Unless U.S. energy policy and industry practice is systemically shaped to intercept and exploit the exponential improvements in clean-energy technology and cost reductions now occurring, the United States could end up with the worst of all situations by 2040: a dystopian grid where energy costs are high and reliability is poor, decarbonization progress is stalled, and the economic gains that have been made over the last century are at risk.
That’s a central premise of Energy 2040: Aligning Innovation, Economics and Decarbonization by Deepak Divan, professor and the founding director of the Center for Distributed Energy at the Georgia Institute of Technology, and recipient of the 2024 IEEE Medal in Power Engineering, and his coauthor Suresh Sharma, a former General Electric executive and the entrepreneur in residence at Georgia Tech. The book explores how new sources of energy are disrupting long-held beliefs and assumptions on how energy should be generated, transmitted and distributed. In the following interview IEEE Spectrumcontributing editor Robert N. Charette talks with Divan about how to align economic imperatives and climate goals for sustainability and affordability.
One of the fundamental themes of your book is that the technological learning curve that has resulted in the rapid reduction in the costs of renewable energy has been sustained for 50 to 70 years and shows no signs of slowing down. You also write that these declines were not predicted by experts in the field just two decades ago. What do you mean by the technological learning curve? What did you find in terms of cost reductions in different types of renewable energy as a result? And why were the experts so wrong in their predictions of renewable energy costs?
Deepak Divan: The technological learning curve is at the heart of our book. We spent a lot of time in the beginning of the book going through the history of why we are where we are because it is important to understand the process and nuances of how we got here. It is quite complicated, but I’ll try to simplify it.
We start at a place where science lagged technology and the market by a significant amount in the early years of the power industry. In other words, the processes of taking technology to market through innovation, through tinkering, through entrepreneurs who were willing to invest, helped create the underlying structure of today’s utility industry.
When the electricity grid was established, it was the Wild West, with every entrepreneur trying to get ahead of the others with their own proprietary solutions. However, it soon became clear that the grid, which was not just a single device but a physically coupled network of a large number of devices, needed to be coordinated and controlled as a whole—very different from most previous technological innovations. Everybody’s appliances needed to work with the same voltage and the same frequency, for instance. So, electricity providers were forced to make everything work seamlessly—challenging in a world before microprocessors and power electronics.
Yet at the same time, the early electricity providers also focused on where the money was, so they ended up targeting those pieces of the market that had best return on their investments. As a result, big, broad swaths of the country, typically rural, were being left in the dark. This helped create the Public Utility Holding Company Act of 1934 that forced more regulation on the electricity industry. It also promised utilities better and more stable economic returns, but in exchange for providing universal access, and so we end up getting the grid that we have right now.
I keep thinking that Elon Musk should not be worrying so much about autonomous cars today. Give me an autonomous inverter first.—Deepak Divan
However, industry regulations also strongly influenced the way electricity providers thought. With the utility industry now regulated, it was not possible to bring innovation to market very easily. Reliability was the most important objective and any new technological innovations that might reduce reliability were frowned upon. As a result, it took 10 to 20 years to bring new technologies to market.
So, the electricity industry went from a fast-moving, risk-taking one to an industry that was very, very slow moving, very risk averse. That was fine as long as technological innovation was also moving slowly.
Over the past two decades, however, something radically changed. Traditional learning curves, where one gained experience over time and the product or service cost went down a modest amount until the next S-shaped learning curve began, started to disappear. Instead, the learning curve across many energy-related technologies and their resultant cost reductions started to happen without much notice over decades seemingly without limit, with few indications when they will ever saturate.
We’ve seen this in microelectronics ad nauseam, for example. We have also seen it reflected in the photovoltaics space, where the learning curves began in the early 1970s. Since that time, there have been hundreds of technologies that have intersected and interconnected to create a 23 percent reduction in price for each doubling in sales volume with no signs that it’s going to slow down.
The same kind of curve is occurring in the battery space because, again, it is micromaterial-based and multiple new smart materials all coming together to give you both more kilowatts and kilowatt-hours. The battery market is now tasting success; it is attracting huge investments, again, with no signs of slowing down.
Why did no one in the energy industry see this coming?
Divan: If we go back to around the year 2000, at that point, solar LCOE (levelized cost of energy) was $850 per megawatt-hour, and batteries were $1,200/KWh. There was nobody in their right mind who thought that that would ever become competitive with gas and coal sitting at around $35/KWh and $50/KWh.
No one believed that the learning rates in solar or batteries, for example, could be sustained. Everyone in the industry thought that the technology was gimmicky and was not really going to be able to scale. After all, solar panels are small little things. How could you compete with a 500-megawatt gas plant?
Additionally, the utilities all used similar 20-year integrated resource planning cycles. So, they were already making investments in terms of what needed to be done and there was not a consultancy in the world who was willing to advise them to say stop everything you’re doing and let’s start moving towards solar. There was no rational basis for that.
The energy industry also believed they had so much economic and policy clout, they could hold off any threat from renewable energy forever.
I do not think the transition to renewables and EVs can be stopped, but I think it can be made extremely messy. —Deepak Divan
A former CEO for PJM, the biggest grid operator in the United States, told me that even in 2010 there was not a single CEO of a grid operator, electric utility, automotive or oil company who thought that electric vehicles, solar power or batteries were going to be cost competitive any time soon.
But by 2015, new energy companies were disrupting energy incumbents’ long-held assumptions. This was reflected by an astounding 97.5 percent reduction in the cost of solar from 2000 to 2022, and this is installed cost! And similarly with batteries, there has been a 92 percent cost reduction over the same period that is just continuing because there are so many new technologies being brought into play.
As to why the biggest companies in the world that are responsible for a huge part of global GDP, have the smartest people in the world and are advised by the smartest consultants in the world, could not see this coming is a fundamental question that we have asked in the book.
One of the implications you discuss is that the distributed energy resources, or DERs, like solar power, windmills and large-scale energy-storage systems are going to change the electric grid from a synchronous generator and inertia-driven system to an inverter-based resource (IBR) rich grid where grid voltage and frequency are not regulated by inertial sources. Can you explain the difference, why and what needs to happen both from a technology perspective to move to a decarbonized IBR grid?
Divan: Getting to an inverter-based grid is one of the things that the industry is struggling with on the technology side. Fundamentally, the existing grid is electromechanical in nature.
There are these big, rotating, energy-generating turbine-driven synchronous machines, and over 100 years we have figured out how to make them work to make the grid reliable. All the simplifications and efficiencies, all the standardizations and designs and synchronous generators that were needed have been figured out and now there is a system that works reasonably well. The grid that has been built in the United States has been called the largest machine ever built, with all these rotating machines possessing huge amounts of rotational inertia, all rotating together in lockstep because of the way synchronous machines operate.
When even a small disturbance occurs anywhere on the grid, all of them continue to operate locked together and to share the power delivered, with the ability to clear any faults as they occur on the system. The entire system is structured around this model. While it is often called a smart grid, there’s nothing smart about it. It’s an extremely good grid but it’s really a passive grid. All the smarts are sitting 15 minutes away at the operator level. So, for 15 minutes, the system has to keep operating until the next command is received.
This enormous machine has several interesting characteristics that make it work well. One is that the grid has a lot of damping built into it. Anytime there is a deviation because of a disturbance on the system, there’s a restorative torque that automatically occurs on it.
Another characteristic is that it is usually thought that frequency is the universal parameter on the system, since all the generators essentially use a power-frequency droop principle to share power equally. However, the problem is that in the synchronous generator world, frequency command is a DC quantity, while the three-phase AC voltages are generated and locked in by the machines’ action itself, not by control action.
Now, as synchronous generators are replaced with inverters, you don’t have any intrinsic rotation or inertia in the system. We don’t have any of the attributes of damping that are automatically built into it. Further, there are now inverters with DSPs [digital signal processors] and FPGAs [field-programmable gate arrays] which allow the measurement of the grid voltage and to act very, very quickly.
For the first time in our history, decarbonized climate-friendly solutions are also lower cost than traditional fossil-energy-based solutions. For the first time ever, what is good for our wallet is also good for the planet! —Deepak Divan
In the early years and all the way until very recently, we only built what we call grid-following inverters. Essentially, the voltage of the grid was taken as given and power was pushed against it. The inverter followed the grid and power could be dispatched per utility command, which worked fine. This has allowed us to scale IBRs in many locations around the world.
The difficulty is as one gets to high penetration of inverter-based resources, the grid is no longer being formed nicely, and so the system can become unstable.
Now there is a need to start thinking about how the grid is going to be formed when we have an inverter-dominant grid. The issue is that one does not have that rotating machine, one doesn’t have that restorative torque, and one doesn’t have the system damping. None of those things are there.
Each inverter thinks it is very smart and it’s going to try to form the voltage based on local information. However, it is also going to have to interact with what another inverter is trying to do to form voltage and what another inverter is doing, and so on. This becomes a problem.
So as these inverters interact with each other, it’s often hard to keep them stable. While we have been able to demonstrate grid-forming inverters and every manufacturer now claims to have one, we do not exactly know what a grid-forming inverter should do, especially when done at scale, to ensure that they do not interact with each other, particularly when millions of inverters are deployed. This creates a challenge.
There is also the concern that each of these inverters is made by a different manufacturer. Some of them were made 20 years back, some were made 10 years back some and these now need to be compatible with what will be made in the next 10 years. They are no agreed standards. Standards are lagging by 10 years or more.
The question is what does one do, if it takes you 10 years to get a new standard out, and given that the rate of solar deployment is so high that in that time some 1,000 gigawatts of PV solar will be deployed, but none of it will be compliant with the future, as yet unknown, standard?
How do you also stabilize the grid in this environment?
Divan: The utilities today have grown up without having to worry about any of these issues. They just focused on how to restore power, how to connect this to that, how to manage the workforce, and so on. Not this dynamic beast which they have few skills in dealing with. In fact, most big electric utilities have few people in their workforce who are skilled in power electronics, because the old system did not need it.
These are very complex issues and part of the challenge is that it is a different operational paradigm than today. We do not have these fundamental issues resolved. The important question, I think, and part of the problem is that nobody can stand in public and say, “Hey, there’s a problem here!”
I keep thinking that Elon Musk should not be worrying so much about autonomous cars today. Give me an autonomous inverter first. That is a much, much more important priority in the near term.
In the book, you were careful to also lay out factors that could derail your energy vision for 2040. Could you discuss a few of them, and what might be done to avoid or minimize them?
Divan: I do not think the transition to renewables and EVs can be stopped, but I think it can be made extremely messy.
Major energy transformations have taken 50 to 70 years, and they have been very messy from a regulatory standpoint. People are pushing back against going to renewables, but I do not think they can win because at the end of the day, everybody is going to respond to the economics and functionality of inexpensive renewables and new holistic solutions.
Even if we in the United States do not do it because of the politics and incumbent resistance, the Chinese and others are going to continue to move the technology along and to drive the prices down. And so you know, you’re going to at some point say, oh, ****, I think we have to adopt this new stuff, because it’s going to seep into widespread use. By then, I am concerned that we will have been left behind.
Nobody can argue with economics of renewables in the future; it is going to drive everything. However, if you do not think about the economics and government policies properly together, they will drive bad outcomes. —Deepak Divan
Another issue that could make things messy is that the utilities do not have the ability to change easily. They must meet their reliability requirements in the near term, which becomes problematic when all these new technologies are coming in. They are not going to absorb these technologies easily.
In addition, the energy load is moving in. Data centers, especially those for AI, are coming online, as well as electric-vehicle charging, heat pumps and green hydrogen. How do you meet those requirements?
It is tempting to say, “Let’s go back to the old days and fire up the gas and coal plants.” While that is not the answer, that is something that easily could happen.
The point I am trying to make is that I do not believe this energy transition can be stopped, but it can be made extremely expensive. It can be made extremely messy and then we will have lost the climate battle at the same time. But it does not have to be so! For the first time in our history, decarbonized climate-friendly solutions are also lower cost than traditional fossil-energy-based solutions. For the first time ever, what is good for our wallet is also good for the planet!
Nobody is laying the difficulties out. Nobody. The hope with writing this book was to start this conversation because we are not seeing anybody addressing these issues holistically.
Unfortunately, most people are unable to act on something that has a long-term benefit but is more expensive in the near to midterm. They will only act in the short term. So, you have to give them a short-term reason for doing something by making it the attractive thing to do financially.
This is very important in my mind. Nobody can argue with economics of renewables in the future; it is going to drive everything. However, if you do not think about the economics and government policies properly together, they will drive bad outcomes.
Who do you hope will read your book, and what are the two or three fundamental messages they should take away and, more importantly, act on and when?
Divan: I think the audience is everybody who is interested in energy in general, including researchers, engineers, policymakers, investors, entrepreneurs and students. People are interested in the topics we raise. Every time I go into a room, I have six people approach me and want to talk about it. They are reading something in the news, and they have only a narrow sliver of information. They are not able to connect all the dots together.
I think part of the problem is that this field is very complex and very nuanced, and when you try to simplify it, you can get to the wrong conclusions. My objective for writing the book was that we really do not hear this line of conversation in the industry. In other words, a holistic view of the problems confronting the industry is required because everything you do intersects with something else.
The utility industry does not fully understand this. When I go to the IEEE Power and Energy Society general meeting, I go to every conference room and I ask a question about the dynamics and scaling of IBRs and distributed systems. Nobody has an answer. This is scary. I mean, this whole industry is there, and they’re absorbing gigawatts after gigawatts of renewable energy and don’t have any idea what the hell is going happen as we move to a distributed energy resources dominant zero-carbon grid (which EPRI has also set as the target for 2050).
Again, oversimplifying is going to lead us to the wrong place, not looking holistically is going to lead us to the wrong place. We have an opportunity where we have alignment between economics and decarbonization for the first time. Let’s not blow it.
This article was updated on 10 July 2024 to correct the units in solar LCOE in 2000 to US $850 per megawatt-hour instead of per kilowatt-hour.
- Engineers: You Can Disrupt Climate Change ›
- How the Green Energy Switch Will Affect Engineers' Jobs ›
- The EV Transition Is Harder Than Anyone Thinks ›
