LIDS research scientist Mardavij Roozbehani spends his days doing what he loves best: pursuing his ideas, taking pleasure in learning, and working to solve challenging and exciting research problems. LIDS has been his intellectual home for many years – he's been here as a graduate student and a postdoctoral researcher – and though he regards industrial research highly, he values the freedom of choice found in academic research. What he's doing with that intellectual freedom right now is working with his colleagues to address one of the big questions of the future: how do we design and operate the next generation of electricity networks in a way that brings about the desired environmental, economic and social benefits? It is a complex problem to work through, but Mardavij is energized by the challenge. "Power systems are very unique and interesting because there is real-time interaction between economics and physics, and the system is safety-critical, with strict specifications and little flexibility. The upcoming changes make these systems only more interesting and more challenging."
Much of the developed world is on the verge of a whole new energy generation and consumption paradigm—a power production system that deploys renewable resources on a large scale, and a consumption system that is interactive and responsive. Many researchers are working on creating new and more efficient energy generation and storage technologies—improving the ways in which solar cells capture the sun's energy or wind farms convert wind power to electricity—Mardavij, however, is working on system-level problems that arise when all the new technologies need to be linked together to create a smoothly functioning system. "The system operation paradigm will change at all levels of transmission and distribution, so will the way electricity markets are run and operated," says Mardavij. One of the principal issues at hand is the level of uncertainty the renewable energy sources introduce into the system. Since the energy supplied is from an intermittent source, how can we ensure the exact amount of power we need will be there exactly when we need it? To do that, the new system will need to be much more responsive to real-time events and changes in the state of the grid. It will take advantage of information technology and real-time communication between energy producers, energy consumers, and system operators to, among other objectives, mitigate the effects of these uncertainties.
"Consider dynamic pricing and demand response for instance," says Mardavij. "This means that the consumers would—based on a price signal they receive, for example, via a smart meter installed in their home—ramp up or ramp down their usage, save a little money, and help the environment." For customers who are more environmentally conscious than they are price sensitive, the signal might also reflect the environmental impact of power use, perhaps in terms of real time CO2 emissions. "The hope is, roughly speaking, to reduce the discrepancy between the total consumption and total renewable generation. And as for the gap that remains, you have to cover that gap with reserve capacity supplied by conventional resources, or maybe storage someday. You want the gap to be minimal, so as to minimize the use of fossil-based energy sources."
In a power grid, at every instant of time, supply and demand have to match precisely, and this is mostly managed by the transmission system operators. Today's system is in a sense an open loop system in which the consumer is a passive agent, with limited means to affect the system and fairly predictable consumption patterns. The future system, however, will incorporate real-time pricing and renewable energy generation and storage at the level of consumers—large or small. For example, a typical consumer might have solar panels on her house. Sometimes her home solar generation system won't meet her needs and she'll need to draw from the grid. Sometimes she'll have a power surplus and want to sell that extra power back to the grid. And there will be thousands or hundreds of thousands of consumers just like her—unpredictable agents with access to a string of new levers to interact with the power grid. One aspect of Mardavij's work involves researching different system operation schemes. A potential solution involves several local control centers that supervise and coordinate the agents within their areas. That way the local control center could coordinate the interactions between each user and the local grid, in a sense analogous to how a transmission system operator works with big power plants.
"At a very high level, what I'm interested in is the sense of what the system architecture should be," said Mardavij, "and by that, I mean questions like how much autonomy should be given to different agents at various levels, what kind of information should be collected and to whom it should be given, how useful different types of information would be, what should the geographical distribution of the local control centers be, what do they need to communicate, and how can they coordinate between themselves and with the transmission system operator?"
The idea of a smart grid is still an evolving concept and it will likely remain so for a while. This, says Mardavij, poses unique challenges to researchers. "It's important to formulate problems at the right level of abstraction, so as to be relevant but perhaps not too specific. When possible, I try to model and analyze the trade-offs between a range of scenarios, rather than focus on a very specific scenario which would make a lot of sense, but might never materialize because policymakers, or evolution of technology or economics dictate something else."
The growing importance of energy has been on Mardavij's mind since his undergraduate days. His interest in energy systems didn't turn into practice, however, until he learned that professors Sanjoy Mitter and Munther Dahleh were forming a research group around the topic. When he heard about the opportunity, he was certain he wanted to come on board. Mardavij wasn't always so sure of his path, though. "Wanting to become an academic researcher was more like an evolutionary process. As a kid, I admired professional athletes a lot," said Mardavij, "For a while in my teens I wanted to become a professional basketball player! It took me a while to realize that that just wasn't in the genes!" Though basketball remains a passion, Mardavij loves the path he's chosen and is happy to be working on important issues and about the positive environmental impact his work could have.