Generating loads of electricity from moving water might soon shift away from the province of behemoth structures like Washington’s Grand Coulee and China’s Three Gorges dams.
Over the last few years, researchers and industry have been chalking up successes developing small-scale, distributed hydroelectric generators to potentially replace their massive forebears, whose footprint causes major disturbances in the environment and communities nearby. These emerging technologies, collectively called hydrokinetics, can turn moving water in rivers, manmade spillways and ocean tides into electricity that gets pumped into power grids.
“For new projects that need to be started, small hydrokinetic distributed networks should be considered as a viable candidate against big dams or other major power production projects,” says Diana Marculescu, a Carnegie Mellon University computer and electrical engineering professor who works in hydrokinetics. “In the long term, these distributed generation projects can become a serious alternative to large-scale hydroelectric, especially in the developing world, where increase in demand will be much larger.”
In an April 2013 report on waterpower, the U.S. Department of Energy forecast that hydroelectric dams and hydrokinetic technologies could provide 15 percent of the country’s electricity needs by 2030.
Huge potential energy recovery
But in the report’s compilation of analyses on hydrokinetic sources, the bigger potential is revealed—1,170 terawatt-hours of electricity is theoretically recoverable in wave energy alone every year. That’s enough to power around 100 million homes. Tapping the energy in flowing rivers without building dams by planting turbines in the water, so-called run-of-the-river generation, could yield another 120 terawatt-hours a year. And converting some of the thermal energy held in ocean water could produce another 576 terawatt-hours a year.
(Power density of the Atlantic Ocean off the Massachusetts coast. Red areas are greater than 1,000 watts per square meter. Courtesy Center for GIS at Georgia Tech.)
Hydrokinetic power companies are beginning to see successes in pilot projects. Verdant Energy, a company that demonstrated tide-driven turbines in New York City’s East River from 2006 to 2009, was issued the first-ever commercial tidal power license in 2012 to generate and sell up to a megawatt of electricity. The company may eventually install 30 turbines in the river as part of their Federal Energy Regulatory Commission pilot license.
Others will soon follow them. FERC has now issued 14 preliminary hydrokinetic permits and another seven are pending. Those projects will have the capacity to produce more than 3.4 gigawatts of power.
“Energy harvesting from water is trapped in an archaic damming paradigm with high up-front costs and ecological impacts,” Marculescu says. “But rivers run to the ocean, and there is an enormous amount of kinetic energy that could be sustainably harvested.”
Making hydrokinetic smart
But for these small power generation units that may one day pepper shorelines and inland waterways to work optimally, they need to smarten up. During project design and turbine operation, managers need to have real-time information about the flow of water at the units. They also need to have an accurate computer model that forecasts changes in water flow rates coming toward the turbines due to upstream weather events.
That’s why Marculescu is leading a team of engineers and computer scientists to develop a toolkit to monitor and control distributed hydrokinetic units. Their tools will help place and operate the generators. “The state-of-the-art model we are working to build is weather-aware and accurate second by second,” she says. “It’s a lot of data we’ll feed into it, but then we could predict what will happen in days or weeks. And then you could decide which turbines to turn on or off and do it as a function of weather and other data that effects the flow rate.”
The system would also provide the brains to direct hydrokinetic-generated electricity onto the local power network, creating a smart component of a smart grid. She says their model will be designed to tell the hydrokinetic units when to feed into the grid, and tell the grid when there won’t be enough power coming out the units so it can find power elsewhere.
“There is a huge benefit to society in this work as we strive to create more sustainable ways to power our lives,” she says. “Small footprint hydroelectric projects could create enough low-carbon energy to power an economy the size of Virginia while minimizing impact to the environment and surrounding communities.”
Power in prediction
The team is going to be building hierarchical models at several scales to analyze river systems, folding in huge amounts of data into each—tidal and river gauge sensor information, temperature and precipitation readings, and hydrological and soil features, among others. There are so many numbers to feed in, Marculescu says, that they still need to work out their system’s architecture. When they scale it up from their current study area to one the size of, say, the Mississippi River drainage basin, will standard computing resources be enough to crunch all the numbers? Or will they need to split these geographic areas up and then network the separate models together?
The project has just begun, and is being funded by a three-year, $1.2 million National Science Foundation grant. Marculescu says the result of their work could also be used to predict coming catastrophic flooding events in great detail. “The dynamics of something like a flood happen so fast that it can take people by surprise,” she says. “But you could use our system to monitor changing conditions and make predictions to say to people, ‘You’ll have this much water in this much time.’”