top of page
Search

Storage, blending and monitoring for reliable recycled cooling water

Writer: Mark Cullens
Mark Cullens
3 days ago
8 min read


Aerial view of a data-centre campus with rooftop cooling equipment and water retention ponds

Contents

Direct answer

Storage, blending and monitoring turn a technically treatable recycled-water source into a dependable cooling utility. Storage provides time to respond to interruptions and quality excursions. Blending allows two sources or treatment products to meet a defined make-up specification. Monitoring proves that the delivered water remains within that specification and initiates a safe response when it does not.


These functions must be designed together. A large tank cannot compensate for an unreliable analyser, and a calculated blend ratio is not dependable if the sources do not mix before the compliance sample. The control philosophy should begin with credible events and work backwards to volumes, instruments, valves and operator actions.


For packaged projects, assign these functions explicitly between the DC-MU treatment module and the selected standard capacity and redundancy arrangement.


1. Assign a clear purpose to every tank

Raw recycled-water storage can buffer utility pressure and allow off-spec feed to be isolated. Equalisation can smooth treatment loading. Treated-water storage can maintain cooling make-up during a process trip. A separate backup-water reserve may cover a longer utility interruption. These are different duties and should not be combined without testing the consequences.


For each tank, document the source, users, normal operating level, usable volume, minimum reserve, refill rate and quality controls. State whether it is allowed to receive off-spec water, whether it is part of an emergency reserve and what happens when the level falls below its operating band.


The municipal recycled-water pillar guide places these tanks within the full utility-to-tower architecture.


2. Size storage from operating scenarios

An arbitrary number of storage hours is not a design basis. Calculate demand and inflow through each event. If a treated-water tank supplies 120 cubic metres per hour and contains 720 cubic metres of genuinely usable volume, it provides six hours only if no other user draws from it and the required pump suction and emergency reserve are protected.


Relevant scenarios include:


  1. A utility interruption during peak summer cooling.

  2. Off-spec recycled water requiring inlet isolation.

  3. One treatment train unavailable at design demand.

  4. An analyser failure that forces conservative diversion.

  5. Transition to a lower-capacity backup supply.

  6. Recovery following the event while the tank is refilled.

Recovery often governs. A tank may ride through a two-hour outage yet take ten hours to return to its normal reserve while meeting live cooling demand. Repeated interruptions can therefore erode resilience.


Use time-based simulations for variable IT load, hybrid cooling modes and phased campuses. Confirm net usable volume after low-level shutdown, freeboard, sediment allowance and fire or other protected reserves.


3. Preserve quality inside storage

Storage can improve hydraulic resilience while degrading water quality. Warm conditions, long residence time and nutrients encourage biological growth. Settled solids can accumulate and later be disturbed. Poor inlet and outlet geometry can leave stagnant zones even when average turnover appears acceptable.


Design considerations include mixing, circulation, shade or thermal exposure, internal finish, access, drainability, overflow protection, screened vents and safe cleaning. Separate tanks or compartments can preserve supply during inspection. The materials and coating should suit recycled water, disinfectant and any pH adjustment.


Use residence-time estimates for early and full campus loads. A tank sized for ultimate capacity may have inadequate turnover during the first phase. Operating ranges, automatic flushing or temporary compartment isolation may be needed.


Disinfectant residual should be monitored and controlled for the stored-water duty, but it must remain compatible with downstream membranes, cooling chemicals and the water-safety plan. Slug dosing without mixing can create local corrosion or fail to protect the full volume.


4. Calculate and control blending

Blending is useful when neither source alone provides the preferred balance of quality, cost and availability. Common arrangements mix reverse-osmosis permeate with filtered recycled water or blend recycled water with an approved potable or desalinated backup.


For a conservative constituent, the blend concentration equals the sum of each source flow multiplied by its concentration, divided by total flow. If a 60:40 blend combines water at 100 and 700 mg/L chloride, the result is approximately 340 mg/L. The same calculation must be repeated for hardness, alkalinity, silica, sulphate and other controlling parameters.


The controlling constituent may change with source conditions. A fixed ratio that satisfies conductivity can still exceed a chloride or silica limit. The permitted blend region should therefore be calculated across the complete feed-water envelope.


Hydraulics determine whether the calculation becomes reality. Control valves need adequate rangeability; flow meters need accuracy across turndown; the streams need sufficient mixing volume; and the compliance sample must be located after mixing but early enough for diversion. Transport delay should be included in alarm logic.


If an analyser fails, the system should move to a defined safe ratio, hold the last validated value for a limited time or divert to backup water. Continuing uncontrolled blending is rarely an acceptable default.


5. Select measurements that support decisions

Every online instrument should have a purpose. Typical utility and treatment measurements include flow, pressure, tank level, conductivity, pH, temperature, turbidity and disinfectant residual. Reverse-osmosis systems add permeate conductivity, differential pressure and recovery. Cooling towers add make-up, blowdown and recirculating conductivity.


Online hardness, silica or organic measurements may be justified when response time and reliability fit the risk. Where they do not, conservative proxy limits and laboratory confirmation may be stronger.


Specify range, accuracy, repeatability, response time, installation, sample conditioning, calibration and maintenance access. An instrument selected for broad range may lack the accuracy required near a critical limit. Air bubbles, temperature variation, low sample flow and dirty cells can produce false readings.


Critical measurements may require redundancy or independent validation. Redundancy should detect disagreement rather than merely display two numbers. Voting logic, deviation alarms and manual confirmation procedures must be defined.


6. Design alarm, diversion and backup sequences

Develop cause-and-effect logic from the failure scenario. A high feed-conductivity event might close the utility inlet, preserve the treated tank, alarm the operator and start the backup source. High treated-water conductivity might divert product to drain or raw storage and inhibit transfer to the cooling tank.


Separate alert, action and trip limits. An alert can trigger sampling or a controlled adjustment. An action limit can modify the blend or tower cycles. A trip limit protects equipment or public health through isolation. Apply appropriate confirmation delays without allowing a harmful volume to pass.


The system must also restore safely. Define the number of compliant readings, laboratory confirmation where required, tank displacement or flushing and the authority needed to return to recycled water.


Backup water may have different hardness, alkalinity and chloride. The transition can change the cooling programme. Dosing setpoints and permitted ramp rates should be part of the sequence, not left to improvisation during an incident.


7. Integrate data with cooling operation

Water treatment and cooling controls should exchange the measurements needed for coordinated decisions. The treatment plant needs actual make-up demand and available tank autonomy. The cooling system needs make-up availability, quality status and expected source transition.


The historian should calculate treatment recovery, blend ratio, tank autonomy, tower cycles and a reconciled water balance. Trends should include chemical use, backup-water hours, off-spec diversions and analyser availability. These measures support WUE reporting without losing the operational detail behind the annual metric.


Cybersecurity and ownership should be agreed. Data exchange need not give one supplier unrestricted control of another system. Clearly defined read and command points, fail states and time synchronisation are more important than an unnecessarily complex integration.


8. Plan sampling, calibration and maintenance

Online monitoring does not remove laboratory work. The programme should state sample locations, parameters, frequency, preservation, laboratory method and response to disagreement. Event-based samples are valuable after an alarm, utility change, treatment clean or unexplained cooling-water shift.


Calibration plans should include standards, frequency, acceptance tolerance and as-found results. An instrument that passes after adjustment may still have allowed an extended period of incorrect control; as-found data determines whether operating records need review.


Provide safe access, isolation, drainage, lighting and sample sinks. Instruments that are difficult to maintain will not remain reliable. Hold critical spares and consumables in line with local lead times.


9. Commission the complete response

Test tanks, blending, instruments and controls as an operating system. Verify usable volumes, transfer rates, mixing, sample delay, valve failure positions and backup capacity. Simulate utility loss, off-spec feed, high product conductivity, failed level transmitter, failed blend-flow meter and loss of one treatment train.


Confirm that alarms reach the correct operators and that written procedures match the automated sequence. Record recovery time as well as initial response. A successful test ends when the system returns to stable recycled-water operation with reserves restored.


The final handover should include a cause-and-effect schedule, setpoint register, calibration certificates, sampling plan, tank inspection programme and operator training evidence.


10. Include human factors and physical safeguards

Automation should make the safe state clear to operators. Use consistent equipment names, readable trends, alarm priorities and displays that show remaining autonomy rather than tank level alone. An operator should be able to see which source is active, whether quality is validated and how long the cooling demand can be sustained.


Manual valves and local controls can defeat a good sequence if their position is not monitored. Identify critical line-ups, use locks or permissions where appropriate and include valve status in pre-start checks. Label recycled, treated, backup and potable systems distinctly throughout the site.


Sampling and chemical handling areas need drainage, lighting, ventilation and safe access in GCC conditions. Instruments mounted in direct sun or inaccessible pipe galleries are less likely to remain accurate. Provide weather protection and maintenance space without creating stagnant sample lines.


Emergency procedures should be short enough to use under pressure. They should state the condition, immediate action, authority, communication and restoration evidence. Train both water and facilities teams because a source transition affects cooling operation as well as treatment.


11. Test resilience against phased campus operation

At low initial load, oversized tanks may turn over slowly and blend valves may operate below their accurate range. At ultimate load, the same system may have limited reserve and little time to respond. Model both ends of the programme.


Modular tanks or operating bands can improve early turnover. Multiple treatment trains and correctly sized flow meters can maintain control across phases. The control system should update autonomy from actual demand rather than a fixed design flow.


Review the strategy when a new data hall is commissioned. Confirm that refill capacity, backup supply, alarms, sampling frequency and operator response remain adequate. Phasing should be treated as a sequence of operating designs, not simply a future capacity allowance.


12. How Crescent Engineering can support the project

Crescent Engineering can develop storage calculations, blend models, instrumentation schedules, controls and cause-and-effect logic, then integrate them within packaged water-treatment equipment and commissioning tests.



13. References and further reading

Storage autonomy, alarm limits and backup sequences must be validated for the specific cooling duty, utility arrangement and jurisdiction.


Related Crescent articles

Continue exploring this topic:


Comments

bottom of page