How Data Centre Can Navigate Heat Wave Challenges

Record-breaking heat waves are making headlines and straining economies around the globe, with no signs of cooling down in the immediate future. Beyond the impact on human lives, extreme heat is taking a major toll on data centre operations and infrastructure across the world.

“If one thing is clear, it is that we need to be prepared for these extreme weather conditions to become more commonplace, so adapting to this inevitability is crucial,” said Karsten Winther, President for Vertiv in Europe, Middle East and Africa (EMEA).

“Customers are becoming more aware of the challenges that extreme heat, for example, brings and are rightly looking at how they can mitigate the resulting risks when it comes to their data centre operations. Our service teams have proved invaluable in providing the guidance and support required to not only manage the immediate impact but to discuss with customers how they can plan ahead to build in resiliency moving forward.”

To keep systems cool and help ensure data centre availability, Vertiv experts recommend these immediate actions:

Depending on the data hall load, run as many units as possible to reduce the load on each cooling unit. Typically, an organization would run all redundant units, but if the IT load is not 100 per cent, one should run enough units to get the per-cooling unit as close as possible. This keeps the refrigerant pressures lower so there is not a risk of trips on high pressure with the high ambient temperature. It also provides a more efficient operating point.
Schedule a regular maintenance plan. Vertiv recommends its customers schedule preventive maintenance quarterly. For compressor-based (DX) units, Vertiv recommends quarterly or more frequent cleaning of the condenser coils. The lack of proper and regular cleaning of the condenser coil can cause more of a reduction in the overall thermal unit capacity than the ambient temperature itself. Additionally, properly maintaining the refrigerant charge level helps to enable the design capacity and operating efficiency to function smoothly. A refrigerant charge that is too low or too high, can cause operational inefficiencies.

“The unprecedented temperatures we’re seeing across the globe are putting increased pressure on data centre operators and infrastructure providers, challenging them to maintain reliability and connectivity,” said Steve Madara, vice president for thermal sales at Vertiv.

He added: “Extreme weather events underscore the importance of a robust preventive maintenance and cleaning plan, tailored to the specific conditions of an operator’s region. Data centre operators may need to be prepared for operating expense trade-offs to compensate for oversizing capacity. Partnering with a qualified data centre services provider is important in order to create a maintenance and services plan that will meet the demands of today’s heat waves and tomorrow’s uncertain climate.”

As the planet continues to warm and the climate crisis intensifies, extreme heat events will become even more commonplace. Short-term actions will need to be accompanied by more long-term, systemic changes in the industry.

“As an industry, we should look to and learn from regions like the Middle East, Africa and Australia, which have already adapted to high temperatures for much of the year,” Madara said. “When we do, it’s clear the status quo won’t carry us into the future. We need to reassess rooftop space planning and design our data centre infrastructures to withstand higher heat loads, rethinking the placement of infrastructure systems in addition to capacity.”

Other long-term action best practices and advice from Vertiv experts include:

Consider different types of cooling systems, including immersion cooling, water-free cooling and evaporative free-cooling systems designed to deliver reliability and efficiency in high-temperature environments.
If relying on solar cells as an energy source, account for the reduction of the efficacy of solar cells as ambient temperatures increase.
Rather than solely relying on historical data to determine data centre capacity – data that fails to account for today’s extreme temperatures – use predictive modelling.