)
It is a complicated time for data center cooling.
The demand for high-density data centers is pushing legacy cooling technology to its limits. Air-cooling systems, which have long served the first generations of data centers, may not be equipped to handle the high heat generated by the processing units that artificial intelligence requires.
While liquid cooling systems may seem like the ideal solution, many facilities still have heat loads, sitting alongside their liquid-cooled AI loads, that are optimally cooled by air.
“Liquid and air cooling systems will need to coexist,” said Matthew Bevan, service and controls sales director at Airedale, a cooling solutions company. “We understand that the redundancy, resiliency and energy efficiency that has been built into our data center products now needs to be designed around hybrid systems. They need to be scalable as ever, but they also need to be adaptable for how liquid and air live together in the same space.”
The pressure is not only on cooling systems but also on building management systems. As data centers increase their equipment density, the associated BMS must process a higher density of information. While a BMS is crucial for monitoring key building performance indicators, legacy systems are made to process data, not predict the complex issues that may arise in high-density environments with hybrid cooling systems.
More data center operators are also turning to modular construction, which allows for maximum flexibility as they scale their facilities to the applications they are ultimately used for. Integrating these modular solutions into a BMS can be very difficult, Bevan said. This is why Airedale is designing its solutions to be as flexible, adaptable and modular as possible.
“A lot of building management systems haven't changed. They're focused on just reading information and handling basic controls when it comes to your chilled water plant,” he said. “With complex chilled water plants, it’s difficult for a standard, off-the-shelf BMS to react quickly and effectively enough to maintain that level of control that you need.”
Airedale’s Cooling System Optimizer acts as an “intelligent orchestrator,” managing the entire cooling plant rather than relying solely on BMS data, Bevan said. It is a smart connected control layer between the cooling system and the BMS that can work in large, complex, variable-flow chilled-water systems, supporting cooling loops of 75 megawatts and beyond, depending on the size of the data center, he said.
“As cooling systems become more complex, there's a need for high reliability and uptime with reduced energy consumption,” Bevan said. “This is why we have built our optimizer around the three core principles of redundancy, resiliency and efficiency.”
Bevan said all three principles are priorities for the optimizer, starting with redundancy. Maintaining uptime to avoid a range of losses that can occur when a system is down is paramount, as is the ability to manage backup equipment effectively when needed. Finally, the optimizer is designed to help minimize energy consumption while optimizing cooling performance.
“You can have very different cooling needs at different times of the day in different parts of the data hall, and we pride ourselves on being able to optimize to satisfy those needs efficiently,” Bevan said.
He added that the optimizer uses programmable logic controller-based deterministic controls, the same controls hardware found in its cooling products. This allows the optimizer to monitor individual product controllers as it oversees the system as a whole to perfectly match cooling demands.
“A PLC-based controller that is working in a distributed architecture allows the system itself to act much more holistically and efficiently than what an off-the-shelf BMS control system may do when it's focused on looking after critical cooling plants,” he said.
Data center operators that are interested in optimizing their cooling systems should work with a partner like Airedale as early in the build process as possible, Bevan said. Airedale works with design teams from Day 1 to integrate its system and works closely with the customer to understand not only its requirements but also its problems.
“Everybody's got a problem they're trying to solve, and the way in which it's being solved often drives requirements,” he said. “We try to understand the problem first, and then we work openly with customers to solve it.”
This article was produced in collaboration between Airedale and Studio B. Bisnow news staff was not involved in the production of this content.
Studio B is Bisnow’s in-house content and design studio. To learn more about how Studio B can help your team, reach out to studio@bisnow.com.
)
)
)
)
)
)
)
)
)
)
)
)