Data center thermal modeling using Computational Fluid Dynamics (CFD) is gaining popularity as a tool for analyzing the effectiveness of cooling within the racks and aisles. CFD provides companies with a detailed 3-D analysis of how cold air is moving through a data center, identifying potential “hot spots” where equipment is receiving too little airflow. Thermal mapping can also find areas in a data center that are receiving more cold air than needed, wasting cooling and energy.
With the push for greater efficiency in data center power and cooling, CFD is becoming an essential tool for many companies, according to Paul McGluckin, a research VP at Gartner. “There was a time when this was an esoteric gearhead kind of thing, but not anymore,” said McGluckin. “It’s well worth the cost to identify potential hot spots ahead of time. If your consultant doesn’t think it’s necessary, get another vendor who will assist you with this.”
Computational fluid dynamics is used to generate flow simulations with the help of computers. It is widely used in aerospace, biomedicine, semiconductors, and developing advanced graphics for movies and games.
“CFD is the best tool for making sure your (data center) model is thorough,” said Pete Sacco, president of PTS Data Center Solutions, who gave a presentation about CFD solutions at the DataCenterDynamics New York event. “We use it as a design tool and also as an operational maintenance tool. It allows me to consider several options for air conditioning, and try alternate approaches that I wouldn’t necessarily try in a client facility.”
Popular products for specialized CFD software for data center HVAC analysis include:
- 6Sigma from Future Facilities, a UK-based consultancy that specializes in CFD tools for mission-critical facilities.
- Flovent from Flomerics Group (FLO), also based in England.
- TileFlow from Innovative Research Inc. of Minnesota.
- CoolSim software from ANSYS, a Pennsylvania maker of engineering simulation products.
Sacco, who uses 6Sigma, said CFD thermal modeling is especially useful in modeling solutions that use several different approaches to cooling – a scenario that is coming up more often as high-density hot spots prompt facility operators to use in-row cooling as well as perimeter cooling using computer room air conditioning (CRAC) units.
“The density is what’s causing everyone grief,” said Sacco. High-density installations are famously less forgiving than lower-density data centers, as temperatures rise quickly when cooling is unavailable. With CFD, Sacco said he “can run a full failure case – let a system fail and see the impact and temperature implications.”
The tool’s real power is in identifying places where cold air is being wasted or is mixing with hot air, reducing the effectiveness of the overall cooling. “Air mixing is the enemy,” said Sacco. “It’s an obvious one, but it needs to be stated over and over. I cannot have hot air mixing in with the cold air I’m paying to generate.”
CFD is also useful for modeling airflow under a raised floor, which is helpful in designing solutions. “We can control airflow using baffles (under floor dividers to direct air flow) and variable speed fans in the CRAC,” said Sacco.
That’s particularly important in open “barn” data center layouts, which tend to be less efficient. “Building massive open floor spaces is less effective than building small rooms,” said Sacco. “Bigger is not better.”