EMP S3E9: Willett Kempton, a pioneer in vehicle-to-grid technology, talks about the state of play for V2G, which promises to become a critically important resource for power grid operators

Back in the 1990s, the University of Delaware's Willett Kempton conducted early vehicle-to-grid (V2G) experiments with PJM, operator of the MidAtlantic region's wholesale power market, testing the feasibility of using electric vehicles to provide regulation services to the grid. In this episode, Kempton speaks to the progress that's been made in the intervening decades to set the stage for today's electricity market, where EVs are just beginning to have enough market penetration to provide enough battery storage at scale to make bidirectional electricity flows between vehicles and the grid a reality – and an important reliability tool for grid operators who will be challenged by steadily increasing dependence on variable, nondispatchable generation resources.

Kempton, professor in the College of Earth, Ocean and Environment and the Department of Electrical and Computer Engineering at the University of Delaware, is also an offshore wind expert, and talks about the wave of coming offshore wind installations domestically, taking advantages of scale economies that have make the technology almost at price parity with the wholesale markets.

Transcript

EMP S3E9: Willet Kempton, University of Delaware, Nuvve
(transcript edited for clarity)

EMP: Welcome to the Energy Markets Podcast. I'm your host, Bryan Lee. So there's no question that electric vehicles and energy storage are critically important elements of the clean-energy transition under way today to address the runaway climate change we are experiencing – runaway climate change which will only worsen in the future. And these two important issues, EVs and energy storage, are joined at the hip in what is called V2G, or vehicle-to-grid technologies, in which EV batteries, like standalone energy storage batteries, can become important frequency balancing resources for grid operators. Balancing is necessary because grid operators must keep the system precisely at 60 Hertz. Any slight variance above or below 60 Hertz and the power grid becomes unbalanced, leading to reliability issues and outages. So the batteries in the many millions of EVs that will be plugging into the grid, in addition to the millions of anticipated standalone battery storage devices, will become increasingly important tools for grid operators, particularly as we become more and more dependent on variable, nondispatchable renewable energy resources. So that is a longwinded setup to introduce our guest today, Willet Kempton, professor in the College of Earth, Ocean and Environment and the Department of Electrical and Computer Engineering at the University of Delaware, where he was an early pioneer in V2G technology back in the 1990s. Professor Kempton is a cofounder and the Technical Advisory Board Chair at Nuvve, a publicly traded company offering advanced V2G technology and electrification solutions. On top of all that he is also a noted expert in offshore wind technology and policy, and Associate Director of the Center for Research in Wind – issues which we should also spend some time talking about today as well. Professor Kempton welcome to the podcast.

WK: Bryan, thank you so much and I appreciate your interest and putting me on. Having a chat. Thank you.

EMP: Well, why don't you go into a bit of detail for our listeners about V2G, bidirectional energy flow, grid balancing, your early experiments with this in what is going to become an increasingly valuable tool for grid operators.

WK: Okay, sure. So, way back in ancient history, 1996, I was looking at ways of providing storage for the grid because we already knew then there would be more and more variable generation. And even without variable generation storage is valuable to the grid for, as you say, balancing and also other services. What we kind of realized starting with thinking about things like batteries for emergency computer backup and stuff, which in aggregate are quite small, you know, I went to an EV conference and said, oh my gosh, here's the big battery we've been looking for, but it's going to be in the garage, you know, or next to the building, or maybe in a big parking lot. So can we use these batteries? And there's a few key numbers that sort of tell you why that works and why it's an important resource. First just understand the size of the battery in an EV. An average, stand-alone house, let's say, uses about one and a half kilowatts. But an average EV charges at seven to 10 kilowatts and some can charge at much higher than that. So, in power measurements, you actually have a huge amount of power per vehicle. Now let's ask how much energy is that – how much power over time? You know, if you've got a, let's say 50 kilowatt-hour battery, which is just a little bit on the large size of what's available today, that seven-kilowatt charge or discharge rate means seven hours of power – seven kilowatts times seven hours, so it's about 50 kilowatt-hours. So, gee, that's quite big, you know, it's way bigger than a computer backup or things that you might have storage for, electrical storage for, otherwise. So, first cut, just looking at the device you say, this is interesting. So but it's in use for driving, you know, they’re in use all the time, you know, so we're not going to be able to use them for grid backups. Well, then the next number that's valuable to know is how much is the typical light vehicle used in the United States? So privately owned light vehicles, your car, are used about 4% of the time. Kind of amazing. So from my point of view, that means, hey, this could be a grid storage battery 96% of the time. So these are kind of basic numbers that tell you why this could be interesting. And then a much more complicated calculation, rather than those ones I just did which you can do in your head, a much more complicated calculation is, if we have a large fraction of variable generation, renewable energy and wind and solar, how much storage does it take to balance that out so that you have, so that you always can turn on and off the lights when you need without saying, oh, no, the wind’s died down, so I can't watch television. As one of our recent political leaders said. Well, that's a model and it’s plus-or-minus a pretty big error bar, but approximately if you had all light vehicles, cars and trucks, personal vehicles, they were all electrified with sort of typical-size batteries, and you had 95% of the grid variable generation, that would be about the amount of storage you need to run the whole grid. So there's various ifs and so forth. Is it that 100% vehicles are going to be electrified, not for a long time. But then again, we're not going to have 100% wind and solar for a long time either, but these kinds of three basic quantities tell you, oh, okay, that could be a match. And of course, when you start looking at costs, the battery is already purchased for transportation. So that means the cost of storage is going down by two orders of magnitude. You still have to have some controls, there's probably some customer service, you have to answer an 800 number when the customer wants to know why their car's not doing V2G today or whatever. But the hardware, you're just adding a few controls, so it's already there. So that's the kind of basic concept and we can talk about early experiments and so forth. But I wanted to get that basic idea out first. 

EMP: Yeah, those are good numbers, and I was hoping you would get into that dataset. But yeah, I mean, I as a reporter back in the ‘90s, caught wind of your early experiments with PJM at the University of Delaware. And why don't you tell us a little bit about what you were doing and how that evolved into what you're doing today?

WK: Sure, we had the idea. We kind of did the numbers. We published some academic articles and Steve Letendre was working with me in in those early days. And he's now working for Nuvve as a policy analyst. But, we then kind of teamed up with some people in electrical engineering at our university and said, hey, you know, let's, let's do this. Let's build it, you know, it doesn't necessarily look that hard. And, and then, early on, we were also talking with PJM, which is the grid operator in the MidAtlantic. So that's about 13 states. I think around an average of 70 gigawatts load served something like that. That's not the peak, it's average. So a pretty big grid operator. And they were interested. So we were luckily also teamed up with AC Propulsion, which had an early EV, the founder of that did the GM Impact design. And they had built a charger, a bidirectional charger into their EVs and their drivetrain. So that meant that we just needed to add controls to manage that. We didn't need to add power electronics that would both charge and discharge the car. So charging is grid-to-vehicle and then discharging his vehicle-to-grid and, of course, the term V2G, it's literally vehicle-to-grid power flow, which is the unique part of it because normally when you charge your car you want it to go from the grid to the vehicle. But of course it doesn't really mean anything unless you have controls in there. It's not of any value to the grid to have you push a button and discharge your car whenever you feel like it. It's of value only if you can do that when there's a need on the grid. So we had the car, we needed the controls and to make it an actual demonstration we wanted a grid operator who would add value for this and they already had markets. So we didn't have to create a new market. The one you mentioned, Bryan, regulation, it's a balancing service. But there's also services like reserves that is a rarely occurring problem where there's some large-scale failure that needs to be compensated for by generators that are ready to go or batteries in cars that are ready to go. Capacity, much less frequently, but sometimes you have to have a contract or the provider of battery or generator provider can have a contract that says when you really need it, you know, you have some really unusual failure, then we'll be there. You know, so that's capacity. And then at the local utility level you've got on-peak/off-peak rates. So that's already again, all these are services that are already there. Although it turned out later we were to realize that it's more complicated to qualify for this than we thought. But the ingredients of our first experiment included all of these. We had a real car you could drive and not just a bunch of electronics on a table wired together. We had a bidirectional flow of power, which gives you the maximum revenue. It’s the most value for multiple reasons. And we had the controls that we built to respond. And we had the grid operator, PJM, who agreed on an experimental basis to give us a signal so we could follow the real signal. So when we charged and discharged, we would be doing the balancing that the grid actually needed at that point. Putting all these things together with some funding from the, I guess it was the state of Delaware, and also maybe Pepco Holdings, a utility holding company, we were able to put all that together and we had a car, you know, you can drive the car to the grocery store and then drive home and plug in and you're getting a signal from PJM. And you're actually helping the grid by providing balancing service. So that was the early experiment. And you know, we thought it was significant. We put out a press release and it got picked up quite a bit. That plus some later alliances we did with some other companies covered in the New York Times, Forbes, Wall Street Journal. So it was considered a significant advance and that was back in the 2008 to 2011 time period. 

EMP: Let's talk about some of the things you found out. A lot of people are concerned about their battery life and they're going to be reticent about participating in this sort of wizardry, if you will, because they're going to be concerned about the impacts on their battery. What did you find out about the impacts on the battery?

WK: Yeah, so we have worked with several automakers about this. And there's also been several really good economic studies on it. Sorry, I would say more research studies is what I mean, which is on the electric chemistry. We haven't done the direct experiments ourselves, but what we found out through colleagues is that for example, Chalmers University, did a lot of cycling back and forth on batteries and testing out, you know, hundreds of different regimes. And it turns out that the wear, when you push all the way to the full 100% charged or all the way to the bottom, is much more than moving up and down in the middle. That is, if you're going from say 30% charge to 60% charge many, many times. There's less wear doing that than going from 95% to 100% a couple of times. Now car designers try to manage that by clipping off the top. So 100% isn't actually a full battery, but still if we manage the state of charge properly, this work by Chalmers University, in particular, in Sweden, shows that there's really minimal wear. Now does that turn out to be true from an industrial standpoint? So we also worked with Honda America and also with their Japanese engineering teams. And so they had they had one of their EVs that they put on the PJM market because we were getting a PJM signal. We could put their car on and they had both controls and the bidirectional charge and discharge capability. This was in a prototype vehicle. It’s not gone into production now. But they were able to run it for several months doing PJM regulation. So they were in a real market running for a significant amount of time. And they measured very carefully the effects on the battery and then put it in their battery warranty model. So it's one thing for a bunch of engineers and electrochemists in the lab to do it. And it's another thing for an OEM to do it with their warranty model, you know, because if they get that one wrong, they've got a lot of liability of replacing batteries. So there's kind of two different ways of saying that's really the best way to do it. And you'd have a split of opinion, but we thought it's important to have both. So Honda's result from that was – and this was published in the SAE World Congress – their result was over the lifetime of the warranty – and it's eight years – they would have 8% wear from driving. If they were doing V2G like they did in our lab the whole time that they're parked, they would have an additional 2% of wear from V2G. So it's not zero, but it's very small. And in fact, it's only 1/4 as much battery wear as you just get from normal aging and use of the vehicle.

EMP: And what are some of the other technical problems that we need to be aware of? We're talking about bidirectional electricity flow. Now, I've got a plugin hybrid, a Chevy Volt. I can't be providing power to the grid from that as much as I'd like to, correct? So elaborate on that.

WK: Yeah, so just a basic technical issue is, can your charger also discharge? And the Volt cannot. I actually talked to the design team for the Volt when they were first putting it together and they evaluated that are we ready to do that or not? They decided not to on their production version but yeah, so the yeah, there's several now cars that do. But you're asking the technical problem. The technical problem is, in addition to charging, does your charger circuitry, and it's really just if you do it from the design phase up, it's really only a few components that you're changing in the charger, but depending on the original topology, but can your charger also discharge, pull the power out of the battery and then you've got to not just dump it on the grid, you've got to have very careful regulation of that power so that you're exactly at 60 Hertz. You don't have a bunch of harmonics. You're not discharging when the grid is down. So there's a bunch of requirements that are encoded in standards and then you can check against those standards and make sure that your bidirectional charger meets standards when it's pushing power onto the grid. So that's a technical challenge. It's a lot of engineering time. So really to make one car it's pretty expensive. But if you're making 100,000 cars a year, like the typical OEM, that's sort of engineering costs that go away really, and then you're asking what's the extra component costs? And that's in the, you know, sort of tens of dollars, maybe $100, $200 at the most. So your costs are very low. But it is a complicated design problem, and there's certifications to do it. But so that is the bidirectional charger as a kind of big technical issue.

EMP: You've used an acronym a couple of times here I'm not familiar with. OEM?

WK: Oh, sorry. Yeah, by OEM I mean, in other words, original equipment manufacturer. So in the automotive space, that would be the big automobile manufacturers. 

EMP: Yeah, that's what I thought you were saying. I just wanted to clarify. And what about metering? I mean, you've got a meter between you and the grid operator. Is the grid operator proactively going and pulling the power out of your car, or is it a passive system? How does that work?

WK: Yeah. So there's two issues you're raising one is the metering and how's that done, and the other is control, so who is deciding when you charge or discharge? So on the metering part, that's a complicated problem, but we think there's a very simple solution. Most of the services that you would want to do with a battery, it's not like a solar panel where you produce electricity when it's sunny, or a wind turbine when you produce electricity when it's windy. It's really something that you control completely. So with a battery system, you charge when they have too much power on the grid or when electricity is cheap. And you discharge when there's not enough power on the grid or electricity is expensive and you're selling it at a high rate. So really, what's important is that you're responding when needed. So, for the ancillary services, that is the services that are regulated by the regional transmission operator, for those services, it's mostly about power at a time when needed. And for the local utility, especially if you're in a residence, probably what they're charging for is energy – that is kilowatt-hours. So you're selling kilowatts when needed. Not when I want to do it and I'm going to push the button I'm selling kilowatts when needed to the RTO. And I'm selling or buying kilowatt-hours from the local utility. So yeah, the RTOs usually will pay you a little bit for energy, but in these faster smart services, I'm only running 10 minutes or something. I don't really I don't need the energy payment. And anyway, I bought that energy a couple of hours earlier because I was plugged in and buying energy and then I discharged later and I’m selling it. So all I need at the local utility level is net metering, and I'll lose a little bit of money. I mean, in V2G operating, I lose a little bit of money in that transaction with the local utility, but if they just treat it as net metering, that's okay, that's good enough. And then at the RTO level, again, it's power. So there has to be a very precise power meter measuring kilowatt-hours and they by the way, they care about time, you know, now we're selling services where if we're more than two seconds in response, we don't get paid as much. And there will be markets in the future – we know at least two RTOS that are planning to do this – which have to be within 500 milliseconds. That all boils down to we're transacting energy, kilowatt-hours, with the local utility, and we're transacting kilowatts in time with the RTO. That's not double dipping, and it's complicated to explain all the reasons it’s not, we can verify that and so forth. But the metering part has this kind of weirdly simple solution. That is, we just keep transacting energy with the local utility, and we transact power with the RTO.

EMP: So what you're saying is it's an apples and oranges thing, right? With the utility, we're dealing with the commodity, the energy commodity, but with the RTO or the grid operator, we're selling a service and they're paying you, compensating you, separate from that.

WK: That's a great way to explain it. A commodity of energy versus a service of power when needed. Yep.

EMP: And I think this is a common misunderstanding, as I read a lot on this issue, where, you know, I'll commonly see a news report talk about how much a battery or an EV battery could provide a typical house’s consumption. When we're not really talking about providing a commodity to the grid. We've bought that commodity. We're not necessarily selling that back. We're selling a different form in terms of that service back. What about the regulation issues? The regulatory issues? What are we seeing there?

WK: So part of this picture is net metering. And there's one state that has very explicitly said net metering should be used for EVs that are providing grid services. You can't just not do anything with your EV and get net metering. But if you're providing grid services, you're qualified for net metering. That's Delaware. In other states, you know, having discussions with the utility regulators or with individual utilities, once they understand it, they're usually amenable to having net metering for this service. But it's kind of a state-by-state matter. Of course, we’ve got 50 state utility commissions. There's examples, but it's not kind of nationwide. And then there's the regulatory issues in the RTO markets and the reason – RTOs are more complicated. People say why are you bothering with that? It's because the value can be quite a bit higher. The RTO, PJM, very early on, you know, kind of hooked us up, helped us measure things, validated what we're doing. Yes, we're meeting all the requirements for PJM markets. But then once we started to say, okay, well, this is something that's going to become commercial. And so we want to figure it out, so that it can be scaled. Then the conversation got a little more complicated, and it's like, well, this doesn't actually follow our tariff, and you know, it really should be done in this different way. And then the lawyers got involved and of course, then it became way more complicated. So like, oh, you've got a car with a battery in it, and it's plugged into a charging station, and the charging station is a resource to the grid. Okay, fine, but your charging cable between the car the charging station, that's a transmission operator. So it's going to have to become a PJM member and operate as a transmission system with the appropriate tariffs. So I mean, that's, that's a real-life example. But there's those kinds of complications that you would never expect. 

EMP: Something about first thing shoot all the lawyers, somebody said? 

WK: I couldn't endorse that. But I can understand why someone would say it. So you were asking about regulatory issues, so that's the PJM level. So then the other thing is interconnection. So you're connecting a device, which is backfeeding to the grid. Okay, fine. We've got UL 1741 in the U.S. to determine the device is okay to interconnect to the grid and backfeed. That's what solar uses that UL standard. But it turns out that that UL standard doesn't expect part of the device to drive around on the road and then plug in somewhere else tomorrow. So you, you know, the UL certification applied to the whole piece of equipment that was at the house and oh, by the way, down in the weeds of UL 1741 it says that all the equipment has to be bolted firmly to the structure. So now we got a car we drive in and plug it in and we want to be interconnected with the local utility and with the RTO. And we're violating several sections of UL 1741. So that meant that we had to develop a standard for electrical devices that could provide power to the grid, but come in pieces, which may be separately certified at different times. So we also at the same time found that the auto manufacturers, the OEMs, were saying, we're not really comfortable with the UL standard. We're regulated mostly by SAE standards. That's in the U.S., at least, SAE standards. We'd rather have an SAE standard or this. So we helped to sponsor and I staffed a committee that developed an SAE standard, SAE J3072, which said, okay, here's some batteries and they roll around and interconnect at different places, but they can have a conversation with the charging station and convince the charging station that those parts of the unit are certified and then the charging station can be the gatekeeper and say, do I get convincing codes from this car that tells me all of that part is going to be certified? And now I know what I've got myself and in fact, I've got a UL or, let's say, SAE sticker on the outside of my box as a charging station. And so those two together, given this new standard that we helped to develop, SAE J3072, can be certified. And we have again a few states that have already recognized that. California is on the way. Delaware, again, does have that explicitly in code now. And it's often something that a utility would just apply for their commission and say, oh, we also want to recognize J3072. So that's not necessarily a difficult process. It would just be part of the tariff making process. But that's another regulatory part, which we've solved important hurdles there.

EMP: And what about NIST, the National Institutes of Standards and Technology, are they part of this convoluted mess you talked about right now?

WK: They certify our, or their standards certify our meters. So we have a meter in the charging station. And that is this meter that we use for the RTO services where it's power that has to be measured, and that has to be certified to within what's called revenue-grade accuracy. So we do use the standards for metering and then we use – we have a certification that we give to the RTO that says this is a certified meter. So when you get 27.324 kilowatts, that's accurate within 0.5%.

EMP: And one of the issues we confront here on the podcast quite often is the disconnect between the state regulation at the retail level and the federal regulation at the wholesale level. I was wondering if, because we're talking about apples and oranges, a commodity and a service, is that less of an issue here than it would be say with other behind-the-meter technologies?

WK: I'm not sure if it's less. That's part – we're partly making that argument actually, again, right now that there's less of conflict than some think immediately. It's kind of immediately I think, because people are used to solar meaning people in the industry, both the utilities and the RTOs they tend to think you must be double counting. So we kind of have to go through that service versus commodity argument. That may make some things easier without any change of regulations. But generally, FERC, the Federal Energy Regulatory Commission, that regulates the RTOs and the interstate transmission, they want distributed energy to happen, because as in this case for storage, it's going to be much cheaper. But it's more complicated. So you're trying to coordinate a bunch of little units out in different locations and even more complicated if they're all behind retail meters. So FERC wants distributed energy to happen because it's often going to be lower cost. And they recognized some of the barriers that there have been for registration of distributed energy as providers to the RTO market. So that's in part because we, you know, we went down we were invited to give a talk at FERC and drove one of our EVs down and showed that actually could do PJM grid services while out in a parking lot down with some of our utility pals. Both Pepco Holdings, and PJM. And so they said, oh, okay, it's actually working here. So, you know, we kind of did a demonstration. I gave a talk on how it worked. And then we commented on several FERC orders when they were at the proposed rule for comment level. So we finally, I think, after a couple of orders that kind of pushed things in this direction, but the RTOs didn't necessarily feel like they had enough guidance. Then we finally got Order 2222, or rather, we got our comments on 2222 in. I think they, I think FERC has now given a pretty clear path for both distributed batteries, batteries that might be in vehicles, and then use of storage behind the meter, which doesn't necessarily need to be in vehicles. But storage behind the meter presumably poses some a little bit unique problems. And then also aggregation. So aggregation, meaning you've got 10 kilowatts here and 10 kilowatts there, and 10 kilowatts somewhere else, well, the minimum amount of power to participate in the ancillary service markets in PJM, for example, is 100 kilowatts or a tenth of a megawatt. So having a 10 kilowatt device, you know, that gets rounded to zero. Even if you have a lot of them. So, aggregation is also key and FERC defined rules for that. So I think with Order 2222, we now have a guidance for the RTOs which will lead to this being possible, theoretically, throughout the United States in all the FERC-regulated RTOs in the country.

EMP: So California looks to be on the leading edge of EV adoption, and has passed laws with aggressive adoption timelines. Nuvve’s CEO, Gregory Poilasne, recently testified before the California Senate in support of SB 233. This is proposed legislation that would establish state goals for bidirectional charging, increase funding for bidirectional infrastructure, advanced interoperability testing and require new EVs sold in the state to have bidirectional charging capability by 2027. He said, “California must approach V2G the same way it has treated rooftop solar or EV adoption – with goals, incentives and fostering stakeholder collaboration.” Do you want to elaborate on that? How key is it for a state regulator of this sort to address these sorts of issues at this time?

WK: Well, I think we're ready for it because a lot of the standards and market options are available now. There's one other standard I'll get to in a minute. But those are now available, and the OEMs – we've been talking to automakers for 15 years, I guess, on this and you know for a long time they were just like, we're just trying to produce EVs that we could warrantee for eight years. You're now making it all much more complicated. Please go away. So I think the OEMs are now more than at a stage where, oh, gee, there's a lot of other companies that are competing with us. Maybe this is an advantage. You know, maybe this is something where we could sell this as an advantage. And that actually doesn't cost that much to add. And maybe we've gotten – our engineers have enough time to figure out how to do that now. You know, so I think we've reached that stage where the OEMs are thinking about it. And then we've got some vehicles that are advertising already their bidirectional capability, the Ford F150, for example, or Lightning. And I think the Hyundai EV6, I think it is, which they’re advertising built-in capability to run – it's not actually running your house but it's running appliances and plug in whereas for F150 it's very explicitly you can keep lights on in your house. There's one more step and mostly has to do with controls and regulation and safety requirements, one more step to run the grid and sell services. You want an aggregator and you want a little bit more standard compliance on the devices, but that bidirectional charger is already coming from OEMs. Not sure if I answered the question which what was your whole question?  

EMP: Well, how important is it for a state to address this suite of issues – this whole bidirectional issue?

WK: Yeah, so given now that the OEMs are prepared to produce chargers that are bidirectional, and there's some Society for Automotive Engineers standards for how to do that, communicate about it. When a state says something like California would in this bill, which is your cars have to be able to do bidirectional power, it's no longer something that's totally crazy that we can't do. Hey, we're just trying to get the batteries to work, guys. It's now more like something you could do and I believe the intent of that bill is that you could do it through either AC or the DC bus. And there are a number of cars that already are bidirectional through the DC bus. So there's a kind of an easier way to do it. And a little bit more complicated when doing it through the AC. We'll see. The automakers may come out and say, you know, we can't do that. We don't want to do it or whatever. But I think that it by creating the bidirectional capability on the vehicles this then you’ve really only got to do the controls, you know, and the standards requirements on top of that, it's cost wise it's almost nothing in terms of components. So I think a long wind up before the swing and hit there, I think it's very important for state regulations to do that. I don't know exactly the wording. Maybe it should be different from California and whatever. I'm sure others will continue to weigh in on that. But that could really push things forward very quickly, I think, having that requirement. And it won't just be vehicles made in California because OEMs sell to a national or global market of course. So it'll either be an option for the vehicle or say, hey, it's not worth having an option, which as a couple of automakers who told me, we're just going to put it in every car and once we design it, just check it out. See that it works. That'll just be what our standard charger is and it'll be bidirectional. So that means, you know, California, a big market in the U.S., then that becomes a global market availability.

EMP: There's a lot of buzz right now about artificial intelligence and how that's going to revolutionize our lives. I would assume there's a role for AI in this sort of construct we've been talking about 

WK: There's a lot of computing that goes on in managing hundreds, thousands, tens of thousands of cars, to synchronize and be cooperative with the grid and do what the grid needs at the right time. A lot of cycles of computer time and processing and hundreds of thousands of lines of code. Very little of that needs AI. And I’ve studied AI formally, I've written language parsing, AI code in LISP, which is an older way to do that. There's only a few things where you probably need AI and that's more specifically would be called machine learning. So you learn you want to you want to learn and the patterns of the user of the automobile, for example, and then say, okay, you know, they say that they need to be charged by seven but we really need to make sure the battery is fully charged by 6:30 because sometimes they're early. So you could be smart about when you charge and discharge. When you make sure you have a full battery by having some machine learning on the on the vehicle. Yes, possibly do that on the market side also. But really, there's people already doing that, maybe they're using machine learning, maybe not. But most of it is just kind of straight, much more straightforward code and implementing standards. You’ve got to do a lot of testing and you’ve got to be able to survive failures in the system and have other things pick up, you know, if all your fancy, complicated stuff fails, you want the charger to charge the battery, you know? So you fail and I've got some patents on this also for EV charging. You have a graceful failure modes where oh no, you can't do that. You're not doing that. You’re not recognizing that but at least charge the vehicle you know. And so you are going to be able to drive even though a bunch of things aren't working correctly or the internet's gone down or whatever. I think there's a pretty limited role for AI. So I don't, I mean, you know, we'll see, others may jump into the space and think it's very important, but it's a lot of computer code. I don't think it needs too much AI. There's so few areas that would be beneficial.

EMP: Well, we're running out of time here. I did want to bring up the offshore wind issue. Anything more related to V2G before we move on?

WK: One other standard activity that my group at the University of Delaware has been very involved in is a standard for low-cost communication between the car and the charging station that has all the capabilities that are needed for vehicle-to-grid. So that is now an SAE standard. There's a third part of that which will add in some standardized V2G functions, but that's sort of a tenth of the cost of the chips of the most commonly used complex standard for charging-to-car communications, that would be 15118. And it's you know, there's like five different chip sets that you can use so you're not relying on one vendor. And it's a little bit slower actually than the most common standard but you don't really need the speed. You're not trying to pass a movie from the charging station to the car. You're just saying you know limit your charging to this level. So that standard SAE J3068 is going to make this much more economical to do for AC charging. That is part of all the pieces that are coming together which means that this is more ready to go mass market. So that's utility interconnection rules, FERC requiring the RTOs to let this kind of battery storage in the car participate in markets and let them be aggregated by companies that can do aggregation like Nuvve, that you mentioned earlier. And then the standards that automakers are comfortable with that are practical for mass usage because you're not adding $400 to the cost of the car to put some chips in to charge the car and charging station. It's more like $20. So that becomes very, very practical. And also it's the best code set for V2G and just standard charging. So that's those standards SAE J3068 and J3702 are the kind of last pieces there that make that whole thing possible to do now in a mass market way. So we'll see if it works out but it's a lot more complicated than I thought when I published the first article on this in 1998. I think the pieces are together now.

EMP: We've kind of been on the same time trajectory in terms of offshore wind as well – at least in this country. We've seen a bit more faster adoption overseas. Delaware, like all Atlantic seaboard states, wants to position itself as a leader in offshore wind and the offshore wind supply industry. Where are we today? It seems that you know we now actually have offshore wind operating off the Atlantic coast. There's building momentum for some in the Pacific too. Are we on the cusp of a groundswell of development? 

WK: Yes, we are on the cusp of a groundswell development. And that's also been a bunch of pieces coming together. You're right. Europe was first. Europe started with 1992 they had the first offshore wind project in Vindeby (Denmark). They use 500 kilowatt turbines, they had 11 turbines, in the ocean, an inland bay, more or less. And then they went through a progression, you know turbines getting gradually larger, two megawatt 3.5 or 3.6. Then, after a long pause, five and six megawatt. Apart from the very small developments that have been built in the U.S. – one off Virginia and one off Rhode Island, all the turbines in the US are going to start at 12 megawatts. You're not going to see anything smaller than 12 megawatts installed. And that's just kind of a small simple measure – 500 kilowatt start versus 12 megawatts start. One turbine, one 12-megawatt turbine is more power than that entire first wind farm. And now we're talking about projects that are gigawatt projects. Not a few megawatt projects. So, and just, you know, for scale, a gigawatt’s about the size of a nuclear power plant, nuclear power plant might run at 90% capacity factor, or a large coal unit. Two large coal units you might be at a gigawatt. So, each of these offshore wind farms is going out is going to be about a gigawatt . Right around 50% capacity factor, so energy equivalent to two of those to be one, one nuke, say. The total accumulated contracts and some state requirements that haven't turned into contracts yet is 40 gigawatts. And we're looking at about $3 a watt, $3 billion per gigawatt, so the queued-up projects, which won’t all get built until 2034 or so but a lot will get built in the next eight years. That's a queue of $120 billion. So it gives you a sense of the size of the industry. Right now we're buying several components from Europe, but that's not going to work because Europe is hitting the accelerator really hard on offshore wind. Their factories and ports are all going to be busy making their own projects. So we really need to do quite a bit of supply chain in the U.S. and that's already started where you get a factory being built. You get a project committed, factories getting built, and then the factory needs to buy steel. So we've got two large steel rolling facilities that are justified to be built on the basis of the offshore wind demand for steel – steel plate. Yeah, short answer, we're going to see a huge explosion and it's starting this year with two commercial-sized projects being built in the Northeast, both off Massachusetts, although one serving New York, and that's one 800 megawatt and one, I forget the exact, it's 100 or 200 megawatts.

EMP: The biggest factor that I saw a little over a decade ago, when I had a brief association with Atlantic Wind Connection, was cost. It was much more costly. Are we seeing now with this larger scale that the costs are becoming more cost-competitive and can rely on the market rather than subsidies?

WK: Yes, short answer. It's been a huge drop in costs. Nobody predicted it. We were predicting costs for the state of New York, for the Commonwealth of Massachusetts, and most recently for the state of Delaware, as part of the Special Initiative on Offshore Wind, which I'm technical adviser to. And we were the low estimate for both New York and Massachusetts. And we now have gotten to the point where they both had bids and accepted bids to purchase power, and we were the closest you know, we were high. We were predicting a drop in prices, because of market visibility. Those larger turbines and other technical innovations and volume basically, experience, learning curve and volume. But we were still high, you know. So the cost since the first projects in the U.S., again, which are just, you know, few megawatts – 30 or 35 or so, total. From that time to now we've seen a price drop by something in the order of 50% to 70% – 50% to 70% drop in price. So we've gone from very expensive electricity to electricity that when you're looking at a project now starting the process now, bidding in maybe two years, building in 2030 or 2031, that's the timeframe you need for these projects, that electricity probably can be at approximately what utilities are paying for bulk power. It’s not going to be cheaper, but with the investment tax credit, which has been extended for offshore wind and nuclear, with that in place we’re looking at costs that are at parity with wholesale. And so you know, we'll see exactly how much steel increases in cost and how much does natural gas increase, too, because that's also going up now, and you're comparing the two of course. But I think we've gone from really expensive power to pretty close to parity in the Northeast. Electric rates are lower in the Deep South and Gulf. So you're still going to have probably higher costs than market in those areas for a while, but the cost curve hasn't stopped going down. We're getting more and more domestic production, more U.S. vessels, ports are getting bigger and now they're getting a lot of work-around supports which means more people handling things more often. So there's much that can still be done to lower costs. And we're, like I said, already about at the price parity level.

EMP: Well, that's a terrific story. I’ve gone through pretty much my notes here. I'll kick it back to you if there's anything else for the good of the order that you'd like to bring up before we sign off.

WK: I think that's good. Great interview. Thanks for having really great background knowledge already because you're coming from the energy sector, it’s not like I'm talking to a journalist who covers many, many things and they don't know that much about energy. So I appreciate that. And also you were prepared, obviously, and had some good questions. So thanks for thanks for being a really great in-depth journalist. Appreciate it.

EMP: Willett Kempton, Uber Guru for all things vehicle-to-grid and offshore wind at the University of Delaware and Nuvve, thank you very much for your time today.

# # #