This Is What It’s Like to Run a Power Grid

You are listening to Shifty, a Heat Maps Weekly podcast about decarbonization and the shift away from fossil fuels. Today we are talking to someone who helped. Keep the grid up. We're talking about what they did today to keep the grid operating, at least in California, what they did yesterday and what they did five years ago. It's all coming up on shift.

Want to accelerate your career in clean energy? Then explore online certificate programs from our sponsor, the Yale Center for Business and the Environment. Whether you're designing policy, unlocking finance, or developing impactful projects, Yale's online clean energy programs equip you with tangible skills and powerful networks. and you can continue working while learning. In just five hours a week, propel your career forward and make a difference.

Listeners of ShiftKey and Shift Key Summer School can now get an exclusive offer. Use the referral code HEAT MAP twenty five and get your application in by the priority deadline to get five hundred dollars off tuition to one of Yale's online certificate programs. in clean energy. It's time to transform your career, your community, and the clean energy landscape. Head to cbey.yale dot edu to learn more.

Join Clean Energy Leaders at RE plus twenty five from September eighth to the eleventh in Las Vegas. As North America's largest event for the modern energy industry, R E Plus brings together innovators across solar, energy storage, EVs, hydrogen, wind, and more. It's where industry leaders tackle today's challenges to shape tomorrow's energy solution.

RE Plus is the epicenter of clean energy leadership, uniting the industry to meet rising global energy demand. And listeners can save 20% with code HEATMAP20 at replus.com.

Hi, I'm Robinson Meyer, the founding executive editor of Heat Map News. And I'm Jesse Jenkins, a professor of energy systems engineering at Princeton University. And you are listening to Shift Heat Maps weekly podcast about decarbonization and the shift away from fossil fuels. Well Jesse. Over the course of this summer we have been running this series we call Shift Key Summer School and we have talked about the very basics of the electricity system from literally what is a watch.

We built up, we went through history of thermal power plants, history of renewable energy. Last week we talked about how electricity markets work. And now we're going to talk about a topic that I think exists at the very edge of my electricity knowledge. Which are these grid operators? Yeah. Who actually keeps all this stuff running, right? Who keeps the grid operating safely, reliably, and affordably every day?

There's a variety of entities that play that role across the country and we are lucky enough to talk to the chief operating officer of one of those grid operators today, the California Independent System Operator. Yeah, so today we're talking to Mark Rothelder. He's the senior vice president and as Jesse said, the Chief Operating Officer at the California Independent System Operator.

He has been there for a long, long time. He was part of the original startup team at California's Grid Operator in nineteen ninety seven and is now the longest serving employee. of California's grid operator. So not only is he senior vice president, he is You know, when we say he keeps the lights on, he literally has kept the lights on for California at this point since nineteen ninety seven. On this show we sometimes talk about ISOs or RTOs.

Or sometimes we talk about electricity markets or local grids, things like PJM, which is the regional grid that covers New Jersey, Maryland, the city of Chicago, parts of Ohio, DC.

We sometimes talk about the New York ISO, sometimes we talk about the New England ISO. They're all grid operators. Just think of them as grid operators. You'll hear us use different names for the grid in this interview, but just keep in your head, we're all talking about the same kind of thing. They're all grid operators. All right, let's get to it. Classes and session. Mark, welcome to ShiftKey. Thank you, Rob. It's a pleasure to be here.

All right, Mark, I think we're gonna start with the most important question. Is it queso or kaiso? It's not the cheese, so I think it's kaiso. Okay. All right. Good. So let's start here. What what is a grid operator? Well, how would you describe what Kaiso is as an entity? Yeah, so uh a grid operator you can probably think about it in the terms of the air traffic controller of the grid, okay? And the air traffic controller doesn't fly the planes.

Um, they're not the passengers in the planes, but they help manage traffic and kind of build security into the system. The same thing with the grid operator. Our measurement of security is we're balancing supply and demand. uh at various scales all the way down to the second by second. And that ensures that we are maintaining frequency of the system, sixty hertz.

the other responsibilities they have is to ensure that we're operating securely in the sense that we are not overloading any line. So as Supply tries to get to demand, it follows laws of physics in terms of where it flows. And obviously the transmission lines and the air that you see outside have limitations how much current can go through them and that dictates the amount of power that can go through them. And so we need to stay below those ratings.

Uh and we do that not only for normal steady state conditions, but also have to be prepared uh if there is a con contingency event. So if a parallel line is tripped or out of service, the flows on the system naturally redirect themselves, reconfigure, and we want to make sure that we don't overload anything above emergency ratings for that condition.

The big challenge for you though is that everything moves at the speed of light. So You have to make sure that if something goes wrong, it's still stable that you know, the system will just automatically adjust based on physics because you don't have time to intervene the way an air traffic controller would in those kinds of circumstances. No, but because it is

Based on the laws of physics we can do calculations and do pre-planning or know what would happen in case of a contingency. So yes, it would move at the speed of light, but we can get ahead of that through simulations and assessments. And then I was going to say that the third part of our responsibility is also being ready for uh a contingency in the case of a loss of a large resource. We usually talk about having operating reserves or contingency reserves.

And so maintaining a level of unloaded capacity that you could dispatch as needed is part of that responsibility. So those are the three main things. Balancing. Security, not overloading any lines, and then maintaining reserves. And then the fourth one is in case you do have some kind of disturbance events that takes out the system, they have the responsibility to help building the system back together, which could take hours to days to do.

Yeah, and hopefully we don't we rarely, if ever, have to do that. It's called black start, turning things back on after the s the grid goes down. So a lot of different critical tasks. Again, everything dealing with physics that moves at the speed of light.

So yeah, let's start here. What is it that you're that a grid operator like Kaiso is doing right now, second to second or hertz to hertz, to make sure that the grid works and the lights stay on? Yeah, so the operators are constantly monitoring both the frequency of the system. We talked about that sixty hertz. If the system is running a little slow, slower than sixty hertz.

It'll basically send signals to resources to increase their output to get back to that balance. If you're running a little fast, they'll say, okay, let's back down some resources and s reduce the production so we can maintain that balance. The other thing the operators are doing are because you're part of an interconnected system, um, they're controlling how much energy is being transferred between them and their neighbor. So they usually schedule that on an hourly basis.

And that also becomes a control point. So it's both frequency and maintaining the interchange with your neighbor. And the two put together is what we call area control air, and that is our control basis, trying to maintain that control area control air as near to zero as possible. So because the grids you know, we've talked about this in the past, I think that we have three big interconnected grids in the United States. Well two big ones and then Texas doing its own thing as usual. Um

And with the exception of Texas, there are multiple grid operators responsible for little fiefdoms within those interconnections. And so the way that you all make sure that everything works together is you basically say, we're gonna fix the exchanges between our regions at the boundaries.

And we're all responsible for making sure those levels stay where they're at. And then once that's fixed, we can kind of all do our own thing locally and just make sure that our frequency is secure. And if you do that, then the system stays stable across the whole region. So C correct. And just to elaborate on that, generally speaking that's done uh at an hourly granularity. There is with the Western

energy imbalance market. We'll get into markets a little bit later, but just to give you a little bit of preview, that ability to balance or change your interchange with your neighbor actually gets reset every five minutes. why this is important to your point, Jesse, that otherwise you would be up within the hour and any deviations, load increase, change in supply,

would have to be absorbed by only those resources in that one balancing area. Now you've got the broader footprint across all balancing areas that can that could help balance the system out. And so that granularity Um you can sell to somebody who's a little short and it all cancels out and keeps things you know. Yeah. Exactly. Exactly. What does the room mean? where this happens looks like. I mean Oh, it's super cool. Is it like uh evil lair with lots of screens and big nozzles?

Well I'm not sure about the nozzles, but yeah, lots of displays, looking at information across the West, you can see transmission lines, you can see voltages, monitoring across the system. And there's probably about six to nine desks out there, and those six to nine desks have various

of responsibility. If someone's watching the generation, what's it's doing and that balancing. There's another desk that's watching to see if the transmission lines are not overloaded and watching that contingency analysis to see if they would overload in a contingency. There's another desk that's doing that interaction with our neighbors to set up the exchange. It's a little dark and so the reason it's dark is that you want them to be focused on the information.

And also with the lights, you don't want them to be distracted too much. So usually comfort lights, uh low, focusing in on the information that they're monitoring. What's your favorite job on the floor? You have you worked in these roles yourself in the past, right? I've never have officially worked on the roles but uh and and I've had the experience of being able to shadow and I I find the generation desk to kind of be

It's kind of like a a little bit of a game where you're trying to constantly balance that and you're looking at your area control error and uh I failed. I had this big I had this big wallet. So this is why they don't let you they don't let you in the room anymore. Exactly. I had this big wall of green for about thirty minutes indicating that I was not maintaining the balance and so that's my memory of failing at trying to do what I have asked them to do.

All right. And it's my understanding that all the the grid operators like I said you have a backup location as well, another secret or maybe not secret, but another backup location where you can perform these functions in case something happens to the primary control room. Correct. Yeah. So it's on a different part of the grid, but close enough that we can get data over there in a quick way. And so we it's a hot backup so we can operate from either location as we need.

All right. So that's the like real time operations, right? Got to keep the frequency of the alternating current at sixty hertz. That means all the generators out there are spinning at the same frequency as well and contributing to that inertia that supports the grid. What are non

synchronized resources playing in this sort of real-time operations. And is this something that uh you think is likely to change going forward as you have more inverter connected resources like solar or batteries or others that become part of that operational mix? Yeah, so you've already alluded to it. So the system has been built around being an alternating current system, which means

Everything that's spinning is synchronized, at least in the Western Interconnection. Something slows down one area, something's gonna happen on another resource because of that synchronism to try to maintain that sixty hertz. Inverter based resources are different because they don't have a spinning mass.

they produce energy and they go through what's called a D C to alternating current conversion. So they start out at D C direct current and then they convert up through power electronics to produce a similar sinusoidal signal. that then joins up with the AC system. We do have many more inverter based resources. This is uh because a lot of the new renewable resources are like solar producing energy in DC.

batteries producing energy in DC, uh wind, even though you think they're spinning, but they're actually producing energy and actually going AC to D C and back to to to A C in some cases. So as a result of that, and you've alluded to that, they don't have spinning mass. So if you have a disturbance that takes a resource out, the inverter based resources have to be designed to respond to that. So that means

you may need to hold back some production so that when you have to produce more to simulate kind of some of that inertia or production, you have the room to be able to do that. If everything's running at full output, you don't have that ability to electronically s deliver more when needed. So I know it sounds complicated. All these electronic integration have largely been worked out and uh th things.

Local control signals, right? That you could they have to be metering the frequency or voltage locally, but they can if they have the right programming, they can basically respond to that similarly to help maintain the grid.

And is that something Kaiso is actively doing now?'Cause I our understanding of the the blackout in Iberia that we've spoken about on a previous episode is that the transmission system operator in Spain is not really taking advantage of inverter based resources to provide uh voltage support or synthetic inertia at that kind of time scale. So I I know a little bit about what I've hear hearing about what happened in Iberian outage and I think that's right.

And that is if you're gonna bring these resources on, inverter based resources, they do have the ability to control voltage. So voltage is kind of the the pressure that keeps the current moving, keeps the power moving, you also have a certain level of threshold of what the voltage needs to be high or low.

And in the case of the Iberian outage it sounds like because of some light loading that the voltage actually started to creep up higher and higher. At some point it start getting into some of the protection devices tripping. and taking it out on high voltage and once that happened it cascaded from there. Yeah,'cause generators start disconnecting. Yeah.

Yeah. So it's important though to understand how those voltage settings are set across your system. But it's also important to and we do have requirements and expectations of the even the inverter based resources providing voltage support capability so they could help control the voltage. Um, and ensuring that they do helps prevent those types of things happening in an uncontrolled way as the Iberian outage experience was. So they're another tool in your toolkit now.

Yep, they are another tool in the toolkit, both from a voltage control perspective, but also from a frequency balancing perspective, and also from a flow control perspective. So we can dispatch these resources to change the flow patterns across the system. So it's not unusual where we get congestion and there's pockets of resources and you say, well you don't want to

reduce the production, but there's times where you have to reduce the production to effectuate and avoid overloading lines and staying secure. So they are like any other resource now. They provide services And that if you hold back some of the capacity, they could also provide contingency and operating reserves. So they are integrated into the system like any other conventional resource.

Uh they are different, so you have to be cognizant of that, how things are set, but we like to call them good uh grid citizens and providing services to the grid. to ensure reliability. What's changing now is that we have ability to store energy Historically, storage has been possible. Usually in a hydro system you can store water and use it some other later time. But now we've got batteries in the system.

that we can charge those batteries during the middle of the day when we have some surplus energy and then we can discharge those batteries in the evening when we need that energy. So we have a about thirteen gigawatts of batteries on our system now. We're a fifty gigawatt system So it's a material addition to the system. And they also, being inverter-based resources, they can operate and move very quickly. So they can provide some of those regulation services to control frequency as well.

So yeah, the batteries are even faster than conventional resources at changing output in response to those control signals that you're sending. So you're great, good great assets in that sense. So you alluded earlier to the idea that you schedule in advance the s flows between

regions as part of the effort to kind of divvy up responsibility for managing the larger grid. There's a lot of other scheduling that has to go on too, right? That you have to plan ahead in order to make sure that in the real time Everything is operated in a secure manner. So why don't you talk through some of those steps that you take maybe an hour ahead of real time? Like what did you have to do an hour ago so that my lights are on?

Yeah, so uh an hour ahead what we start doing is for those who are not in what we refer to as the energy imbalance market, we will set up exchanges with our neighbors. We minimize the cost of meeting what our forecast demand is. Uh and we do that and we can set up our interchanges based on bids that are offered for that exchange with our neighbors. So

leading up to the hour, we will exercise those bids and basically set up those interchange schedules. And that's actually happening about anywhere from seventy five minutes to forty five minutes before the hour starts. And so that's kind of done at the market level, the operators are watching that and those interchanges are handed off to the operators and they do a kind of a a a checkout with each other. So the balancing area operators are calling each other and saying, Hey

I I see you we've cleared about six hundred megawatts with you. Is that what you see on your side? Yes, six hundred megawatts. Got it. We're good. That's what we're good for. And everything kind of goes from there into the real time control system. if they don't agree, they they basically have to remedy that and come to an agreement

Everybody's gotta work together on this. Can't be off doing your own thing. Um w what else happens throughout the day? You've gotta schedule power plants as well, right? Particularly large thermal power plants that take a long time to get running. So now we're in the realm of what we call commitment of resources. So when you're in real time, there's a certain amount of resources, maybe fast start resources you could start because they're quick start, you can bring'em on for use.

But there are other resources. It takes time to get the steam, to produce the energy, to produce electricity. And those you start getting out in the realm of four hours, eight hours start times, depending on if they were already up and running at some time. So we do look out as much as four hours to do some mid range startup of resources, optimize commitment of those resources. And then once we're there committed, then we'll do the fine tuning dispatch.

from there through the real time market. When I talk about the market, I want to make sure it's clear that the market is doing the same thing that the operators are trying to do. So it's trying to meet the demand at the lowest cost, but it's also doing it subject to constraints that it sees. So we see the network We see the flows in a network and if we see congestion because a flow is exceeding the limit or approaching the limit, we can redispatch the system

in an optimized way to minimize costs, subject to that constraint. We call that security constrained economic dispatch. Um we also call something w about security constrained unit commitment. So we may commit resources also to respond to not just balancing the system but also making sure the flows on the system are secure. So all those things, same thing doing. And then lastly,

the market is also procuring ancillary services. So we've maybe holding resources back and compensating them for that reserve to support the third feature of our operational responsibility, that is maintaining operating reserve.

Do you want to build critical skills that are transforming the clean energy sector? Then discover the Yale Clean and Equitable Energy Development. Online certificate program from our sponsor, the Yale Center for Business and the Environment. This fully online five-month program is designed for working professionals who want to build projects that serve both people and the planet. You'll dive into energy fundamentals, project finance, and key phases of development and operations.

while learning how to center community engagement. In just five hours a week, you'll learn from leading experts, develop practical skills, and gain actionable insights. You'll engage in video lectures, live sessions, and a real-world group project And Yale will ensure that you have the expertise to advance clean energy solutions that benefit communities and drive economic resilience. Listeners of ShiftKey and ShiftKey Summer School right now can get an exclusive offer on these programs.

Use the referral code GEATMAP25 and get your application in by the priority deadline to get$500 off tuition to one of Yale's online certificate programs in clean energy. To transform your career, your community, and the clean energy landscape, head to cbey.yale.edu to learn more.

I had a question which is we just kind of started to get into it, but I we've had a lot of conversations about resources and bringing Power plants online when we need them and dispatching batteries when needed and the kind of minute to minute work of keeping the grid up versus the hour to hour work. And one thing I don't think we've talked about explicitly and this may be a noob question, but I'm not. Jesse, so I'm allowed to ask them.

Exactly, exactly. One thing we haven't talked about really is markets. Kaiso does have a market. You can go and see the projected curve for demand on a certain day. So Is the market for wholesale electricity in KISO resolving almost independent of the operation, the moment to moment operation, and then you're kind of filling in the blanks? Or do you know what is supposed to dispatch when based on the the electricity market?

I wouldn't say uh doing it independent or separate from the operation. I I would say it's doing it in support and maybe in coordination with the operation. So there are operators who have ultimate control over what resources do. And they can direct that. But the market is actually informing the operators of what the best way to dispatch or commit resources are.

uh at different timescales, whether it be five minute, fifteen minute, hourly, day ahead, that's the market function. And if you didn't have the market function doing that and and there are jurisdictions that don't, what w traditionally would happen is you would be doing bilateral trading. So you'd actually literally be calling someone and saying, Hey, I'd like to buy some energy or can you produce some more energy from this resource? Go ahead and do that.

Um, they would still have the automatic generation control. That's kind of the last piece of the puzzle. But the exchange of energy, the trade of energy is more of a bilateral. In our market, we do that through a centralized optimization that again tries to minimize the cost subject to certain constraints. And that is all happening automatically informing and supporting the operations and the operators.

Yeah, I mean we skipped past this kind of quickly at the beginning, but as you mentioned, you're the air traffic controller, not the pilots of the planes. Kaiso doesn't own power. Plan. This is different from other regions where we have traditional vertically integrated utilities that might also be in charge of both transmission and owning their own generators.

in most of the United States now, but not all of it, we have these independent organizations like KISO or ERCOT in Texas or PJM here in New Jersey that are responsible for all of these market operations acting that don't own the power plant. So I think the question is

how do you make sure the power plants want to do what you want them to do? And that's where the markets come in, right? Is you're basically trying to align the economic incentives that these generators have with what you need from them from a system operation and economic dispatch perspective.

You're you're exactly right, Jesse. So when we do that dispatch, we not only produce and tell them where to dispatch to on their curve, but also we give them a price that they're going to be settled out for that energy. And because we've optimized based on the bids that they've offered, um, under normal conditions, you would expect to see a price at their location.

to be at or greater than the bid offer that they offered, assuming they were dispatched to that level of their operation. So it's Yeah. They're happy. The price could be higher. They're even happier with that. That's a profit. Um they are driven largely by economic incentive uh to follow that dispatch instruction. Yeah. On the day-to-day basis, what markets does Kaiso run? Jesse referenced a few minutes of time, but

We probably break it up into day ahead and real time markets. The real time markets is what we described already, what was happening all the way from an hour ahead to five minutes to fifteen minutes. And then at the five minute level, the market stops, and then that's where some of the automatic traditional generation controls takes over. But that's fine tuning on top of the tuning that we've already done, the balancing through the market.

backing it up, we run a day ahead market and we co optimize over twenty four hours in the day ahead. And then we also produce prices day ahead. So all the energy that is produced at that level is actually settled at the day ahead price. And so that's the market side producing the price that is then used for settling the energy produced by those resources that are offering into the market.

And you need that day ahead schedule because of kind of two things, right? One is the speed at which generators can turn on and off. Right. And the second is the fact that you can't instantly change your output levels, right? Especially for thermal power plants. A battery might be able to go from fully charging to fully discharging in a few seconds, but a thermal power plant

big hung of spinning metal is gonna take some time to speed up or slow down. And so If you're thinking about changing levels of demand throughout the day, you're thinking about how do I manage battery levels throughout the day, you're thinking about how do I ramp up and down my power plants throughout a day, you gotta plan all that stuff out in advance so the generators Yeah.

basically prepositioned to be able to do what you need to do in real time operations. That's a function of that day ahead market. Right. That is correct. There are larger units that need twenty four hours, thirty six hours to start. Uh and so that day ahead market allows us to make those decisions. Some of the fuel has to be scheduled ahead of time. So a natural gas generating resource, um, there's a separate market that is actually scheduling the gas, so the gas is at that location in time.

In their gas network, yeah. Yeah, so th while electricity moves at the speed of light, gas moves about twenty to thirty miles per hour. you have to schedule that through the pipeline in advance so that you know that you have that when you need it. So the day ahead market provides that set of schedules so that you can do some of that pre planning, scheduling of your gas fuel supply in time for actual use.

And then the reason you have to settle the market day ahead and pay people if they follow that schedule is again just make everybody incentive compatible or make all these generators actually do what you want them to do. You're saying I need you to be operating on this schedule. If you do that, I'm going to pay you enough to cover your costs.

If you actually show up and do what you said, you get exactly that much and you're hedged against any uncertainty like the fact that maybe actually in the short term prices are lower or prices are higher in some cases than we thought because the system conditions change from the schedule. So basically

If you follow the schedule, you get paid the day ahead price. Deviation from that schedule then is gonna get settled at whatever price you need in the short run to get them to change their behavior on that shorter time scale. Absolutely. And let's just be honest. Nothing is ever perfect. Day ahead, you're forecasting your conditions. Those conditions change.

intra hour, you have different patterns that deviate from the hourly expectation. So you still have deviations and imbalances that you have to still optimize for in real time. But you're exactly right. you've cleared probably ninety percent of the energy by the time you get to real time using that day ahead market, much less volatile because you have the biggest set of

capability to choose from. And then there's probably a little bit more volatility in terms of the prices in real time as you deal with the real time deviation. Right. The closer you get to real time, the less less degrees of freedom you've got to actually change things'cause of the physics. The way I like to think about it is like you're increasingly fine tuning your position, right? Day ahead you're setting a twenty four hour schedule at hourly blocks.

when you get to the middle of the day you might recommit some of those units, change their schedule based on the changing forecast if they're able to respond fast enough, but not all the generators are, so some of your degree of freedom has already fallen away by that point. Then you go into your real time operations and you're trying to set things at a five minute block.

But of course, you can't just do five minutes'cause we even within that five minutes, everything's changing second to second, and that's where things like frequency regulation services and even the physical inertia and automatic generating control all play increasingly fine grained roles so that Supply equals demand all the time, everywhere. That's the key. Where do most dollars change hands in the electricity market? Is it mostly in the day ahead market?

the volumetrically the majority is going to be through the day ahead market. And and we can't ignore that there's also bilateral markets that are still happening kind of in parallel with the day ahead market. So you can make a contractual deal with someone to say, I want price of energy at this price and I'll pay you the differences

from the day ahead price, contract for differences. And so you can enter into bilateral arrangements that hedge your exposure to day ahead and real time prices, but that's an addition to but volumetrically the majority of the megawatt are clearing and settling through the day ahead market. But there are those side hedges that are happening outside the market to insulate or hedge against the price volatility, whether it be day ahead or That's beyond your Right. That's beyond our role.

The generators do. Yeah. That's market and instruments that they have with each other and I it's bilateral. We don't settle those. we we actually have some mechanisms to facilitate them called inter scheduling coordinator trades. We can help facilitate those, but they don't have to use those. When you say you Can enter a bilateral agreement. You mean me as an energy user, like a data center or a factory vr and a energy generator?

Yeah, so I'm talking about a supplier who's represented by what we call a scheduling coordinator or a consumer represent represented by a load serving entity or a scheduling coordinator representing that load serving entity. So I would love to say you as an individual consumer can do this. we're operating at the wholesale level. So our market is dealing with people who are buying and selling large volumes of wholesale electricity. We're not in the retail market. And in fact The retail

prices sometimes can be correlated to our prices, but oftentimes those retail rates have nothing to do with the prices that we produce on an hourly basis. It's the wholesale market. that the buyers and sellers at the wholesale level that are settling at our market price. So that's an important distinction. And when I say you, it really means the wholesale s providers and consumers. So for our listeners, like that would be the utility or retailer that you buy your power from, right? Like the

P G and E or in my case, P S E N G here in in New Jersey, or Pepco in Southeast Fenton, or whoever. Those are the kind of traditional utilities, but also some competitive retailers in some markets. Not California. California has community choice aggregators as another entity, right, that does this that are basically counties or county like entities that have taken on this responsibility for their residence.

So lots of different ways in which this is structured throughout the country, but basically somebody's buying power on your behalf as an end consumer, unless you're a really big like maybe Google has their own trading desk and does their own you know, buying and selling on their own behalf. But most people do this through some other what you call the load serving entity is that intermediary.

Yeah. Hey Jesse, can I just go back to the air traffic controller'cause I just want to make sure so we made it clear that you made it clear that we don't own the generation, we don't we're not the load serving entity, we're not the utility providing retail service to the consumer. Um, the th third piece of that that we kind of glossed over is the transmission lines themselves. We don't own those either.

Um, we plan for them and we'll probably get into the planning phase uh in in a little bit, but we are effectively we're operating them in the sense that we are allocating space on that transmission system as we're optimizing the system. So it we're providing equal access, open access to those lines. not just for utility owned generation, but any generation that is in that wholesale market. Um and we do that in a nondiscriminatory way.

That's the independent part of the California independent system operator. That's the I independent system operator and that's an important distinction between what would be done under a vertically integrated utility. They own the transmission line. and they w how they allocate that out. But there's not a the there's not a mechanism for optimizing the use of those lines. And that's the role we play through the market is optimized allocation of the space on that on those transmission lines.

Maybe to make this a little bit more concrete, walk through like how you're orchestrating the generation fleet. Like what is the typical mix of resources that you're calling on at different times of day on a typical California day? Like kind of start at let's start at 8 A.m. and move through the day. Okay. So if it's like today, it's a moderate summer day.

Um the there'd be some thermal resources, gas resources that would already be on, probably near their minimum load, which is probably about thirty, forty percent of their full operating capability. And they would be sitting there waiting for dispatch instructions as the load increased. And I talked about the morning because

uh people start turning lights on. This is when the load starts to increase in that morning hour. So to balance a system as that load increases relatively quickly, you're gonna have a combination of probably solar starting to come up and produce naturally because the sun is coming out. Uh you may have a little bit of wind production starting to increase because the wind's starting to blow because the temperatures in the system are driving that wind.

Um, if that's not enough energy, we're dispatching probably thermal resources, probably doing some exchanges through the Western Energy Imbalance Market with the neighbors. And then you get to about probably nine o'clock, ten o'clock and things stabilize and then

what ends up happening, at least in our system, is you start to see solar production continue to go up, but the load is not increasing. It's kind of flattened out. We start to probably see some backing off of thermal resources that were brought up

during that morning load pull and now we're starting to back off on those and maybe even getting to the point where we have surplus energy in the middle of the day. We're exchanging and maybe exporting some of our energy to our neighbors because we have surplus. We're probably starting to see batteries charge up in the middle of the day because now we've got this cheap

surplus energy. And this is going to probably go on till about four o'clock, five o'clock in the afternoon when the traditional peak of the day is, and this is when the highest gross load is. Um and then we start to see another dynamic happen and that is, at least in our system, the sun starts to set and then the solar production starts to decrease.

What's interesting about that is as the solar production decreases, it happens over about a three, four hour period, and it's a relatively fast ramp out of those solar resources. So what comes back in the load is not dropping. And in fact if you think about It's rising often, right? It's actually still rising because some of the load that was previously served by behind the meter rooftop solar, that load is also coming back on the system because the s solar production is decreasing.

So so to again, to rebalance a system and keep that balancing straight, what do we have the we we have to start ramping up a couple of things. We start to turn maybe what was exports around and we start importing energy from our neighbors.

um we start discharging the batteries that we just charged up earlier and now we start to discharge those. And to the extent we still need other energy, we probably have a combination of uh thermal gas resources that were bringing them off their minimum load, dispatching them up during the day, and probably some hydro resources that are able to be dispatched during the day. Between six and seven PM we hit what we call our net peak.

We call it net peak because it's the load, gross load minus wind and solar production. And that tends to be the most critical time when we need since the ramp out of wind and solar is more solar, that's is the highest where we need other resources to be available and dispatched. And so once we get through that net peak, come around six thirty, seven o'clock.

things just start to gradually turn around and then we're ramping out over the rest of the day the thermal resources, the interchange and the hydro resources that we previously dispatched up to get to that net peak. And this all starts over again the next morning. Yeah. Yeah. And when you talk about that choreography, most of that's being set in the day ahead market. I mean, at this point, you know the contour. I I would say even as a

as a flyby user of the Kaiso website, I kind of know what that choreography looks like. But as that changes over time, that's all being mostly set in the day ahead market. Absolutely. And the other benefit of the day ahead market is that you can see the entire picture all at once. And so you can optimize that

for that entire picture. And that's really powerful because now at least for batteries where you have four hours of battery storage, you can get really good at knowing when's the value of that energy to store, when's the value of the energy to discharge at a very good level looking over the entire day. You don't have as well visibility when you get to real time.

because your outlook is only looking out an hour, maybe an hour and a half out. So you don't see the full picture of the day ahead, you can see the full picture and that's the best time to set up those gross patterns of operation, including the battery charging and discharging. And by the way, you alluded to it. Thanks for mentioning the Kaiso dot com. You can go to Kaiso dot com and get our app. And you can watch over the day all these things happening.

And you can see the generation moving up, you can see the reserves being used, you can see the batteries being deployed, you can see the interchange, you can see the prices. And uh me as uh closer into it, I have a lot more information. But I just love watching the app to just see how things are operating. I can see things that The true grid nerds are checking the Kaiser or PGM apps every every hour.

I actually am sad because when I lived in PJM we had an app, but I don't think New York ISO has an app. We just have a data website. Sometimes the day ahead market comes in above where actual demand is. Sometimes it comes in below where actual demand is. Does Kaiso have a preference? Do you like to run? Does Kaiso like to run the system? Do you try to always overshoot day ahead? Do you try to undershoot?

we try to get it right and we're always wrong. Okay. Um I I'll say it that way. There is an asymmetry. Being able to keep the lights on is asymmetrically uh detrimental if you don't have enough supply. If you have too much supply, y it's you're not jeopardizing serving everybody.

you what you're then doing is well maybe you have to back something off, maybe you have to curtail some resources and maybe there maybe even some wind and solar resources. And as a result of that, we would try to mitigate for that by ensuring we have enough reserves um ensuring we'd have enough unloaded capacity to cover that uncertainty on the upward direction. So when maybe we we call it the residual unit commitment process. Um we do try to forecast

what a high confidence level forecast is. So it's not your typical fifty-fifty forecast. Maybe we're going towards a ninety percent tile load forecast and ensure that we have enough capacity, at least available online or startable in real time to meet that higher level forecast. That's all built into the residual unit commitment to ensure that we've mitigate those risks of being under supply.

And a lot of that's done through that reserve market where you're basically paying generators to s that would otherwise want to produce at full value because the market price is higher than their cost of production. to basically stand back a little bit and not use their full production so that they have extra capacity available in case the load is higher than projected or the wind is low is lower than projected or something falls offline that you needed in the schedule.

You have to basically pay them to stand by, otherwise it wouldn't be incentive compatible anymore. They would want to crank up their generator to full in order to cash in on the day ahead or real time markets. Right. And on the reserve, spinning reserve, operating reserve, they can explicitly bid to deliver or to offer that capacity and provide that capacity. And so the price that's produced is actually both the marginal cost of those bids

that are taken, but also the opportunity cost. And the opportunity cost is if you have to hold them back like you just described from producing energy that would otherwise have been economic energy and being paid a higher price than their bid. that difference between their bid and what the price, marginal price for energy is, is what we call opportunity cost. And that opportunity cost is built into the marginal price of holding the reserve.

Join Clean Energy Leaders at RE Plus 25, happening from September 8th to the 11th in Las Vegas. As North America's largest and most comprehensive event for the modern energy industry, RE Plus brings together innovators and decision makers across the full clean energy spectrum. including solar, energy storage, EV infrastructure, hydrogen, wind, microgrids, and more. RE Plus is where the industry's brightest minds gather to tackle today's most urgent challenges.

and shape the solutions of tomorrow. So come to the epicenter of clean energy leadership and drive collaboration and innovation to meet the world's rising energy demand. As a ShiftKey listener, you'll save 20% on a full conference pass. By using the code HeatMap20. Register now at re-plus.com. That's re-plus.com.

All right, let's go way back in time in our Wayback Machine. You can't build a power plant tomorrow. Takes a bit of time to to construct a power plant and get it connected to the grid. So the other big function that you play is helping the system plan ahead.

over years often, right? Whether that's interconnecting new generation, building transmission, or running I guess in this case you don't run the r bilateral reliability auctions. That's done on a bilateral basis between load serving entities and generators, but many other markets run what we've talked about in the past on the show, a capacity market where they're trying to basically procure ahead a certain amount of capacity as well.

Talk about some of the key steps you have to take maybe three or four years ago to make sure the system today has the mix of resources online and installed that you actually need.

I think what you're referring to is resource adequacy. At least in California, we don't have a centralized capacity market, but we have a bilateral uh resource adequacy process. And what that means is that we talked about load serving entities, utilities, Choice aggregators all being load serving entities, they have a responsibility to either have under their control or contractual control or actual physical control enough capacity to meet their um monthly peak forecasted load. Okay.

Plus a margin. Right. Plus a planning reserve margin. So now we're talking planning reserve margin, not a operating reserve margin. Reserve margin is usually somewhere between now probably seventeen to twenty two percent, depending on how you want to calculate it. The planning reserve margin is covering a couple of things. One is Look at year ahead, month ahead, you still have uncertainty about what your load forecast is going to be. And it's not unusual for load to be

four to ten percent off on any given day relative to your forecasted monthly peak. Okay. You also have to have enough planning reserve margin to cover that additional capacity to cover our operating reserves. So we talk about that's probably about six percent at any given time that we have to hold. Uh and then there's just uncertainty in supply. You think you have a control of a resource, it trips off, it's not available, it's on outage.

So you have to make up for that. So all that put together is your planning reserve margin. You plan for these uncertainties and you build that into your capacity that you have control over. And once you identified and have shown those resources that you have arranged to m to be able to meet your peak load plus planning reserve margin. Now those resources are obligated to offer into our daily market. That's the concept of resource adequacy.

That's not getting at the build what's built. These are resources that exist already and they're contracted in sufficient quantities to meet that forecasted peak demand. You do that go leading into the month.

for the full peak load, but you only have to do that for at least in California, about ninety percent of your requirements. So you you cover about ninety percent of your total peak load plus reserve margin, but you have some ability to ge fill that up, fill the tank up with additional capacity as you go through the month.

But in that year ahead, a certain amount of that capacity does have to be in localized areas. So because load centers are constrained by transmission that can get into the area, it's not just anywhere that capacity can be anywhere, a certain amount of that capacity has to be in localized areas and those have to be shown for the full localized area requirements. um as part of the annual process.

The third piece of this, and this is probably unique to California, is a certain amount of this capacity also has to be dispatchable. It can't just be base loaded, non dispatchable. And so we calculate on an annual basis what our expected monthly three hour ramps are gonna be, how much movement and we're up to now about eighteen thousand megawatts over three hours where we need dispatchable capacity, firm dispatchable capacity that can ramp over three hours.

And so again, in the resource adequacy program, a portion of that total capacity needs to have ramp dispatch capability, and that's the flexible RA portion. And your job in this f process,'cause those contracts are bilateral, your job is to define what that ramping requirement is, right? And as well as some of those define those local deliverability areas by

Understanding the grid and its constraints. Yeah. Correct. In other markets it's a little bit different. Like you know, we talked recently about the PJM capacity market where prices have exploded in the last couple of cycles. They run a centralized process. Instead of bilateral deals between individual loads serving entities and generators, they run this centralized capacity market.

which is trying to serve a similar function. But everybody pools their bids together, all the supply, all the demand, and then you get several clearing prices, potentially one for each of these deliverability zones that are defined in the region as well by Uh but that's a function that the grid operator in those regions that have centralized capacity markets, they perform that role as well. And that's not universal. A handful of the grid operators do that.

So it's gonna get a little bit nerdy here because we kinda glossed over what the planning reserve margin is and it seems like a magic number that I told you how it comes about. In reality th that planning reserve numbers should really be analyzed from a perspective of um and this is a industry target You want to have enough capacity so that you ensure that you don't have insufficient supply m more than one time uh one day event.

over a ten year period. And you can do simulations based on the makeup of your resources, the forecasted load, and identify um how much capacity would be needed to meet that one in ten year loss of load expectation target. That's really the reliability target, if you want to say, the service reliability target.

Now, how much of that actually is becomes the planning reserve margin, that's somewhat of a policy decision of how close you get to that. But that would be a measurement that ties the capacity amount and the planning reserve margin back to the service level r reliability and ensuring you minimize the potential of having insufficient supply over a statistical period of time.

And and if you're listening to this and you're saying, Hey, wait a minute, my power has gone out way more than one time every ten years, that's probably because the outages at the distribution network level, the mass majority of localized blackouts are caused by trees falling on power lines or other kind of weather related damages. So what you're talking about here is the sort of system wide supply at the state level shouldn't be short.

on supply more often than one time every ten years. And if you are, you have to institute these sort of rolling blackouts, right, to bring the system back into balance. Which I think happened a couple summers ago, the most recent one. One rolling blackouts or other emergency measures. Our last time that we've had to go to these measures would have been August of twenty twenty, uh for a couple of days. Um

So yeah, you want to avoid that. It's disruptive. Um, and that's why you want to do some rigorous analysis to ensure your resource adequacy capacity is sufficient to minimize those risks. There's power on the grid today. We've talked about what happened to get power on the grid today. We've talked about what happened that y yesterday they got power on the grid. We've talked about what happened a year ago. What happened five or ten years ago to make sure we have enough power today.

So now you're talking about in resource planning, transmission planning, integrated resource and transmission planning. And this is where you start from the longer range forecast. What is the first driver is what is your load, what's your forecast of demand ten years from now? Okay. That's that's the main driver. And then you answer the question, well, what type of resources do I have, will I have How will they

produce, when will they produce? What's their pattern production? And what do I need to add to the system in terms of resources to meet the increasing uh trajectory of load growth. This is now the utilities, load serving entities, uh making those longer term investments, making power purchase agreements with developers who are are interested in in developing new generations.

The third piece of that is, okay, so now you have a resource plan. How do you ensure just like how do we operate to ensure the supply can get to the demand? We also have to do in the same way in the planning horizon five, ten years out. How do we know we have the wires? to support those resources where they're located to deliver to the demand where it's consuming.

And so this is where ListaKiso has a role. We are the transmission planner for our system and we will identify transmission upgrades that are needed in the ten year horizon. to support the what is the expected resource mix to meet the expected load growth that is forecasted by others. So that all comes together to ensure we have the wires, we have the supply to meet the growing demand. Um

What's what makes some of this challenging now is uh the amount of load growth is changing. Large data centers are part of the mix. How and when and how much will they um consume? Will they actually develop? That's all kind of a question. Um we have fuel switching or conversion from using let's say gas for heating to using electricity for heating, mobility, electric vehicles. So electrification is also changing the amount of load growth we're expecting in the future. So it's

Whereas historically it's been that one one point five percent, that may not be the pattern going forward. We may have higher load growth because of these other factors that may accelerate the need for new supply, new transmission in the near term. Oh yeah. retirement of resources. So some of the older resources are kind of end of life, no longer able to produce when needed and should be retired. Well, if you take

six hundred megawatts out of the system because of a retirement, but you still have that same load. So you need to make up for some of those retirements and you need to do that in anticipation of those retirements. You can't wait for the to retire and then just go build a resource. It takes too long to build the replacement. So the other piece of this is as a result of those resources that want to offer themselves to load serving entities

There's oftentimes a large pool of resources that want to develop new resources. And what they want to do is they want to interconnect to the system. And we there's a mechanism and a process for those resources in advance of having contracts with load serving entities, they come to the grid operator us and say, Can I interconnect here? I'm having I have an idea, I want to interconnect here and then

We as the grid operator as part of our overall planning process, we say, yes, you can interconnect there, no problem, no no upgrades needed. Or we may say, if you want to interconnect there and you want to have deliverability

under resource adequacy in the future, you're gonna have to do some upgrades to ensure it's deliverable. Or there may be some reliability upgrades just because of the local constraints and need for the local area that is required. So We identify those upgrades and oftentimes the interconnection processes there's much more

request for interconnection than there is demand for those resources. And so it it's something that's been in the media for a while now is that these overheated interconnection cues. And we're trying to work through that. We've got some mechanisms to address those h overheated communities, more supply. But it's a good thing. It it means that we've actually got competition

at the supply level so that when low serving entities are seeking to make power purchase agreements with these resources, you're not limited. You've got a wide range of competition which helps drive down prices in the end for that capacity. So that's how it's supposed to work in theory. But uh this interconnection process is notoriously slow in the United States or other parts of the world probably too. Again, I'm more familiar with, you know, PJM, but it can often take

like three to five years to get through that process. And then that's not even necessarily the time it takes to actually build the physical infrastructure that is required to interconnect. You still gotta build a substation, maybe you have to build a new transmission line. Um why is this so slow and so hard? Just pose that question.

uh par part of it is it does take time to do those studies, but th that's not the only element. It takes time to get permits to be built. It takes time to get the financing and get the power purchase agreements for those resources. Um, it takes time to secure the land to build those plants, um, and do they really have site control?

That's why it takes time to develop a a power plant, but why does it take time to actually get your interconnection request met and and build the transmission we need? Yeah. So there it's it's the more you have in the queue, the more challenging it is to do those studies on a timely basis and so If we have a reasonable number, we can get through those things in a for us, probably a year cycle of planning. If we have a large influx of interconnection requests,

we've had to do it over a two year cycle just because of the amount of studies that would have to be done to do that. We do attempt to automate those studies and we have mechanisms to kind of cull the interconnection queue so that uh if ones are not viable they drop out earlier. And we've also added a mechanism to do that through tying it to the commercial interest. Are there load serving entities who actually have commercial interest in the development of these requests?

and we try to focus in on those as a primary starting point so that we have alignment between what is ultimately going to be procured and developed aligns with what we're studying and we're we're efficiently studying those resources. So this is aligning the process with the overall interest in developing and the interest in with those resources.

So the last question is we've talked a lot about what the grid does, kind of on a day or weekly, or monthly, or yearly basis. What is your day to day like as someone who oversees the grid? Yeah. So my oversight and my role as chief operating officer is four verticals. One is market policy design, so we're working on

market rules of the future. I have r also infrastructure planning. So we talked about the transmission planning piece, the interconnection piece. That is also under my ultimate responsibility as COO, but we have a vice president of of infrastructure planning. Operations. So we talked a lot about operations. The operations includes both the um the grid operations that we talked about, balancing secure operations, that piece of it, but we also settle the market.

And that's part of our operation. So the money flow uh and ensuring those generators are getting paid and the demand is getting charged and balancing all those accounts. That's under our operational responsibility and ultimately my responsibility. And then the third vertical I have is so all this obviously takes a lot of technology, computer applications to do that. That's also under my responsibility as chief operating officer. So my day in various ways is to uh get engaged in those

different verticals at some level, whether we're developing some policy going forward, we're having discussions about that. Whether we're where we are in the planning process, we're having discussions about that. I'm uh constantly looking at the application, looking to see what's operationally it's going on, more so on a hot day than a normal day.

And then lastly, I mean I'm w again working with my technology team to make sure everything is running smoothly and the cu applications are secure, supporting the reliable grid operations, but also secure in the sense that they're not exposed to any kind of information security or or threat from outside. So that's my day. Last question, but whenever there's a really intense ramp, I remember a few years ago it was a very, very hot day in California.

And there was a lot of solar and then the question was are is natural gas gonna ramp up fast enough in the late afternoon to meet the roaring demand from air conditioning that we knew was coming? And it did. It worked great. Whenever there's these kind of nail biter moments in the grid, Are these things where you know that

the technical capability is there in the system. You've already programmed it and so it's already been rehearsed to some degree. And so it's just making sure nothing goes wrong and you're standing there on the dashboard. You're standing there in the control room. Washing the dashboard, making sure it's all working. Or Is this so closer maybe? to let's say a mid twentieth century Chicago commodities floor.

And you're on the phone with a generator being like, Listen, I know it's really hard, but like can you give us every last electron? You know, like you gotta turn everything up. Like we're gonna figure out a way to make it work, but you gotta have everything ready. Yeah. So mostly the former and not the latter. That was my Hollywood dramatized. The latter, maybe.

It's not like uh it's not happened the latter. I mean there's been times where the stress conditions are enough that I'm maybe making some calls or my staff are making calls to see what else is out there, working with the state. to also what they can do. They have some additional emergency measures, strategic measures they can take. So we're coordinating with the state.

uh authorities as well. But generally speaking it is well rehearsed. We've got plans of all sorts of different scenarios and we have things that we will play, the playbook of what we will do in certain conditions. And those work very well. The market is operating automatically. Um and in general is able to deliver when it's needed and able to identify when we are insufficient, then we have to take some of those additional measures. But again, when we get to those points of additional measures

The operators have the playbooks. They're doing that. I'm not on the phone directly with the generator asking them to do something. I may be doing some informing and coordinating with the state and others, maybe media to inform them what's going on so that everybody's got an idea of what's happening at the operational level and what they can do to conserve power, flex alert or whatever. That's my role. We have a lot of trust in our operators and they do a great job and they can manage that.

Well Mark, clearly you have a lot on your plate, as does the rest of the team at Kaiso keeping the grid reliable for all of us, or at least those in California. So thanks for taking some time to join us for Shift Key Summer School and help educate our listeners on how the grid works. Thank you so much. If you have thoughts on this week's episode, you loved it, you hate it, you can always email us at shiftkey at heatmap.news or just me personally at Rob R O B at Heatmap.news.

Fifty is a production of Heatmap News. Our editors are Gillian Goodman and Nico Loricella. Multimedia editing and audio engineering is by Jacob Lambert and Nick Woodbury. Our music is by Adam Cromwell. Thank you so much for listening and see you next week.