Cooling Data Centers in a Crisis
OTHER PARTS OF THIS ARTICLEPt. 1: Strategies and Tactics For Successful Data Center UpgradesPt. 2: Ensuring Reliable Power to Data CentersPt. 3: Commissioning Data Center Systems To Achieve Performance GoalsPt. 4: This Page
Planning upgrades to data centers is complex and potentially costly, give the big-ticket products that go into such projects and the long-term implications of their success or failure. But as maintenance and engineering managers take part in the planning of large-scale, long-term upgrades of the HVAC and power systems serving these mission-critical operations, they would be wise to also pay attention to data centers’ potential short-term needs in the event of an emergency.
When managers consider the impact of utility outages, they tend to focus their attention on electrical outages. If a data center loses electricity, operations shut down. For this reason, many areas that house essential operations have dual electrical feeds or alternative power from standby generators. Mission-critical operations might even have an uninterruptible power supply to eliminate even minor power disruptions.
The problem is that many managers and others involved in planning have not paid the same kind of attention to the impact on these operations if a facility’s cooling system should fail. Facilities that house telecommunications or large computer operations demand cooling year round. Just as with the loss of electricity, the loss of cooling to such equipment would force managers to shut down systems and operations to prevent damage to sensitive components.
The need for emergency backup cooling systems for data centers has led to the development of a range of portable cooling systems that are designed for temporary use as a backup system in an emergency or as a supplemental system to meet temporarily high loads.
Self-contained, fully mobile units mounted on wheels are available with cooling capacities of up to 5 tons. The units are small enough to fit through doorways and can operate practically anywhere in a building. A flexible duct carries heat from the units through a window or into an unoccupied space nearby.
Larger-capacity units mounted on trailers can be set up outside a facility. The units deliver conditioned air to the space through a flexible duct and can be air-cooled or water-cooled, depending on the system’s size and the availability of local utility connections. Some units come with their own generator if the facility does not have sufficient electrical capacity. Managers must have a plan in place before a crisis hits. Trying to figure out the right system for a situation and then locating it and getting it in place after an outage occurs simply will not work.
Managers need to match units to the needs and conditions of the areas that require immediate cooling. They can start this process by identifying critical areas in facilities that demand cooling. For example, computers, telecommunication, and other sensitive electronic equipment might require cooling 24/7 to prevent overheating and damage. Even certain areas where cooling is strictly for comfort might have critical needs, particularly if the loss of cooling would result in significant operating losses for the facility.
To accurately size a portable unit, engineers need to perform load calculations for the space, taking into account heat gains from exterior walls, windows, people, equipment, and surrounding areas. But managers can use several guidelines to develop quick estimates of cooling requirements.
Managers next should must identify factors related to the specific application. Does the area have enough power for the portable cooling unit? What voltage is the power supply? How and where will the unit reject heat? What size do portable units need to be to fit in the space? Are there weight and noise limits? What other modifications does the space require to integrate the portable cooling unit to a particular application?
Managers must address all of these issues well in advance of an actual cooling outage.