Power disturbances cost the US industry tens of billions of dollars yearly in lost data, material, and productivity. Recent global events affecting power supplies have demonstrated more than ever the need for facilities depending on mission-critical loads to be able to be up and running as soon as possible. To that end, a host of facilities are updating their UPS systems with the latest technologies. Choices abound and opinions differ as to which systems excel: battery, flywheel, or microturbines.
Hybrid UPS
In December 2009, Syracuse University opened its Green Data Center in a 12,000-square foot space in response to an increased demand for greater computing capabilities and data storage. The facility is designed to use 50% less energy and produce fewer greenhouse gasses than traditional data centers and incorporates an integration of technologies that had not previously been used together in a data center. The foundation: 12 Capstone Hybrid UPS Microturbines, creating the first onsite power system to integrate C65 (65-kW) microturbines directly with a dual-conversion UPS to provide power for mission-critical loads.
The low-emission natural gas microturbines are the center of a trigeneration-combined cooling, heating, and power (CCHP)-system that further boosts the data center’s energy efficiency. BHP Energy, a Capstone distributor, integrated the design of the CCHP system so the microturbines produce electricity and supply heat and cooling power to the data center and a nearby building.
Traditional data centers rely on power from the utility and have banks of batteries keeping servers and equipment running during a short power loss, with a standby emergency diesel generator typically used for longer outages.
“The way we do it is power comes out of the generator, which is directly connected to the turbine, and we take the AC [alternating current] power and convert it from AC to DC [direct current] and then DC to AC,” says Jim Crouse, executive vice president for Capstone. “The power electronics we use to do that are similar to the power electronics you’d find in a commercial UPS system in any of the manufacturers-we use the same IBGT technology. Because of that, we were able to add another power module to our standard microturbine design. The new product is what we call a hybrid UPS, which is a fully functioning UPS system with a microturbine built in.”
At Syracuse, the Capstone microturbines enable the data center to be isolated from the utility, yet draw on it for backup power. A 40-ton EnerSys battery bank provides sufficient power to carry the maximum load for 176 minutes in unusual catastrophic situations. IBM, a project partner, provided $5 million in design services, support, and equipment, such as cooling doors utilizing chilled water to cool each server rack independent of its neighbors, thus reducing cooling and energy costs.
Two onsite 150-ton Thermax absorption chillers convert exhaust heat from the microturbines into energy to chill water used to cool the racks and the entire building. A Cain Industries heat exchanger can be used at the same time as the chiller to produce hot water to heat peripheral areas of the data center and the building next door. The system also employs free cooling when outside temperatures are low. A heat exchanger produces chilled water directly from a rooftop-cooling tower. IBM also provided computer equipment that operates from a DC-powered distribution system, eliminating traditional power loss associated when converting AC electricity from the utility to DC to power servers. The high-voltage DC equipment, including ACDC rectifiers, is from Validus DC Systems.
The hybrid system “provides the power conditioning and power quality that a data center or a power-quality-need customer may have, while at the same time allowing them to operate cogeneration for peak shaving or use it as backup power, if that’s what makes the most sense for their installation,” says Crouse. “The goal was to have a UPS that was equal to the UPS systems in today’s marketplace with the flexibility of power generation built into it.”
The combination of the various technologies is a first, says Crouse, adding that Capstone company officials believed the economy was “ripe to commercially develop and move the product to the marketplace.”
Syracuse University officials favored its operational flexibility, Crouse notes. The university has onsite fuel storage so if the natural gas supply were interrupted for any reason, university officials can operate the microturbines.
“It’s important for a data center to have available onsite fuel storage in case of interruptions the same way you’d get an interruption if someone were to run into a telephone pole,” he points out. “If something happened to the natural gas supply, you’d need to have a backup onsite fuel source. We’re able to operate the microturbine both on natural gas under emergency and switch over to propane fuel.”
Maintenance for the hybrid UPS is similar to Capstone’s standard products, Crouse says. “The batteries are external to the microturbine, and they would be serviced like any other typical UPS battery would be serviced. The microturbine package is identical to a standard microturbine package except that it has one additional power electronic module.”
Data centers are the largest market for this technology, Crouse says. A Homeland Security building also is utilizing it in conjunction with other power quality technologies.
“It’s really for any application where the customer needs power conditioning and power reliability,” he adds.
The system’s footprint sets it apart from others, Crouse says. “One of the significant differences is our system is designed to be installed outdoors. The footprint is similar for a traditional UPS, but our system can be installed in a parking garage, in a basement or outside instead of taking up very expensive data room space, which is different from a more traditional UPS that needs to be installed indoors.”
Facility managers contemplating the installation of UPS are those seeking a product that is equal or better in reliability. Typical questions they ask, says Crouse, include: “They want to start evaluating the economics-can we do cogen? Does it make more sense to do peak shaving? What are the economic and environmental benefits and efficiency?”
Another benefit is that if Capstone installs the system in a configuration of a central battery bank similar to what was done at Syracuse University, “it allows the customer to use the DC bus in addition to the AC bus and to run direct DC loads within the data center, which is an area of growing interest,” says Crouse.
Flywheel UPS
A flywheel power system stores kinetic energy in a rotating mass, immediately converting it to DC power. It connects to the DC bus of a UPS, charging from the bus and returning power when the bus voltage drops below a threshold level. Information Technology (IT) infrastructure services provider Terremark chooses flywheels for it company’s backup power at all of its sites.
One of the biggest reasons: “It takes up a lot less space because you don’t have to build another room for battery strings,” says Ben Stewart, Terremark’s facilities engineering senior vice president.
“I can go down a list of reasons we don’t use batteries,” he adds. “First of all, the batteries take up space. We’re in the colocation business, so we build large data centers for people to relocate their data centers into. If we’re using up space for our infrastructure, that’s space we can’t sell to our customer.”
Having to condition rooms to 77 degrees for batteries and keep them well ventilated consumes energy, Stewart says. “Data centers are very big on energy efficiency, so by getting rid of the batteries, you get away from the energy costs of having to condition the rooms. The flywheels do not have to be conditioned at all.
“Batteries last three to five years and each battery string is about $100,000,” he continues. “Flywheels can go 20 years without having to do anything more than replacing a bearing once in a while, so the total cost of ownership is quite a bit less.”
Stewart goes on to say batteries need to be monitored on an ongoing basis, “because when one cell goes, you can lose the whole string, so there’s a lot of ongoing maintenance associated with batteries. There is no such maintenance associated with a flywheel.”
In its newer installations, Terremark is using Active Power technology. While the company used to use rotary UPS systems from a Dutch company, it is now using static UPS with a flywheel.
“It’s just like a battery UPS, but we take the batteries off and use the flywheel to spin little DC generators to produce the battery power that then goes to the IBGT inverters in the UPS to produce the AC power,” says Stewart.
Terremark is getting 15 to 20 seconds of backup with its flywheel system, Stewart says. “The only thing you need UPS for is to ride through while your generator’s getting started and stabilized. Today’s standby generators have to come up within 10 seconds. You don’t need any more than 10 seconds of UPS because by then your generator should be up and running.”
A generator that doesn’t start right away is not going to start in 10 to 12 minutes, explains Stewart. “You’re not really buying yourself any time with the UPS. You really don’t have 10 to 12 minutes of backup time with the battery-based UPS. Today’s data centers have such high power densities that they heat up very, very quickly. When you’re on UPS, all you’re doing is keeping your servers, your storage, and your network devices hot-you’re not turning any of your cooling equipment. After about a minute, most data centers overheat, and you have to shut them down anyway.”
Stewart says the flywheels are rated for seismic activity. “I’d be more worried about batteries in a seismic zone than a flywheel, because in bouncing those batteries up and down, if their one of their cases crack, they leak their acid, the battery leaks all over the other batteries, and you lose those batteries,” he says.
Flywheels are two to four points more efficient than battery-based units, Stewart says. “There are some battery static UPS units out there claiming they’re 99% efficient, but all they are doing is bypassing utility to the equipment and using a very high-speed static switch to switch their UPS in at the last second. That’s not really an efficient UPS.
“You can do that with a flywheel as well, but using a flywheel as designed and not using the static switch, they’re about 98% efficient-more than any of the other battery ones on the market by two to four points or so,” he adds. “It’s not that dramatic, but it is a little more energy efficient.”
The number of discharges a flywheel can withstand is “infinite,” says Stewart. “With a flywheel, you’re spinning a mass of metal, using that kinetic energy to produce DC power,” he says. “You’re not doing anything to it. With batteries, they only have so many discharge cycles in them, and if you’re in an area with a lot of power disruptions and you’re going through batteries a lot, you’re not going to get three to five years out of that battery string before you replace them.”
In planning a UPS system, the most critical factor is redundancy, he notes. “Any data center has to have a fair amount of redundancy so that if something does happen like a generator doesn’t start or a UPS fails, you’ve got enough redundant components to keep your data center alive.”
“The lower-tier data centers that don’t have a great deal of redundancy might have a hard time justifying the flywheels, because you only have 15 to 20 seconds of backup and maybe if you had a minute or two, you might be able to do something. If you’ve got good redundancy, flywheels are not a concern whatsoever. There’s really no difference.”
VYCON manufactures a third-generation flywheel. “It’s either a direct replacement for lead acid batteries-you can remove the batteries from the application and attach a flywheel to your static UPS system-or they work in parallel with batteries for certain types of applications,” says Dann McKeraghan, vice president of sales.
VYCON’s flywheel design considered previous designs on the market and was developed with an eye to reducing maintenance costs, says McKeraghan. “The amount of downtime to maintain the system is pretty much eliminated and the cost associated with maintenance is very low.”
VYCON has removed mechanical bearings from its flywheel design. Those that use bearings require about eight hours of downtime when they have to be changed every 24 to 36 months, says McKeraghan. “Customers don’t want that downtime built into something they buy,” he notes.
The design also incorporates an increase in power density; the system can provide up to 300 kW of power. “Our flywheels can be in parallel capacity and have additional runtime, as well, when attached to UPS systems,” says McKeraghan. “It’s a flexibility for customers who have to buy a large system from day one because that’s where they’re going to grow.”
Case in point: colocations facilities.
“They can buy just the amount of storage they need via the flywheel and can add additional flywheels later if they need more runtime or capacity,” says McKeraghan. “Batteries are a little more tricky when you want to add new batteries to old battery strings. It’s pretty much a “˜no-no’ unless you’re absolutely forced into it because the new batteries tend to age quicker and look like the old batteries within 18 months.”
Flywheels offer a smaller footprint, McKeraghan says. “It depends on the UPS size, but it’s typically anywhere from a third to a half amount of the space the big UPS system. It may take up 15 to 20% of the space of a large wet cell battery plant, for example. In general, flywheels require less space.”
The biggest demand for the systems is any mission-critical operation that cannot have interruptions, McKeraghan says. Healthcare is a large piece of the company’s business, as are traditional data centers and pharmaceutical and industrial manufacturing operations.
“The device can go into a harsher environment than a battery can; it can be hotter or colder and not affect the operation of the flywheel,” he says, referencing the 77°F temperatures needed to maintain batteries.
Casinos are another large market for the systems. The UPS and flywheel systems are being used to protect the slot machines.
“That’s a large revenue stream for those companies and it’s important to keep them online as well,” says McKeraghan.
The VYCON flywheel offers about 20 to 25 seconds of runtime, McKeraghan says. “The intent is to provide enough runtime for the customer’s generator to come up, get online, and become the input source to the UPS after a power outage. In one respect, it’s a bridge to the generator. The customer needs to have a generator that’s maintained and functional that will last probably five to 10 years. The automatic transfer systems that have been associated with them have become extremely reliable. That has made this 20- to 25-second bridge technology to the generator appealing to a lot of the customers.”
The upfront costs of a flywheel can average 20% more than a typical battery, says McKeraghan. “But on the return on investment [ROI], the cost of ownership is about a three-and-a-half to four-year period going forward,” he adds. “You”˜re not replacing flywheels like you would batteries.”
Many VYCON customers consider the system a “green” technology because they’re not having to deal with lead-acid in their workplace, says McKeraghan.
“The elimination of ventilation and eyewash stations are another aspect of the benefits,” he adds.
Maintenance is another issue VYCON’s end-users consider, says McKeraghan. “Batteries are very maintenance-intense. At the very minimum, most customers will do four preventative maintenances a year on batteries to check connections, and there are hundreds of connections. Flywheels take about a 15-minute maintenance window with no downtime, so the reduced maintenance aspect related to the flywheel is a big consideration for most customers looking at this technology.
“You don’t replace it like you do a battery every four years, and it takes up a smaller footprint and frees up an area for customers to use that space for additional revenue stream by adding computer equipment or servers.”
Beyond Batteries
Six years ago, a battery failed on one of the battery strings of the UPS system at Beth Deaconess Medical Center in Boston, MA. Computers and switches had to be brought back online in sequence over several hours, stalling business operations.
Following that incident, it became apparent to hospital management that its two data centers needed a reliable and sustainable backup power system. The 550-bed adult medical-surgical teaching hospital is affiliated with Harvard Medical School and services 250,000 patients and their families annually. The hospital is a proponent of cutting-edge green technology with a goal to increase green purchasing 15% by fiscal year 2012.
Plant Manager Ty Dell designed a 7,000 square-foot Renaissance Center Data Center 1.5 miles from the hospital campus. A smaller data center (3,000 square feet) is housed at the main campus. He chose two sets of VYCON VDC clean energy flywheel systems to back up two 225-kVA UPS systems based on a desire to reduce the data centers’ carbon and physical footprint and save on maintenance costs.
Each flywheel is rated at 275 kVA for 6.4 seconds of runtime. With a 400-kW computer load, the hospital gets 30 seconds of flywheel runtime while waiting for the generator to start up during a power outage. The data centers support all of the vital clinical systems that need computers to schedule and serve patients, including medical records files, as well as functions such as the pharmacy, laboratory, chemotherapy suite, and operating rooms.
The hospital’s two data centers have reduced utility consumption by consolidating servers, replacing older, less energy-efficient devices, and reconfiguring equipment despite growing electronic demand. The data centers also are improving thermal management by reducing under-floor cables, using perforated inserts, shutting down an unneeded air-conditioning unit, and making other adjustments as needed.
The task of the Washington Department of Ecology is to protect, preserve, and enhance the state’s environment, while promoting proper air, land, and water management. The department’s services include enforcement actions to obtain compliance with environmental laws and regulations when voluntary compliance cannot be achieved, assessment of environmental conditions and trends, education efforts such as classroom curriculum guides and teacher workshops, permitting, environmental review, site cleanup, and spill response.
The Washington Department of Ecology depends on its data center’s critical servers and systems to remain running at all times in order to make available on an ongoing basis the environmental databases and monitoring systems. The Washington Department of Ecology’s aging UPS had been maxed out to capacity, and a solution was needed that would address the need for the latest technology, provide an ability to expand the data center as equipment needs grew, and offer a high level of availability. The system also had to have a small footprint to fit into the same equipment room with battery banks and cooling units. After putting specifications out to bid, the system that was chosen was an Eaton 9395 UPS, a 225-kVA unit. It is one of Eaton’s largest systems.
The UPS is available in models ranging from 225 to 1,100 kVA and operates at 99% efficiency with an Energy Saver Mode. The double-conversion topology isolates output power from all input power anomalies and delivers completely conditioned sine wave output during the most severe power disturbances. Another feature is the Variable Module Management System, which uses techniques similar to virtualization to only bring online the amount of power within the UPS needed to support the data center efficiently. It also can be configured with an inherent redundancy option. An Eaton field technician can isolate and service a module while the other carries the load with no need to go to bypass. Its scalable architecture enables the department to adapt to future changes in load requirements without requiring the purchase of a new UPS. For example, a 275-kVA module can be added to achieve N+1 redundancy or additional capacity.
Ed Spears, product marketing manger for Eaton’s Critical Power Solutions Division says the company is noting several trends in UPS systems, particularly the data center market. “One of the obvious ones is the emphasis on green, efficient, and sustainable practices and sustainable materials in the data center,” says Spears. “We make sure that we have products that offer things like smart certification for LEED [Leadership in Energy & Environmental Design] compliance, because a lot of the red tape in obtaining LEED points toward the gold or platinum certification for a facility.
“We’ve seen changes in the UPS system in the last five years, particularly to markedly improve the machine’s efficiency to make sure we’re not contributing unnecessarily to the extreme power costs” he adds. “To operate data centers today, we need to be part of the solution in reducing power costs, and not part of the increase or the waste of the power that occurs quite dramatically in most data centers.”
Spears notes that some 40% of the power that comes into the data center gets thrown off as heat. “We see data centers utilizing air-side economization, where they’re using as much outdoor air as possible without air-conditioning,” he says. “That means that the UPS needs to produce a minimum of heat on its own in its operation to make sure that we don’t cause the need to add extra costs in air-conditioning and cooling capability to a data center.”
Another trend is occurring in the type of data center end-user. “In the past, where we might have supported a lot of small- to mid-sized data centers with small- to mid-sized UPS systems and those data centers are under the ownership of the end-user client, now we find ourselves supporting larger data centers that are colocation centers where they are leasing out power and cooling capacity along with physical space and networking to a whole range of clients,” he says.
In that case, reliability is the ultimate concern, Spears says. “It’s good to be efficient, but it’s also good to make sure if you’re a colocation operator signing a service level agreement that basically dictates that power and cooling must be available with a high degree of reliability, or you’ll be assessed a huge cost as the owner of the facility that you have reliability,” he says.
While in the past, facilities may have had one large UPS or one large air-conditioning system, today’s concerns are focused on having plenty of capacity no matter how much or how fast growth occurs, Spears says. “People are saying, “˜I need a UPS that meets my need on day one and then is scalable and modular to grow with me as I grow and when I grow, without disturbing the operations in the data center,'” he adds.
The scalable, modular gives more flexibility and cuts upfront and installation costs, Spears says. In a cloud infrastructure, there is an emphasis on how the UPS system is deployed. “Where we used to put in multiple module highly redundant UPS systems, some of our customers are saying they’ve got a data center here, but they have another data center that is a mirror-type arrangement or a disaster-recovery type data center, and if they have a severe problem in the first data center, they don’t really need redundant UPS systems because they’re going to switch all of the processing to the second data center,” says Spears. “It becomes less necessary to have redundant UPS systems.”
Criticality is a dominant concern these days as well. Case in point are data centers that support such websites as Amazon.com where purchases are being made and “they need to go to a super secure data center which has a lot more reliable infrastructure and backup than the normal data center for browsing their site,” says Spears. “That changes the way we proportion and provide UPS systems there.”
Virtualization is another trend Spears sees. “It’s saving tons of energy and resources by virtualizing servers and storage where possible, but that means the critical applications being performed in the data center are constantly on the move from one piece of hardware to another or they’re consolidating many critical processes on the same server,” he says.
“We’re starting to see the need to be able to do that kind of thing with our power,” he adds. “If I’ve got a data center where 20% of the servers are idle at any point in time because I’ve virtualized onto other servers, it would be nice, if there was a power outage, to immediately order a shut-down of the idle servers and concentrate all of the remaining UPS and battery power on the remaining servers that are functioning. That means we have to know where the processing is happening in the data center so that we can direct our power to those pieces of hardware that are actually critical at that point in time.”
Eaton works with companies like Microsoft and VMware to be able to be part of the virtualization efforts, “so we don’t end up leaving a bunch of unused power on the table,” says Spears. “We’re moving power away from or shutting down idle servers. We’re concentrating that power on active servers. We’re arranging the power to support the more critical functions during a power event like a power failure or a thunderstorm, and we’re doing workload balancing, not just at the processing level, but at the power and cooling and infrastructure level. The idea is we will waste less by concentrating power where it’s needed.”
The former attitude was that reliability and availability trumps all, Spears says. Today, data center operators are being pressured to operate their centers more efficiently than in the past.
“That pressure translates all the way down to the server equipment vendor, the IT equipment vendor, and the power vendor to make sure that not only are we extremely available and reliable, but we do that in a way that provides better efficiency than we have in the past,” notes Spears.
One of the newest UPS systems to have come on the market comes from Mitsubishi Electric Power Products. The 9900B Series UPS System has a power rating of 750 kVA and is designed to offer large data centers and mission-critical facilities maximum efficiency greater than 96% and a 30% smaller footprint than typical UPS modules with a similar power rating.
The 750-kVA UPS module can be configured in a parallel system of up to eight units using Mitsubishi’s parallel technology. It was designed as such for an eye to system reliability, eliminating any single points of failure associated with other types of UPS systems.
Updates and Expansions
Langan Engineering & Environmental Services in Elmwood Park, NJ, has been experiencing rapid growth in servicing real estate development sectors. In five years, the company grew from 200 to 560 employees and needed to update its data center. Company officials sought to do so in an environmentally responsible manner. Amplifying the problem was the fact that the company’s previous UPS units were at maximum capacity and were continually being overloaded.
The mission: to install a UPS system that would support efforts to reduce data center energy consumption and enable expansion for blade servers and virtualization.
In early 2008, the company chose Emerson Network Power’s Liebert NX with Softscale technology and Liebert Services Remote Service Delivery. The system has provided Langan with a 97% efficiency level and has decreased downtime by 20%. The one single unit has replaced seven older units.
Additionally, the system can be sized to current requirements and scaled with a software key as needs change. Modules with different ratings can work in parallel to handle higher capacities or offer redundant operation. Start-up with the unit also includes one-year remote service delivery with monthly equipment health reports, alarm retrieval and warehousing, escalation management, and continuous emergency response from Emerson Network Power’s Liebert Services.
Michael Breen, IT Manager for Langan, says after considering other options, he chose the Liebert NX and the Soft Scale technology, “because it gave us the ability to go up in capacity from 40 kilowatts to 80 kilowatts for essentially the same money. I am able to get a lot of potential growth on that. I can’t predict our power usage 10 years from now. All I can assume is it’s going to go up even with virtualization, and we’ve been very successful in decommissioning and virtualizing our more than 100 servers.”
The new system has been a source of relief for the company, where utility power problems include power interruptions and disturbances. To address the environmental initiative, the company installed 14 blade servers and implemented a virtualization strategy to decommission legacy servers and reduce the data center’s energy
consumption.
Breen has planned it so that if one of the company’s remote sites goes down as the result of a weather-related event, the company can easily take the virtual image of the affected servers and get them going again at its disaster recovery site. Three years into the installation, Breen gives the system an “excellent” rating.
“It’s definitely exceeded my expectations,” he says. “The main benefit I see is that it is a totally sophisticated technology that provides me a boatload of information and knowledge in order to keep my data center functioning.
“Ultimately, it’s all about minimizing downtime,” he continues. “We have almost 700 employees in 14 locations. We have central applications such as our Microsoft Exchange e-mail servers and our Microsoft SQL database servers here in Elmwood Park, New Jersey, and when we go down, even if it’s on a Saturday or Sunday because they’re working in Abu-Dhabi on Saturdays, or if we go down late at night in our California offices, it hurts that, too. I’m judged by downtime.”
Several years ago, the $80-million company made a significant investment into its infrastructure in New Jersey, where half of the employees work. More than a half-million dollars was invested into a new Caterpillar 600-kW generator as part of the company’s disaster recovery and business continuity program, which Breen was tasked to create upon being hired. With the purchase of HP C7000 blades, three-phase electric was required.
“You can buy them single-phase, but for energy savings, it’s more efficient and better to do three phase electric,” points out Breen. “The heat load went up the minute I bought the blades and I had to buy an air-conditioner for them.”
The existing units did not provide enough power, so Breen purchased a 12-ton Liebert DS Precision Cooling System and then purchased the Liebert NX with Softscale technology. Disaster planning is critical, Breen points out.
“Bad things can happen, and as much as you want to trust the local utility systems, we wanted to take that control out of their hands,” he says.
Breen tests the system twice a year by switching off power to the automatic power switch with successful results. “The Liebert NX has been able to make that transition because when you switch from alternating current to direct current, which is the batteries, you’re going to have a blip in time where you lose power,” he says.
“I can handle about 55 minutes’ worth of battery time, although it only takes me seven seconds to switch to generator power,” he adds. “That gives me so much flexibility that I don’t have to take down servers, because servers, when they go down, take at least 30 minutes to go down and back up again.”
The decreased downtime has provided a satisfactory ROI, Breen says. “We have an average billing of $100 per hour. With 500 billable employees, we’re looking at almost $50,000 per hour in downtime. Our ROI is measured back to downtime.”
Breen also favors the energy savings he’s getting from the Liebert NX, based on its 97% efficiency rate compared to the previous system’s 93% efficiency rate.
There were internal electrical challenges in installing the system. As the company grew, “the electrician had essentially put every printer, PC, and coffee pot on the electric panel in our server room,” says Breen. “We had to rewire the entire building going back to the data center and put in a new electrical panel in order to isolate the electrical breakers on our fourth floor here. You need isolated electrical panels for your data center, which is common sense.”
Breen says the reliability of the Liebert NX is “outstanding.” Emerson provides quarterly maintenance on the system to ensure the batteries aren’t overheating.
“We’re paying more than $5,000 a year to have them come in, but we made the investment into the Liebert NX and the Precision Power Center and the batteries, and we feel that is money well spent,” he says. “A good return on our investment is to maintain the system and tweak it when we see problems.”
And one did occur during a quarterly inspection. “Kudos to the Liebert technician who notified us that one of our three phases of power was much higher in current than the other two,” says Breen. The problem was traced to a server, and the company was able to rebalance the three phases.
“Three years later going into this we’re still making adjustments because, in a data center, everyone plugs in anywhere they want, and they’re not looking at the big picture of trying to spread that power out,” he adds.
Breen says he’s a “big believer” in maintenance. “A lot of people don’t want to spend the money for continuing maintenance, but you really have to bite the bullet, because your company is riding on this facility and you made an initial investment. My advice is that you’ve got to spend the money to keep the system in peak performance.”
Another feature of the Liebert system that Breen likes is that he gets an e-mail alert when the system goes on battery power. Recently, he received an early-morning Saturday e-mail indicating the data center was on battery power.
“No one knew that the building had lost power,” he says. “I was able to mobilize my electrician to the site, and ultimately, we were able to find the bad electrical feed from the fire, just because of a simple e-mail alert that the Liebert NX was able to deliver to me and start the repair process. We were able to minimize downtime just by a simple e-mail alert and notification; that’s huge to me. Who’s going to be there at 2 a.m. or on a Saturday, Sunday, or holiday?”
Emerson Network Power recently rolled out a new product, the Liebert APM On-Line UPS. The transformerless UPS allows capacity increases with FlexPower hardware assemblies with no additional floor space required and operates at a 94% efficiency at loads of 50 to 100%.
“It uses the similar double-conversion technology that is in NX, but we’ve condensed it and allowed the modularity so we can grow solutions to fit the customers’ needs,” says Gary Anderson, business development manager for AC power products for Emerson Network Power.
The Liebert NX can be used throughout a gamut of different solutions, says Anderson. “It can be used in a network closet or some smaller areas, all the way up to a data center in a row of racks. The APM is really designed to be used in a data center in a row of racks, but grow rack by rack as needed.”
Additionally, the APM is scalable in that end-users can add modules to add capacity. “It’s got a very small footprint with the ability to have on a smaller ratings to have batteries integral to the unit,” says Anderson. “We’re trying to minimize our footprint and minimize that construction cost on a cost-per-square-foot basis.”
In evaluating choices for a UPS system, Ed Spears, product marketing manger for Eaton’s Critical Power Solutions Division, says end-users should consider a life cycle analysis: The process examines the pieces early in the supply chain and factors in how far it’s being shipped for assembly, the amount of energy needed to manufacture and test the large UPS system, and how it compares to previous designs.“Once the product is manufactured and tested, I have to ship what is a fairly heavy piece of equipment, including batteries, to a different location,” says Spears. “Then I’m going to install it, deploy it, and have it operate as a UPS system for as much as 15 to 20 years.”
Other questions he poses: “How much does that cost me in power, and how can you show me as the vendor what you’re saving me that I can drop to my bottom line? At the end of the service life of the UPS, the assumption is that I will not just be simply dropping that in the Dumpster somewhere—I’ll be recycling those parts.
“How recyclable are they?” continues Spears. “It’s certainly true that batteries are one of the most recycled products in the world, which is very good news, but you also have the rest of the UPS.
“For example, we make transformerless UPS systems, which saves us a huge amount in terms of weight and footprint,” he adds. “But, it also means we’re cutting about 60 to 80% the amount of copper and iron and varnish, and those are the kinds of things we’re using inside our power electronics. That’s a big change for us over the last five to 10 years.”
Facilities have to stay on top of electrical needs, states Breen. “You’re going to have to ask them to make the run a little bit longer to another electrical panel and keep it out of the data center,” he says.
Gary Anderson, business development manager for AC power products for Emerson Network Power, notes that rapidly changing technologies in that market mandate backup power such as UPS systems. “We’ve also see a number of hospitals requiring UPS, especially as everything becomes digitized in the hospitals,” he says. “You need to have power there all of the time, especially with the equipment. It’s not so much while the equipment is running, but it’s a revenue saver for hospitals because the reboot time on a lot of the equipment is actually quite long.”
Most hospitals have generators that need to be up and running by code in 19 seconds, says Anderson. “But even within that 19 seconds, it will knock out all of the equipment, and it can take a half hour to reboot all of the equipment.”
More facilities that are computerdependent are embracing UPS systems, he points out. “It’s become more of a revenue situation of the risk and the cost of going down.”
In some high-tech industrial processes, such as chip manufacturing, the cost of restarting the process can reach a half million dollars, Anderson says. “It makes a lot of sense to have a UPS so you can avoid that. The UPS also provides some power conditioning to the load, which can minimize fluctuations in manufacturing processes.”
Additionally, there are the hidden costs, such as lost productivity during downtime, Anderson adds.
Still, there are challenges in implementing the technology. “We need to understand what are the loads that are going to be put on this,” says Anderson. “You can’t plug every plug in the building onto the UPS—otherwise it will become much too large, and it will become an efficiency issue. We want to right-size UPS loads to the equipment.
“UPS has an inherent efficiency—92 or 95% in most cases,” he adds. “We want to get the right product for the right need. Once we have that in place, we have to put the correct infrastructure in. There is some footprint. How long do you want it to run? Is it tied to a generator? What kind of redundancy do you need? You can have one UPS, but if the UPS fails, then what happens?”
Additionally, facilities need to have a strategy to service the UPS to make sure that the batteries are going to be active and alive, says Anderson. “Unfortunately, battery technology hasn’t improved significantly over the last 100 years, and most of the UPS batteries that are in use are lead-acid batteries much like what’s in your car,” he says. “The claimed life is 10 years. Most people get maybe five years out of them.”
Redundancy is important in installing UPS systems, Anderson points out. “If a company has a load of 100 units and gets a UPS that provides for 100 units, people want redundancy to have another UPS there as well that can do 100 units. In that sense, both UPS systems are actually sharing a load of 50 units. Some UPS systems are more efficient at the higher loads near 100%; some are more efficient at 50% at half load and are meant to share. We need to have the right unit for the right customer, for the right redundancy level.
“A lot of companies don’t need to have two UPS systems if they are protecting an office building—one is fine,” adds Anderson. “The chances of that one failing is pretty small. If you have a data center, you want to have two UPS systems, because you’re going to feed two different loads in different directions. Then it becomes more of a design concern than how it operates. People just look at the operation, and they make assumptions about what they need upfront when they should really think it through a little bit better.”
As for affordability, different factors come into play with respect to capital expenses and operational expenses, Anderson says. “There are a lot of people who aren’t limited in capital expenditures; they can ask for a lot of money in the first place, but they might be constrained in their operating expense down the road. There are some cases where they are very limited in their capital expenditures, but they can use operating expenses costs a little bit more freely.