Many challenges face the effort to green the grid. Some of the most familiar include the variability of wind and solar resources and the need to assure generation from other sources when they are not available; the current technical limits on storage batteries; the inadequacy of existing transmission networks to connect utility-scale renewables in sparsely populated areas to electricity users in commercial, urban, and industrial centers; access to rare earth metals and other materials used in solar panels, batteries, and other new technologies; and uncertainty around whether electric vehicles will take or contribute power to the grid and under what scenarios. Importantly, however, the nature of the power system is changing and the underlying logic for controlling and optimizing electric power may not work well in the future.
These challenges are less familiar outside the world of engineers, likely because they are so highly technical and involve lots of equations. For example, in control rooms around the world, this equation is front of mind:
ACE = (Ta – Ts) – 10B(Fa – Fs)
But explaining what it means involves concepts of frequency and load control, balancing areas, control errors, and more—too many details to engage in viable policy discussions. Yet the viability of this fundamental control algorithm itself is at stake in the grid of the future. Perhaps it is time for policy-thinkers and policymakers to observe the efforts of power-systems engineers to tease out technical approaches that will make sense on next-generation grids. At the heart of the research is the shift from spinning machines to power electronics.
Interconnected power systems that developed throughout the 20th century depend upon mechanical spinning machines that generate electricity and transmit it across transmission lines. The underlying arrangements for sharing and controlling power on these networks rely on the ability to add or subtract power from systems very, very quickly—assured, in part, by the availability of spinning reserves. In addition, operators keep the lights on at the right voltage and at the right frequency through automated adjustments to generators that take place within seconds of frequency or voltage deviating from plan. If a frequency or voltage deviation is too large, it will affect system stability, as well as equipment on the network, and could result in a blackout. The system itself sends signals to each machine to speed up or slow down automatically, thereby maintaining stability.
The above equation captures this concept; it measures the difference between planned and actual system behavior and power sharing and introduces a variable that allows individual systems to help each other maintain stability across a network. Importantly, human power-system operators can call for further adjustments as needed—for example, last year during Winter Storm Uri, when frequency dropped below a safe number, grid operators in Texas, Louisiana, Oklahoma, and other states called for load shedding to avoid a total collapse of the grid.
Over the past 40 years, we have added new types of devices to deliver and take power to and from the grid—solar panels, wind turbines, storage batteries, electric vehicles, and so forth. Such devices interface with the system through an apparatus called an inverter. Other than wind turbines—which do spin but are limited in particular ways—the generating sources are not spinning machines, and the inverters are composed of semiconductor-based power electronics—also not spinning machines. Today, on large interconnected systems, inverters are “grid-following.” That means they watch out for and follow the voltage that is currently set by large spinning generators. But as more and more renewables penetrate power systems, and the share of traditional generators declines, so too will the presence of actual spinning machines decline. At some point, the stability and the reliability of the grid may be at risk.
“Grid-forming” inverters—that is, inverters programmed to establish and maintain frequency and voltage—will likely replace spinning machines as the dominant apparatus on grids in the future. But engineers do not yet know how to make a variety of grid-forming inverters work and play well together. This will be a transition of scale as well as type of apparatus, moving from thousands of generating units connected to the grid to millions of connected power electronics devices. Further, inverters will come from many different manufacturers and connect many different types of power-generating and power-using facilities to the grid, with no standards yet in place. This is a chaos scenario engineers fear.
Anticipating this future, the U.S. Department of Energy issued a request for proposals in late 2020 for “a new consortium dedicated to developing control technologies for a modernized electric grid,” and now the Universal Interoperability for Grid-Forming Inverters (UNIFI) Consortium, led by the National Renewable Energy Laboratory (NREL), is underway. On July 21-22, 2022, the UNIFI Consortium met to contemplate technical control of power on future electrical grids. Eighty engineers—and one energy historian who has written about the grid—gathered in person for the first time to discuss a 5-year initiative to bring order to the potential chaos of the next-generation power grid. Ben Kroposki, the consortium’s organizational director, explained that a successful initiative will facilitate a shift from the 20th century’s synchronous machine-based grid to a 21st-century inverter-rich grid. The goal of the work: to ensure that grid-forming inverters, from any vendor and connected to any device, are interoperable and will sustain a stable, reliable, economical, and secure power system.
Figure 1 — UNIFI Consortium in Golden, Colorado, on July 22, 2022
For two days, presenters covered the purpose of the initiative, the organization of the consortium, the objectives of each segment of the initiative, and case studies of small systems that are already integrating and testing grid-forming inverters. Attendees included academic engineers, power system operators, U.S. Department of Energy representatives, vendors, and utility engineers. Individuals hailed from every continent except Antarctica. In other words, although this initiative is U.S.-based and U.S.-focused, it is truly international in scope and potential influence.
To set the tone, Deepak Divan, the John E. Pippin Chair Professor and director of the Center for Distributed Energy at Georgia Tech and a Georgia Research Alliance Eminent Scholar, stated unequivocally that our power system is on the cusp of fundamental transformation. He asked how the assembled diverse group could reach consensus on universal concepts, on terminology, and on how to test, validate, and disseminate their findings.
Over the course of the meeting, the UNIFI Consortium participants described each of the major thrusts of the initiative: modeling and simulation, controls, hardware, integration and validation, demonstration projects, standards development, intellectual property management and domestic products, education, workforce development, and communication and events. While the topics sounded familiar to this non-engineer, often the details floated into the realm of acronyms and equations that were entirely foreign. The non-engineering subjects, however, generated the most energetic discussion, especially intellectual property and patents, education, workforce development, and communication. More significantly, Ryan Quint, the director of engineering and security integration for the North American Electric Reliability Corporation, drew the participants’ attention to the need to inform and coordinate with regulators and policymakers. Once the UNIFI Consortium agrees on guidelines and standards, the policymakers and regulators will have to design markets and rules that embrace those concepts.
On Day 2, the consortium members spent the afternoon touring the NREL labs, both at the Golden and Flatirons Campuses. At the latter, they observed computer simulations of grid-forming inverter controls during various types of system interruptions. On the way out, one of the hosts pointed to a notebook beside one of the test engineers. On it, someone had written out a more extensive version of the ACE equation:
Figure 2 — Hand-written ACE Equation
The test engineers explained that prior to our arrival, on an unrelated matter, they had been discussing how contemporary frequency control works on the grid. Indeed, algorithms for power control will be at the crux of the UNIFI Consortium’s work and should be on the radar for those concerned about a sustainable power system in the future.
During the last century, engineers, system operators, academics, manufacturers, politicians, regulators, and customers competed, collaborated, and negotiated to develop the grid as we know it today. Development of the algorithm for frequency and load control (the ACE equation) that allows for the world’s largest interconnected machines to work together harmoniously took place over the course of three decades. Historical research illustrates the complexity and contention that was inherent in the process. Today, time is short, but the past will frame how the next generation of algorithms unfold.
In the relatively near future, the legacy infrastructure of our power grid, including the standards and approaches for optimization and control, will face a transition. This will not be an evolution from a simpler to a more complex interconnected system. It will be an evolution from one set of complexities to another, matching the dynamic behavior of electricity itself. As ideas emerge for future power markets, and future rules to ensure stability, reliability, and resilience on the grid, it will be important to consider technical implications: How might reliability standards limit or encourage integration of inverter-based resources? How might the presence of grid-forming inverters upend the viability of markets? And how quickly will these become pressing matters as investors develop more wind and solar generation, politicians incentivize electric vehicle purchases, and manufacturers design and sell their own versions of power electronics? Let’s keep an eye on the UNIFI Consortium and similar initiatives around the world, and look for the opportunities to speak across the technology/policy aisle.
 ACE (Area Control Error) is the difference between how much power is meant to be exchanged between subsystems on an interconnection and how much is actually exchanged—ideally zero. (Ta – Ts) represents the difference between actual and scheduled power flow over a tie line. 10B(Fa – Fs) represents the difference between actual frequency and intended frequency (60Hz in the United States), multiplied by 10 times the frequency bias. The frequency bias, B, is a variable related to the energy of inertia of the rotating masses on the system, and determines how much one subsystem will assist another to restore stability when the frequency changes.
 See Peter R. Hartley, Kenneth B. Medlock III, and Elsie Hung, “ERCOT Froze in February 2021. What Happened? Why did it Happen? Can it Happen Again?” (working paper), Rice University’s Baker Institute for Public Policy, Houston, Texas.
 U.S. Department of Energy, “Solar Energy Technologies Office Fiscal Year 2021 Systems Integration and Hardware Incubator Funding Program,” https://www.energy.gov/eere/solar/solar-energy-technologies-office-fiscal-year-2021-systems-integration-and-hardware.
 Universal Interoperability for Grid-forming Inverters (UNIFI) Consortium, https://sites.google.com/view/unifi-consortium/home.
 Julie Cohn, The Grid: Biography of an American Technology (Cambridge, MA: MIT Press, 2017).
 For presenter slide shows, including the author’s, please visit https://sites.google.com/view/unifi-consortium/blog.
 Cohn, The Grid: Biography of an American Technology.
 See Julie Cohn, “Bias in Electric Power Systems: A Technological Fine Point at the Intersection of Commodity and Service,” in Electric Worlds/Mondes électriques: Creations, Circulations, Tensions, Transitions (19th–21st C.), eds. Alain Beltran, Léonard Laborie, Pierre Lanthier, and Stéphanie Le Gallic, NED-New edition (Peter Lang AG, 2016) 271–94, http://www.jstor.org/stable/j.ctv9hj6hk.15.
This material may be quoted or reproduced without prior permission, provided appropriate credit is given to the author and Rice University’s Baker Institute for Public Policy. The views expressed herein are those of the individual author(s), and do not necessarily represent the views of Rice University’s Baker Institute for Public Policy.