Selecting, maintaining UPS

Engineers who design and specify systems should be cognizant of the various conditions that facilitate optimal lifecycle performance for an uninterruptible power supply system.

By Ed Spears, Product and Training Manager, Eaton Corp., Raleigh, N.C. May 1, 2008

Engineers who design and specify systems should be cognizant of the various conditions that facilitate optimal lifecycle performance for an uninterruptible power supply (UPS) system. Among the factors to consider: selecting a UPS uniquely suited to a particular customer and the facility; ensuring that environmental concerns are addressed; overseeing the system’s integration with control systems; and outlining appropriate service requirements for ongoing UPS health.

Because today’s MEP designers face a variety of integration issues, it can be beneficial to consider a power chain offering, in which a single vendor packages an integrated system to include products such as surge suppression devices, switchgear, HVAC, power distribution, equipment racks, and generators. In addition to providing significant savings for a customer, a packaged solution also offers a more coordinated approach that is easier to maintain. The up-front capital cost also may be lower.

When specifying a UPS, attention must be paid to right-sizing and load matching requirements. Typical guidelines recommend that a UPS be loaded to no more than 80% of its rated capacity, which primarily is to allow headroom for unplanned additional loads. In parallel systems and dual-bus architectures, the UPS system often will operate at less than 50% of load, oftentimes at about 40%. If the system is a 2N+1 configuration, the system can be at 30% or less. Because some models are quite inefficient at this level, be sure to choose a UPS that delivers good partial-load efficiency.

Furthermore, if the load characteristics are different than typical IT or data-processing loads—such as medical imaging, motors, lighting, and HVAC—then the UPS should be oversized to allow for the varying nature and changing magnitude of these loads, as well as for the inrush of motors and large transformers.

Traditional UPS systems used in today’s 2N data centers have an efficiency profile that is high on the efficiency curve—at near 100% load. However, in the typical operating range of 30% to 50%, there is a significant drop. Today’s newer high-efficiency designs offer a “flatter” efficiency curve, and maintain excellent efficiency levels, even at lower operating levels, often seen in 2N and dual bus data centers.

Many times, IT and facilities managers have different requirements for the same UPS. While an IT manager may seek a UPS that is reliable, easily maintainable, has simple network management protocol, or Web-based monitoring, and offers high efficiency for minimized power and cooling costs, a facilities manager will request a unit with a low initial cost, ease of access for installation and service, and Modbus communications and monitoring capabilities. Select a UPS model that allows both facility monitoring systems and IT data center monitoring systems to interface with the unit at the same time.

How many do I need?

A key question that often arises in the early design stage is whether to use one large, centralized UPS and battery system, or several UPS systems distributed among the various sections of the data center. There is no single answer to this question.

Although a centralized system limits the number of devices to monitor and eases maintenance with a single battery system, a distributed system allows power protection equipment to be deployed physically and electrically close to the load it supports. This approach enhances reliability, and allows the IT manager more visibility into the critical power products that ensure system availability.

Regardless of the type of UPS chosen, it is imperative to ensure proper environmental conditions for the unit, including temperature, humidity, and ventilation. While the electronics systems that comprise UPS systems are exceptionally tolerant of temperature and humidity variations, it is prudent to maintain conditions of 70 to 80 F. Additionally, if the UPS has internal batteries or adjacent battery cabinets, they have a more restrictive temperature requirement, ideally 77 F. If kept at a consistently higher temperature, the battery service life may be dramatically impacted.

When it comes to ventilation, most installations will rely on air conditioning in the space where the UPS is located to remove the accumulated heat. Deploying a highly efficient UPS that produces less heat addresses this concern, and saves ongoing cooling and electricity costs.

When integrating a UPS, the unit also should be evaluated for compatibility with a facility’s existing or planned building monitoring system (BMS) or building automation system, which may use various protocols for communication with security, fire detection and suppression, critical power, and cooling equipment. Additionally, simple relay contacts may be used with almost any BMS to communicate the key status items for the UPS. Sophisticated systems that solely monitor the condition of the UPS’s battery system also exist, offering advanced warning if a battery or battery array becomes weakened.

Once a system is selected and installed, the user should pay close attention to service requirements for the UPS and its battery system. An effective preventive maintenance plan will save time and money by minimizing business interruption and the costs of downtime, and enhance overall return on investment by extending the life of the critical power equipment (see Figure 1). Preventive maintenance also is crucial to achieving maximum performance from the UPS system.

Author Information
Spears has 28 years of experience in the industry in UPS systems testing, sales, applications engineering, and training.