Energy Storage’s Seasonal Challenge

President-Elect Joe Biden’s clean energy plan aims to make America’s electricity system carbon neutral by the year 2035. To reach its goal, the plan will seek to develop the nation’s clean energy infrastructure, and expand the role of wind and solar power. Yet renewable energy presents certain challenges, one of which is to ensure that electricity is available even when wind and sunshine are scarce.

In recent years, grid-scale batteries have emerged as an increasingly economic way to address the variability problem, or intermittency, of wind and solar output. In fact, over the last two years demand for grid-scale energy storage has accelerated, particularly in the Southwest, where batteries are increasingly used to balance daily ebbs in solar generation.

Yet as renewables become a larger part of America’s energy mix, the challenge of balancing intermittency will grow exponentially. Eventually, storage could be called upon not only to even out daily fluctuations in energy output, but seasonal variation as well.

Kleinman Center research associate Oscar Serpell explores the potential for grid electricity storage, in its many forms, to meet the seasonal balancing demands of a low-carbon electric grid. He also looks at the limitations of today’s energy storage technologies, and at the advances that may be needed to enable dramatic reductions in carbon emissions from the electricity industry.

Full Transcript

Andy Stone: Welcome to the Energy Policy Now podcast from the Kleinman Center for Energy Policy at the University of Pennsylvania. I’m Andy Stone. President-elect Joe Biden’s Clean Energy Plan aims to make America’s electricity system carbon neutral by the year 2035. To reach its goal, the plan will seek to develop the nation’s clean energy infrastructure, and expand the role of wind and solar power. Yet renewable energy presents certain challenges, one of which is to ensure that electricity is available even when wind and sunshine are scarce.

In recent years, grid scale energy storage has emerged has one option to address the variability of wind and solar output. In fact, over the last two years, demand for energy storage has accelerated, particularly in the southwest where batteries are increasingly used to balance daily ebbs in solar generation. Yet, as renewables become a larger part of America’s energy mix, the challenge of balancing intermittency will grow dramatically. Eventually storage could be called upon not only to even out daily fluctuations in energy output,  but seasonal variation as well.

On today’s podcast, we’ll explore the potential for grid electricity storage in its many forms to meet the seasonal balancing demands of a low carbon electric grid. We’ll also look at costs, and at alternatives to storage. My guest is Oscar Serpell, research associate here at the Kleinman Center whose recent work has examined the potential for seasonal energy storage. Oscar, welcome to the podcast.

Oscar Serpell: Thanks so much for having me, Andy.

Stone: It’s great to have you on here finally after so many years working together.

Serpell: Yeah.

Stone: So let’s go ahead and jump into this. So we hear a lot these days about electricity storage. Can you give us a quick overview of why storage has drawn so much attention and come to the fore?

Serpell: Sure. So I would say that storage is really providing two related but distinct services, both of which are really contributing to its importance on the energy transition. The first is that storage is allowing for the distributed or wireless use of electricity in a way that was really not possible until fairly recently. This allows us to carry around smartphones, these pocketable computing devices that would have just a couple of decades ago really passed for supercomputers. And storage now allows us to carry those devices around all day on a single charge.

And in the same vein, storage has allowed for a revolution in transportation. Automobiles in the US account for about a third of all carbon emissions. And the only reason that we’ve been able to create competitive electric vehicle alternatives is because of recent progress in storage technologies. I mean, without advanced lithium ion cells, you’d have EVs that could really only travel a fraction of the distance on a charge.

The second service that storage offers us is load balancing. Being able to move electricity basically through time from the moment it’s generated to a moment when it’s really most needed. For all of their environmental benefits, renewable energy sources such as wind and solar, they do have their drawbacks. Specifically, they will always to some extent put us at the mercy of mother nature, no matter how efficient or cost competitive they become.

At a given location, the sun does not always shine and the wind does not always blow. And this means that renewable energy is in some ways kind of inherently less flexible than fossil fuels. But of course that doesn’t make our transition towards renewables any less important or necessary. So because of this inherent intermittency of renewables, the larger share that wind and solar make up in our energy mix, the more variable our supply of electricity is going to become.

California is a great example of this, where the introduction of lots and lots of solar to the regional grid has created a phenomenon known by many as the duck curve, in which the capacity factor of solar, which is basically just the percentage of a solar panel’s nameplate capacity that’s actually being generated at any given moment, that changes over the course of a day as the sun rises and sets. And it reaches a peak sometime around midday, early afternoon.

Meanwhile, demand for electricity often peaks in the late afternoon between sort of four and six p.m. This is when everyone is getting home from work, or at least under non-COVID circumstances that’s when everyone would be getting home from work, starting to do chores around the house. And unfortunately this coincides with solar radiation levels beginning to dip as the sun sets. So this pattern causes an energy surplus at midday, and possibly even an energy deficit in the late afternoon, early evening.

This daily oscillation, when you graph it, it can sometimes look like a duck, hence the name. Energy storage, the importance of it is that energy storage is really the most direct method of smoothing these kind of daily curves. If you can store some of the surplus energy in the middle of the day and provide it in a few hours later in the evening, you can shave off the peaks and fill in the troughs of this duck curve. And if the share of renewables increases, the importance of storage and other load balancing techniques is going to increase.

Stone: So you just mentioned that there are other technologies out there for balancing the intermittency of wind and solar. What are some of those other technologies?

Serpell: That’s right. So batteries, specifically when we’re talking about lithium ion batteries, they’re really a fairly new concept when it comes to grid-level storage and load balancing. The technology is in many ways kind of more well suited for applications such as laptops or EVs. In fact, I think it’s about 95% of existing grid storage today uses a much older technology, and that is pumped hydroelectric storage, or pumped hydro. This is the process of using surplus electricity when it’s available to pump water uphill to a storage reservoir. After that, the system can operate just like really any other hydroelectric power, where the stored water is released and the force of gravity turns a turbine used to generate electricity when it’s needed.

Of course, these systems don’t rely on cutting edge technology, they’ve been around for many years, and they have really been the only viable solution for grid storage until very recently. And these gravity powered systems, they do have a few other advantages. For one, they last a very long time. A regularly used and well maintained lithium ion battery, it might last you for seven or eight years before its storage capacity degrades considerably. By contrast, a pumped hydro system could last 30, 40, maybe even 50 years or more and still maintain most of its original capacity.

For example, there’s a facility down in Virginia called Bath County, the Bath County pumped hydro station, that’s been operating continuously since 1985. And as of today it’s still considered, or kind of claims the title of the world’s largest battery, which depending on your definition of battery, it really is in respect to its energy storage capacity. It’s got about 3,000 megawatts of power, and 24,000 megawatt hours of energy capacity.

So that’s kind of the main alternative for grid storage at the moment. But that said, pumped hydro also has its limitations in that it’s much more land and water intensive than lithium ion batteries, and it can only be cited in areas with appropriate topography and elevation, which makes it much more limited in application.

Stone: So you just talked about some of the limitations specifically related to pumped hydro. But in general, what is storage particularly good at, and what are its general limitations?

Serpell: Storage is really good at helping variable renewable energy sources like wind and solar help to mimic fossil fuels in many ways. Fossil fuels are marvelously flexible energy sources. Obviously they have other issues, but just in terms of the flexibility they offer, they really are hard to beat. They can be produced, refined and stored more or less indefinitely without loss or degradation. They’re extremely energy dense. And they can be transported anywhere when and where they’re needed.

Storage technologies can help a sustainable energy system retain some of these advantages that we currently enjoy from our largely fossil fuel based energy system. Unfortunately, storage, especially existing battery storage technology, it’s far from a perfect solution to this need for flexibility. Lithium ion batteries have several limitations. They can certainly be improved through future research and development, but they’ll likely continue to be a barrier for their application in some storage needs.

One limitation being their lifespan, which I’ve already briefly mentioned. Lithium ion batteries are really only designed for a few thousand charge cycles. And for our phones or even an electric vehicle, that’s really plenty. But when you’re thinking about balancing an entire electricity grid for decades to come, those short lifespans will increase the cost of upkeep considerably. Batteries are also — they’re not terribly well equipped for longer duration storage.

Batteries experience something called self-discharge and this basically means that if a battery is left full, and we’re talking specifically about lithium ion batteries right now, after a few weeks or months it’s going to have lost a sizable chunk of that energy storage capacity. How large that chunk is really depends on the length of time it’s been left, and the specific technology being used. But in some cases and at sufficiently large scales, that self-discharge could represent a significant loss of stored energy. And batteries also really don’t like being left empty for any significant length of time. This causes them to degrade more quickly.

Stone: So as we get more renewable energy, the intermittency of wind and solar energy will need to be balanced over longer and longer periods. And eventually actually over seasons rather than just days. Why is this?

Serpell: Yeah, so this gets back to kind of what I was just saying. When we think of energy storage systems, it’s really important that we observe the distinction between power and energy. A battery’s energy capacity is the total amount of energy put out as electricity that a system can store. The power of a battery is the rate at which it can charge and discharge that energy. Today, because batteries are primarily being used or relied upon to meet short, kind of one to four hour dips in electricity supply, the power output of a storage system is somewhat more valued currently.

We want to know that when that duck curve starts to curve upwards, there’s sufficient power from storage available to meet increased demand. But we also really require it to provide that power for only a relatively short period of time. And that’s why most battery systems are rated with about a four hour duration. However, as deployment of intermittent renewable like wind and solar increase, the duration for which we’ll need that storage is also going to increase.

And this is kind of challenging to describe without visuals, but if you imagine the duck curve, and any listeners who want a visual can just google the duck curve, if you imagine the duck curve and you imagine stretching it out so that instead of a single oscillation each day, we’re really looking at a single macro-oscillation over the course of a whole year.

Here’s another way of thinking about it. Solar panels are going to generate more electricity in the summer months. There’s less cloud cover, sun is more intense. But solar radiance is still going to vary over the course of a single day in the summer. But the total energy generated per day in the summer is going to be higher than it would be in the winter.

So if you were to take the daily total generation from solar, and graph each day for an entire year, you will see a curve not totally unlike the curve you would see over a single day, where it rises during the summer and drops in winter. Only now it’s over a much longer time horizon. And over that longer time horizon, you will need long duration storage to help balance these daily generation totals.

Stone: So in your recent report you point out that the drive to electrify everything will create new electricity demand patterns that will make it even more critical that electricity be delivered reliably at all times. How will the shift to electric home heating, as an example, further complicate the shift to a clean grid?

Serpell: Up to a point, we’ve really only been discussing the load variability caused by the intermittency of renewable energy. And by the way, I’ve been focusing mostly on solar because of the duck curve example, but we do see a fairly significant variation in wind capacity as well over different months of the year. And in many ways, wind is even sometimes less predictable over days and weeks than solar is.

But basically the intermittency of renewables is only one driver of variability. Variability can occur on the supply side from renewables, but it can also occur on the demand side. The problem is as we begin to electrify more and more end uses of energy, demand patterns for electricity are going to change. So you mentioned heating, which was the focus of one of our recent papers that I wrote with my research colleagues, Amy Chu who is an assistant professor at Mills College, Benjamin Perrin who is a doctoral student here at Penn in the Department of Material Sciences, and Garesh Sunkar [?] who is a master’s of business administration at Wharton.

So for example, in the northeast we use a colossal amount of energy to heat our homes and businesses in the winter. Until recently, electrifying heating was far too expensive and very inefficient to really be a viable solution for everyone. But today, thanks to developments in technologies such as air source and also ground source heat pumps, it is now fairly economical to convert your home to electric heating.

These systems in fact can be quite a lot more efficient than existing oil and gas heating systems that they’re replacing. And of course these systems sever the direct connection to the use of fossil fuels. However, just because electric heating has gotten more efficient and affordable doesn’t mean that in aggregate this shift would not have very profound impacts on demand patterns for electricity.

Right now in the PJM Footprint, which is the kind of area where we are now in Philadelphia, for example, the yearly peak in electricity demand is very consistently in the summer months. However, we modeled what it would look like if you electrified all heating in PJM, this is residential and commercial heating, using the latest and best air source heat pump technology. So these are very efficient systems.

But still it significantly shifts the annual peak demand for electricity from summer to winter. And unfortunately, this coincides with the period of lowest solar radiation and lowest capacity factor for solar power. So this is a real problem, and it could create an even greater disconnect between when energy is produced and when it’s required, thus amplifying the importance of storage.

Stone: So if I can summarize that, so it sounds like not only do we have this issue of as we have more renewables we need to balance those renewables seasonally rather than daily. But also what you’ve just described here is that we are actually going to have more demand in those off seasons for particularly solar power production in the winter, which only amplifies, as you just said, the need to have some sort of storage to get us through those long cold winter months when everybody potentially, for example, here in Philadelphia in the future, is going to be electricity to heat their homes.

Serpell: That’s right. And just as the supply variability is going to depend on the grid mix of generation, the demand variability is going to depend on what end uses we electrify. So we looked specifically at heating, but you could also look at what the seasonal impact of widespread electrification of automobiles would look like. For example, vehicle miles traveled are somewhat higher in the summer than they are in the winter.

So if that were to translate over to electricity demand for electric vehicles, you could potentially see some of this variation in seasonal demand to start to balance each other out. But that would require much more sophisticated modeling that what we were able to do.

Stone: It sounds like a lot of pieces will have to come together in the puzzle to make that happen too.

Serpell: Absolutely.

Stone: For example, a vast EV network. I want to touch a little bit further on that report that you did that involved Philadelphia Gas Works. And in that you researched the economics of storage under a scenario, as we just said, where home heating, for example, here in Philadelphia is actually electrified. Tell us about that study and what you found, and about the economics of a low carbon grid, and the way to balance that grid here in the city.

Serpell: Yeah, we did look at the costs of electrifying, and to be honest it’s really not pretty, right? I mean, once you factor in the duration of storage that would be needed to balance these monthly and seasonal variations, achieving it using only, for example, lithium ion battery technology, it becomes a real, real challenge. And again, this is because what you would really want for this scenario would be a storage system where the ratio of power to energy capacity, remember we kind of outlined that, was much smaller than the typical battery system that you see today.

So using the ratios of existing lithium ion battery technology, or existing lithium batteries in use today, you would need to deploy many times the power output that you would ever need just in order to get the energy capacity needed to deal with these seasonal variations. So in an optimization model that we built actually for the whole PJM region, we saw a consistent over building of renewable capacity before seasonal storage capacity was really ever considered.

And this was true for lithium ion batteries, it was true for pumped hydro storage, and it was also true for reversible fuel cell systems using hydrogen as a storage medium. These were the three storage technologies that we looked at. In other words, the model opted to over build generation and accept the economic losses of curtailment before ever really relying upon storage to shift load from one month to another.

Stone: That’s interesting because what you’re essentially saying here is that the economics of seasonal storage really don’t work if it is indeed cheaper essentially just to over build wind and solar to make sure that there’s always enough even when the wind isn’t blowing very strongly.

Serpell: Yeah, they — I would say they don’t work at the moment. But again, these are challenges that we’re going to be facing a little ways down the road. And depending obviously on the developments and further innovation in the storage space, that could very well change. One of the big challenges with seasonal storage actually has nothing to do with the technology limitations, of which as I’ve outlined there are plenty, it has to do with how storage is valued in the market.

For example, you could envision a very simple variable price structure in which storage providers charge their systems when electricity is cheap, representing a surplus of electricity available, and sell that stored electricity when prices are high, representing a period of high demand. The revenue generated by the storage provider in this set up would be entirely derived from the profit made between the electricity price at charging and the electricity price at discharging.

The problem with this system is that it basically incentivizes storage providers to charge and discharge their systems as frequently as possible, and provides no compensation for the standby time in which they’re either storing energy or are ready to receive surplus electricity. Looking to the future, if we need storage to help balance monthly and seasonal load variations, we’re going to need a market structure that incentivizes storage providers to effectively keep their systems on standby until they’re really needed.

Stone: So that’s interesting, so if you’re cycling those batteries or whatever the type of storage may be infrequently, then there’s really much less revenue is essentially what you’re saying. So new market structures need to be come up with to accommodate that.

Serpell: Right, right. Because once you make the capital investment in a storage system, if you’re trying to run it as a business, you want to run it as often and as frequently as possible. And that’s really not what we’re going to need storage to do down the road.

Stone: You know, there’s another aspect you brought up in the report that I thought was very interesting and which doesn’t get a whole lot of discussion, and that’s the issue of land use. You looked at the land use requirements for these different types of storage, and some of them were pretty dramatic. And that also factored in, I believe, into the decisions or the conclusions you made in the report. But tell us about those land use considerations.

Serpell: So land use is a really tricky one. On the one hand, the US is a very large country with few constraints on available land. And at least compared to, for example, many European countries, there’s really not much land constraint in the US. And even in Europe, there is technically plenty of land available to support the necessary renewable energy generation that will be needed.

But that said, it’s really not that simple. Land rights, land ownership, community resistance, and also concerns over preserving ecosystems and productive agricultural land, mean that citing new renewable generation, it’s going to be far from straightforward, especially once you consider the gigawatts and gigawatts that will be demanded by a fully renewable electrified energy system. So in our model, we did calculate the land use of the various scenarios using each of the three storage technologies that we looked at, but we didn’t assign a cost to this land use.

So the optimization model calculated land use, but it did not use that land use to give recommendations. I think if we had included a cost for land use, we may not have seen as much over deployment of generation as we did. This is because solar panels, and especially wind power, take up a lot of space. And in some runs of the model, it was deploying enough solar and wind power to cover Virginia.

Stone: And that’s dramatic, and that’s got to be expensive.

Serpell: Right, and this is because it was over deploying generation to avoid having to deploy storage. So when it comes to a comparison of the actual storage technologies, lithium ion and reversible fuel cells, they do take up significantly less space than the current prevailing grid solution which would be pumped hydro systems. And they’re much less dependent on local geography. But compared to the land use demanded by the generation, really all three storage technologies represent a comparative reduction in land use. So the more value you place on land use, the more that will favor deployment of storage.

Stone: So let’s kind of cut to the chase here. We’ve talked some about the challenges to storage, particularly for seasonal purposes, some of the land use considerations. Given all of this, and the cost considerations again, what role should we realistically be casting for storage overall within our electricity system, again as we reduce the carbon intensity of the system?

Serpell: Yeah, so I don’t mean to avoid this question, but I think it’s going to depend so much on the technology developments and the cost curves of storage over the next decade or so. Like I said a couple times now, many of the challenges with storage that we’ve been discussing, they really only become a major consideration once a large majority of the electricity is being produced by solar and wind power. And once our energy system has also become significantly more electrified.

So it’s kind of hard to say exactly what the future limitations of storage will be once these kinds of seasonal load balancing services reach their height of importance just because the technologies could have improved significantly by that point. I mean, if storage technologies continue to improve at the rate that they have been doing so over the last couple of decades, then it’s still very possible that they could offer something approaching a fossil fuel level flexibility when we need it, and for a fully electrified economy.

The technology is not there right now, but it very well could be in the future. And in general, I would strongly caution anyone from betting against the electrification. The world has really slowly been electrifying since the days of Edison, and storage technologies are just allowing this process to extend far beyond what was possible just even a few short decades ago.

Stone: So let’s say for a moment the costs continue to be a big problem. What other options should we be looking at for cost effective, low carbon balancing of a future grid? Again, that has a high penetration of renewable energy.

Serpell: Yeah, so in the event that some or all of the existing limitations of storage cannot be overcome in time, we probably do need to really consider the role of hydrogen and of renewable fuels that can act as basically flexible substitutes for existing fossil fuel demands. Even if this is a much less technologically mature solution, electrify everything as a concept is very elegant in its simplicity, and of course a large amount of electrification is absolutely going to be necessary like I just said. But we shouldn’t discount the challenges that come with it.

It’s not just about the end use practicality or economics of electrification, it’s really about the impact that that additional load would have on peak demand and on the overall variability of demand, and what that means for balancing a future grid. We’re likely going to need hydrogen or some form of renewable fuel for things like aviation, long distance aviation, certainly some types of industry. So it may be a good opportunity to also consider what role those technologies could play for things like residential heating.

Then of course natural gas peaking plants coupled with carbon capture and sequestration, they could be a back up method for sustainably dealing with these seasonal peaks and demand if storage or renewable fuels don’t pan out. But we’re going to have enough carbon capture to deal with as it is just to stabilize the climate from the emissions that we’ve already produced and will be producing. We really need to avoid I think relying upon it wherever possible.

Stone: Well, it’s interesting that you bring that up, because natural gas really is the default right now, right?

Serpell: Right.

Stone: And it really it is the balancing solution as you point out, you’re going to have to have some way to clean that up within the context of a low carbon grid.

Serpell: Right.

Stone: Carbon capture and storage being the most obvious, but that’s not cheap either, right? And you have to have access to geologic storage, which means either you’ve got those gas peakers close to the storage or you’ve got a long pipeline to take the gas, the carbon dioxide from those peakers to the geologic storage location. So all these solutions are challenging, right?

Serpell: Yeah, and also peakers, they’re kind of — in the kind of scenario we’re talking about, they’re going to face many of the same challenges from seasonal variation that face storage solutions, right?

Stone: In terms of economics?

Serpell: In the terms of economics, right. Because it’s going to be tough for any operation, storage or peakers, to survive when they really may only be needed for a month out of the year. That’s not a good business model for anyone really.

Stone: You know, I wanted to bring up one other point, and just kind of emphasize something here about storage. So we’ve talked about the limitations of storage, again for seasonal purposes. But despite those seasonal challenges, storage again really does hold huge potential in cases where shorter duration is just fine. And tying in gas and storage all together, some recent reports have come out that show that storage is becoming cost competitive with natural gas generators, and balancing the daily intermittency of solar output.

And also the Federal Energy Regulatory Commission, the FERC, which regulates wholesale electricity markets, obviously recognizes the value of storage because it has implemented new market rules to accelerate the growth of storage, again in electricity markets. I wonder if you could just tell us a little bit more about that.

Serpell: Absolutely. Yeah, I completely agree. We are very quickly approaching I think a point where daily intermittency from renewables really becomes much less of an issue. I think already in places like California where they’re already facing these challenges, with careful demand side management, aggregation, making sure that here’s a diverse renewable generation mix, and using the smart deployment of short duration storage, I am fairly confident that we can effectively move — we can effectively deal with the daily intermittency of renewables.

It’s really the seasonal variation that I see as more of a long term challenge. And Order 841, it’s an essential step in that direction, right? Part of the challenge for large-scale deployment of grid storage, as I said, is the market structures that are in place. We really need to make sure that the market accurately values the service that storage providers are providing, and that includes the standby value that storage systems offer even when they’re not actively charging or discharging.

Stone: So let me ask you a final question here, and just kind of sum up. So we’ve obviously got a big challenge ahead, right? We’re headed for an electro system that will be cleaner, larger, and an even more essential part of our lives if we’re talking about everything being electrified. At the same time, we’ll need to balance this intermittency of these resources. So what policy solutions, additional policy solutions, might there be to address the challenges presented by a low carbon grid?

Serpell: Yeah, so I think developing robust and fair market incentives for storage providers is going to be so important if we’re going to build our storage capacity to a point where it can really provide grid wide load balancing. I also think that it’s important to acknowledge that more and more individuals are becoming effectively small scale storage providers through the set up of home backup systems, the purchase of electric vehicles. There are more and more batteries entering the system.

And finding policy, market, and technology solutions to effectively aggregate all of those storage systems while also compensating owners is going to be a real challenge, but a very important one to solve. And lastly, I would just stress that we shouldn’t forget how far we still have to go before these storage technologies can support renewable generation to the point where we can have a renewable electrified energy system that can function with the kind of flexibility that we’ve really come to take for granted today from fossil fuels.

You know, a 70%, 80% renewable grid, it might not seem anywhere close, it might seem like a lifetime away, but it really isn’t. And we’ll be there before we know it, and we need to make sure that storage solutions, or if necessary renewable fuel solutions, are ready to support that future grid. And this needs to be a major focus of policy innovation moving forward.

Stone: Oscar, thanks very much for talking.

Serpell: You’re very welcome, thanks so much for having me.

Stone: Today’s guest has been Oscar Serpell, research associate with the Kleinman Center for Energy Policy. Today’s episode will be the last before we take a short break for the holiday season, but of course we’ll be back in January with a new year of episodes covering the world of energy policy. In the meantime, check out our archive of episodes on the newly redesigned Kleinman Center website. Or to get all the latest from the Kleinman Center, including research, blogs, and events, sign up for our monthly newsletter again on our web page. Thanks for listening to Energy Policy Now, and have a great day.