March 31, 2015

Adding renewable energy to power grid requires flexibility


Solar panels, wind turbines, electric vehicles and other green power sources are proliferating rapidly, but their reliable integration into the existing electric grid is another story.

A study led by Eilyan Bitar, assistant professor of electrical and computer engineering, offers a comprehensive reimagining of the power grid that involves the coordinated integration of small-scale distributed energy resources. The study, commissioned by the Power Systems Engineering Research Center (PSERC), asserts that the proliferation of renewable energy must happen at the periphery of the power grid, which will enable the local generation of power that can be coordinated with flexible demand.

Bitar’s study outlines a new architecture to enable what he calls a grid with an intelligent periphery – a version of the so-called smart grid – along with coordination strategies and mathematical models to simulate how such a reorganized grid would work.

“The uncoordinated proliferation of distributed energy resources will wreak havoc at scale,” Bitar said. “Certain components of the legacy power system will fail; the existing distribution infrastructure isn’t equipped to accommodate, for instance, a large number of electric vehicles plugging into the grid at the same time under the same transformer … but, imagine taking all these new resources and coordinating their control.”

The way the power grid works now, large plants deliver power to substations, where electricity is provided on demand to homes. The fluctuating electric loads at each substation interact over a complex transmission network, but nearly everything that happens below the substation level is left uncoordinated. Electricity is an example of a commodity with inelastic demand – in the U.S., people are used to having it all the time, whenever they want. Bitar thinks that mindset will need to change slightly.

All the talk of the smart grid wouldn’t be nearly so complex if solar and wind, for example, were a reliable supply. But those resources are variable, and the gaps must be compensated by traditional bulk power generation, effectively defeating the purpose of a renewable source.

In an intelligent grid, this variability in supply would be balanced through the coordination of flexible distributed energy resources at the periphery of the system. Power would be produced locally and consumed locally, giving rise to self-sufficient communities or cities, called microgrids. Such an approach would decrease the need to transmit bulk power hundreds of miles to counterbalance fluctuations in renewable sources.

The architecture of such a system, which requires sensors and actuators in appliances, electric vehicles and the like, isn’t the hard part, Bitar said. The hard part is the design of algorithms to efficiently manage the deluge of information produced by those sensors in order to coordinate the simultaneous control of millions of distributed energy resources on fast time scales.

Much of this coordination will involve using flexibility in demand to compensate for variability in supply. For example, electric vehicles, considered a viable alternative to internal combustion engines, need to plug into the grid in order to charge. But because they’re battery-based, the demand for their charge is flexible – for example, utilities could offer monetary incentives to electric vehicle owners willing to shift their charging patterns. Coordinating that flexibility with a variable resource like rooftop solar could lead to increased penetration of this renewable energy resource while supplying clean power to electric vehicles.

PSERC, of which Cornell is a member and also its founding institution, is a consortium of universities and industry partners working to solve emerging problems in the U.S. power system. Bitar presented a subset of the research results at the IEEE Conference on Decision and Control in December 2014.