The Smarts Of Smart Grids

By Jason Ganzhorn

On August 14, 2003, at 12:15 p.m., a sequence of events began in Ohio that led to the famous "Northeast Blackout of 2003." Ultimately, 256 power stations shut down. New York City and most of the rest of New York state lost power, along with significant portions of surrounding states and the province of Ontario, affecting an estimated 55,000,000 people. The root cause of the blackout was later found to be the shutdown of a single power plant in Ohio and improper maintenance of high-voltage power line paths. What seems like it should have been only a minor problem escalated into a catastrophe due to swiftly cascading failures as larger and larger shifts in power load forced the shutdown of many power stations to avoid damage.

Detractors of the current power grid structure claim situations such as these highlight the vulnerability of the system to accidental failure and targeted terrorist attacks. Many believe that the solution is to make the power grid "smarter" by allowing energy suppliers to have more control over how they distribute power through two-way communication with users. The U.S. Department of Energy announced on Nov. 24, 2009 that it would be exploring this "smart grid" technology by funding projects in several locations around the country, the largest of which is the Pacific Northwest Smart Grid Demonstration Project, with 14 test sites in Idaho, Montana, Oregon, Washington, and Wyoming.

This demonstration project will involve 60,000 metered customers and 112 megawatts of power assets. Washington State University, the University of Washington, and 12 utilities from around the Pacific Northwest, including Seattle City Light and Portland General Electric, have partnered with the Bonneville Power Administration for this project. Battelle, a company involved in the development and commercialization of new technologies, operates the Pacific Northwest National Laboratory for the U.S. Department of Energy and will be coordinating the project from that laboratory.

This is not the first project of its kind. Another 1-year demonstration project called the Pacific Northwest GridWise Demonstration Project, which utilized much the same technology but on a smaller scale, was launched on the Olympic Peninsula in March 2006 and ran for about a year. That project was made up of two studies, one testing consumer respond to demand-driven prices, the other testing the effects of devices that could scale back power usage in high-demand scenarios. Each of these studies involved less than 200 people/households, but the results achieved were satisfactory enough to encourage further funding of similar, larger projects like this latest one. The project report states that several goals were met, including the reduction of peak load, high consumer acceptance of contracts which varied price according to load, and successful use of the Internet as a control mechanism for distributed generation.

There are several goals involved in the current project. First, power suppliers want to measure the effect on client power consumption when clients have information relating to their power usage and when the price for power is changed according to demand. Secondly, power suppliers are interested in developing grids that can draw from several sources of highly variable power, such as wind and solar energy, while maintaining a steady output. That is in fact one of the University of Washington's goals in implementing the smart grid project on their campus. Third, methods for sending messages to clients to scale back on their energy usage in case of overly high demand are also being tested.

The two universities involved in the project are deploying pilot projects as test-bed sites. The funding for this project allows the universities to upgrade their meter system. In the case of the University of Washington, there are currently only 9 electrical meters on the central campus, hindering efficient tracking of energy usage. In the course of the deployment, more than 200 smart meters will be installed. This project also allows the universities to conserve energy and educate college students about their energy usage habits, all while providing an excellent opportunity for researchers to test experimental technology to improve the granularity of smart grid information. In exchange, these sites will provide test data to the Department of Energy.

As an example of how these project sites will be implemented, in Seattle, the University of Washington and Seattle City Light are working together to simulate supplier-client interaction. Simulating the supplier, Battelle will test sending messages to its client, the University of Washington. These messages will tell the UW to reduce power consumption if the grid is under high demand. These messages would be useless if the university had no control over its power usage, so the University of Washington will install smart meters on several campus buildings. These meters will track energy usage and integrate with the existing building automated control systems responsible for operating building ventilation, heating, cooling, and lighting systems. Administrators will be able to send messages to the meters to adjust parameters such as building temperature in order to decrease energy usage. Another bonus from the meters will be the ability to track spikes in energy usage which often indicate malfunctioning equipment.

This interaction between Seattle City Light, Battelle, and the UW is a microcosm of the project as whole. Like Seattle City Light, Pacific Northwest National Laboratory (PNNL) will monitor all 14 test sites and test sending electronic messages ("transactive control" messages) to each of them. Norman Menter, the project manager for the University of Washington, gives an example message the university might receive: "We're experiencing high demand for electricity in a certain area of the Northwest electrical grid. We need to reduce consumption; would you reduce consumption 2 percent on your campus if the price of electricity were dropped a quarter of a cent per kilowatt-hour?" He explains that, based on that information, the administrators at UW could then make a decision to decrease power usage or not.

Another procedure that could be tested by the PNNL is the transfer of excess power from one site to another. As the power grid is currently implemented, transferring large amounts of power across the United States is not very efficient. This is because many of the sub-grids were constructed independently and are connected to each other at only a few nodes. However, large-scale power transfer across the nation is desirable. According to Menter, "You can generate PV [solar] power most reliably here in the West between one o'clock and four o'clock in the afternoon," and "if you could use that one-to-four o'clock West Coast slot of PV power and send it all the way across the country, which in theory you can, you could meet the peak evening demand for power on the east coast." Smart grid technology promises to increase and strengthen the interconnections between grids, to facilitate this kind of large-scale, long-distance energy transfer.

In addition, the technology provides opportunities to increase the security of the system. "The power grid is one of the nation's most critical cyberphysical infrastructures," says Tadayoshi Kohno, a professor at the University of Washington who specializes in security. "There's a large, growing interest in security for the nation's power grid."

One way the security of the system will be improved is that power generation will be decentralized into smaller generation stations, reducing the impact of the loss of a single power supplier (from either natural or man-made causes). As it stands now, there are generally few connections between various sub-grids – these chokepoints are also vulnerable to attack or disaster and also hinder the easy transport of energy.

However, care must be taken to ensure the security of the updated grid system from the ground up, or there could be the potential for even more dangerous attacks as the flow of information over the power grid increases. "One of the things we've learned," says Kohno, "is that if you're creating something from scratch, it's oftentimes easier to make sure that it's secure than if you're trying to take an existing system, add the Internet to it, and retrofit security back onto the existing system. We have an expression saying, ‘Security is better built in than bolted on.'"

IOActive, a computer security services company which has been doing research on the vulnerabilities in the smart meters typically used in smart grid implementations, agrees. According to a recent press release, they strongly suggest "independent third-party security assessments of all Smart Grid technologies that are being proposed for deployment in the Nation's critical infrastructure."

Smart grid technology promises many things: increased energy efficiency, better integration of alternative energy sources, and opportunities to increase the security of the power grid system. However, it is not a cure-all solution. It will require consumer participation to live up to its full potential. The technology involved can increase efficiency on its own, but the real strength is the information gathered which can be then used by consumers to monitor and control their energy usage. There are also several concerns about how this will play out in real usage scenarios; if power companies really have the ability to remotely turn off or lower power to consumers' appliances, those consumers may well take issue. Ultimately, that is what this project is all about: hashing out the problems that could arise before dedicating an entire national power grid to the upgrade.

Jason Ganzhorn is a senior studying computer science at the University of Washington.

Images

Top: The Electricity Infrastructure Operations Center (EIOC) control room main display. As part of the Pacific Northwest National Laboratory, the EIOC serves as a platform for researching, developing and deploying technologies to better manage and control the power grid. Photo: Pacific Northwest National Laboratory.

Bottom: The Grid Friendly Appliance Controller is a sensor installed in an appliance that can detect when the electricity grid is under stress and then turn the appliance off. Photo: Pacific Northwest National Laboratory.