In today’s interconnected energy world, it’s not easy for islands and remote communities, cut off from the ready energy supply of big grids, pipelines and superhighways. Witness the international drama last winter when Nome, Alaska became ice-locked and only secured fuel because of an elaborate sea effort by Russia and the US.
Dependent on the outside world for fossil fuels, places like Nome face not only shortages, but also sky-high energy prices. Some residents of Nome spend 45% of their income on energy. Similarly, Hawaiians pay more than twice as much for electricity as Californians.
Islands often address shortages by building more power plants, particularly wind farms, but that can increase costs even more. So a group of researchers from Carnegie Mellon University’s Electric Energy Systems Group (EESG) recently tested an alternative approach in the Azores Islands, a way to use existing infrastructure more intelligently before building new capacity.
“This is important because it’s not all about capacity; it is not all about building generation. It is about what we call in our research, ‘just-in-time and just-in-place,” said Marija Ilic, a CMU professor and EESG director.
Using resources at just the right time and place requires digital automation and communication tools, aka smart grid, to tweak power plants and wires for maximum output and stability. The team experimented with this approach on the islands of Flores and San Miguel, forecasting the system’s future performance, and then automatically balancing supply and demand down to the minute to employ the cleanest resources possible.
Dance on the Wire
To accomplish their task, the researchers instituted a careful dance, an effort that precisely timed the operation of conventional power plants, wind energy, flywheels, hydroelectricity, demand reduction, and even plug-in electric vehicles to eke out maximum production.
Wind power – a headache for grid operators because of its variability – was at the heart of the experiment. The CMU researchers tackled this variability problem initially as other grid operators do. They used weather forecasts to predict how much electricity wind turbines would produce a day ahead, a few hours ahead and even ten minutes ahead and called upon hydropower or demand response to make up for wind power losses.
But for all of the advanced forecasting and scheduling, this approach has a glitch; the wind is fickle. In the last few minutes, it might suddenly defy predictions and gust or drop off, throwing off the grid’s balance. “Ten minutes ahead you can predict where the wind would be, but in real-time you will have real fluctuations and if you don’t balance it, the frequency will be all over the place and voltages will be unstable,” Ilic said.
To resolve the problem, the researchers developed highly sophisticated computer applications and automation controls to adjust the operation of a fly wheel storage device, sending just the right amount of power between the fly wheel and grid in those last crucial moments. The fly wheel is expensive to operate, so the smart grid application was designed to call upon it only at the very last moments. Less costly ways exist to stabilize frequency and voltage on the wires; fly wheels make sense only when disturbances are large and sudden, according to Ilic.
“We had to develop very intelligent controls for when these unexpected events happen, models that didn’t exist before. This is a major innovation,” she said. “We have shown here for the first time what type of control logic you have to have.”
Lessons For the Rest of Us
What they found may offer important lessons for an urban island, like Manhattan, which is not remote, but suffers the same problem of high electricity costs and detachment from power supply. New York City’s dense population leaves little room for new power plants or delivery wires, so the city needs to use its existing infrastructure to maximum advantage. Even more important, the research offers promise for the US as a whole, as it increasingly integrates wind power into its grid. So the next step for Ilic and team is to see if their application works on such a large scale.
Ultimately, the project could change the way US and European policymakers plan power portfolios, according to Ilic. Green energy policy now centers on specific, somewhat arbitrary targets set by government. New York, for example, wants 30% of its electricity to come from green sources by 2015. The challenge for states like New York then becomes: Can we make the goal?
But to Ilic such goals put the cart before the horse; building capacity before knowing how much the existing system can handle. “We asked the question differently. If you were to put in more renewable energy, how would you manage it and what is the pay off?”
Ilic added: “We must revisit how we operate and plan the system, and then identify what is most effective. There is a lot we can do before we go and build.”
