Exploring ERCOT Pt. 2 with Brandt Vermillion (ERCOT Market Lead @ Modo Energy)

- It sounds complicated, and it is complicated. But it serves a certain purpose.

- It's very distributed, the idea of having all these different nodes rather than larger aggregated zones as you allow for just more complex management of--

and more granular management of the system itself. And it allows for attacking a congestion problem with a lot more nuance for example. The main idea behind LMPs is really they're there to help resolve congestion and protect all elements of the power grid, or the bulk electric system while still providing power to consumers in the most cost efficient manner.

- Hello, everybody. This is Quentin and Brandt. Welcome to the transmission podcast. And today we're doing a second episode start with Brandt talking about Ercot. Answering all the questions that you guys have sent in that we didn't manage to answer in our first episode together. Now, if you'll forgive us, I'm just going to tell you about something that's changed at Modo that might be of interest to you.

So we've just launched our Modo energy platform in Ercot in Texas. You can access all of our data, and our benchmarking, and our research, all of it. It's awesome. You can do it on platform.

We've worked with some of the biggest asset owners optimizers and operators in Texas and the world to build this, and we're really, really proud of it. So if you haven't seen it already, come talk to us. If you're a bank, if you're an investor, if you're an asset owner, if you're an optimizer, you've got to see this thing. And that's it from the Modo energy plug, and we'll get stuck straight in.

[MUSIC PLAYING]

- Hello, Brandt.

- Hey, Q. How's it going?

- Yeah, very well. Good to be sat here with you again. Welcome back to the podcast.

- Yeah, thanks for having me back on.

- So everyone who's listening to this, we've got Brandt in again. Brandt now works for Modo energy, but used to work at Ercot in the control room. And so we're going to be talking about lots of detail and characteristics of the Texas power market. Why each part of that matters for battery energy storage owners and operators. So this is another niche episode but it's all about Texas stuff, it's about Ercot.

And we did have Brant on a few weeks ago to do an intro to the Ercot market, an intro to the control room, how Texas manages power, and some of the stuff around what happened with storm Uri a couple of years ago. We're recording this on, what is it. It's the 12th of January 2024. And we're about to have a really cold snap this weekend. And next week so we're going to see what happens there.

If you're listening to this, and it's already happened, hopefully the lights will stayed on. And this is a cool episode because essentially we're going to run through a lot of questions that have come in. And so my job as an interviewer is pretty straightforward. I just need to ask you two questions and poke you a little bit. But yeah, if you're listening to this and you've heard Brandt's voice before, yes, this is the same guy.

And if you haven't heard it before, do go back and listen to the first part of this two parter and you'll get a bit more background. But for just to take it from the top before we get stuck into the details a reminder, what is Ercot, Brandt?

- Yeah, so Ercot is the entity that manages the flow of power in the state of Texas, or at least in the large majority of the state of Texas. Serving roughly 90% of the demand in the state. And what that means essentially is they're the air traffic controller of the power grid, essentially. In the sense that they're trying to balance supply and demand ensuring that frequency stays at 60hz. And yeah, just basically overseeing the actual clearing of the market itself as well.

- And I'm going to jump straight in with some of the detail, now. So in Ercot in Texas, there's something called locational pricing, or locational marginal pricing. LNPS, sometimes, if you're listening in Europe to this, there is a move in some countries and some systems in Europe to move towards locational pricing. Lots of other parts of the world have already got it. Sounds complicated, and it is complicated, but it serves a certain purpose.

So we're going to talk about the question that came in is, what is LNP? And what is locational marginal pricing? And how does it work in Ercot? So Brandt, how does it work in Ercot?

- I think the easiest way to think about it, right. Is you have probably what you would traditionally think of as your regular power price, if you're coming from Europe and just the non LNP method of doing of formulating prices. So you have your general system reference price. And then from there, you're familiar with the idea of congestion, right?

So you have--

- Never heard of it.

- Right, so the idea of a transmission element being overloaded, or being at risk of becoming overloaded is something that a governing body or an entity a system operator like Ercot is looking to manage, or any other system operator.

- A transmission element. As in an overhead line or--

- Line or transformer.

- A line or a transformer.

- Generally would be the infrastructure, yeah, most in question there. But basically the way to think about that is the way to help alleviate this these potential overloads. Or these active overloads potentially if an if a line or a transformer has more power than it can handle flowing across it at any given point in time, is the way the market is designed you have prices all across the system. And at a settlement, sorry--

- It is a hard thing to explain. We're actually going to keep this bit in, because this is good context for how hard it is to explain this thing.

- Yeah, so--

- That's start on the top, right. So we've got you've got the whole of Texas with a big transmission system across it that Ercot manage. And there is not one price across Texas right, there's lots of different prices.

- Yeah, so the crux of it is, yeah. Like you're saying, there are hundreds or even thousands of different reference prices across the system. And the way those are derived are if there is a currently overloaded transmission line or transformer or potentially at risk transmission line or transformer, then you have--

the economic dispatch has the ability to increase or decrease prices relative to that system reference price to essentially incentivize generation or just flexible resources on the system.

So that could be a demand resource essentially as well to either raise output, reduce output, increase consumption if it's a battery that's charging, or a demand resource that has flexibility or decreased consumption. And that's functionally the main idea behind LNPs, is really they're there to help resolve congestion and protect all elements of the power grid or the bulk electric system while still providing power to consumers in the most cost efficient manner.

- And to give price signals to encourage certain behavior.

- Yeah, both to encourage certain behavior and in real time. And also, if it's over a longer period of time--

so if you have recurring congestion issues, so the same type of issue is happening on a regular basis theoretically that would also incentivize actual buildout of additional infrastructure to help alleviate that issue. So yeah.

- So if in one part of Texas it's often congested. And so the prices are higher in one part of Texas, you might think I'm going to build a gas peaker there or I'm going to a gas powerplant or something like that there, to take advantage of that price difference.

- Yeah.

- And so, why is it called locational marginal pricing?

- Yeah, so the idea of marginal it goes back to that system reference price that we're talking about. So so generally, I think in the UK you guys typically call that like your short run marginal cost, is that right?

- Yeah.

- We don't have to get all the way into it.

- Well, so yes and no. But yes, it's through a different--

coming at it from a different angle.

- Right because it's not centrally dispatched.

- Yeah, it's not. Yeah.

- So, but--

if we were to imagine that it was, you could probably think of it that way. Where your next megawatt, or your next amount of demand--

- In the merit order.

- --that needs to be procured, essentially costs this much. And that's generally what I mean by saying this system reference price, slash, system marginal cost, slash system lambda is typically the term used in Ercot.

- And how many of these are there? Is it--

you said hundreds or thousands or what?

- Yeah, I don't remember if it's more on the order of hundreds or if it is actually in terms of 1 to 3,000 different reference prices. It's very distributed the idea of having all these different nodes rather than larger aggregated zones, is you allow for just more complex management of--

and more granular management of the system itself. And it allows for attacking a congestion problem with a lot more nuance for example.

So for an example, if you have just one transmission line that's potentially at risk of becoming overloaded, you might only have a small group of generators that are actually really driving the potential congestion on that line. And most of that's just because they're electrically quite--

sorry, electrically quite close in the sense of they're only like one or two stations away, or maybe a little bit more.

But essentially they're not super far in terms of--

there's not a lot of--

there aren't a lot of pathways in between them to for the power that they're generating to travel across. So again, think of this, so there's different angles in terms of the question of what's marginal mean. And one of the angles is being--

if we think about a generator that's sitting at 50MW.

And it increases its output to 51MW. Essentially, a percentage of that increase can flow over this potentially at risk transmission element.

That's what's called a shift factor in Ercot, which is actually a part of deriving our locational marginal price.

- A shift factor?

- A shift factor, yeah. So and that's essentially to say that if say, an increase in that generators output is potentially harmful to the constraint or the overloaded element in the sense of it makes the overload worse, then it would have a herding shift factor, and in the sense of if the economic dispatch is going to curtail that generator, basically each megawatt that it curtails that generator by it's getting x percentage of relief.

So if it has an 80% shift factor, it's getting 0.8mw

of relief for every one megawatt that unit is curtailed. And so that's a piece of deriving that locational marginal price. And then the other piece of it is basically just, how severe is this overload? How far away are the generators that are going to be used to help resolve this potential overload? Yeah, so those are really the two biggest pieces, because if you have generators that are further away, their shift factors are going to be weaker, they're going to be less.

So basically you have to curtail more generation to resolve this issue and if you already have a transmission element that has a potentially larger overload, so say it's overloaded by as much as 100 MVA if it's or megawatts, we'll just keep it simple. Basically a larger transmission element.

- On the big old transmission line. Big line.

- Big overload.

- And you want to stick an extra 100MW down it, which you really shouldn't at this at this point.

- Right, so yeah. Basically, it's at risk of becoming overloaded if you lose another element somewhere nearby. And to resolve that, we have to back off quite a bit of generation and raise a bunch of generation elsewhere. And that derives the other term in the locational marginal price formula which is called a shadow price, it's a very ominous term.

- Shadow price.

- Yes.

So basically, that's those two components make up the basis between your--

if you're a generator, your locational marginal price, or your resource node's price relative to the system's overall reference price. And if there are multiple constraints currently active on the system. So multiple transmission elements that are potentially at risk of becoming overloaded, technically, the way that the economic dispatch works in terms of generating these locational marginal prices is it's going to look at every single act of constraint.

It's going to look at your generators impact on each of them. So basically your generators shift factors across each of them. And then it'll consider those shift factors against the shadow price, and essentially it does a big summation across all of them.

- To each generator or each unit's contribution to the level of overload.

- Yeah.

- And then applies--

it's like a weighting to that. And then shifts the price at that node up or down.

- Yeah, based on whether it can help or whether it helps or hurts the overload.

- And when you say we, just to be clear, so when you're talking about we have to back off that generator, or we have to increase, that's essentially Ercot's control room has to do that. Because Ercot is essentially dispatch system where the control room is deciding who gets turned up, and who gets turned down.

- Yeah, except it's all automated.

- But sorry, I mean yes. Yeah.

- But yes, it is all coming over the ICC, or it's all communications network.

- So what are those people in the control room even doing then? No, we did that--

- Making sure that it's all working right, that's all it is.

- Just pressing a button.

- Keeping an eye on it. But it's a fun job.

- What's this--

can we talk about congestion then in more detail.

- Yes.

- Yeah. I think most people will understand what congestion, the concept of what congestion is, right, which is where you have to move power from one place to another, you have an overhead line or a transformer, or you have a whole load of kit, equipment. And that equipment is rated to a certain number of megawatts that say--

and you don't want to go over that, you don't overload it.

So there's some congestion there's lots of little bits that are adding up together. Congestion is fundamental to how the system works over here, isn't it. It's fundamental to pricing, to optimization, to trading. So let's start with how much congestion is there, if you can answer that question?

And why is it so significant in Ercot, and where does congestion often appear? I'm aware that's a bit of a trick question, but--

- Yeah. Yeah. So I guess we'll say how much congestion is there? Yeah, I mean it's a tricky one, I guess for--

and I'm going to throw a number out, but it's still like hard to tether that to--

well, what does that actually mean?

- But conceptually it's a weird question.

- Yeah.

- Have a go.

- I'll say for example, I know in 2022, I think the constraint that caused the most additional cost to the system--

in terms of--

and by additional cost, I mean you have to redispatch all this generation to help resolve this congestion. And as a result, that usually means that potentially more expensive generators are filling the gap of this curtailed generation. So that essentially becomes this term called congestion rent. And the idea of congestion rent, yeah, that's the term.

- Is that real? Is that like a colloquial term? Like that's just congestion, right. Or is it actually this thing is called congestion?

- Well, I don't know about in an academic sense. But I know it's a term used in the independent market monitor.

- All right, so cool, cool.

- It's somewhat legitimate. Yeah, so essentially the I think the number one constraint in like 2022 was 300, or $350 million of basically additional system cost, which is basically what congestion rent is. So if you think of that as like a layer on the overall dispatch stack or the overall marginal cost stack, I don't know exactly what that would be in terms of percentage of the overall system cost. It's probably not like gigantic but--

- It sounds like a big number.

- But it is, I mean, functionally like $350 million a lot of money, right.

- But then that's the market doing the thing, right? The market is designed to have those differences. The market is designed to encourage solving the congestion problem with price signals, and encouraging behavior. So a guess is a good thing.

- Yeah, I mean, I think as long as it is rectifiable by the system operator on a given operating day, as in the economic dispatch is able to redispatch that generation. And I said, I don't know exactly what it is as a percentage of system cost, but I'm fairly confident it's not a gigantic number. So you could say it is on some level a necessary evil, because not only like you're saying does it--

is it the system working as the way it's intended to, but that the next step of the system working the way it's intended to. We were saying earlier, is it also helps things out in the long term.

Where it's you're helping identify like where the greatest needs are for transmission projects, which are really expensive to build often get passed to the consumer which isn't necessarily ideal. And then on the other hand it's a great tool for like somebody who's maybe like a battery developer or something like that to help. That's a really good starting point in terms of identifying like, where do I want to put my battery.

Well, here's this congestion issue, which it's a mutually beneficial thing where it's the battery can come in and be very helpful to the system in terms of helping to resolve this congestion issue. But also it's good for the battery because it's good for your overall price spreads in terms of what you're seeing on a given day.

- It's a codependency.

- Yeah.

- It's a codependent relationship.

- So I mean, it is--

it's, I think, a very interesting market design and it's definitely--

I'd say it's probably the closest thing that you're going to see to in some ways free market economics in power grids which--

- I love it.

- Usually don't really have those necessarily. So it's definitely it's a fun one to look at.

- I mean, there are lots of trade offs. And you get lots of big numbers like $350 million. You get all these headlines, which I think make people sometimes squint a little bit. But on balance across, I just think it's a magnificent system, but that's coming at it from a sort of European mindset.

All right, and I'm also aware that I'm totally screwing you over here, because I've got these bullet points. And half the questions I'm asking are not on the bullet points. All right, large flexible loads. Question came in about that, where to start? What is a large flexible load?

- Yeah, so that's the term that's been assigned to, I'd say, primarily data centers and Bitcoin mines are the two largest denominations of what a large flexible load would be. But generally it's large as in, on the order of tens or even hundreds of megawatts of additional demand on the system all located at one point.

- In the olden days this used to be a factory.

- Yeah.

- And in the new days, this is Amazon Web Services, data center.

- Yeah. Supercomputing, or Bitcoin miner, yeah. What, which, kind of the same.

- But which is exciting in a different way.

- Yeah, I mean it's definitely--

Texas is the heart of innovation, you could say, but yeah.

So I guess going back to that whole idea of large flexible load, so yeah, you said in the old days you'd have a factory. And that's more just large load. Now, it's the idea of have the flexible component to it, which is really interesting idea in the sense that you have basically these really large demand resources that theoretically there--

they have this they have their own strike price, right.

In terms of what they are willing to pay to consume power from the grid. And they're looking at again, they're looking at their locational marginal price and theoretically the system in reference price if those two aren't diverging at any given point in time. And they're saying know we are not willing to pay more than x amount of dollars per megawatt that we consume.

And you know, anything above that, we will curtail our consumption basically to minimal power levels like 5% of their overall regular operating consumption because that's just part of their business plan. But at the same time it goes hand in hand with the natural market design that Ercot has in the sense that the system is designed to have prices go high in times of scarcity is the fundamental piece of that market design. So it's another, you know, incentivized participant in the market, right.

- So sometimes it's called--

it's a bit like demand side response DSR, or what other names are there for it. It's large place--

large consumers of power that can switch off or curtail or turn down for periods of time and get paid to do so. And I guess what's special about locational pricing is that if you're going to build a Bitcoin miner, say you want 100 megawatt Bitcoin miner, you choose where to put that in the world or within Texas in places where there are low prices, which is places where it's OK for the system to have that there it.

Kind of makes sense right? You wouldn't be at the middle of a load center like Houston or whatever, and put it put a Bitcoin miner in the middle of there. But you might do it out where there's excess wind or the other side of constraint or something like that. It really is boomtown for Bitcoin miners over here.

And so all of these folks are participating in the market through--

a bit like load flexibility. So what does that mean? Does Ercot dispatch them directly? Do they get turned off directly from Ercot's control room or from the algorithm?

- I would say as of now, no. But that's still something that's in the process of being ironed out in terms of legislation. Yeah, there's a whole task force at Ercot in terms of the stakeholder process that has been developed to basically handle the question of how do we integrate these multiple gigawatts of large flexible loads that are coming online and in close proximity to each other in terms of time. And how do we basically design a new classification of resource for them to operate within.

So you know, Ercot has a classification called like a controllable load resource. But historically that's pretty small resource for it's basically one side of a battery, right. So this is essentially just another technology type in a way where it's just having that means that you have to in some level design a whole new set of rules, because it's just like a different way of operating and some of them don't necessarily have the capability to have, as far as I know, they don't necessarily have the capability to step from full consumption, to partial consumption, to zero consumption.

I think some definitely do. And it's probably a case by case basis. But I think in the sense of it's not, it's not as standardized as building a natural gas powerplant, or even at this point a wind farm, or a solar farm, or battery. So it's just more in the infancy of things. And I think that's part of what makes it really interesting.

- I bet out of the data centers have got backup generators anyway, they can probably ride through one of those periods, like with diesel perhaps. And then--

- Yeah, I think there's different approaches to it where some of them are co-located with a wind or solar resource, and they might have an offtake agreement with them. I think there's really probably double digit different types of approaches if you were to look at both data centers and Bitcoin mines, and look at all the different--

basically, from when they've come online and across different companies the way they're kind of approaching trying to solve this problem.

- You think for a Bitcoin miner it'd be pretty straightforward as well, because you kind of hashing out to figure out to--

what every 10 minutes there's a new block on the blockchain, right. So surely you can just switch that off and switch it on whenever you want, but I guess there's a whole load of comms and soft star and various other things.

- Yeah, and I think that's why when they're curtailing from 100% to--

it's not to 0%, it's usually probably like 5%, just to make sure that you're preserving you know, your barebones functionality, cooling and all that, so yeah. It's definitely it's an interesting thing to see begin to take off. And it's another thing that it's very much, like this is, at least in North America, it's very much happening in Texas and not necessarily anywhere else. So it's another piece of what makes Ercot an interesting market.

- Because since they banned mining in China, however you can do that. But you had all these Bitcoin miners, these companies just getting all of their hardware and flying it out to Texas, right. Because they used to move between, I think they used to move seasonally between when there was loads of hydro, and like the--

I'm going to get this wrong, but essentially low fossil fuel power prices and then hydro.

- Yeah.

- But because they're pretty mobile, they could just stick--

I saw some pretty funny things on a substack, some funny images of all these Bitcoin miners on planes flying to Texas. It's good for the economy down here, I guess.

- Yeah.

- But a little bit controversial, isn't it?

- Yeah, yeah. I mean, I think you definitely would get a variety of responses depending on who you ask.

- I'm on the side of the miners.

- But I don't think--

I think in the sense of I'm coming at it from my viewpoint being more so on the grid side of things.

- Yes, I think I see the potential for another useful and flexible resource on the grid. And I tend to think that's a positive overall, so yeah. In that sense, that's my quick take on large flexible loads on the whole.

- OK, so we just went down a rabbit hole trying to explain ancillary services. And we got lost. I'm going to start again, so we've had our first go, which we've edited out. And now here's the second go. So question is, can you explain ancillary services in Ercot, over to you Brandt.

- Yeah, so I think we'll break it into two primary buckets. So the first one being you have your dynamic frequency service, which is basically just keeping frequency in line with 60hz, and managing the small deviations away from 60 that you have. And the second one being, you have your various reserve services, which are made up of different types of ancillary service products that each have their own criteria as to when they're deployed, and then how resources get qualified to carry them, or to have those have responsibility in the form of these different ancillary service types.

And among those, you have a frequency reserve service, and I know I already said that you have one service that's maintaining frequency, but that's more for those smaller deviations like I was saying. And you have another one that's there to act as a backup in the case of--

you lose a large generator trips offline, and you lose hundreds of megawatts of generation in a given point in time.

And that might cause frequency to deviate pretty far away from your set point of 60hz, right. And so at that point your regulation service that's just trying to manage those smaller deviations, isn't necessarily going to be enough to help recover frequency, so you need another product or another ancillary service to be deployed quickly to help recover that frequency.

And now, with it's a new change in Ercot that happened last year in June, they implemented the Ercot contingency reserve service, which basically took the place of another ancillary service that is actually still active but it's basically deployed less often.

- Every TSO has a different problem.

- So it's a whole web, but yeah, without getting too lost. Yeah.

- Let's go a la carte. OK, so what have we got. We've got rig up, rig down. Let's do those ones first.

- Right which is your small deviations like we were talking about. And then next you have recovering the large frequency disturbances and as of now, ECRS, or Ercot contingency reserve service is a primary one that does that, so below, I think it's 59.91hz

you'll see the CRS, Ercot contingency reserve service get deployed. And then an even larger frequency deviations, and under different sorts of other criteria you might--

- Worse the contingency. This is the next level.

- Yeah, yeah. The naming can be tricky to get your head around too, but there are certain times where you'd have the response of reserve service or PCRS also be deployed to help recover frequency. Now, moving away from frequency, there's also the notion of, we need to have reserves at the ready in the case of scarcity conditions. And that can look a couple of different things.

So the primary one is what we would probably traditionally think of in terms of scarcity conditions, which is just we have a ton of demand on the system at a given point in time on a given day and we're struggling to meet that demand with the generation that's freely available and online and operating right now. So we have the--

this basically additional contracted demand, which is procured in the day ahead market to act as reserves that can be called upon in those traditional scarcity conditions.

And that is served--

those the services that generators can have responsibility for and be providing that type of service is again PCRS, so it's still the same service, it's just a different criteria under which it's deployed. And those are going to be generators that can ramp up to full capacity, I think it's full capacity, in 10 minutes. So they go from being offline to at full output in 10 minutes. So that's going to be like true peaker gas plants, as well as, batteries mostly.

Those are going to be the two primary components.

- Still quite slow though, isn't it? 10 minutes is still quite a long amount of time long.

- Yeah, so that's why--

- Time, long time.

- Yeah, so that's why you have--

that's going back to their control room, that's something that they're looking at, right? So in terms of have your various forecasts of, we know we have a load forecast and we see that load it's currently ramping, this is where we roughly think it will be in 10 minutes and 30 minutes time. Similarly you have wind and solar forecasts. Since you can't actively dispatch wind or solar, you just relying on where it is at any given point in time.

And if this gets into the second aspect of these reserve services where you might have enough generation to meet demand over a longer period of time, like over the course of like an hour or two. But you might be in a transitory period where--

and we call this the best summation of this whole idea is a net load ramp. So net load being--

- Net load ramp?

- Yeah, net load being have the load component, which is just your overall system demand. And then the net component being that you're taking the wind and solar generation out of that stack. And so what you're left with is your net load, because that's--

those are the aspect you can't necessarily control as a system operator are your demand and your intermittent generation resources.

And so, from there you have this remaining demand that you need to meet by your dispatchable generation stack. And so if that is a quickly moving target and in particular, if it's quickly moving up, which is usually driven by potentially an actual demand increase. But in particular, a decrease in wind or solar. And in particular--

- Cloud.

- Yeah, yeah. Honestly, the biggest the most common example of it being usually just at the end of the day when the sun is setting you're going from quite a bit of solar output to zero solar output and as little as an hour and a half usually as the sun sets across the state of Texas. And so yeah, to make this whole long winded point to wrap it up, you might have a period of time where--

- Oh, by the way I'm here for it, you go. You go.

- Over a short period of time, you might have a short period of time being like 10 and 30 minutes, you might have quite a big ramp up in this net load term, and we call that a net load ramp and if it's happening very quickly and at a large magnitude. You might, despite the fact that you have, theoretically, you have enough capacity available on a given day, some of that generation might not be able to ramp up and come online quickly enough to fill the gap that's being left behind by the wind and solar generation, slash the additional gap that's being created by a demand increase.

So you might be saying, all right, well, we're planning on having this large conventional generator come online at the top of the next hour, but they have to actually ramp up and get ready to be unlined before they're there. And so to fill that gap we have these reserve products and that's going to be, again PCRS, and lastly your non-spin reserve service or non-spin.

- Non-spin.

- Non-spin, which is just what it sounds like where--

maybe not what it sounds like, but explain the naming on that one.

- Please write in to our P.O. box with what you think it sounds like.

- Yeah, well, I'll go back to the naming in a second. But the idea basically is you have generators that can actually ramp up to full output in a short amount of time. And kind of fill in that gap. So those are your two periods of time where you're going to be deploying your reserve ancillary service products.

And yeah, just to throw in the why it's called non-spin, I guess, is the idea is at least before batteries were commonplace on the grid, it was going to be exclusively carried by dispatchable resources which typically in Ercot in the last decade or so would mean especially fast responding ones would mean natural gas resources that are--

probably call them peaker plants, they may or may not be called that really at this point but because it's not necessarily at peak anymore in terms of demand at least.

But yeah, so the idea of non-spin just being that you have these fast responding conventional generators. And when these generators are online they're going to have an actually physically spinning mass within them, so when they're offline--

- They're not spinning.

- They're not spinning.

- They're not spinning like spinning reserve or whatever that you have in some markets, but we're not going to go into that here.

- Yeah.

- While we're talking about spinning things, how does Ercot manage inertia?

- Yeah, that's a good question. So--

- Also not on the list, but here we go.

- Yeah, no we'll jump into it. Yeah, so basically Ercot has minimum system inertia targets at any given time that they have determined are necessary from a frequency control standpoint. And what I mean when I say that is, if you have lower than x amount of system inertia, it means that if you were to lose a large generator, so again, if you have a large generation trip, that could potentially mean that your system frequency will deviate further from 60hz than if you had more system inertia at any given time, right.

- So you want lots of spinning things on the network so if something trips, everything keeps spinning.

- Right. So--

- At 60hz, or however many 1,000 RPM.

- Exactly. Right, so that's the idea is basically they have.

They have an idea of where they need to be at for different levels of demand. And so in the control room for example, if you were to go below a certain inertia target essentially, you're going to start to see alarms essentially going off on your wall board right, so on your big screens at the front of the room, you'll see an alarm for system inertia a warning to say, hey, we're getting a little low on system inertia, maybe, and if it's to a point where it's severe enough where there's actual risk of like hey, if we were to lose one of the larger generators online on the system right now we would experience such a large frequency decline.

Basically it's larger than they'd prefer to have to recover from in the sense of you might have you might have--

might get to the point of load tripping off if frequency is that low.

- And so largest loss of load coverage. By the way, is it the inertia alarms?

Are they calculated or are they measured?

Are calculating inertia based on I have this many--

I've dispatched this many gigawatts and I know it's spinning plant? Or is it measured inertia somehow? And there are a few different ways to measure it?

- Yeah, that's a good question.

And honestly I'm not totally sure what the answer is. But I think it depends on whether different generators can provide different amounts of inertia, because if not, then you could basically just say, well, I think that would be the case given the fact that I think variable output would mean a different amount of inertia from a given generator. I'm not totally sure about that, it might also--

- I'm a little bit fascinated by measuring inertia, but that's a bit of a geeky topic.

- Yeah. It might also--

- Not going to go into right now.

- Yeah, yeah, yeah.

- But anyway, so there's alarms on the board. And they go off if there's not enough inertia, and then I guess you guys have got to bring on some more conventional plant.

- Yeah, exactly. Ultimately, if inertia was low enough to the point where there would be concerns about frequency recovery in the event of a large generator loss, and to cover for that I guess, to have another form of contingency planning I guess, you would bring another conventional plant online just to essentially boost your system inertia.

- All right, do you know what we haven't talked about on this list is real time optimization. I think this question is related to well--

this is an abused phrase, isn't it? Core optimization. But I think we're talking about how some markets like California or Australia, you can co optimize in ancillary services with batteries. Why can't you do that in Ercot?

Actually, what is that? And then why can't they do it in Ercot?

- Yeah, so real time co optimization of ancillary services, right. So what that means is--

so in Ercot right now, all ancillary services and all the responsibility for them is procured and assigned in the day ahead market. And the real time co optimization essentially would just mean that rather than doing that all in the day ahead market, you'd be doing it in real time, in line with your economic dispatch.

So essentially it would just mean that with every dispatch interval so every 5 minutes when Ercot runs its economic dispatch at the same time you'd be saying, well, we have this amount of each ancillary service type that we're looking to procure, we have generators that are offering and to have responsibility of different ancillary service types at different price points. And similarly are offering to generate or consume if it's a battery, certain amounts of energy again at different price points.

And so Ercot's economic dispatch is going to look at all those bids and offers across both energy and all these ancillary services, and basically just try to come up with the least cost solution to fulfill all of its ancillary service procurement targets, and also make sure it's procuring enough energy so you're meeting current system demand. And essentially why that's desirable is, the conditions of the day ahead market are not the same as real time, right?

And so the closer you are to real time, the more information that generators are going to have. And they'll theoretically be able to position themselves in a better way, and at the same time Ercot may be able to end up with a cheaper overall solution in terms of dispatching the whole market.

- You should get more liquidity, because there's more information closer to real time. And so the end consumer or control room should get a lower price on average, you'd think.

- Yeah.

- But it's more risky, isn't it? Because day ahead you can lock it in and you're going to get the--

I don't know, reserve product that you want, or the ancillary service that you need. If you're doing that every 5 minutes, what if you don't have the ancillary services you need? I guess, is that the reason why you can't do it?

- I think it's more of just it's really complex to switch systems that are legacy systems that are designed a certain way.

- Oh what an anti-climax. I thought it was I thought it was some risk management system.

- No, because I think you have--

I mean, theoretically, the risk would probably be more so that you'll end up with more expensive ancillary services in some cases then you might if they were only procured on the day ahead market. How that squares with actual cost of energy as well. It's hard to say. I mean you might still end up with an overall cheaper solution even if you were to compare in a vacuum your price of ancillary services with and without real time co optimization.

You might end up with higher ones in real time potentially given the fact that you might have generation that's holding itself back in certain scenarios at least, that wants to preserve that flexibility. I still think it's--

especially with where Ercot is at in terms of its storage buildout and things of that nature. I think you have a lot of generation that's looking to procure ancillary service responsibility and looking to carry that in real time just because they want to have that secure revenue stream.

And I don't think that would be going anywhere with real time co optimization. So yeah, I mean, I think overall it would be likely to help make prices overall cheaper rather than being a potential risk of we don't have enough answers--

like, we don't have enough in terms of offer volumes to actually fulfill the procurements we need. And I think that would be backed up in terms of what the offer volumes in the day ahead market look like.

And like we're saying, that might not necessarily carry over to real time 100%. But I think there's enough of a buffer there where I don't really think that would be a concern.

- Now, Brandt, on to everybody's favorite question. So what is your contrarian view?

- Yeah I was going to what I'm going to go with here is the idea that I think 2024, one of the big headlines for batteries is going to be ancillary service saturation. And what ultimately I think that means is that the top battery on moto's 2024 year end leaderboard will have gotten over 50% of its revenues from energy arbitrage. So yeah, to summarize that a little bit is that typically when batteries are coming into any energy market, a lot of their initial revenue stream tends to be from ancillary services just because it's a secure form of revenue and it also allows them to cycle less and protect the lifespan.

Those are the primary reasons I'd say. But once you have enough of this battery energy storage buildout that occurs, and you have more batteries competing for these ancillary services, you have more volume being offered in than is procured, which ends up suppressing prices. I think that's something that we've talked about on the last episode, but that being said, I think--

and the reason why I guess going back to the actual take itself, that 50% feels a little bit contrarian is that last year I think energy arbitrage revenues made up something like 7 or 8% of total battery revenues in Ercot.

So for your best performing battery to be getting over 50% of its revenues from energy arbitrage would, I think, be quite a significant leap. And so yeah, I mean that's functionally why I think it's a novel concept I think in terms of batteries operating in Ercot.

- Operationally, it's completely different way of running a battery to--

I mean, you're doing rather than--

I don't know, the switching on here and there for a bit of frequency response. You're cycling these things really hard, you're running them at higher power for longer durations that you've got thermal management to deal with. And it's what batteries were designed to do, right? This is the whole point they're supposed to be moving power around, not just little fluctuations in response to frequency.

- Yeah.

- I think it's very exciting.

- Yeah, I guess I guess really the crux of it is--

I think it's--

the point is that the juice is worth the squeeze I think in this case.

- The juice is worth the squeeze, love it.

- Is that not a British colloquialism?

- I don't know. I don't know. I've been talking so much about squeezing avocados recently, for some reason this topic has come up loads of my life over Christmas. I don't know why. And so anytime anyone talks about squeezing things, I'm thinking about avocados.

- Fair enough. Fair enough.

- Well, actually, OK. So separately, it's a thought experiment. This is why it came up. And you can thank Jack, moto Jack for this, is how many--

it's like an interview question, how many times in an in an avocado life on average do you think it gets squeezed?

And it can get because everyone looks everyone squeezes it when they get in there. But it probably gets squeezed to be categorized when it was first picked. And so that's a question we're trying.

- That is such a Jack question.

- That's such a Jack question.

- Yeah

- But yeah, we should probably leave it there, we've gone well over time.

- Yeah.

- Brandt, I want to say a massive thank you for coming in and talking about squeezing avocados. And yeah, if you're listening to this, thanks for all the questions and the feedback that comes in. It's wicked, we absolutely love it. And hopefully we did a good job of going into more detail.

- Yeah.

- Until next time.

- Thanks for giving me another chance to hop on. Till next time.

- Thanks man.