If you want to transmit a lot of power over long distances, it is cheaper to use direct current (DC) rather than alternating current (AC). But AC from a generator needs to be converted to DC for transmission and then back again to AC by the time it comes out of the wall socket. That’s the Achilles’ heel of long-distance power transmission. A “switch” that can handle large amounts of power is a key element of DC-to-AC conversion technology — and new tech for making a better switch is a subject of keen focus for GE researchers and others.
Currently, DC-to-AC conversion happens in large buildings that sit near major cities or population centers. The details are enough to keep power engineers up at night. The first step is to turn the DC power from the incoming transmission lines on and off rapidly, which yields a rough approximation to the up-and-down waveform of alternating current that is then smoothed for use. A single switch that can handle a thousand amps of power at 300,000 volts doesn’t yet exist, so engineers instead use small switches made of semiconductors — the same delicate material that makes up the computer chips in a smartphone. Since each semiconductor switch can handle only 3,000 to 7,000 volts, engineers string hundreds of them together in a series, like Christmas tree lights, to handle the whole load. The trouble with this arrangement is that it’s finicky and expensive to ensure that the system will continue to operate when one of the switches fails or when lightning strikes nearby.
A big switch that could handle the high-voltage, high-amperage transmission lines would be a much better solution. Tim Sommerer, a physicist at GE’s Global Research Center in Niskayuna, New York, has been looking for several years into just such a switch — one using charged gas, or plasma, instead of semiconductors. Plasma switches are used in some X-ray machines and mercury-arc valves (an early technology used to convert DC to AC); Sommerer and his team are adapting them for use in the demanding environment of power transmission. A prototype operating in GRC’s labs consists of two electrodes separated by helium gas. When a large voltage appears between the two electrodes, the gas conducts electricity from one electrode to the other, like a well-controlled lightning bolt.
To open and close the switch, researchers place a wire mesh between the two electrodes. By controlling the current through the mesh, they can trigger an arc between the two electrodes. “You have a hard time stopping it from conducting,” says Sommerer. “But we control it partly with the geometry of the chamber, and we can also stop the flow by pulling current out of wire in the [mesh].”
The current in their prototype flows for less than a millisecond at a time, but that’s enough to test the concept, improve the control of the current flow and glean insight into how well these switches will perform. “You can learn a fantastic amount in a millisecond,” says Sommerer.
So far they’ve run the machine at a hundred amps at 100,000 volts and are now working toward 300,000 volts. But one day plasma switches could lead to big cost savings in the power industry. “It now costs $400 million to convert 3 gigawatts from AC to DC transmission and back again,” says Sommerer. “We expect that this technology could cut that cost in half.”