This audio was created using Microsoft Azure Speech Services
As the needs of data centers evolve, the ways they work have changed with new networking methodologies and increases in server density. The facilities to house these operations are also changing, and that involves cooling needs. Fortunately, one of the major driving forces is reducing operating costs so most technologies are improving in this regard. When making a cooling system selection, data center host companies need to look at four major factors: financial, facility constraints, qualitative, and regional.
One of the most significant changes is a shift in cooling technologies for large data centers, defined for purposes of this discussion as 1 MW and higher. In a new webcast, Data Center Cooling with Large Capacity Systems, Joe Capes, director of business development for Schneider Electric’s cooling line of business in the Americas, addresses these shifts as more users move away from traditional and expensive vapor compression technologies as the primary solution.
The word “economizer” has taken a higher level of importance in the discussion. That’s not a new term, certainly. The idea of leveraging cool outdoor air to reduce compressor loading is widely practiced, but has not moved pervasively into data center cooling approaches. That is changing, particularly for large systems, as the term “free cooling” comes into the equation.
Of course no cooling approach is truly without cost, but Capes says that when the mindset changes from traditional compressor-based systems, to an economizer that gets an occasional assist from a vapor compression cycle if needed, you’ve turned an important corner. Capes explains that the costs of such a system can be much lower once you get to the critical mass of IT load around 1 MW. One major reason is that legacy cooling systems are designed to run at full capacity, but rarely do so in normal operation. Studies show that most data centers typically run in the 28% neighborhood, and distributed low capacity technologies lose efficiency when running under partial load.
Fans associated with large economizer systems can ramp up and down and maintain a much more linear relationship between cooling capacity, energy consumption, and efficiency. Even when ambient temperatures get too high to allow an economizer to deliver its highest efficiency, in some areas, a simple indirect evaporative cooling system can delay the need to kick in a vapor compression cycle.
The nature of a facility will affect the practicality of using an economizer. The ducts necessary to move the amount of air required can pose challenges when undertaking retrofits in existing buildings. Most air handlers are mounted on the roof, and if the data center is on a lower floor of a tall building or in the center of a large footprint, there can be major constraints. Air quality and moisture content can also enter into the discussion if there is rain-forest humidity or heavy particulate levels.
It’s also no surprise that regional considerations enter into the discussion. Capes mentions the situation in Vancouver will be much different than Miami, although there really isn’t anywhere on the planet, including somewhere as hot as Dubai, where indirect evaporative economizer technology loses all of its effectiveness. For a detailed review of the 4 critical factors you should consider when selecting a large capacity cooling system, please view this webcast, but dollar for dollar, for systems greater than 1 MW, an economizer-based cooling system using hot aisle containment will generally deliver the lowest overall cost.
