http://www.rdmag.com/ShowPR~PUBCODE~014~ACCT~1400000100~ISSUE~0904~RELTYPE~CVS~PRODCODE~00000000~PRODLETT~BG.html
Feeling the pinch, the growing computing industry looks for innovative ways to conserve and handle energy. Their efforts could ultimately help other industries.
Paul Livingstone
Attendees at the 2008 Intel Developers Forum examine the Wireless Resonant Energy Link, which can send enough energy over short distances to light a 60 W bulb. Image: Intel
In Spring 2007, residents and businesses in Sydney, Australia, were asked to power down their lights for one hour to call attention to the significance of the consumption of electricity. Called Earth Hour, the event gave environmentalists a rallying cry and actually caused a measurable drop in local electricity usage. The event has quickly gone global, and in April 2009 hundreds of millions of people participated.
Largely a symbolic gesture, Earth Hour nevertheless attracted the efforts of India, which conserved about 1,000 MW in Dehli, and the Philippines, which saved more than 600 MW. In most parts of the world, however, electricity use goes far beyond the light bulb. Electrical devices are proliferating as the microprocessor pervades our lifestyles. As a result, the computing industry is feeling the pressure to reduce energy consumption. Phenomenal growth in both the high-end server markets and the consumer-level markets (the age of sub-$100 laptops will soon be upon us), and has prompted R&D into the ways we use and consume power. An entire sub-industry of "green" data solutions has grown to include big players like HP, IBM, Google, and Microsoft, which have revamped the design and operation of their sprawling data centers.
At Intel Corp., Santa Clara, Calif., researchers are looking for ways to efficiently supply power to computer chips. It's not just the chip that attracts attention now; customers are thinking of energy use in the value equation. Users are no longer worried as much about the speed of their computer as they are about other features, such as power use and, in the case of laptops, weight and complexity. As laptop use grows, and with the proliferation of mobile Internet devices, researchers are finding out that a lot of the ways we power these devices are both wasteful and antiquated.
"On a percentage basis, information technology represents about 2% of the total power demand in the world. Most of telecom or ITT, in fact, represents a small percentage of the U.S. domestic energy picture, and this is probably a high-water mark," says Justin Rattner, vice president and chief technology officer at Intel Corp. "But just because it's 2% is no reason to ignore or minimize the role we have in energy use."
It's not just about conserving energy, says Rattner. Computers will have the unique ability to impact the entirety of energy management, offering an opportunity to exert a beneficial influence on the other 98% of the power consumed. To this end, Intel is pursuing an aggressive blue-sky research program into a variety of energy-conservation tools that could help both the bottom-line of the company and the bottom-line of its users.
Cutting Out Electricity's Middlemen
Adapted from technology pioneered at MIT, the functional prototype for Intel's Wireless Resonant Energy Link was demonstrated at a recent event in New York City in 2009. Image: Intel
In its approach to power management in personal computing, Intel takes its cue from Toyota Motor Corp., which is at the head of a successful wave of battery-electric automotive power systems. Instead of treating certain systems, such as the brakes and the electrical systems, as separate entities, they are treated as a contiguous, interoperable unit.
The same thinking has colored Intel's view of computers. The chip might be efficient, but what about the power supply? The data bus? The hard drive? The high-RPM DVD drive is rare, for example, but it's always a drag on the system.
According to Rattner, electrical inefficiencies in computing are rampant. For example, a computer is designed to request short burst of 30 W of power or more. However, a typical machine consumes less than 10 W most of the time, leading to excessive power consumption.
"If you look at where the losses occur, they are clearly losses between the point source and what I would call the A/C outlet or wall power. Half the power gets lost between generation and the wall socket, and a lot of it has to do with the way the grid is operated and its efficiency," he says.
There's another 50% loss from the wall to the pins on the chip and that's due to the primary AC/DC power supply, Rattner says. One or more stations of voltage regulation are required to achieve correct levels for the computer. These conversions occur at fairly high current, and each one only retains about 80 to 85% of the power.
"There's a surprising if not frightening amount of power that gets lost to heat," says Rattner, and chip also loses energy to heat.
Hybrid Power Platform
Lithium-ion batteries offer reasonable storage and high charge/discharge rates, but they are heavy, don't last a full work day, and the cathode degrades over time. Until a new type arrives, the trick is making batteries last longer.
"We looked at Toyota's hybrid cars and how they do stuff and looked at them for inspiration," says Chris Wilkerson, an Intel researcher involved in the Hybrid Power Platform initiative, a project that aims to increase the mobility of portable Internet-ready devices like laptops.
"The idea is to integrate ultracapacitors to power delivery system," he says. The first idea is the battery hot swap feature, which will mitigate the downside of a laptop's sleep mode: "Right now, when a laptop goes into hibernation so the battery can be switched, the system writes everything onto a disk and writes it back when it needs it again."
This unnecessarily eats time and energy. Normally, compact capacitors can store about 1 µF. The ones Intel plans to use may range from a quarter Farad up to 2 F, enough to supply about 5-10 W for a second. This will not only streamline the hibernation mode, it will supply enough power to keep the laptop running while a new battery is placed in the computer.
"Because the ultracapacitor is a secondary power source and also a battery, we can do a better job of managing how we get power from the battery. It provides us an opportunity to spend battery life more efficiently," says Wilkerson.
For example, lithium-ion batteries deployed in laptops are designed to supply very high outputs—up to 50 W—but the computer rarely needs that. Rare power peaks hurt battery efficiency and 10-15 W is typically all that the computer needs. Instead of expecting the battery to supply high-demand power, an ultracapacitor can fill that role, and perhaps pave the way for methanol fuel cells, which do not have enough discharge capacity to be a sole power source.
Part and parcel of this system will be tracking software that allows users to monitor their power consumption metrics and provide feedback via an interface.
Wireless Resonant Energy Link
Intel researchers are now able to identify and harness sources of power for consumer electronics devices, making them less reliant on the electrical grid, and also recycling that energy to power smaller devices. Image: Intel
At one end of the table is a large piece equipment that generates electromagnetic energy, attached to a copper coil that looks like a 1930s radio antenna. At the other end of the table is another copper coil attached to a 60 W light bulb. When the magnet is powered up, the light bulb shines brightly. There are no electrical cords to be seen.
It's called the Wireless Resonant Energy Link (WREL), and it's Intel's take on medium-range induction technology pioneered in Massachusetts Institute of Technology laboratories in 2006-2007. Intel researchers say this might free us all from the wall socket.
"We're excited the possibility of cutting the last cord," says Josh Smith, an Intel researcher who has been very involved in the WREL project, which began by looking at multiple forms of wireless energy transmission, including radio propagation and ultra-high frequency RFID.
Intel researchers first built a system called Wireless Ambient Radio Power (WARP) that was intended to supply long-distance energy to small devices.
"Imagine a TV tower with a fixed amount of energy. The further away you are the same amount of energy is spread out over a geometrically larger area. The energy drops off as a square of the distance," says Smith. Despite the rapid decrease in available power, the transmitter can send enough—50 µW—to fuel small devices, such as a hygrometer, at a distance of 4 km. This is obviously not enough to power most consumer electronics, but according to Manny Vara, technology evangelism director at Intel, a lot of useful devices don't need much power.
"We know that you can trickle-charge something from the environment in multiple ways. In and of themselves they might not be enough to power something useful like a future cellphone, but if we combined it with other energy harvesting technologies we might be able to do it," says Vara, referring to solar cells and harvesting prototypes that pull energy from body heat, motion, or friction.
"We might be able to give you enough electricity from your surrounding that you would need to only plug in your phone every few days. Eventually, you might never need to plug it in," says Vara.
Intel also investigated another approach called WISP, or Wireless Identification and Sensing Platform. This is essentially a long-range ultra-high frequency RFID system. But again, it does not supply enough wattage to support power-hungry devices.
At the other extreme is inductive coupling, which has reared its head in the consumer sector. Think electric toothburshes. Wireless gadgets on display at electronics shows rely on this system, but it has a very short range, almost a contact.
WREL relies on a similar principle, but the range can measured in feet or tens of feet. The device works from the notion that coupled resonant objects are able to exchange energy efficiently. The wine glass shattered by a singer's single note is an excellent example of acoustic resonance, but the WREL depends instead on the oscillations of electromagnetic waves sent between copper coils a short distance apart. The receiving coil lies within a non-radiative field generated by the other coil, which is attached to a power source that produces about 7 MHz waves. Because most of the power picked up by the receiving coil remains bound to the vicinity of the sending unit, little energy is lost.
"When you think in terms of the wavelengths compared to the structure you are looking at, they are quite large. The receiver is seeing almost the same piece of wave as the sending unit," say Smith.
The advantage that Intel sees is that power is reasonable—tens of Watts—and the distance is adequate for use in offices or homes. For example, researchers have envisioned installing a power supply coil in a room's floor or flat-screen television, then installing small coils in devices—they could receive power from anywhere in the room. Potentially, materials other than expensive copper could be used.
"What we're doing now is trying to optimize the efficiency as high as possible, figuring out wh
Compared to conventional wall socket power, says Rattner, that's not a bad efficiency. "No question, wireless energy is a very exciting development. I would say within the next four to five years it wouldn't be a surprise to see these technologies start arriving in the home." he says.at power level we can do this at and meet all relevant guidelines and regulations," says Smith.
—
0 件のコメント:
コメントを投稿