As hardware densities increase, traditional cooling methods struggle to keep pace without consuming vast amounts of power, so we have engineered state-of-the-art cooling systems to drive energy efficiency.
Deep dive into the cooling at our new data centres
Posted on in News
Advanced cooling
Airflow with fan wall topology
Traditional raised-floor cooling often creates uneven temperature distribution and limits the physical height of rack installations. We bypassed this limitation by implementing a ‘fan wall’ or ‘through-wall’ airflow topology. This method is sometimes known as a ‘room-flood’ approach.
Instead of pushing air up through floor tiles, we use the full height of the room to distribute cooling air. The cool air passes gently through a diffusion wall that runs the length of the data hall. This wall acts as a precise physical boundary, effectively separating the facility’s grey space from the white space where the IT equipment resides.
Flooding the entire room from floor to ceiling eliminates thermal dead zones. Every single server rack receives a consistent, reliable supply of air regardless of its position in the hall. This design dramatically improves the baseline efficiency of the entire facility.
Low-velocity airflow
Once the cool air cools the IT equipment, the resulting ‘hot’ return air is directed straight up into a ceiling void.
This is often achieved through a dedicated chimney mounted on the top of the rack. The hot air flows into the ceiling void and travels back into the chiller units. By physically isolating the hot return air from the cold supply air, we prevent thermal mixing and reduce the workload on our cooling infrastructure.
Because we maintain strict hot and cold separation, we can utilise low-velocity airflows. The fans operate at lower speeds simply because the air moves slowly and purposefully. This allows us to fully leverage the efficiency of our chillers, making the whole system vastly more efficient and drastically reducing our overall energy consumption.
Intelligent and automated climate control
Managing the climate across a data hall requires constant, precise adjustments. We use sophisticated automated systems to monitor and modulate our cooling in real-time.
Pressure differential sensor matrix
Located securely in the service corridor, our N+1 310kW fan wall cooling units provide the primary cooling. To control these units, we installed a comprehensive pressure differential sensor matrix.
This matrix constantly monitors the difference in air pressure across the environment. It dynamically matches the volume of air being consumed by the IT equipment with the exact fan speed required. Because this process is completely autonomous, the system never works harder than necessary.
Temperature sensor matrix
Air pressure is only half the equation. We also utilise an automated air temperature cooling control system, governed by a vast temperature sensor matrix.
These sensors measure temperatures across the entire white space. If a specific zone experiences a spike in heat load, the system autonomously modulates the cooling units to address it instantly. This granular level of control ensures optimal operating conditions while eliminating wasted energy.
Concurrently maintainable infrastructure
To guarantee continuous uptime, our chilled water circuits are designed to be concurrently maintainable. This means our engineering teams can perform routine maintenance, upgrades, and repairs on the cooling infrastructure without interrupting the system’s operation. The redundancy built into the pipework and valves ensures that cooling water continues to flow seamlessly.
As a result, our clients experience absolutely zero temperature elevation in the white space during maintenance windows. The IT equipment remains completely protected, and operations continue without a hitch.
Maximising efficiency with free-cooling chillers
Moving on to the heavy-lifting equipment: a robust bank of N+1 780kW ‘free-cooling’ chillers reside securely in a dedicated plant compound. To mitigate risks during summer heatwaves, these chillers are specifically rated to operate in external temperatures up to 40 degrees Celsius.
We employ dynamic chiller sequencing to extract maximum energy efficiency from this setup. The autonomous control system constantly evaluates fan speeds and internal compressor activity. It then decides the absolute most energy-efficient way to run the chillers based on real-time data.
The system reacts dynamically to external weather conditions and internal data hall loads. For example, it might selectively engage specific compressors to lower the required fan speeds, thereby reducing the facility’s overall energy consumption.
Optimising free cooling hours
Compressor operation is highly energy intensive. To combat this, we integrated a chilled water system optimisation control system.
This system’s primary goal is to maximise our ‘free-cooling’ hours. Free cooling leverages the naturally cool outside air to chill the water, bypassing the need for power-hungry compressors entirely. By aggressively maximising these non-compressor hours, we dramatically shrink the carbon footprint and operational costs of the facility.
Next up … FRN2
The advanced cooling topologies, intelligent automation, and robust resilience detailed above represent our new baseline for all ongoing and future infrastructure developments. Already operational in our live MCR2 facility in Manchester, these high standards will be fully implemented at FRN2, our four data centre currently under construction in Farnborough.
If you are looking to house your IT infrastructure in a facility that prioritises both maximum uptime and environmental sustainability, we invite you to contact us to learn more about the capabilities at our London edge and Manchester sites.