Separate process wastewater from sanitary sewage
A Pietermaritzburg data center needs a treatment strategy designed separately for industrial cooling-tower blowdown and sanitary sewage. Blowdown is a controlled purge from the cooling circuit: evaporation leaves dissolved salts, treatment chemicals and other additives behind, so periodic purging controls cycles of concentration and limits scaling and corrosion. It is an industrial process stream, not domestic wastewater. The sanitary-sewage design must instead be based on domestic flows, wastewater characteristics and the applicable sewer-acceptance route. The project team should therefore establish separate flows, treatment objectives and approval pathways for cooling blowdown, sanitary sewage, and secondary streams such as water-treatment backwash, equipment-cleaning water and air-handling condensate.
The distinction is not merely semantic. RSP Engineers’ 2025 process-wastewater guidance describes cooling-tower blowdown as the principal process-wastewater stream in many large data centers and notes that its volume is typically greater than sanitary flow. Blowdown can carry concentrated minerals, biocides, corrosion inhibitors, anti-scalants and elevated temperature, whereas sanitary sewage requires a domestic-wastewater design. Equipment selected for one stream should not be assumed suitable for the other.
For Pietermaritzburg, the project team should confirm the sewer authority’s acceptance conditions separately from any environmental or direct-discharge assessment. Sanitary sewer acceptance should be tested against domestic sewage flows and characteristics, not against cooling-tower chemistry. A useful process-separation starting point is provided in the data-center wastewater treatment planning guide, but its treatment logic must still be checked against local source water, chemical package and receiving-route requirements.
Characterize the blowdown before choosing equipment
Equipment selection should begin with a measured cooling-water and blowdown balance, because the same conductivity can behave differently when silica, hardness, antiscalant, biocide and temperature conditions differ. Request the cooling architecture, evaporation and drift rates, target cycles of concentration, calculated daily and peak blowdown, water-treatment chemical package, operating hours, makeup-water source and anticipated change in any of those inputs. These data determine whether the project is primarily controlling a discharge stream, replacing make-up water, or doing both.
The sampling programme should cover pH, conductivity or TDS, silica, calcium, hardness or alkalinity, chlorides, sulfates, phosphates, suspended solids, residual chlorine and the specific biocides in use. Include copper, zinc and other metals where corrosion control is relevant. Sample across representative operating conditions rather than relying only on one clean-loop result. The chemical inventory must identify active ingredients, dosing ranges and whether chemicals will continue to be applied if treated water is returned to the cooling system.
RSP Engineers’ 2025 guidance gives a pH range of 6.0–9.0 as a frequently encountered example. It is not a Pietermaritzburg discharge limit. The controlling requirement must come from the receiving authority or the reuse specification; neither can be inferred from the equipment catalogue.
| Input to obtain | Why it changes the treatment decision |
|---|---|
| Cooling architecture, evaporation and drift | Defines the water balance and expected blowdown flow. |
| Target cycles of concentration and chemical package | Controls scaling, corrosion, biofouling and the composition of the purge. |
| Representative water analysis | Sets the barriers needed before RO, reuse or discharge. |
| Make-up-water source and quality | Determines whether reuse can replace municipal or borehole water and what treatment is needed. |
| Discharge route and authority conditions | Defines the applicable South African approval and acceptance basis. |
The water-demand context is significant. IDE’s 2026 article states that a 100-megawatt data center can require up to 2 million litres of water per day, although that figure is not a Pietermaritzburg facility estimate. HydropureWater’s 2026 guide, citing UGA Extension TP-121 dated May 2026, reports that 75–90% of data centres globally use water-based cooling and that an analysis cited by TP-121 attributes approximately 57% of the water to potable municipal or groundwater sources. These figures explain why even modest reductions in net withdrawal may matter, but they cannot size the local train.
Compare discharge-first and reuse-first treatment concepts
The first project decision is the receiving route: authorised discharge or internal reuse. For discharge-first treatment, map the intended receiving route and obtain its authority conditions before selecting equipment. The assessment must address pH, thermal load, suspended solids, residual biocides, metals and conductivity where applicable. A receiving sewer or surface-water route should be treated as an authority decision, not assumed from the presence of a drainage connection.
For reuse-first treatment, define the required cooling make-up-water quality before selecting membranes. Evaluate filtration or clarification, reverse osmosis, disinfection and polishing against the cooling system’s scaling, corrosion and biological-control requirements. Storage, distribution and makeup substitution also need verification. Permeate that meets a generic reuse specification may still alter corrosion, foaming or biocide performance in the cooling loop.
Higher cycles of concentration reduce the blowdown volume, but the burden shifts to chemical control, pretreatment and membrane protection. IDE’s 2026 article reports that conventional brackish-water RO commonly reaches 75–80% recovery before scaling becomes unmanageable. HydropureWater’s 2026 guide also cites UGA Extension TP-121 from May 2026, which states that closed-loop cooling can use up to 70% less water than open evaporative cooling, at the cost of greater electricity demand and system complexity. These are comparative findings, not a guarantee for the proposed Pietermaritzburg facility.
| Decision point | Discharge-first concept | Reuse-first concept |
|---|---|---|
| Primary objective | Meet confirmed receiving-route and authority conditions. | Reduce net freshwater demand and return treated water to the cooling system. |
| Core design input | Permit or acceptance conditions, discharge flow and thermal and chemical loads. | Make-up-water target, cooling chemistry, recovery demand and storage arrangement. |
| Key risk | Designing equipment before the receiving route is confirmed. | Producing permeate that does not remain compatible with the cooling programme. |
| Typical treatment logic | Equalisation, pH control where required, removal of selected contaminants and authorised discharge. | Filtration, membrane treatment, polishing, disinfection where required and protected storage. |
| Water-balance effect | Maintains the discharge route while protecting infrastructure and receiving conditions. | Reduces freshwater withdrawal when the return flow is accepted by the cooling design. |
U.S. terms such as NPDES, POTW and Clean Water Act requirements in the 2026 HydropureWater guide describe international examples, not South African law. The project team must confirm the applicable South African environmental, water-use and municipal discharge requirements with the competent authorities. Regional examples in the regional data-center blowdown treatment guide and the Ahvaz data-center wastewater case study should be used as decision prompts, not as substitutes for local approval.
Build a treatment train that matches the chemistry
An auditable treatment train follows the source segregation and treatment barriers. A typical sequence is source segregation, equalisation, filtration or clarification, chemical conditioning where required, membrane or polishing treatment, disinfection where required, storage, and final reuse or authorised discharge. Every arrow should correspond to a measured contaminant or operational requirement; otherwise the train contains equipment without a defensible design function.
A multi-media pretreatment filter for RO protection can remove suspended material before the membrane stage. HydropureWater’s supplied product literature states that the multi-media filter can provide SDI suitable for RO protection, but the membrane manufacturer’s feed envelope remains the controlling specification. It should also be evaluated as a side-stream filtration barrier for the cooling loop where that operating objective applies.
Ultrafiltration can provide a stronger barrier against suspended solids, colloids and bacteria before sensitive equipment. HydropureWater’s supplied specification describes a 0.03-micron PVDF membrane, an operating range of 2,000–40,000 L/h and a 300 ppm turbidity acceptance limit; these are catalogue parameters, not universal project values. Feed flow, turbidity, recovery and required log reduction must be checked before assigning UF to the train.
An industrial RO system for cooling-tower reuse can demineralise blowdown and reduce salinity before reuse. HydropureWater offers automatic PLC-controlled systems with recovery rates up to 95%, subject to feed chemistry and scaling limits. IDE’s 2026 article reports conventional brackish-water RO at 75–80% recovery and describes an industrial case operating at around 95% through a dedicated high-recovery architecture. That case does not establish a guaranteed local recovery. Local scale analysis, feed testing and pilot work are required before adopting the higher value.
Disinfection must follow the cooling-water and by-product strategy: the 2026 HydropureWater guide identifies chlorine dioxide and UV as alternatives. If equalisation or clarification produces sludge, a plate-and-frame sludge dewatering system may be required, with jar or pilot testing used to establish solids production. A PLC-controlled chemical dosing skid should be assigned only after pH correction, antiscalant or biocide requirements are confirmed. Further chemistry guidance is available in the cooling-blowdown chemistry and treatment guide.
Plan the footprint, utilities and operating obligations
The installed footprint follows from flow, treatment barriers and storage—not simply from the number of treatment units. Equalisation and storage tanks, pumps, membrane or filter vessels, chemical storage and dosing, electrical supply, controls, drainage, spill containment, vehicle access and maintenance clearance must be shown on the site plan.
RSP Engineers’ 2025 guidance compares the footprint of a simple pH-neutralisation system at a few hundred square feet with filtration and chemical precipitation at several thousand square feet. These figures indicate the scale of the planning difference, not a quotation for the Pietermaritzburg project. Equipment arrangement, gravity drainage, bunding and maintenance access still require design.
Operations extend beyond treatment. The monitoring plan should cover flow, pH and any reuse or authority parameters, with laboratory support, operator training, chemical consumption, membrane cleaning and solid-waste or concentrate handling. Sampling points and automatic monitoring should be included where required.
Each waste stream needs an owner and destination: sanitary sewage, industrial pretreatment sludge and reverse-osmosis concentrate are not interchangeable. The concentrate route must be accepted before an RO system is frozen, because its quantity and chemistry affect the recovery target and operating cost.
HydropureWater’s 2026 guide gives a 20–34-week procurement-to-commissioning range for a standard modular skid. Use it as an early planning benchmark, not a guaranteed project schedule. Custom chemistry, authority approvals, site integration and performance validation can change the delivery programme. The RO energy-consumption guide provides a separate engineering basis for evaluating connected electrical demand.
Convert site inputs into a supplier-ready scope
A comparable bid requires one controlled process basis. Issue the cooling design, daily and peak blowdown, water analyses, chemical package, target recovery, reuse demand, discharge route, utility limits, footprint and operating envelope to every bidder. Require each bidder to state its pretreatment assumptions, membrane type and sizing basis, definition of recovery, reject-water destination, chemical consumption, automation, redundancy, monitoring, commissioning tests and exclusions.
Performance guarantees should be tied to the supplied feed-water envelope rather than catalogue maxima. Require a fouling and scaling mitigation strategy, replaceable-element pricing, membrane-cleaning assumptions, chemical storage requirements and a clear allocation of operating responsibilities. A recovery percentage without a stated feed composition and concentrate route is not comparable.
Complete site validation and authority consultation before freezing equipment. A change in the cooling chemical package, makeup source, target cycles of concentration or discharge route can alter pretreatment, disinfection and concentrate handling. Treat HydropureWater’s 20–34-week modular-skid range from the 2026 guide as planning information only, with the project schedule confirmed after design, approvals, integration and performance testing.
Frequently Asked Questions
How should RO and UF be sized for cooling-tower blowdown?
Size them from the measured feed-water envelope and the required permeate or reuse duty, not from peak blowdown alone. UF sizing needs turbidity, flow and required barrier performance; RO sizing needs conductivity or TDS, silica, hardness or alkalinity, sulfate, temperature, antiscalant limits and the concentrate destination. Obtain the membrane manufacturer’s feed envelope and use site testing or pilot data to verify the selected recovery.
What recovery rate should a supplier guarantee?
Specify recovery with a defined feed composition and calculate the associated concentrate flow. IDE’s 2026 article places conventional brackish-water RO commonly at 75–80% recovery, while HydropureWater’s literature states that its PLC-controlled industrial systems can reach up to 95% subject to chemistry and scaling limits. Neither is a universal site guarantee. Require the bidder to demonstrate the proposed recovery across the complete feed envelope and identify any special high-recovery architecture.
What should be included in the cost estimate and bid comparison?
Request an itemised scope covering treatment equipment, tanks, chemical storage and dosing, electrical and controls, monitoring, sludge or concentrate handling, installation, commissioning, training and operating costs for chemicals, membrane cleaning and consumables. No research-based quotation for this Pietermaritzburg scope is available, so a defensible budget requires the site flow, analyses, recovery target, reuse or discharge route, utility limits and footprint from this article’s supplier-ready process basis.
How should a supplier be selected without misreading a South African discharge requirement?
Select on demonstrated performance within the stated feed envelope, scale-control strategy, reject-water route, redundancy, operator support and commissioning evidence—not only recovery or catalogue capacity. Ask bidders to identify every assumption and exclusion. Require the project team to obtain current South African environmental, water-use and municipal sewer-acceptance conditions from the competent authorities; U.S. NPDES, POTW and Clean Water Act terminology must not be used as local acceptance criteria.
