), humidification drains, once-through system flushes during commissioning, DI/RO concentrate from polishing loops, and the small domestic sanitary load from on-site staff. Fire-water test drains, generator coolant flushes, and DI/RO concentrate from humidification or polishing skids arrive episodically but carry metals or pH excursions that must be screened. If the Busan site co-locates with a seawater auxiliary loop in the Ulsan-pilot mould, screen wash and intake backwash are added — high-volume, high-TDS streams that should stay on a separate coastal permit rather than enter the freshwater CTBD chemistry. The design implication is straightforward: every stream on the inventory needs a tank, a pipe, and a documented discharge path before the P&ID is frozen.
Korean Regulatory Floor: Water Environment Conservation Act, KECO Limits, Nakdong TMDL
The Water Environment Conservation Act, administered by the Ministry of Environment and enforced locally by Busan Metropolitan City, sets the discharge envelope; site-specific effluent limits are negotiated at the permit stage, and the indirect-discharge path to the Busan municipal sewer is governed by a Busan Environmental Corporation pretreatment contract. For 'B' zone industrial discharge to public waters, typical targets are BOD ≤120 mg/L, SS ≤120 mg/L, total N ≤60 mg/L, total P ≤8 mg/L, pH 5.8–8.6, and fecal coliform under the relevant basin rule. The design driver that separates Korea from many other jurisdictions is the impaired-waters / TMDL overlay: in the Nakdong basin during dry-season low flow, sulfate is held to <500 mg/L in the effluent, which is what pushes operators off straight-to-sewer BWRO and onto NF + precipitation. MSIT/MOTIE hyperscale siting guidance updated through 2024–2025 increasingly requires a water-stewardship and reuse plan as part of approval, so reuse is now a permitting asset rather than a CSR footnote. The table below summarises the operational limits a Busan process design has to clear.
| Parameter | Typical KECO 'B' zone limit | Nakdong basin / dry-season driver | Design implication |
|---|---|---|---|
| BOD | ≤120 mg/L | — | Biological stage unnecessary if CTBD-only; required if domestic waste co-mingled |
| SS | ≤120 mg/L | — | Dual-media filtration easily clears; pre-NF target is <5 mg/L |
| Total N | ≤60 mg/L | — | CTBD negligible; relevant only to sanitary stream |
| Total P | ≤8 mg/L | — | Anti-scalant selection must avoid phosphate-based chemistries |
| pH | 5.8–8.6 | — | CO2 trim on blended NF permeate is the standard correction |
| Sulfate | — | <500 mg/L (TMDL, dry-season low flow) | Forces NF + gypsum precipitation; BWRO alone cannot meet at 6 cycles |
| Temperature | Site-specific sewer contract | — | Equalisation tank attenuates chiller-trip excursions before discharge |
| Heavy metals | Restricted under pretreatment contract | — | Exclude generator coolant and fire-test drains from auto-routing |
CTBD Chemistry at 3.5 vs 6 Cycles: Why Conventional RO Hits a Wall
At 3.5 cycles of concentration, a representative Nakdong-derived CTBD carries 42–55 mg/L TSS, 0.5–1.3 mg/L iron, ~80–110 mg/L calcium, ~30–45 mg/L magnesium, and sulfate already in the low hundreds of mg/L, with silica and calcium approaching their conventional scaling thresholds (Brown & Caldwell, 2025 IWC). At 6 cycles, the same ions land at roughly 1,400 mg/L Ca, 370 mg/L Mg, and 9,500 mg/L sulfate in the NF reject — well past the 75–80% recovery ceiling that conventional BWRO can hold reliably (IDE Tech, 2026; Brown & Caldwell, 2025 IWC). The reason is kinetics: sparingly soluble salts such as silica, CaCO₃, and CaSO₄ hit their induction-time limits faster than NaCl, so pushing recovery past ~80% with conventional RO triggers membrane scaling, more frequent CIP, and unstable operation. Korean source water is generally softer than seawater, but the Nakdong dry-season baseline already carries 30–80 mg/L sulfate, which magnifies the concentration effect once cycles climb. The design decision becomes binary: stay at 3.5 cycles with BWRO and accept a high freshwater draw, or push to 6 cycles with NF + calcium-sulfate precipitation and unlock a 40–60% freshwater offset. The table shows the chemistry shift between the two operating points.
| Parameter | CTBD at 3.5 cycles (NF feed) | NF reject at 6 cycles (precipitation feed) | Design consequence |
|---|---|---|---|
| Ca | ~80–110 mg/L | ~1,400 mg/L | Lime dose scales with reject Ca; Ksp of CaSO₄ is the limit |
| Mg | ~30–45 mg/L | ~370 mg/L | Drives clarifier sizing; affects sludge volume index |
| Sulfate | ~150–280 mg/L | ~9,500 mg/L (7,600 mg/L soluble) | Sets lime stoichiometry at 0.8–1.6 mg Ca(OH)₂/mg soluble SO₄ |
| Silica | ~10–20 mg/L | Concentrates in reject | NF membrane choice (e.g. Veolia DK-400) limits silica rejection |
| TSS | 42–55 mg/L | — | Dual-media + greensand to <5 mg/L pre-NF |
| Iron | 0.5–1.3 mg/L | — | Greensand polishes to ~0.1 mg/L pre-NF |
| Recovery ceiling | BWRO 75–80% | NF 90% + precipitation loop | Recovery jump is the freshwater-offset lever |
Recommended Process Train for a Busan Data Center
The defensible block flow for a 2026 Busan campus is a front-end equalisation, a media + cartridge pretreatment train sized for the NF, a three-stage NF at ~90% recovery, a calcium-sulfate precipitation loop on the ~10% reject, CO₂ trim, and ClO₂/UV disinfection before reuse or compliant discharge (Brown & Caldwell, 2025 IWC; IDE Tech, 2026). Equalisation splits into an unmixed working tank at 3 hours HRT and a jet-mixed off-spec tank at 22 hours HRT to absorb swings from cooling-tower cycling and chiller trips. Pretreatment uses acid trim for pH control, a dual-media and greensand filter for CTBD pretreatment to drop TSS to <5 mg/L and iron to ~0.1 mg/L, and a 5 µm cartridge filter sized at 3–5 gpm per 10-inch element for monthly or less-frequent changeout. The NF skid runs three stages at ~10 gfd flux, 90% recovery, with concentrate recycle and permeate backpressure for turndown; Brown & Caldwell's 2025 IWC screening identified Hydranautics PRO-XS2, Veolia DK-400, and Veolia DSL NF8040 as the leading candidates for high sulfate rejection with controlled silica passage. The reject stream goes to a lime reactor at 0.8–1.6 mg Ca(OH)₂ per mg soluble sulfate (chemistry-program dependent), a seeded clarifier with high sidewater depth, a filter press for gypsum sludge dewatering, and CO₂ trim on the blended NF permeate to bring pH into the 5.8–8.6 envelope. PLC-controlled chemical dosing for acid, anti-scalant, and lime keeps stoichiometry inside the treatability window. Disinfection uses an on-site ClO2 generator for cooling-loop residual control with a UV sterilizer as a chemical-free polishing step ahead of reuse or outfall. Where the make-up target is adiabatic cooling or humidification — both more sensitive to TDS — a polishing industrial RO system for CTBD recovery or post-NF polishing is added after the NF permeate. The DOE FEMP guidance on on-site wastewater treatment as a federal water-saving opportunity documents the same NF + precipitation + reuse logic for U.S. federal sites under comparable freshwater pressure.
Process Comparison: BWRO vs NF + Precipitation vs Closed-Loop Hybrid
The technology choice for Busan is not between good and bad — it is between three recovery envelopes and three permitting positions. BWRO is the simplest, lowest-CAPEX baseline at 75–80% recovery, fits a small colocation suite with low make-up demand, and is the right answer if freshwater supply is unconstrained and the discharge point is the municipal sewer with no TMDL pressure. NF + calcium-sulfate precipitation is the default recommendation: ~90% overall recovery, sulfate compliance against the <500 mg/L Nakdong driver, modest CAPEX premium, and proven by Brown & Caldwell's 2025 IWC treatability work on data-center CTBD. Closed-loop hybrid pushes 90–95% recovery, eliminates surface discharge entirely, and is the strongest water-stewardship story to bring to MSIT/MOTIE and Busan Metropolitan City, but adds a meaningful CAPEX premium and the largest sludge-handling duty. The IDE Tech MAXH₂O Desalter is a vendor example of a high-recovery CTBD architecture at ~95% recovery with ~1 mg/L silica permeate, cited here as a reference design, not a recommendation. The comparison below is the table a Busan buyer should be able to hand to procurement.
| Criterion | BWRO | NF + gypsum precipitation | Closed-loop hybrid |
|---|---|---|---|
| Overall recovery | 75–80% | ~90% | 90–95% |
| Cycles supported | 3.5–4 | 5–6 | 6+ |
| Freshwater offset | Baseline | 40–60% vs BWRO | Highest; near-zero draw option |
| Effluent sulfate | Risk >500 mg/L at high cycles | <500 mg/L compliant | No surface discharge |
| CAPEX direction | Lowest | +15–25% | +40–60% |
| OPEX driver | Membrane replacement, energy | Lime, sludge hauling, anti-scalant | Sludge hauling, energy, chemistry |
| Sludge handling | Minimal | Gypsum cake, non-hazardous | Gypsum + blowdown solids, highest duty |
| Permitting posture | Standard indirect discharge | TMDL-defensible, reuse-eligible | Zero-discharge, strongest stewardship |
| Best fit | Small colocation, freshwater-rich | Busan hyperscale default | Water-stressed flagship sites |
Sanitary and Auxiliary Streams: Keeping Them Off the Critical Path
On-site domestic wastewater at a Busan campus is a few hundred people at most and is best handled by a buried package plant sized to local site constraints rather than a civil MBR. Fire-system test drains and generator coolant flushes are episodic but carry pH excursions and trace metals; route them to the same equalisation tank with a manual pH screen and bypass to off-spec storage when out of band. Seawater-side streams from a Busan co-located auxiliary loop, where they exist, should bypass the NF train entirely and discharge under a separate coastal permit; do not let high-TDS seawater backwash contaminate the freshwater CTBD chemistry, or the precipitation stoichiometry will be wrong. A buried package STP for on-site domestic wastewater is the conventional Korean site solution because it frees up building footprint and sits below the visual envelope of a campus that has neighbours. The goal of this section is simple: keep the small streams from becoming the streams that delay commissioning.
Economics and Risk: What the Busan Buyer Should Pressure-Test
Directional CAPEX: NF + precipitation adds ~15–25% over a baseline BWRO skid of equal permeate capacity; closed-loop hybrid can add ~40–60% but unlocks a zero-discharge permitting path that is increasingly valuable to MSIT/MOTIE. The dominant OPEX lever is the freshwater offset of 40–60%, which compounds against Busan municipal potable tariffs and against any future Nakdong basin allocation cut. Sludge handling is the second lever: gypsum cake from the filter press is non-hazardous in Korea but requires a licensed waste hauler, and Busan Metropolitan City acceptance should be confirmed before tender, not after commissioning. The single biggest process risk is the interaction between the cooling-tower chemistry program and the downstream precipitation chemistry — Brown & Caldwell's 2025 IWC treatability data showed that the anti-scalant and corrosion-inhibitor selection can shift sulfate solubility by ~20% and double the lime dose in some cases. Coordinate anti-scalant, corrosion inhibitor, and biocide selection with the NF/precipitation design from day one; retrofit is always more expensive than co-selection.
Frequently Asked Questions
What wastewater and cooling blowdown treatment does a data center in Busan, South Korea need?
A Busan data center needs an equalisation tank, dual-media and greensand filtration to <5 mg/L TSS and ~0.1 mg/L iron, 5 µm cartridge filtration, three-stage nanofiltration at ~90% recovery, calcium-sulfate precipitation on the ~10% NF reject with lime at 0.8–1.6 mg Ca(OH)₂ per mg soluble sulfate, CO₂ trim, and ClO₂ or UV disinfection, designed to keep effluent sulfate <500 mg/L under the Nakdong basin TMDL (per Brown & Caldwell 2025 IWC; Water Environment Conservation Act).
Why push cycles of concentration past 3.5 in Korea?
At 3.5 cycles a conventional BWRO tops out at 75–80% recovery because silica, CaCO₃, and CaSO₄ hit induction-time scaling limits; NF + gypsum precipitation lets a Busan operator run 5–6 cycles at ~90% overall recovery and unlock a 40–60% freshwater offset against the Nakdong basin's dry-season scarcity (per IDE Tech 2026; Brown & Caldwell 2025 IWC).
What are the KECO effluent limits a Busan data center must meet?
Typical KECO 'B' zone discharge limits are BOD ≤120 mg/L, SS ≤120 mg/L, total N ≤60 mg/L, total P ≤8 mg/L, and pH 5.8–8.6, with the Nakdong basin sulfate TMDL holding sulfate to <500 mg/L during dry-season low flow, which is the design driver behind the NF + precipitation train (per Water Environment Conservation Act; Busan Metropolitan City permit).
Can a Busan campus be designed for zero liquid discharge?
Yes. A closed-loop hybrid at 90–95% recovery, with NF permeate as make-up and no surface discharge, eliminates the discharge permit risk entirely and is the strongest water-stewardship story to bring to MSIT and Busan Metropolitan City, at a directional CAPEX premium of ~40–60% over a baseline BWRO skid (per IDE Tech MAXH₂O reference architecture, 2026).
What role does reclaimed water play for a Busan site?
Reclaimed water from Busan Eco Delta City and adjacent STPs is a viable non-potable make-up source that reduces pressure on the Nakdong basin and supports the water-stewardship narrative that MSIT/MOTIE hyperscale siting guidance increasingly requires as part of project approval.
