by Jeff St. John
August 07, 2020

One of the most complicated challenges of a 100 percent renewable power grid is how to replace the inertial stability provided by the spinning generators that the modern grid is built to serve.

The kinetic energy of these massive rotating machines works like a shock absorber to keep grid frequency from dropping too fast when demand exceeds supply or rising too fast when supply exceeds demand. Without this stabilizing force, power grids could face a greater risk of frequency excursions that could force generators offline or cause cascading outages like the 2003 blackout that affected about 50 million people across the Northeast U.S. and Canada’s Ontario province.

Solar panels, wind turbines and batteries, by contrast, use inverters and power electronics to convert direct current (DC) output to alternating current (AC) at the frequency of the grid they’re connected to, either 60 hertz or 50 hertz. While these inverters can react nearly instantaneously to alter grid frequency, they lack the inertial connection to grid power that makes spinning generators so valuable as stabilizers. 

Setting up inverters to augment or mimic that inertial stability is a highly technical challenge. “We’ve designed and controlled the system to take advantage of the characteristics of synchronous generators,” explained Daniel Brooks, vice president of integrated grid and energy systems at the Electric Power Research Institute.  

“When you start to displace those synchronous generators with power electronics-based, inverter-based resources, now you’ve got a problem with the way you control the grid for reliability reasons, because the inverters have an entirely different set of characteristics and capabilities.”

Hitting the 100 percent renewable goals now in place in many states and countries will force utilities, grid operators and regulators to solve this problem eventually. But many wind-power-rich transmission grids are starting to face the problem already — and coming up with solutions. 

Primary frequency response and "synthetic inertia"

Advanced inverter capabilities are already serving essential grid reliability needs on the wind and solar-rich transmission system of Texas grid operator ERCOT, for example. “ERCOT recognized that the amount of wind coming onto its system was displacing central station synchronous generators, and became concerned about frequency stability,” Brooks said. In worst-case contingencies like several power plants being forced to go offline simultaneously, that could put the entire system at risk. 

So in 2012, ERCOT started requiring all new interconnecting generators, including wind and solar farms, to be able to provide “primary frequency response,” or be able to increase or decrease real power output immediately to stabilize frequency. “If you have energy behind the inverter, it will inject instantaneously into the system. That’s not the same as inertia…but it has the same effect of balancing the supply-demand imbalance.”

The result, according to this 2018 ERCOT presentation, has been a major improvement in its primary frequency capability, as well as reduced need for “secondary” frequency regulation services provided by batteries or fast-responding demand-side resources. 

Canadian grid operator Hydro-Québec TransÉnergie has taken a similar approach to managing its wind-rich system, but it uses the term “synthetic inertia” to describe how it has required wind farms to respond to frequency fluctuations since 2010. “They basically modeled the requirements to mimic what a synchronous generator could do,” Brooks said, albeit through the same fundamental action of increasing power output to stabilize under-frequency events that ERCOT demands of its generators. 

A 2015 transformer failure that forced offline about 1,600 megawatts of generation on Hydro-Québec’s 40,000-megawatt grid showed that this synthetic inertia capability was able to stabilize grid frequency as well as synchronous generators could be expected to do. But it also revealed the potential for a “double-dip” in frequency from wind farms recovering from this instant energy injection, leading to revisions to its requirements.  

From synthetic to “virtual” inertia: The “grid-forming” inverter 

All of these approaches are designed around a power grid that’s still having its primary frequency set by synchronous generators. But what happens when inverter-based generators are providing the majority of grid power? 

At the theoretical endpoint of this transition, inverter-based generators can shift from “grid-following” systems that match their operations to frequencies created by spinning generators to “grid-forming” systems that actively create the frequency of the networks they’re serving. 

That’s how global grid giant Hitachi ABB Power Grids approached the challenge it faced at the Dalrymple ESCRI (Energy Storage for Commercial Renewable Integration) project in South Australia. The 30-megawatt/8-megawatt-hour battery system sits at a substation serving several towns on a peninsula that also hosts about 90 megawatts of wind farms, with only a single connection to the mainland. The battery allows transmission grid operator ElectraNet to provide grid stability and prevent outages in case of lightning strikes or other disruptions to the line. 

The ESCRI project can also provide its services to the greater South Australia grid, which suffered a massive blackout in 2017 after storm-caused faults led to wind farms shutting down, driving grid operator AEMO to require wind turbines to alter their control settings to ride through similar situations in the future. Since then the region has seen some of the world’s largest battery projects, including Tesla’s Hornsdale project and General Electric’s Solar River system, come online to provide frequency support for the remote and wind and solar-rich grid. 

But unlike those grid-following battery systems, ESCRI “had to take an approach that wasn’t in line with how the larger interconnected bulk power systems work,” Stephen Sproul, a senior engineer at Hitachi ABB Power Grids, explained. Because the average 5 megawatts of demand from the peninsula is so much smaller than the 90 megawatts of wind serving it, “when we isolate from the grid, we have to drop off about half the wind farm instantaneously,” he said. 

“To transition from the grid to this islanded formation, we had to operate it in something we call grid-forming mode,” he said. The ESCRI battery’s inverter and Hitachi ABB Power Grids' “virtual synchronous machine” control system sets the frequency and voltage for the peninsula’s grid. Unlike the fast frequency response being provided in Texas and Quebec, this type of grid-forming capability requires inverters to act more quickly than measurement devices are capable of recognizing and responding to grid conditions, as the graphic below indicates. 

In this sense, the ESCRI system is acting much like an low-voltage microgrid using its own inverter-based generators.  Hitachi ABB Power Grids has multiple remote microgrids operating around the world, and the ESCRI project “is a good example of how we’ve applied some off-grid lessons to an on-grid project,” Sproul said. 

There are significant technical challenges to integrating grid-forming inverter operations into a larger power grid, he noted. “It’s a tradeoff. if it’s too desensitized, it’s not fast enough to form the grid, but if it’s too sensitized, it’s always fighting the grid.” But the ESCRI project does demonstrate that grid-forming qualities could aid in renewable integration.

Some early-stage uses could be replacing synchronous condensers, which are essentially unpowered motors linked to the grid to provide voltage stability for renewable projects seeking interconnection to power grids to ride through faults. “That’s an old technology being used in a new application, but it’s a large sunk cost,” he said. A battery connected to a grid-forming inverter, by contrast, could also use its stored power for various grid-serving, revenue-generating services, much as the Dalrymple system has done: “ESCRI has almost paid itself off in a couple of years.” 

Grid-forming inverters at transmission and distribution scale 

Just how to define “grid-forming” versus “grid-following” inverter operations is a subject of much debate among power systems engineers, EPRI’s Brooks said. 

At a high level, the difference can be boiled down to, “does the inverter resource depend on a strong grid in order to be able to provide its services? Or is it able…to be a voltage and frequency source even when there’s a disturbance that causes the grid voltage to drop very low?” 

Creating the modeling and control architecture to put "grid-forming" inverters to use in the broader grid control systems that manage multiple power plants and other grid-stabilizing resources is a technically challenging proposition, he said. But it could be very helpful in “weak grid” conditions, such as those faced by grid operators in sections of their transmission systems where renewable energy outweighs that provided by traditional generators. 

Over the past two years, EPRI has been working with ERCOT and the Southwest Power Pool, the grid operator operating in the U.S. Midwest from South Dakota to Oklahoma and the Texas panhandle, and transmission owner American Electric Power, on ways to determine how the region’s growing wind farms could help set frequency and voltage on stressed portions of the grid. 

“There are places in the system where if you interconnect an inverter-based resource, that resources is going to have the potential to cause instability,” because “they’re not able to set grid frequency and voltage when there’s a problem.” One solution “is to re-tune the programmed response of those inverters to support those systems, rather than to cause a problem.” 

In a nod to Hitachi ABB Power Grids’ work on microgrids, Brooks highlighted similar efforts to enlist inverters to stabilize low-voltage distribution grids. EPRI is working on an Energy Department-funded project called “Solar Critical Infrastructure Energization system,” or Solace, built around the concept of a grid-forming solar PV inverter. 

Partners including the University of Texas Austin, utility Austin Energy, nonprofit Pecan Street, inverter maker Yaskawa Solectria Solar and grid technology giant Schneider Electric will collaborate on the project, which will eventually target the rooftop solar and battery-equipped Mueller neighborhood in Austin, Texas as a test bed. 

“The project is basically looking at how you can use distributed resources, primarily inverter-based resources, to provide an increased resiliency to critical infrastructure on the system,” Brooks said. In essence, it’s an attempt to replicate a stand-along microgrid across utility-controlled distribution circuits, using grid-forming inverters to provide the source of frequency and voltage to keep the system energized during wider outages. 

The same system could be used to help restart the grid, he noted. A similar project by U.K. grid operator National Grid is exploring the potential for distributed energy resources to assist in what are termed “black-start” operations. Of course, this interaction brings the “added complication of making certain that the provision of that service up to the bulk system doesn’t result in an operating problem on the distribution system,” such as a burst of transmission system frequency-stabilizing power overloading low-voltage circuits.

But from a conceptual standpoint, the megawatt-scale inverter operating the ESCRI system and the much smaller inverters connecting rooftop solar systems to the distribution grid aren’t that different, he noted.

“When you’re talking about a microgrid or an islanded portion of the system…those distributed resources end up looking like bulk system resources.”