Charged Up and Spinning the Power
The flywheels-versus-batteries debate continues uninterrupted, along with demand for “bigger and better” backup.
By David Engle
Uninterruptible Power Supplies (UPSes), already dominant fixtures at information technology (IT) data centers, are being reconfigured for even larger roles. All do pretty much the same two things, namely, provide clean power protection to critical loads, and “bridge power” in the event of a utility failure. The latter task means giving instant, seamless energy backup whenever the grid current falters, enough to keep critical systems running until standby gensets kick in. All do this—yet differ radically in how they store energy.
In the UPS world, there’s a well-known competitive dichotomy between flywheel power—in which kinetic energy is present in a spinning rotor—and chemically charged batteries. Each has its relative or perceived advantages and disadvantages.
|Photo: Emerson Liebert
Onsite storage cabinet with a battery recharger and shelves housing spares close at hand.
A modular, three-phase UPS boasting efficiencies of up to 93.4%; a high 0.9 output power factor; and lightweight, compact, “smart” and “green” design.
|Photo: Emerson Liebert
Battery spares are basically storage cabinets that load fresh, or recharged, batteries close at hand.
A decade ago or so, there came a major innovation on the turning-wheel side of the business: the former relatively ponderous, relatively low-rpm (revolutions per minute), seated rotor type systems were challenged by new, frictionless maglev flywheels. The latter brought an end to bearings and gave extremely high rpm, for stronger power at a lighter weight. The achieved results were: (a) much longer product lifecycles; (b) lower total cost of ownership (saving about $10,000 per flywheel every few years in maintenance costs, according to one flywheel maker), achieved in part by eliminating possible bearing failure; (c) energy conversion efficiency of about 98%, much better than batteries; and (d) vastly reduced space requirements than rooms full of batteries.
Another major advantage touted by flywheels vendors is their superior response to frequent minuscule outages. According to the Electric Power Research Institute, about 80% of power disturbances last less than two seconds, and 98% percent last less than 10. Flywheels easily handle these many brief snags; in contrast, batteries suffer more wear and tear from repeated charge-discharge cycles. They also tend to need much more monitoring and maintenance.
But the knock on flywheels, at least in some minds, is the exceedingly short power “bridge” they provide—typically, a mere 15 seconds, or a bit more if chained in parallel. Nearby, an idle diesel must spring to action and take charge in that short time. Automatic transfer switches, if present, must also first detect the utility power outage severity and decide to engage—if these and other elements work correctly.
“Nowadays,” says Lee Higgins of Active Power (Austin, TX), “most generators will fire up in about seven to eight seconds.” It should be noted that Active Power was the first vendor to develop flywheel energy for commercial applications, and the first to win a patent for tying a flywheel to a UPS.
So, it’s 15 seconds—or else. What if something goes amiss and generators fails to engage in time?
This dire prospect has preserved the market’s appetite for batteries, “just in case,” despite evident flywheel efficiencies. In reply, flywheel maker Active Power answers in a way that typifies the competitive back-and-forth between rival technologies. In this case, a few years ago Active Power added a new feature to its flywheel, called GenSTART, “as part of a broader solution” to UPS backup reliability, explains Higgins. Essentially, GenSTART takes the turning flywheel generator energy and turns it into a secondary “ignition spark,” of sorts, “to jumpstart the standby diesel.
“A lot of times,” he continues, “one of the main points of failure for standby diesel generator is the starting battery that’s attached to it.”
This reportedly weakest link often results from lack of adequate battery condition monitoring. In any case, GenSTART, powered by the spinning wheel armature, “will actually help provide cold cranking amps to go ahead and fire up that diesel in the incident,” says Higgins. GenSTART thus isn’t a battery replacement but a backup redundancy: even if the battery should fail, the rotor won’t.
Apart from the tussle between storage media, the decisive and pervasive trend in UPSes these days is clearly towards bigness. This is largely driven by explosive IT industry growth occurring in both mobile Web and cloud computing. As for the latter, enterprises are shifting huge processing loads off the desktop and onto remote IT servers via the Internet. Expansion in both numbers and size of data centers has thus been remarkable. Today, notes Higgins, smaller centers are being supplanted by medium or large ones, and even with supersized “extreme scale IT, for the Yahoos of the world,” he says.
All of this is a powerful stimulant on innovative UPS designs, functionality, features, engineering, and the pace of product development. The following brief survey of introductions highlights some examples.
Battery UPSes: Keeping Pace by Increasing Voltage
Not to be outdone by their flywheel rivals, makers of battery-equipped UPSes are also meeting the expansive market demand. Troy Miller is the manager of business development for the power quality products division of S&C Electric Company, a century-old supplier to utilities. He too notes the dramatic mega-sizing of data centers, which, he says, “are getting more and more power-dependent and much larger, as we start to have proliferation of smart phones and all the transactions being done and stored on the Internet. The sheer size and scale of these things has grown tremendously.”
For a comparison, 10 years ago the phrase “large data center” meant something like 2 MW; today, S&C commonly services bid proposals in multiples of that, often up to ten-fold, and occasionally more.
Besides bigness in scale and power load, what’s also emerging is a trend toward loading more functionality in UPSes. Miller explains that, traditionally, UPSes only backed up servers; now though, “When you get larger like this, you can back up the entire facility.”
And this is increasingly very desirable for data security. Data centers demand a critical infrastructure to maintain cool rooms and to power their robust monitoring and control systems; hence, there’s a strong case for UPS backups being applied to these too, says Miller.
The practical way of doing so, though, is to step up the bus to medium voltage—the strategy that S&C now pursues. Using 4,160 up to 25 kV, he says, “you can back up everything. You can have a very, very large UPS running at medium voltage, so you can protect the entire data center . . . as opposed to just the servers, or the ‘white space’ as the IT industry calls it.”
S&C thus offers its PureWave-brand UPS at this level, from 4 MW up to 20 MW, serving sites in this fastest-growing segment of the IT market—although there remains, he says, plenty of overlap for medium voltage systems down in 2–4 MW land.
In expounding for a moment on the medium- versus low-voltage options, he continues, a second key factor to appreciate is: “When you’re at low voltage you require copper to carry the current, which costs more money, using larger wires.”
Also, the scale needed to serve big loads on low voltage adds too much complexity. “As you’re daisy-chaining 30 or 40 of these things together to get a 20-megawatt unit, you have a breaker before each one,” he says. “And each one of those is a connection point that can heat up or fail over time. All of this is eliminated by wiring for medium voltage.”
Still another benefit of stepping up voltage power is the possibility of installing fully redundant master controls and enabling on-the-fly “hot swap” between them. This means having, says Miller, “two complete master control units that are always available, one running and one on standby. So if there’s a problem on one, you transfer over to the other immediately without any interruption.”
On the other hand, installing medium-voltage circuits in data centers and building-wide UPS systems is still relatively new. At present, Miller notes, S&C is the only UPS system-maker able to serve the higher-load 20-MW range. The vast majority of products “still operate in 700-kilovolt-amp chunks that operate at 480 volts,” he says.
On the question of flywheels-versus-batteries, one might naturally wonder if having stacks of batteries in Miller’s huge, facility-sized UPS backup array doesn’t add an intolerable burden in terms of floor space and power storage to pay for and maintain—not to mention huge operational costs in kilowatt-hours.
In reply, he says that in fact, to avoid this the S&C UPS battery system has been rethought as well. Hence, “You don’t have to burn power to keep it running all the time.” He explains: “Ours is actually an offline UPS. For most of the time it is in what we call standby mode. Only if there’s a problem, then it starts to operate. So the efficiency is much greater.”
This contrasts markedly with “the traditional double-conversion UPS that you have underneath your desk. This is the exact same technology as the vast majority of UPSes that are out there. It makes noise all the time. It’s running constantly,” i.e., burning energy but doing no real work.
If that’s the case, how reliably does the “idle” battery activate and engage? In reply, Miller cites the specific electrical engineering standards for fast-transfer battery power that it meets (IEC 6240-3 class rating).
“Basically,” he says “we have to make the decision that there is a problem, then switch over to battery within two to four milliseconds. And that happens every time.”
Flywheel, Batteries—Or Hybrid?
The either-or issue over storage also recently spurred a breakthrough compromise.
In August 2011 flywheel-maker Vycon introduced its marriage of the two technologies, called the VDC XEB Energy Storage System. Thus, users who liked flywheels for the high efficiency, etc., but needed longer backup times (in umpteen minutes, depending) can have both.
In this combination, the maglev flywheel actually fits especially well by serving as the first line in the backup sequence: in that role, it effectively shields the batteries from suffering excessive wear from the numerous tiny but draining interruptions noted above. As for it operation: The VDC XEB wheel connects to a direct current (DC) bus, receiving charging current to provides clean DC power during discharge, and it runs at about 99.4% efficiency. Upon detecting an outage lasting more than a few seconds, the VDC XEB “gently” transfers to integrated backup generators. The presence of batteries gives another layer of time insurance for sites with a sequence of things happening, such as automatic transfer-switch time delays, or synchronization requirements for multiple generators. The tight coordination of so many elements in a 15-second timeframe might otherwise go awry.
For added runtime and redundancy, Vycon’s hybrid can be paralleled, up to four. It’s compatible with other major three-phase UPSs brands.
Patents cover a flywheel hub that is made of aerospace-grade steel, a high-speed permanent magnet motor generator, a contact-free magnetic levitation system, and a touch-screen performance status display. Real-time monitoring software handles diagnostics, log files, adjustable voltage, RS-232/485 interface, alarm status contacts, soft-start pre-charge from the DC bus, push-button shutdown, DC disconnect, remote monitoring, and Modbus and SNMP communications.
Hybrid UPS technology enables the flywheel to charge and discharge “at very high rates for countless cycles without degradation throughout its 20-year life,” the flywheel maker claims.
Capstone’s Hybrid UPS/Microturbine
Vycon’s dualistic solution is neither the first nor the only multi-storage system or hybrid UPS; rather, it follows close on the heels of a unit from Capstone Turbine Corporation in Chatsworth, CA. In this dramatically innovative product, backup redundancy comes neither from batteries nor flywheels, but from a spinning turbine. Capstone’s dual-conversion UPS was introduced in 2008 and gained its first adoption a year later.
The turbine power concept here boasts far superior efficiency to the conventional two rivals, because the backup power is not spinning or sitting idly, but, rather, is typically trigenerating.
|Photo: Active Power
Active Power CleanSource UPS with flywheel rotor.
Capstone vice president Jim Crouse explains that “In many situations, a hybrid UPS can be used in a combined-heat-and-power application . . . making it extremely efficient,” since the turbine yields power, cooling, and heating. The latter can work year-round, he adds, either driving an absorption chiller to help meet a data center’s cooling needs in summertime, or providing winter heat.
In this, it is also flexible: According to a particular site’s need, it can work in any of three modes: as a standard UPS (more-or-less), a high-efficiency cogenerating plant, or an emergency backup generator like a conventional diesel (hence, replacing what is normally a virtually unused genset with a very productive, working asset).
Capstone’s hybrid UPS complements its UpSource microturbine plant, aimed at providing straight-out prime power for data centers; in this role, Capstone claims that it requires less maintenance and yields lower cost of ownership than traditional battery-backed UPSes.
Along with the above UPS innovations, several other notable ones have come forth recently; the following list, based on announcements, is by no means exhaustive, and doesn’t include countless ancillary elements being introduced in backed-up power trains.
Liebert’s NXL, and Spare Battery Concept
Emerson Network Power, the UPS branch of Emerson company, has eight UPS-related business divisions, one of which is the well-known Liebert brand. Liebert has recently introduced its newest and largest UPS, called the NXL—reflective, again, of bigness in market demand. And a second innovation this year: “Liebert Battery Spares.” Basically, it’s a storage cabinet to be hold fresh batteries or recharge old ones, close at hand.
There are other changes and modifications afoot as well. First is the press for greater efficiency, especially given the ever-increasing power output. For years, the standard in a double-conversion UPS was the acceptable loss of about 4–6% of its energy in the conversion. When UPSes were modestly sized, this was considered a reasonable price to pay for optimized protection.
However, as UPSes consume ever-more energy, losses from even “normal” inefficiencies mount. Several correctives are now being applied. Firstly, in the latest models, even if optimized for availability, they can still match the efficiency of earlier systems by applying better controls. And secondly, in cases where a load may not be quite as critical, or the utility power is thought to be quite reliable, the 4–6% sacrifice can simply be turned off, in favor of peak efficiency.
Next, in terms of optimizing availability regardless of efficiency losses, tweaking the system configurations can improve this. For example, Tier IV data centers (as defined in Uptime Institute standards) generally use a minimum of 2 (N + 1) systems, supporting a dual bus to eliminate single points of failure, system-wide. This entails using two or more independent UPSes, each able to bear the full load. Each provides power to its own network. This peak-availability strategy enables 100% concurrent maintenance IT equipment to be positioned as close to the input terminals as possible. However, it may of course suffer a bit on UPS efficiency at low loads. Scalability also becomes more complicated.
Author's bio: Writer David Engle specializes in energy-related topics.