PNNL proves just how smart a smart grid can be

2009 Green 15: National lab teams with IBM find green success in a real-world smart-grid deployment

There's been plenty of talk in recent years about upgrading the country's aging, inefficient, unreliable electricity-delivery system to a smart grid: an intelligent system that promises to make energy flow more freely, more reliably, more efficiently, and at lower costs to both consumers and businesses. Such a system certainly has support; President Barack Obama included funding for the development of a smart grid in his proposed stimulus package.

Talking about a smart grid is easy. But demonstrating its effectiveness and value in a real-world implementation is a significant challenge. Yet the Pacific Northwest National Laboratory (PNNL), with the assistance of IBM, took on the task last year with its GridWise Olympic Peninsula Project. "[We] wanted to show both the depth and breadth of smart grid functionality by integrating both demand response and distributed generation to affect multiple levels of the power grid at the same time," says Robert Pratt, the GridWise program leader. "This helps the power company to manage delivery and reduce power outages in an overloaded power grid."

[ See the full list of this year's InfoWorld Green 15 award winners. | Subscribe to InfoWorld's free weekly Green Tech newsletter. ]

The homes and businesses of participating customers were equipped with multiple intelligent utility devices, from thermostats to water heaters. The devices were tied to the PNNL power-delivery system using Internet-scale Control Systems (ICS), a prototype interoperability framework from IBM. Customers programmed their thermostats and water heaters with tools on the Web, providing preference information such as whether they were willing to have their home or office temperature automatically adjust in case of a power shortage.

The underlying ICS software took customer inputs, combined it with real-time grid-sensor data, generation capacity, and market-trend information, and automatically adjusted each device as necessary. As a result, the software helped keep the grid running smoothly by dynamically and intelligently responding to stress in the system, kicking in only when there was an energy shortage and curtailing only those loads that consumers were willing to shed.

The capabilities of the system didn't end there. The software also purchased electricity from providers based on real-time prices that fluctuated with supply and demand. As a result, those users willing to shift high consumption to off-peak hours -- thus easing stress on the grid -- also reap savings thanks to lower energy bills.

One of the biggest challenges PNNL faced was engaging customers to participate in the pilot program, Pratt says. "We needed to create a no-lose value proposition ensuring that residents and businesses understood that they had nothing to lose and could only gain from their participation."

[ Learn more about the benefits of the smart grid. ]

As an incentive, PNNL offered the 112 participating residences and businesses a 10 percent discount on their utility bills. Making the Web interface easy to use also contributed to the project's success, Pratt says.

The project yielded up to 50 percent reduction in short-term peak distribution loads and an average of 15 percent decrease in overall peak loads over the course of one year. PNNL projects that if all customers were engaged in reducing peak loads at this level, peak electricity prices would drop substantially. Furthermore, about $70 billion worth of new construction for generation, transmission, and distribution systems could be avoided over a 20-year period.

PNNL and IBM also gleaned valuable lessons about adjusting the "throttle" of electricity generating stations to meet fluctuating demands. If those adjustments are made on a continual basis, it results in excessive wear and tear on equipment, as well as wasted fuel.

"By setting a limit on demand by purchasing power every five minutes, we found we were able eliminate these fluctuations completely at times," says Pratt. "This is especially important as the U.S. continues to incorporate large amounts of wind resources. Then, we are not only managing changing customer demand but also fluctuating wind resources. Managing the grid is becoming increasingly complex as we add renewables, and smart-grid approaches like this will really help."

Since project completion, energy-use management programs modeled after the Olympic Peninsula Project are being considered for implementation both across the United States and abroad, Pratt says. Meanwhile, IBM is drawing on the results from the project to advise the Obama administration on incorporating smart utility grids into future infrastructure projects that are under consideration.

Copyright © 2009 IDG Communications, Inc.