The fragility of the electric grid was demonstrated once again as Hurricanes Harvey and Irma caused widespread power outages. However, it doesn’t take a natural disaster to bring the grid down — our aging electric infrastructure is increasingly vulnerable to various types of failure. While detectible power outages account for only a small percentage of those events, surges, sags, noise, harmonics, load fluctuations and other power anomalies occur on a daily basis.

For data center operators, this is a major concern. The quality of the electricity flowing through power lines varies substantially, and those fluctuations can harm sensitive IT equipment.

Uninterruptible power supplies (UPSs) serve as a primary shield against these anomalies. As the name suggests, UPSs automatically supply backup power when they detect a sag or outage. However, the primary function of the UPS is to provide a reliable source of conditioned power with a regulated voltage level.

Two types of UPSs are commonly used in data center environments:

  • Line-interactive. If the voltage from the electric grid or generator falls outside specified high or low threshold values, the UPS uses transformers to bring the voltage within the acceptable range. This prevents frequent transfers to backup power, which reduces battery life significantly. However, the inverter that converts DC to AC is always engaged, providing a smooth transfer to battery power when an outage is detected.
  • Online. These systems isolate equipment from power fluctuations by continuously converting the AC power supply into DC power, then converting the DC power back to AC power at the ideal voltage. There is no transfer when AC power fails because it is not the primary power source. A transfer occurs if the inverter power path fails.

The choice of UPS has implications not only for the quality of power supplied to equipment but for the overall energy-efficiency of the data center. UPS systems are part of the data center’s power overhead — that is, the power consumed by non-IT equipment. Line-interactive systems are inherently very efficient. Online UPSs, on the other hand, provide about 94 percent efficiency in normal operating mode, and about 98 percent efficiency in eco mode. However, line-interactive systems are not practical for loads greater than 5kVA, and online UPSs provide superior power conditioning.

Whatever UPS you choose, proper maintenance is extremely important. According to the Cost of Data Center Outages study published every three years by the Ponemon Institute, UPS failure is the No. 1 cause of data center downtime, accounting for 25 percent of all such events.

Battery failure is the primary cause of UPS problems. The electrochemical cells found in UPS batteries deteriorate over time, and a number of operational and environmental factors can shorten their life. Batteries degrade more frequently they are charged and discharged. A full discharge also decreases battery life, so many units automatically shut down with some capacity still in reserve. UPSs are deigned to operate in an ambient temperature of 68 to 77 degrees.

At minimum, a preventive maintenance program should involve visual inspection of the equipment for loose connections or signs of corrosion, and cleaning of the UPS enclosure. Network-based monitoring and management software can provide visibility into the status of multiple UPS systems, and facilitate setup, configuration and controlled shutdowns.

Rahi Systems has expertise in the design of data center power infrastructure, and offers UPS solutions from leading manufacturers. Let us help you select and implement the right UPS systems to protect your IT equipment, minimize downtime and maximize energy efficiency.

Sydney: +61 (2) 8488 4700 Melbourne: +61 (3) 8488 7390