the film forum library tutorial contact

'Robofish' Lives to Tell Tale of Turbine Turmoil by Jonathan Brinckman The Oregonian, May 5, 2000

"Sensor fish" boldly swim into harm's way
to help scientists learn how to make dams friendlier to young salmon

Tom Carlson holds in his hand the nearest thing yet to a talking fish.

Six inches long, stuffed with $5,000 worth of electronics, it's a crash-test dummy designed to take a wild ride straight through the massive generating turbines of the Columbia River's dams.

Millions of young salmon pass through the turbines each year on their way to the Pacific from spawning grounds in the Northwest's rivers and streams. Hundreds of thousands die, battered by the whirling blades, torn apart by the turbulence, smacked against metal surfaces.

In their struggle to reverse dwindling salmon runs in the Columbia River Basin, scientists measure turbine death rates by periodically sending live salmon fingerlings through the turbines. But live fish can't tell scientists what they experienced during the three-minute trip through the dam.

Mechanical fish can.

Each of these robofish comes equipped with a rechargeable battery, data-storage chips, instruments that measure acceleration and pressure, and wires poking from a rubber tail. Each also carries a rescue kit: two golfball-size balloons triggered to inflate after the fish gets through the turbine, and a miniature radio transmitter that sends out a beacon.

Teams of biologists in boats downriver from the dam use radio locators to triangulate on the signal. After retrieving an ersatz fish, biologists reconstruct its account of the voyage by plugging the dangling wires into a desktop computer.

The "sensor fish," developed over two years at a cost of $590,000, illustrates the lengths the federal government will go to make its dams less lethal to salmon. The robofish's "experience" will help scientists understand how to manipulate the hydropower system to make it friendlier to migrating salmon.

The U.S. Army Corps of Engineers, which operates the dams, spends $3 million a year on trucks and barges that carry half the 150 million ocean-bound juvenile salmon past dams. Metal grates covering two-thirds of each turbine entrance keep most fish that stay in the river out of the spinning machinery. They go over the spillways or through fish-bypass systems.

But about 20 percent of the young salmon and steelhead that remain in the river go through turbines. And about 10 percent of them die at each dam they pass.

That's too many, federal biologists say.

To Carlson, however, the puzzling thing about the turbines is not how many fish are killed. It's how many survive.

Turbine mortality
Biologist calculate that about 5 percent of the young salmon that go through turbines are killed immediately, mostly from hitting metal structures or being ripped apart by turbulence.

Another 5 percent are snapped up by predatory fish after leaving the turbine. Disoriented by their swirling three-minute trip, the young salmon are easy prey for northern pikeminnow and other fish that lurk below the dams.

Carlson, a scientist at the U.S. Department of Energy's Pacific Northwest National Laboratory, has spent hours at a Corps of Engineers research center in Kentucky watching thousands of colored beads go through scale models of turbines built of transparent plastic. The beads illustrate what happens to fish in turbines where they are overwhelmed by the powerful current and can't control their position.

Far more than 5 percent of the beads bounce off walls and posts in the model turbine, suggesting that that percentage of fish would die. Why are only 5 percent of the fish killed? The answer, Carlson said, must be that fish usually hit structures lightly, with glancing blows.

"If I were to reach out and touch your eyeball, it wouldn't cause any problem," Carlson said. "If I were to reach out and poke you hard in the eye with this pen, there's a chance of causing some real damage."

The sensor fish program, which Carlson developed, is partly designed to figure out which turbine parts the fish hit the hardest. With that information, engineers will be able to redesign turbines to make them less harmful.

One idea is enlarging the internal posts to make them less like a knife edge and more like a rounded barrel.

Redesign saves fish
Turbine redesign already has paid off. Two new turbines installed at Bonneville Dam last year have smaller gaps between the turbine blade and the turbine housing. Narrower gaps reduce the turbulence, which means that fewer fish are ripped apart.

Tests with live fish show that the narrow gaps halve the death rate of fish killed going through the turbines. Results from the sensor fish program point to another possibility: increasing the survival of fish once they get past the turbine.

In a startling finding, trials have found that dramatic swings in pressure and acceleration continue for minutes beyond the 10 seconds it takes fish to get past the turbine blades. Turbines, it turns out, create turbulence that continues down the concrete tubes that lead from the dam.

When fish finally shoot from the pipe, they've been spinning for minutes. They haven't had a chance to reorient themselves and aren't prepared for the gantlet of predators.

Carlson's trials suggest that the answer might be to lengthen the exit tubes. That would smooth out the flow.

"One of our goals is to deliver the fish from the turbine environment so it is capable of coping with what it meets," Carlson said. "If we manage the turbine environment, we have a good chance of managing what happens downstream."

To Carlson, sensor fish are simply a way of gathering data from the dark, mysterious inner world of turbines.

Sitting in his Portland office, he displays graphs showing how pressure and acceleration change as fish approach the turbine blades, rush through the spinning vortex and continue through the concrete exit tube. He talks easily about the complex physics of the journey, the reliability of the data.

But what about the rubber tail of the slightly gummy reddish robofish? If the sensor fish can't be steered, why does it need a tail?

"The tail itself?" he says, then looks up with a smile. "That's more touchy feely."