Why Diyarbakir changes the data-center water equation
Diyarbakir's hot, dry summers push evaporative cooling systems harder than temperate-zone benchmarks, raising both makeup demand and blowdown per MWh of IT load — the same demand pattern IDE Tech describes for a 100 MW facility that can use up to 2 million liters of water per day (IDE Tech, 2026). Surface and groundwater in the upper-Mesopotamian basin are already allocated across the GAP region's agriculture, the Diyarbakir municipal network, and industry, so a freshwater-only cooling strategy carries the same social and permit exposure that the IDE Tech analysis attaches to water-stressed US basins (IDE Tech, 2026).
The engineering response is to treat water as a managed resource inside the fence: size a high-recovery reuse train before committing to a freshwater supply, the way the IDE Tech commentary argues for cooling-tower blowdown as a strategic resource rather than a waste stream (IDE Tech, 2026). The design envelope that follows is therefore constrained by four questions: which streams must be treated (cooling-tower blowdown, RO reject, sanitary wastewater, possible on-site process wastewater), what end uses are acceptable (cooling makeup, irrigation, zero discharge), what the seasonal source-water envelope looks like, and which Turkish discharge rule applies to the residual.
What the wastewater and blowdown streams actually look like
Cooling-tower blowdown is the largest and most concentrated wastewater stream on a Diyarbakir site. At roughly 4 cycles of concentration it represents 25–30% of makeup water (Genesis Water Technologies, 2026); pushing cycles higher shrinks the volume but raises the dissolved-solids load. Genesis Water Technologies documents the typical blowdown envelope as TDS of 1,200–6,000 mg/L (4–8× makeup), concentrated calcium, magnesium, silica and alkalinity, suspended solids of 10–50 mg/L from corrosion products, biofilm fragments and airborne dust, and accumulated treatment chemistry (biocides, scale and corrosion inhibitors, dispersants) (Genesis Water Technologies, 2026). Sanitary and facility wastewater is a smaller, lower-TDS stream that still requires biological treatment before any reuse or sewer discharge. RO reject and membrane concentrate form a third stream that must be planned for even on a reuse-oriented site, because Genesis Water Technologies shows RO recovery on blowdown realistically sits at 50–85% (Genesis Water Technologies, 2026). Diyarbakir's municipal supply is not seasonally flat — summer evaporation concentrates hardness and silica in the Dicle basin — so a one-season water analysis is not enough to size the membrane system; request a quarterly dataset before finalising any recovery or ZLD assumption.
| Stream | Typical flow share | TDS (mg/L) | Key species | SS (mg/L) | Notes |
|---|---|---|---|---|---|
| Cooling-tower blowdown (4 CoC) | 25–30% of makeup | 1,200–6,000 | Ca, Mg, silica, alkalinity, treatment chemicals | 10–50 | Volume shrinks, TDS rises as cycles rise (S2) |
| Sanitary / facility wastewater | Site headcount dependent | Low relative to blowdown | BOD, COD, pathogens | Variable | Needs biological step before reuse or discharge (S2) |
| RO reject / concentrate | 15–50% of RO feed (50–85% recovery) | Higher than feed | Concentrated scaling ions | Low if UF precedes | Must have a downstream fate (S2) |
Treatment objectives: reuse, compliant discharge, or zero liquid discharge
Cooling-tower makeup reuse is the highest-value, lowest-risk objective. Genesis Water Technologies reports 60–85% recovery for this end use on blowdown, and the same publisher's alternative-water-source guide confirms 70–90% on-site blowdown recovery is realistic when the train is properly designed (Genesis Water Technologies, 2026). Compliant discharge is the fallback when reuse is blocked; Genesis Water Technologies notes discharge fees in stressed regions now run $5–15 per 1,000 gallons and that some jurisdictions enforce TDS caps below 1,500 mg/L, which together often make discharge more expensive than reuse even before any scarcity premium (Genesis Water Technologies, 2026). Zero liquid discharge — 95–99% overall recovery per Genesis Water Technologies — is technically proven but capital-intensive at $3–8 million for a typical data-center ZLD, and should be reserved for sites facing a hard discharge ban or fully internalised water-cost stress (Genesis Water Technologies, 2026). On-site trains can be staged: reuse first, then concentrate polishing through MVC only if discharge remains constrained, since Genesis Water Technologies describes RO + MVC as a typical 85–95% combined-recovery configuration (Genesis Water Technologies, 2026).
The technology menu: filtration, membranes, evaporation, ZLD
Side-stream filtration is the cheapest insurance in the train. Genesis Water Technologies cites $50,000–$200,000 in CAPEX for typical data-center installations of self-cleaning spiral or multi-media filters at 10–25 micron, and ties this directly to higher cycles and cleaner downstream blowdown (Genesis Water Technologies, 2026). Ultrafiltration protects the RO membranes: 0.01–0.1 micron pore size, 90–95% recovery, operating at only 10–30 psi, removing suspended solids, bacteria and biofilm fragments that would otherwise shorten RO cleaning intervals (Genesis Water Technologies, 2026). Reverse osmosis is the workhorse: 95–99% salt rejection, permeate TDS of 10–50 mg/L, operating pressure 150–400 psi, recovery 50–85% on blowdown, with CAPEX of $250,000–$500,000 for a 50,000 GPD unit and OPEX of $1.50–$3.00 per 1,000 gallons (Genesis Water Technologies, 2026). Nanofiltration sits below RO — 75–150 psi, 70–85% recovery, permeate TDS around 30–50% of feed — and is the right choice when hardness and multivalent ions, not total TDS, are the limit on cycles (Genesis Water Technologies, 2026). Mechanical vapor compression is the bridge to ZLD: distillate TDS below 10 mg/L, 95–98% recovery on the concentrate, CAPEX of $1–3 million for a 10,000–30,000 GPD unit, and 15–25 kWh per 1,000 gallons of distillate (Genesis Water Technologies, 2026). A crystallizer plus dryer completes a true ZLD train — 95–99% overall recovery, residual solids below 1% of the original blowdown volume — at $5–$15 per 1,000 gallons OPEX and the highest CAPEX in the menu (Genesis Water Technologies, 2026). The final lever is chemistry and dynamic RO operation, the approach behind the IDE MAXH₂O architecture that combines controlled precipitation of sparingly soluble salts with dynamic RO cycling to push overall recovery past the conventional 75–80% BWRO ceiling toward ~95% in real installations (IDE Tech, 2026).
| Unit operation | Role | Performance band | CAPEX (typical) | OPEX (typical) |
|---|---|---|---|---|
| Side-stream filtration (10–25 μm) | SS, biofilm, corrosion-product removal | Enables higher CoC | $50,000–$200,000 (S2) | Minimal (S2) |
| Ultrafiltration (0.01–0.1 μm) | RO pretreatment, biological barrier | 90–95% recovery, 10–30 psi (S2) | Project dependent | Permeate backwash, quarterly CIP (S2) |
| Reverse osmosis | Dissolved solids, hardness, silica removal | 50–85% recovery on blowdown, 10–50 mg/L permeate TDS, 150–400 psi (S2) | $250,000–$500,000 for 50,000 GPD (S2) | $1.50–$3.00 / 1,000 gal (S2) |
| Nanofiltration | Partial softening, lower energy | 70–85% recovery, 75–150 psi, 30–50% feed-pass TDS (S2) | Project dependent | Lower than RO (S2) |
| MVC brine concentrator | Concentrate minimisation, ZLD bridge | 95–98% recovery, distillate <10 mg/L TDS (S2) | $1–3M for 10,000–30,000 GPD (S2) | 15–25 kWh / 1,000 gal distillate (S2) |
| Crystallizer + dryer | True ZLD, solid waste | 95–99% overall, solids <1% of blowdown (S2) | Lift on top of MVC (S2) | $5–$15 / 1,000 gal (S2) |
Putting it together: a realistic treatment train for a Diyarbakir site
The baseline train for a Diyarbakir site is side-stream filtration followed by UF pretreatment and an industrial RO system, with the RO permeate blended into cooling-tower makeup and the concentrate routed either to further treatment or to compliant sewer discharge under the applicable Turkish permit. Where Diyarbakir's hardness profile is the constraint on cycles, add nanofiltration upstream of the RO so the unit operation fights the limiting species directly. For a high-recovery train, integrate controlled precipitation of silica and calcium carbonate with dynamic RO operation, targeting the ~95% overall recovery the IDE MAXH₂O case demonstrates with permeate silica around 1 mg/L (IDE Tech, 2026). A ZLD train adds MVC and a crystallizer to the RO and is justified only when the site cannot discharge at all or when the avoided water cost fully internalises the $5–15 per 1,000 gallon discharge alternative (Genesis Water Technologies, 2026). On-site sanitary and facility wastewater is best handled separately in a packaged biological unit sized for actual headcount, with membrane biological reactor or packaged sewage treatment units typical for this duty, and antiscalant and biocide dosing aligned with the downstream membrane chemistry. Plan spare RO and UF elements into the procurement schedule from day one; a Diyarbakir site with seasonal silica swings will see shorter cleaning intervals than the vendor's steady-state case.
| Train | Unit ops | Expected overall recovery | CAPEX band | When it fits Diyarbakir |
|---|---|---|---|---|
| Baseline reuse | Side-stream filter → UF → RO | 50–85% on blowdown (S2) | $250,000–$500,000 for 50,000 GPD RO (S2) | Default for sites with a permitted discharge path |
| High-recovery reuse | Side-stream filter → UF → RO + controlled precipitation + dynamic RO | ~90–95% (S1) | ~$1–3M incremental over baseline | Hard freshwater allocation, ESG reporting, no discharge path |
| Full ZLD | RO + MVC + crystallizer | 95–99% (S2) | $3–8M (S2) | Hard discharge ban, fully internalised water cost |
Decision framework: choosing the right train for your site
Step 1 — fix the design water balance. Define cycles of concentration, freshwater source, target makeup TDS, and Diyarbakir's summer evaporation factor before any equipment is sized. Step 2 — fix the discharge constraint. Identify the sewer or receiving-water TDS, hardness and biocide limits, and whether any TDS <1,500 mg/L cap from the comparative data applies locally (Genesis Water Technologies, 2026). Step 3 — score the trains against your balance and constraint: baseline reuse (lowest CAPEX, 50–85% RO recovery on blowdown), high-recovery reuse (~$1–3M incremental, ~90–95% recovery, IDE Tech case), and ZLD ($3–8M CAPEX, 95–99% recovery, $5–$15/1,000 gal OPEX per Genesis Water Technologies, 2026). Step 4 — stress-test the decision against water-cost escalation and a hypothetical discharge ban, because hyperscale operators are already designing for those scenarios (IDE Tech, 2026) and Genesis Water Technologies documents the same on-site reuse logic in the Texas hyperscale case (Genesis Water Technologies, 2026). Step 5 — only then engage suppliers with a sized basis: feed flow, feed TDS, target cycles, permeate volume, and concentrate fate. Comparative case studies in water-stressed cities — data center blowdown treatment in Sanaa, data center blowdown treatment in Khartoum, and data center blowdown treatment in Curitiba — are useful references for how the same train behaves under different climate envelopes. For the sanitary side, a comparative read of COD/BOD removal technologies helps set the biological-unit selection apart from the membrane train.
| Step | Question to lock | Data input required |
|---|---|---|
| 1 | Design water balance | Cycles, source, makeup TDS, Diyarbakir summer evaporation |
| 2 | Discharge constraint | Sewer/receiving water TDS, hardness, biocide limits (S2) |
| 3 | Score trains | CAPEX/OPEX bands from S2, recovery ceiling from S1 |
| 4 | Stress test | Water-cost escalation curve, hypothetical discharge ban (S1, S2) |
| 5 | Supplier engagement | Feed flow, feed TDS, cycles, permeate volume, concentrate fate |
Frequently Asked Questions
What CAPEX and OPEX should a Diyarbakir data center plan for cooling-blowdown reuse versus ZLD?
For RO-based blowdown reuse, plan $250,000–$500,000 CAPEX for a 50,000 GPD unit and $1.50–$3.00 per 1,000 gallons OPEX; for full ZLD, plan $3–8 million CAPEX and $5–$15 per 1,000 gallons OPEX (Genesis Water Technologies, 2026). The deciding check is whether your avoided freshwater plus avoided discharge fees exceed the incremental OPEX of the higher-recovery option over a 5–7 year horizon.
How do I choose between a local EPC, an OEM, and a water-treatment specialist for this train?
Score each supplier against three verifiable inputs: a referenced high-recovery cooling-tower case at ≥90% recovery, named performance data for the membrane and MVC units, and willingness to size from a quarterly Diyarbakir water analysis rather than a single sample. The IDE Tech MAXH₂O case study (95% recovery, ~1 mg/L permeate silica) and the Genesis Water Technologies on-site reuse references are the kind of evidence to ask for (IDE Tech, 2026; Genesis Water Technologies, 2026).
What recovery can a conventional BWRO actually hit on cooling-tower blowdown?
Conventional brackish water RO plateaus at 75–80% recovery on cooling-tower blowdown before scaling becomes unmanageable; high-recovery designs that combine controlled precipitation and dynamic RO operation reach ~95% in operating installations (IDE Tech, 2026). Recovery on blowdown feed specifically is realistically 50–85% on a single RO pass (Genesis Water Technologies, 2026).
Why does a one-season water analysis fail for Diyarbakir membrane design?
Diyarbakir's summer evaporation concentrates hardness and silica in the source water, so a single wet-season sample understates the scaling load the RO will see in August. Request a quarterly dataset covering at least one full summer before finalising membrane selection, antiscalant dosing set-points and the concentrate-disposal basis (Genesis Water Technologies, 2026).
