Bumping along in Ingram's old pickup truck, we pull into Hoeganaes Corp.'s steel plant near Gallatin, Tenn., outside Nashville. Overhead, power lines are coming in from a 988-megawatt coal-fired power plant 8 kilometers away. Big graphite cylinders wait in boxes—replacement electrodes for the plant's arc furnace, in which scrap steel is melted down to make powder, primarily for the auto parts manufacturers that are legion in this part of the country. There's metal scrap mixed in with the gravel on the plant's driveway, which doesn't make the ride any smoother in Ingram's modest truck.

The SuperVAR machine itself, at first glance, isn't much to look at: it's just a shipping container sitting on a flatbed truck. In the substation just behind it, I hear birds chirping—"faux birds," that is. The sounds are recorded songs of distressed birds played to scare away real birds that might fly into the station and short out circuits, explains Ingram, a lanky, thin-lipped Alabamian with a dry sense of humor.

PHOTO: American Superconductor Corp.

Exciting the grid: David Madura, chief engineer (left) and Hank Valcour, principal engineer, check out SuperVAR, which is housed for testing in a container on a flatbed truck.

We are joined at the machine by Charles ("Chuck") Stankiewicz, vice president and general manager of American Superconductor's power electronics unit in Middleton, Wis., which analyzes transmission grids and sells SuperVAR machines. Also on hand is David Madura, product manager and chief engineer for the SuperVAR machine at American Superconductor, as well as other representatives of the company and TVA. Stankiewicz has flown in from Wisconsin and Madura from Massachusetts, and Ingram himself has just driven 3 hours from Chattanooga, to show off and explain what they all obviously think is a very hot item [see photo, "Exciting the Grid"].

This steel plant is an ideal place to test the SuperVAR, because its huge arc furnaces put enormous and fast-varying inductive loads on the grid, consuming large quantities of reactive power erratically. The arc can vary between extinction, at zero current, to the short-circuit condition, caused by scrap contacting the electrode, says Ingram. Physical movement of both the solid charge pieces and the melt cause variations in the arc length—fluctuations that can occur many times a second. Superimposed on these effects are variations caused by mechanical vibration of the electrode and its supporting structure.

Conceptually, the SuperVAR machine is simple enough. The rotor is a large thermos bottle, about 2 to 3 meters long, containing liquid neon at 27 degrees above absolute zero. In principle, a SuperVAR could be built to run on liquid nitrogen at 77 degrees above absolute zero, but current densities would be lower and the number of wire windings greater. The neon comes from a cryocooler in which the primary loop runs on gaseous helium. Inside the thermos are the rotor coils, made from American Superconductor's first-generation superconducting wires.

SuperVAR can work two ways. It can provide compensating reactive power continuously to the grid as needed. A sensor provides readings of how much current is lagging voltage, and the control unit tells the exciter to inject the needed compensating reactive capacitance. Or the voltage can be set at a suitable level in the rotor, so that it spontaneously adds the compensating reactive power when the grid voltage starts to drop.

The SuperVAR setup at the mill is designed, says Ingram, to provide accelerated test conditions. Running on a 13.8-kilovolt circuit that feeds Hoeganaes's 50-megavoltampere arc furnace, the SuperVAR machine has been in operation for 3000 hours. It has gone through more than 2300 melt cycles, in which 140 million kilograms of steel have been melted. That's the equivalent, says Ingram with a note of irony, of melting 100 000 Porsche Boxsters.

Hoeganaes, the owner of the mill where SuperVAR is being put through its paces, bills itself as one of the world's largest makers of ferrous metal powders. As we stand next to the SuperVAR machine, a scrap bucket carrier rolls into the mill, destined for the arc furnace. It takes 60 tons of scrap to make a standard load, and a melt takes place—the big electrodes dipping down into the furnace—about once an hour.

Stepping into a building behind the trailer, where monitoring devices and controls are located, we wait for the next melt. After a while, there's a big rumble, and voltage levels dive drastically—by as much as 20 percent—on the meters we're looking at. We step back to the trailer to listen for the SuperVAR to kick in, but the machine is almost noiseless. Nevertheless, the whole structure is vibrating from the enormous, instantaneous torquing of the SuperVAR in response to the melt operation going on in the adjacent building. It's a punishing environment, and it's exactly what SuperVAR's designers had in mind for this initial test.

Besides running almost flawlessly so far, the SuperVAR has several inherent advantages over the standard synchronous condensers that do not rely on superconductors. Because there is no heating in its rotor, there are virtually no thermal stresses and no losses in its field coils. Also, thanks to the large air gap between the stator and rotor coils—a feature that is possible because the current density in the superconducting coils is so high—the machine is able to inject on demand lots of current and, hence, lots of reactive power.

Exploiting the large gap between rotor and stator, SuperVAR's designers were able to surround the rotor coils with an aluminum conductive tube or shield. It's this feature that enables the machine to react so vigorously to drops in reactive power on the grid. Basically, when there is a transient on the circuit running through the stator, currents in the shield spontaneously oppose what's happening in the grid, says Bruce Gamble, providing a shorthand explanation. Gamble is American Superconductor's director of engineering for supermachines.

So SuperVAR is expected to be a cost-effective source of dynamic reactive power. What's more, its footprint is much smaller than those of equally rated conventional machines—a very important factor, especially in urban grids where substation space is scarce. In many places, because of the somewhat haphazard way the grid has been reorganized in the era of restructuring and deregulation, reactive power is not always provided efficiently.

To provide a sense of the potential market, Stankiewicz and Ingram note that TVA, with generating capacity of about 30 gigawatts, has 259 capacitor banks with a total rating of 9.4 gigavoltamperes-reactive (GVAR). Scaling to the whole United States, which has more than 700 GW of generators, there might be as much as 250 GVAR of transmission capacitors—all candidates for replacement by superconducting synchronous condensers. Of course, not every situation will be suitable for a SuperVAR-type device, but even a modest slice of the global transmission market would be a big market indeed.

If the market develops as SuperVAR's boosters expect, it's unlikely that American Superconductor will own it all forever. But the company was well positioned to devise the first prototype. The innovation was a happy conjunction, really, of several corporate developments. One was its commercialization of first-generation superconductor wire manufacturing technology, which the company accomplished by working closely with researchers at Oak Ridge National Laboratory, near Knoxville, Tenn. Two years ago, American Superconductor opened a production line for its first-generation wire in Devens, Mass., where it is turning out about 400 km of it each year.

Another development was American Superconductor's acquisition of a leading firm in Wisconsin that makes the big semiconductor devices used in power electronics. That division now makes power electronic converters and devices based on insulated gate bipolar transistors (IGBTs) that are used not only in SuperVAR and other reactive power products but in fuel cell­powered buses to be tested in Winnipeg, Man., Canada, and in wind farms being built by Denmark's Vestas Wind Systems AS, the world's top windmill manufacturer.

Finally, American Superconductor has had considerable experience and success with rotating superconducting machinery in the context of the U.S. Navy's motor demonstration program. The company has so far built and delivered a 5-MW superconducting motor for the Navy and is presently making a 36.5-MW propulsion motor. These motors have some important similarities with the SuperVAR machine. The Navy's imprimatur, notes Ingram, was a critical factor in TVA's decision to proceed with American Superconductor on SuperVAR.