In September 2016, gale force winds in the state of South Australia damaged major transmission lines. But the storm wasn’t the only cause of the subsequent blackout affecting 850,000 customers, almost the entire state. It was a cascade of effects that occurred, in part, because the state’s renewable energy sources didn’t know how to handle unusual operating conditions.
Dominic Gross, a new assistant professor in the University of Wisconsin-Madison Department of Electrical and Computer Engineering, is working to make sure that similar problems don’t occur as more and more green energy is added to power grids across the world.
Gross, who received his doctorate from the University of Kassel and spent four years as a postdoc at ETH Zurich, says for decades power generation has been dominated by large synchronous machines that connect fossil fuel generation to the grid. While those sources are vilified for their carbon footprints, they do have some advantages over renewables. “Their behavior is homogenous and very predictable,” Gross says. “Roughly speaking, all of these devices behave the same.”
That means power system operators can control how much electricity is produced at any given time. Moreover, synchronous machines instantaneously balance generation and demand by using their rotating mass as an energy buffer and autonomously adjust their power generation to keep the grid frequency and voltage steady.
Grid operators do not have that kind of control when it comes to renewable generation, like solar and wind, which produces energy more erratically and requires a steady grid frequency to lock onto. “The challenge becomes how to integrate all of these new sources and how to control this system where you go from a couple of hundred or thousand generators to millions of devices that need to contribute to reliably operate the grid,” Gross says. “You have this problem of variability—so you don’t know exactly how much energy we are going to have tomorrow.”
But Gross’s research looks at the problem at much faster time-scales; he’s interested in how to keep the grid functioning over the course of seconds or milliseconds when there may be disruptions. “The way they are operated right now, renewables just inject whatever power they’re generating and they don’t care about grid stability,” he says. “On the millisecond scale or second scale, we need to balance power generation and demand to maintain grid stability.”
The problem in South Australia shows just how critical seconds can be when it comes to grid stability and why the power electronics and algorithms that connect renewable energy to the grid need a redesign. When the large wind farms that produce much of the electricity in the area detected a problem, the turbines powered down, dropping 450 MW of electricity generation—a quarter of the state’s production—in just 7 seconds. That disruption caused the South Australia grid to begin pulling large amounts of electricity from traditional power sources in the neighboring state of Victoria. And that surge tripped the transmission line connecting the states, leading to the massive blackout, which lasted two days in some areas.
Gross says that the shift to renewables is happening fast, meaning there’s an urgency in figuring out how to better integrate photovoltaic and wind power into existing systems. All over the world, from Ireland to Australia, grid operators are already seeing more erratic system frequencies due to renewable integration, leading to expensive emergency interventions and, in worst case scenarios, blackouts. “These changes are happening a lot more quickly than people thought they would,” Gross says. “When I started working on this four years ago, some said these problems were 20 years off. And then a year later we had the first blackout because of this.”
Gross is currently investigating ways to control the electricity from renewable sources so they can pump energy into the grid in a way that helps to actively stabilize the network instead of disrupting it. “This is where my work comes in, redesigning these controls so that all these decentralized generation sources actually form the voltage waveform at their point of connection,” he says. “This is called grid forming control, which basically means that they can bring up the grid and they don’t require a big power station that they can lock onto.”
UW-Madison is a great place to pursue his research. “You have this enormous expertise on the power electronics side here and also power system expertise,” he says. “It fits very well with what I’m doing. There’s a chance here to work with both sides and move this forward.”