You’re reading an outdated article. Please go to the recent issues to find up-to-date content.
A growing fixture in today’s energy-dependent businesses, uninterruptible power supplies (UPS) deliver a continuous stream of electric power, which counteracts the effect of potential shutdowns that may occur through an outage or blackout.
Different degrees, or levels, of backup protection exist and can be improved upon, depending on the makeup and execution of your customer’s system. To mitigate the impact of brief power losses, UPS systems are installed to provide improved reliability. Under normal circumstances, the UPS system, which is connected to a facility’s electrical service, acts as a power conditioner to remove any power spikes, transients, voltage sags or other power anomalies that are present in the incoming line.
This filter effect is accomplished by a double conversion of alternating current (AC) power to direct current (DC) power and then back to AC power. The UPS system typically uses a battery bank that activates when power levels fall below a certain threshold. The batteries provide DC power that goes through a rectifier that converts it to AC power, which then feeds the load. If utility power is lost, then immediately without the load seeing more than a few milliseconds of disturbance—if any at all—the batteries provide power to the UPS module that then feeds the load.
“If a business is concerned about downtime, then they would almost certainly have a UPS system installed to protect file servers, the main computer system or critical communication equipment. In the last 10 years, almost every business has installed one of some size,” said Jim Mackey, president, Evergreen Power Systems, Seattle. “As we become more dependent on computers for accounting data, for communications and for basic operations, we can’t afford downtime.”
The desired result is uptime. Uptime is defined as the amount of time an infrastructure is on, up and uninterrupted. Many factors affect uptime, including unreliable utility power, climate, the environment, machine or system failure, and human error. Even though most power losses are only seconds in length, the loss can add up to millions of dollars of lost revenue for businesses.
While some offices have less stringent requirements for desired uptime than data centers or financial centers, the goal for many businesses and facilities is to achieve high nines, a term that refers to the percentage of downtime in a calendar year. Eight hours of downtime a month equals 96 hours in a year, or 98.9 percent of uptime (or one nine).
Three nines (99.9) amounts to nine hours of downtime a year. Five nines (99.999) amounts to five minutes of downtime in a calendar year. That is considered the highest level that can be achieved, though the Department of Homeland Security is aiming for a future level of seven nines, 99.99999, three seconds of downtime a year. Electrical contractors are participating in the effort.
“The bottom line to achieving high nines is to have a fully redundant system with good solid generator backup by redundant UPS systems with full bypass and ties, as well as being backed up by generator systems with good fuel capacity, with no single point of failure throughout the distribution. Sometimes you’ve got a single utility feed but everything secondary to that utility feed is backed up,” said Gil Hammond, project manager, Evergreen Power Systems. “It’s also important to have good design and construction of the UPS system, creating a sustainable kind of environment, one that can be maintained and have sufficient capacity on the redundant system so that it can be maintained without endangering the load.
“That’s really the secret,” he continued. “You have to have very rigorous quality control systems in place to minimize the human element and to control maintenance of equipment.”
UPS systems created by a variety of companies come in many sizes and types for different applications. Small UPS systems (up to 1,500 watts) are typically used to back up important desktop computers, small file/print servers, networking equipment, and point of sale devices in small office environments, such as law firms, physicians’ offices, small retail stores, bank branches, etc.
Midsize UPSs (1,501–10,000 watts) protect larger application servers, critical networking equipment that delivers advanced services like voice over Internet protocol (VoIP), and even small computer rooms housing up to five equipment racks. Larger UPS systems (above 10,000 watts) are most often used for data center protection, large retail store or branch protection, or even entire facilities protection.
Companies or institutions choose UPS systems depending on their needs.
“It’s a question of how much does a company want to pay for and what is it going to cost if they go down,” said Steve Carlini, director of product management, American Power Conversion Corp. (APC). “We call it multiservice level offerings. The customer will tell us what their availability needs to be and we’ll give them the solution that best matches that.
“For small and medium data centers, we offer our general purpose Smart UPS VT. With it, if one of the power modules fails, a company would have to call a service technician to come in and repair it. Companies that can’t afford the downtime could choose the Symmetra PX, our scalable, modular redundant UPS. It has modules so that if one of them fails, the power is shared between all of the modules. That system can still operate while the power module is changed, a task that takes only a few seconds when done by a trained user.”
In the healthcare environment, facilities typically install redundant systems, backing up some of the information with a UPS system. Specific equipment or departments within hospitals call for particular systems.
“There’s variations of our UPS that we use for blood analyzers or MRIs, machines that have very high inrush currents so they’ll pull a lot of power initially while they’re operating,” said Paul Marcoux, director of NCPI Science Center, APC. “Then they’ll settle. You have to design special UPSs for those applications. Then there are regulations that must be complied with depending on the type of patient care, for example, critical patient care UPSs or vicinity patient care UPSs.
“For industrial settings, we also have industrial grade UPSs with different internal components and different external components, which will have thicker sheet metal, drip shields and air filterings. Some are watertight,” he said. The company builds the UPS systems to correlate with different grades classified by the National Electrical Manufacturers Association (NEMA).
Ready and available
Whatever the application, companies often want to improve the level of availability.
“When you are talking about improving the level of availability, there’s multiple different ways you can do that,” Carlini said. “You can do it with adding run time, with adding power modules for availability, then you can go into higher levels of availability of adding a generator backup. Beyond that is a generator backup with an N+1 architecture or an N+ as many power modules as you want.
“Then there is a two-band system, which is multiple generators feeding multiple UPSs,” he continued. “The highest availability is 2 N+1, a system in which you have two scalable modular UPSs that are redundant that feed down different power paths to the critical load.”
And, availability can be affected by the system design.
“How the systems are wired also supports the higher nines,” Marcoux said. “In some instances, they can have two separate circuits feeding these, or two independent power supplies. That wiring gives you a higher availability. So you may have utility A and a separate utility coming into the building to support it, utility B, a 2 N-type configuration.
“The reality is that in that type configuration, the entire system can be operated on the A or B power system. And that A and B type design is taken all the way from the utility coming into the building all the way down to the independent server, which has a dual-input, corded power supply so it can take an A or B. If there’s a failure in any of the infrastructure above that, from the server on back, it doesn’t impact the customer at all and that’s how the high nines are maintained. The UPS is one of the components within that system. And the UPS is itself internally redundant.”
Installed and monitored
Once a UPS system is installed, maintaining high nines can depend on the practices related to monitoring and maintenance. Installation, servicing, or other changes to a UPS system in some facilities can be done in scheduled maintenance windows if there is no redundancy built into the system.
“At some 7-by-24 facilities that cannot schedule shutdowns, that do not have alternate backup plans, and do not have system redundancy to support system maintenance or revisions, the personnel are required to work these systems hot and to take the necessary safety precautions, a very risky practice that is not conducive to the high nines goal,” Hammond said.
In highly critical environments, the addition of sophisticated monitoring systems can allow for observation of conditions that may not be otherwise obvious to a maintenance crew. These might include UPS battery monitoring systems, run time and loading trend data collection, transformer monitoring, and automated breaker position information.
“Typical electrical equipment installations don’t have those kinds of issues,” Hammond said, “but if you are in charge of monitoring a raised-floor environment, it is a special situation. There’s constant change going on in a data center. There is always new equipment being upgraded and installed. Anytime someone touches a piece of equipment or gets under the raised floor to deal with cabling, there’s a chance of something happening. The quality control protocol may include the practice of limiting unauthorized access to electrical and mechanical equipment, or the raised floor environment itself.”
The human element may be the bottom line in achieving high nines.
“Creating redundancy for databases and servers can help companies but achieving high nines is not about increasing redundancy,” Hammond said. “With a redundant system installed under normal circumstances and without an act of god, it’s pretty bulletproof. Where problems arise is where anyone touches the system. It’s shown in virtually every case where there’s a issue, it’s human error.”
Achieving high nines calls for vigilance throughout. EC
CASEY, author of "Kids Inventing! A Handbook for Young Inventors" and "Women Invent! Two Centuries of Discoveries that have Changed Our World," can be reached at [email protected] or www.susancaseybooks.com.