Control of Electric Energy Systems
In 2010 Aleksandar Stanković, PhD, joined the Department of Electrical and Computer Engineering as the first Alvin H. Howell Professor in Electrical Engineering. Stanković earned an MS in electrical engineering from the University of Belgrade, Yugoslavia, and a PhD in electrical engineering and computer science from Massachusetts Institute of Technology. He was a professor in the Department of Electrical and Computer Engineering at Northeastern University before joining Tufts. Stanković has worked in the area of electric energy systems for twenty years and is currently focused on control, estimation, and dynamic modeling of these systems. “It’s one of those areas that naturally connects with many other fields, such as mechanical engineering and environmental engineering,” says Stanković.
“Electric energy is unique in the way it is a form of energy that is easy to transmit over large distances, it can be transformed into just about any form of energy, and it can be very finely controlled,” says Stanković. “But it has one weakness, which is storage—electric storage is not very good. Whenever you look at your almost empty laptop or cell phone battery you’re reminded of that.” Speaking about long-term storage of large amounts of electric power, Stanković says “the world has poured a large fortune into battery research, and it’s not there yet, so I think it is a genuinely hard problem, and the more useful strategy in the near- and mid-term is to try to avoid storage.” One solution is to use electric power for what it is good for—transmission, transformation, and control—and complement it with other sources of power that are better at storage, such as natural gas. “My case is that by controlling these energy systems in a coordinated way one can achieve major advances in performance,” says Stanković.
“A way to think about an energy system that I find useful is to think of it as a four-layer structure,” says Stanković. From the bottom up, the layers are Materials Flow, Energy Flow, Information Flow, and Capital Flow (see diagram below). The current energy system is unsustainable. The earth’s materials (gas, coal) are finite, so the input to Materials Flow is unsustainable. The output from using these materials (carbon dioxide, nitrogen oxide, other pollutants) is also unsustainable.
Energy Flow draws on nonrenewable materials (NRE) and renewable energy sources (RE). Unfortunately, the input of renewables into current U.S. energy consumption is very small: approximately 3% hydro, 4% biomass, and 1% wind, solar, and geothermal combined. Renewable energy input is likely to remain small until the market (the Capital Flow layer) makes renewables more competitive with nonrenewables, possibly by adding environmental and societal costs to the price of nonrenewables. Energy Flow also draws on the world’s water resources because water is needed to transform materials into energy. “In many places in the world, you can have water or you can have energy, but you cannot have both,” says Stanković. Energy Flow is currently unsustainable, and getting worse as energy demand increases throughout the world.
Multilayer Energy System
Four main layers—matter, energy, information and capital flows.
“The interesting part right now is the coupling between information and energy—the cyber-physical system,” says Stanković. “You have essentially two networks that are connected at many points, but traditionally computer scientists and electrical energy engineers have not worked together.” Capital Flow is the overriding top layer of the energy system. The work needed to integrate information flow with energy flow—and achieve major advances in performance—is limited by the amount of capital invested in this work.
Stanković likens the energy systems work ahead to moving from the Middle Ages to the Industrial Age. A major challenge is to develop electronic controllers that can support the critical timing needs of energy flow. For example, an electromechanical perturbation in the Pacific Northwest could travel to Los Angeles in 4 to 5 seconds, where it could cause a major power outage. Stanković and his research group are working on “smart” controllers that could basically “ride out” a power surge and continue functioning normally. Alternatively, local controllers could be programmed to transiently lower their demand for power. The Stanković group is also developing and testing a control system that would distribute information processing throughout the energy system, so that even if some local controllers went down, those that remained active could continue to support the electric grid. “That resiliency is important, and it’s enabled by this information layer,” says Stanković.
Stanković works primarily at his computer on estimation, model development, and simulation, but he also has a lab where models are tested physically, using semiconductor switches, power electronic converters, electric motors, etc. “I discovered early on that even simple experiments are great at sorting out good assumptions from bad assumptions,” he says. One simulation project his lab is currently working on involves probing an energy system for which they have limited data because much utility information is proprietary. “We can run quick bursts of computer simulations, where you can set initial conditions, run a short burst, and then change the conditions and run another burst,” says Stanković. “In that way I can explore the local space in a rational way, and I can know whether the overall system is stable, without asking for sensitive information.” They can use this simulation framework to probe computer code and answer analytical questions. “These models are statistically justifiable because you are introducing the least of your own thinking into what the model really is,” says Stanković. “And you can answer analytical questions without ever knowing what equations are behind the code.”
Stanković is part of a multi-institution collaborative research group that is working to create a fully monitored and dynamically controlled transmission grid. A current project is to look at the local control laws for energy users and determine which laws help—and which laws hinder—the overall system. “It’s going to take a long time to change the entire system to improve it,” says Stanković. “We’re just starting the process, the conversation. I think Tufts is a good place to do it.”
For more information, please go to http://www.ece.tufts.edu/people/faculty/bio/stankovic.php.