Emergency Cooling Strategy: Bridging Chiller Failure with Storage and Free Cooling

Emergency Cooling Strategy: Bridging Chiller Failure with Storage and Free Cooling

The emergency cooling tank is a physical asset; the emergency cooling strategy is the operating plan that uses it. When a chiller trips, the question is not just whether you have reserve water, but how the plant rides through the event without exceeding supply-temperature limits. This article covers the strategy, with the tank as one component among free cooling, sequencing, and controls.

An emergency cooling strategy bridges a chiller failure by combining a reserve water tank, available free cooling, and automatic sequencing so the IT load stays within its thermal envelope until backup capacity returns. The tank buys minutes; free cooling and smart sequencing buy more, and the controls decide the hand-off. The strategy matters as much as the steel.

The first minutes: stored cooling

The reserve tank (covered in the companion article) supplies chilled water immediately, holding supply temperature while the plant reacts. Its value is the time it buys, so the draw path must be automatic and the volume matched to the bridge window. The Dubai fire-water tank (2025, 652 m3) is a reserve water tank we delivered for another critical site, cited as analogous capability, not a data-center project.

• Tank supplies cooling in the first minutes.
• Draw path must be automatic.
• Volume matched to bridge window.
• Buys time, not the whole event.

The middle: free cooling

Where climate allows, free cooling (economizer or waterside) can carry much of the load once the chillers are down, extending the reserve. The hand-off must be automatic and the free-cooling path must be sized for a meaningful fraction of load, not just partial relief.

• Free cooling extends the reserve.
• Automatic hand-off to economizer.
• Path sized for meaningful load.
• Climate-dependent value.

The controls that make it work

Sequencing logic must detect the trip, open the reserve draw, invoke free cooling, and stagger backup chiller start to avoid a power spike. A resilient tank is useless if the BMS does not know how to use it. Coordinate the tank level, temperature, and pump status into one failure routine.

• Detect trip and open reserve draw.
• Invoke free cooling automatically.
• Stagger backup start to limit spike.
• One integrated failure routine.

Reliability engineering behind the tank

The tank itself must survive the event that triggers it. Bolted steel tanks we have delivered, such as the Ecuador installation that remained stable after a 7.8 magnitude earthquake in 2016, show how shell, anchor, and foundation are engineered for resilience. That project is an analogous resilience reference, not a data-center delivery.

• Shell, anchor, foundation engineered for event.
• Seismic and wind resilience proven.
• Reserve must survive the trigger.
• Engineering, not just volume.

Honest boundaries on our references

We have no data-center emergency cooling project in this library. The reserve and resilience tanks cited (Dubai 652 m3, Ecuador seismic 2016) are analogous critical-infrastructure deliveries offered as evidence of our ability to build resilient, monitored storage, not as data-center projects.

• No data-center emergency project in library.
• Citations analogous, not claims.
• Resilient tank engineering is real.
• Strategy design needs your spec.

Frequently Asked Questions

What is the difference between the tank and the strategy?

The tank is the reserve asset; the strategy is how the plant uses it, plus free cooling and controls, to ride through a chiller trip without exceeding supply limits.

How long can free cooling extend the reserve?

It depends on climate and the free-cooling path size; in cold climates it can carry most of the load, in hot climates it adds little. The tank still sets the floor.

Do controls matter more than tank size?

Both. A larger tank buys more minutes, but without automatic draw, free-cooling hand-off, and sequenced backup start, the reserve is unused during the fault.

Can one tank serve both emergency and TES?

Yes, a shared stratified tank can serve both, but the emergency draw path and isolation must remain dedicated so tariff shifting cannot drain the reserve.

How resilient are bolted tanks to earthquakes?

Engineered bolted tanks with proper anchors and foundations can survive severe events; one we delivered in Ecuador stayed stable after a 7.8 magnitude quake in 2016.

Have you done this for data centers?

Not in our project library. We cite reserve and resilience tanks at other critical sites as analogous capability, not data-center deliveries.

Emergency cooling is a strategy before it is a tank. The reserve water buys the first minutes, free cooling extends it, and integrated controls decide the hand-off to backup capacity, all without exceeding the IT thermal envelope. Size the tank for the bridge window, but invest equal effort in the sequencing and resilience of the shell itself, because the asset must survive the event that triggers it.

Suggested CTA: Send us your failure scenarios and climate data and we will help design a reserve tank plus the draw and free-cooling logic for your cooling continuity plan.