RICHLAND, Wash. -- As stewards of the U.S. electrical system, grid operators have a well-deserved reputation for caution. With the mandate to provide reliable, affordable energy as their bottom line, they typically favor time-tested approaches over radical innovations and can be slow to embrace new technologies until they've stood up to lengthy scrutiny.
But the past two decades have seen some rapid shifts in U.S. electrical systems, with power consumption plateauing even as new fleets of renewable power stations come online. With more changes on the horizon, operators have had to move quickly to keep pace, using modeling and simulation to work out the complexities in the new power system.
"We're heading toward a grid that, far from being a one-way street, is constantly balancing the needs of distributed renewables, electric vehicles, home [energy] storage and traditional generators," said Henry Huang, a power systems engineer with the Pacific Northwest National Laboratory. "The challenge going forward is how we can analyze and operate a grid at that level of complexity."
Located in the town of Richland, Wash., PNNL is one of a small collection of U.S. laboratories sketching out the map of the future U.S. power grid, a system that will be larger and far more interactive than our own.
At a very basic level, running such a grid is analogous to coordinating flight patterns from an airport control tower, Huang said. Every airplane is outfitted with sensors that communicate the vessel's location and status to its arrival destination. Using high-performance computational models, control tower operators can chart overlapping flight paths well ahead of a plane's scheduled departure.
Similarly, operating an advanced, distributed energy grid requires a lot of feedback from the system itself. Smart meters, synchrophasors and other tools are spread out across the grid and tell operators where the flow of electrons is beginning to ebb or flow. The real difference is speed. Where commercial airlines travel at around 500 to 600 miles an hour, the grid's traffic moves at the speed of light, Huang said.
Electricity, by the numbers
The technology has proliferated quickly. "I think I've seen more technology change in the last five years than in the previous 15 or 20," said Joe Rigby, chairman of the board at the regional energy holding company Pepco Holdings Inc. "By this time next year, our operation" -- which serves about 2 million customers in the Washington, D.C., and Maryland area -- "will have 1.4 million meters out there."
Nationally, about a third of U.S. household have a smart meter, and the number is rising.
But "smartening" the grid is only half the battle. Interpreting the flood of data that comes out of those devices may be the bigger challenge.
"There are billions of components we have to model, everything from the power plants and transmission lines to home appliances," Huang said. "So really, a more apt comparison to what we do might be trying to model a chemical reaction at the atomic level -- trying to map the changes in a billion interacting molecules."
Part of PNNL's work is to upgrade the traditional algorithms that utilities have used in the past to regulate energy flow. This is not simply a matter of adding more server banks or computing power but an intensive effort on the part of engineers to coordinate the physical components of the grid with the software controlling them.
This itself can be a challenge. The U.S. electrical grid is, by many estimates, the largest interconnected machine on Earth, cobbled together over more than a century. Its components, manufactured at different points in history by different makers, don't always "speak the same language," to borrow a phrase from grid operators' lexicon. Standardizing the language of the grid will take time, and in the meantime, models must find a way to accommodate uncertainty, as well as data flows, in their analysis.
Much of this work involves looking ahead. "President Obama said the U.S. has a goal of 85 percent clean energy by 2035, and that's the goal we're planning for," Huang said. Unlike physical electrical systems, PNNL's virtual models can scale up renewables to any level in order to study the effects of such a move on other power infrastructure.
For that to happen in the real world, Huang said, "the grid will have to change."
"Right now, to add the same levels of wind power as, say, Denmark, would cause you a lot of problems. But if we calmly recognize the need for new ways to look at and operate the power grid -- how to run our systems in real time rather than a step behind, how to handle the uncertainties -- I think we should be OK."