Roger Anderson: Making a Smarter Power Grid

Columbia Engineering’s collaboration with Consolidated Edison Company of New York (ConEdison) has the potential to not only eliminate the kind of brown-outs Brooklyn residents endured this week but also to influence the way engineering is taught. 

Anderson has also done extensive study and research in Geophysics. He has been quoted frequently in recent weeks by numerous publications about the oil spill in the Gulf of Mexico. 

Roger N. Anderson is the ConEdison Senior Scholar at Columbia Engineering’s Center for Computational Learning Systems, as well as an Instructor in the Department of Earth and Environmental Sciences. He has developed a patented Adaptive Stochastic Controller (ASC) technology along with senior staff associate Albert G. Boulanger that will be part of ConEdison’s Smart Grid project.
Columbia’s role is to produce the ASC for the New York City Smart Grid Demonstration Project, funded by the federal Department of Energy (DOE), to help build a smarter, more efficient, more resilient national electric grid. Last November, the DOE awarded ConEdison $45 million for its Secure Interoperable Open Smart Grid Demonstration Project — one of 16 projects selected nationwide. The New York Public service Commission will provide additional matching funds to make it the largest Smart Grid project in America.
Under the Smart Grid, Anderson says, “Things will happen too fast for optimal decision making by humans — even from the highly skilled operators at ConEdison.” The Smart Grid will use Columbia designed ASC computer control to respond more quickly to power surges and emergency outages so that power can be distributed to where it’s needed, using photovoltaics, wind, battery storage, and other green sources from areas with less demand.
“We are trying to make the electric grid smarter, much more like the Internet, so that control systems can route power to where people need it, when they need it, and from green sources as much as possible” says Anderson. 
The ASC is an algorithm-based, machine learning system that will allow the Smart Grid to calculate, on a minute-by-minute basis, the cheapest, most efficient and most reliable ways of delivering electricity to consumers. It will also allow the Smart Grid to anticipate and respond automatically to problem events such as weather changes and equipment failures in order to prevent power outages.
“The whole idea of the smart grid is to have enough computer intelligence so you can control [the grid] better and make it more stable,” Anderson says, adding that such a system would have surpluses of portable, distributed generation and storage. “Instead of building more power plants and substations, utilities will use the existing power infrastructure more efficiently by being able to move it around and supplement it with green sources.”
Such technology is also being developed in Europe and China, he says, and an international working group has been established to share best-practice ideas and technologies.
The impact on engineering schools could be fundamental, Anderson says, further expanding the burgeoning movement toward multi-disciplinary approaches and away from traditional departments such as Electrical, Mechanical, and Civil Engineering Columbia’s Center for Computational Learning Systems, led by director David Waltz, is just such an interdisciplinary center of Research and Development. With more than one hundred such centers university-wide, Columbia is a national leader in such “Silo-Busting,” according to Anderson.
“We’re going to see the kinds of computational skills that are now being used in business and medical professions entering the Smart Grid profession," Anderson says. “We must see that kind of multi-dimensional skill set development [in engineering schools] as well.” 
500 W. 120th St., Mudd 510, New York, NY 10027    212-854-2993