Recovery guarantees and measurable water savings for data centres
Contents
Direct answer
A blowdown-recovery guarantee should describe sustainable net performance across an agreed feed envelope. It should cover product flow and quality, recovery, availability, energy, chemicals, cleaning and residual streams. A separate measurement-and-verification plan should prove the site’s actual water and cost saving relative to an approved baseline.
Process recovery and site saving are not the same. A plant can meet its guarantee while product is spilled, tower demand changes or another source increases. Clear boundaries, calibrated meters and normalisation for cooling load and weather prevent disputed claims.
For repeatable delivery, translate these requirements into the performance schedule of the DC-BR packaged module and prove its interfaces through the battery-limits, FAT and commissioning framework.
1. Separate equipment guarantees from project outcomes
The equipment supplier controls treatment performance within defined boundaries. The owner controls product demand, tower operation and many site utilities. The utility or cooling-chemistry provider may control additional inputs.
Create two linked documents. The process guarantee defines what the recovery system will produce and consume. The measurement-and-verification plan defines how site withdrawal, discharge and cost change will be calculated.
This separation supports the business case without assigning unmanageable risk to one party.
2. Define the feed and operating envelope
Include normal, upper-percentile, maximum and short-duration upset values for flow, temperature, pH, conductivity, hardness, alkalinity, silica, chloride, sulphate, suspended solids, organics, oxidant and other controlling parameters.
State which cases require full performance, derated operation or feed diversion. Define sampling method, laboratory, instrument accuracy and resolution of disputed results.
Include utility ranges, cooling-water availability, air temperature where relevant and allowed influent chemical residuals. Guarantees based on one clean sample are not adequate for blowdown.
3. Guarantee net product and residuals
Net product excludes internal recycle, backwash, displacement and off-spec water. Define the compliance point, pressure and quality limits. If product is blended, state whether the guarantee applies before or after blending.
Guarantee sustainable recovery as net compliant product divided by measured blowdown feed over the test period. State allowed tank-inventory correction.
Define concentrate, sludge, backwash and cleaning flows and key compositions. A system that produces the promised water but breaches the approved residual route has not met the project requirement.
4. Set availability and turndown requirements
Availability should exclude only agreed owner-caused events and planned maintenance within a stated allowance. Define start-up, standby, cleaning and recovery time.
Set net capacity with one required train or common component unavailable. Identify shared pumps, analysers, chemical systems and controls that can defeat nominal redundancy.
Specify stable turndown for early campus phases. Cycling, stagnant trains or frequent flushing can reduce both availability and net recovery.
5. Specify energy, chemicals and cleaning
Measure electrical energy at a defined boundary and normalise to net compliant product. Include feed, transfer, filtration, RO and residual pumps as agreed.
Guarantee or schedule chemical consumption for each feed case. State membrane and consumable assumptions. Cleaning frequency should be tied to normalised performance limits and representative operation.
Record cleaning waste and downtime. A high recovery achieved through frequent cleaning may fail the lifecycle case despite meeting a short test.
6. Establish the savings baseline
Use a pre-project period that covers representative heat load, tower cycles, source quality and weather. Correct meter faults and exclude clearly documented abnormal events.
Define baseline make-up and discharge as functions of cooling duty where possible. If the tower is optimised at the same time as recovery, separate the saving attributable to increased cycles from recovered product.
Freeze the method before commissioning. Changing the baseline after results are known undermines credibility.
7. Design the metering and data model
Meter blowdown feed, product, concentrate and major backwash or waste streams. Record make-up sources, tower blowdown, tank levels, heat load, weather, conductivity and recovery availability.
Select meter accuracy and range for the expected turndown. Provide calibration, totaliser retention and time synchronisation. Tank-level changes can create apparent production or loss during short tests and should be corrected.
The historian should calculate instantaneous and period recovery, net return, energy intensity and site saving. Flag missing or invalid data rather than silently substituting estimates.
8. Conduct acceptance and reliability tests
Factory tests prove fabrication, controls and simulated sequences. Site tests prove hydraulic capacity, product quality, residuals, metering, diversion, standby changeover and integration with tower operation.
Test normal and limiting feed where safely available. Apply agreed temperature and quality corrections. Challenge analyser failure, high conductivity, low tank level and recovery after a trip.
A reliability run should last long enough to observe stable normalised performance, backwash and ordinary cleaning demand. Record operator interventions and excluded periods transparently.
9. Verify savings after handover
Use monthly and annual periods, normalised for cooling duty and relevant weather. Reconcile site make-up, blowdown, product, residuals and inventory. Compare with the approved baseline.
Report product actually reused, not only produced. Include energy, chemicals, cleaning, disposal and availability so financial saving can be calculated.
Review differences between guaranteed process performance and achieved site outcome. The response may involve treatment optimisation, product-demand controls, tower operation or metering—not automatically a supplier failure.
10. Allocate remedies and change control
Define retest, correction, performance damages or other remedies proportionately. Identify owner obligations for feed, utilities, operation and maintenance.
Changes in tower chemistry, source water, product use or discharge limits require technical review. State how guarantees are adjusted or retested.
Maintain the guarantee schedule, test records, calibrations and operating data as part of the proof and assurance framework.
11. Protect data quality and auditability
Assign each measurement an owner, calibration frequency, accuracy and valid operating range. Store raw totals as well as calculated KPIs. If a meter is replaced or scaled, retain the change record and closing total.
Time synchronisation matters when flows and tank levels change rapidly. Misaligned data can create apparent loss or production. The historian should use a common time basis and flag communication gaps.
Define rules for missing data before the performance period. Short gaps may be estimated from an agreed method; extended gaps may invalidate the period. Avoid filling missing values silently.
Laboratory samples used for guarantees need chain of custody, methods and agreed laboratories. Preserve split samples for disputed results where practical. Online instruments support control but do not replace the agreed compliance analysis.
12. Align guarantees with operation and maintenance
Supplier performance assumes defined operation. State owner obligations for feed availability, utilities, consumables, calibration and timely maintenance. Training and procedures should be complete before the reliability run.
Guarantees should not encourage harmful operation. A recovery target that drives the plant above its validated scaling margin can threaten membrane life and residual compliance. Use multiple limits and a priority hierarchy.
After acceptance, retain the same KPIs for operational review. Compare actual cleaning, energy and chemical consumption with the guarantee schedule and investigate trend changes early.
Service agreements can include periodic performance review and optimisation without obscuring accountability for defective equipment. Define reporting, response time and access to operating data.
13. How Crescent Engineering can support the project
Crescent Engineering can define performance schedules, metering, acceptance tests and M&V calculations, then support commissioning and ongoing optimisation against the agreed baseline.
14. References and further reading
Contract language, baselines and remedies should be reviewed for the project’s technical and commercial requirements.
Related Crescent articles
Continue exploring this topic:
The business case for cooling-tower blowdown recovery in AI data centres
How far should cooling-tower cycles of concentration be increased?
Filtration, softening and high-recovery RO for cooling-tower blowdown
Managing silica, hardness and salinity in cooling-tower blowdown
Brine minimisation and discharge control for blowdown recovery
Retrofitting blowdown recovery into an operating data centre
Comments