What's Really Holding The Grid Together - Fluence

I'm your host, Ed Porter. Welcome back to transmission. When something goes wrong on the grid, it needs a sudden surge of electricity to shut the fault off safely. A spinning turbine can throw 3 to 5 times its normal output in that instant. Most inverters can only match 1.5 times what they are already producing. And inverters are what are replacing those turbines at scale. As renewable generation grows, that gap becomes a problem. And this mechanism is one of several examples where the grid does more than just move power from A to B. This episode gets technical and I'll make no apology for that. Inertia fault. Current voltage control. The physics that keeps the grid online when something goes wrong. It's the layer sitting underneath every headline about blackouts, data center power deals, and the renewables build out. To get to grips of power markets, you need to get into this detail. Ben Braun, chief engineer at Fluence, has spent his career working on the power electronics from batteries to the grid. We explore what's really holding the grid together. As old power plants retire and what happens when thousands of inverters all react to the same event in the same preset way? Want to see what stability, Pathfinder and inertia markets are making for batteries? Asked Ko, Modo Energy's AI analyst. Sign up is free link in the description. Let's jump in. Hello, Ben. Welcome to transmission. Hi, Ed. Thanks for having me here. Oh, pleasure. And as ever, let's get straight into it. So what is one thing that people get wrong about grid forming inverters? One thing. Or would I be allowed to list a couple of things? You can cheat. All right, let me. I'll try to keep it short. I'll list a couple of them, and then maybe we can pick one to go into in more detail. The very first one that I like to point out is that grid farming is so much more than just the power converter controls, so a lot of people are naturally drawn to it. That's where it starts. But that's not where it stops. You have to think about the plant holistically. If you want to operate in grid forming, and probably throughout the conversation, we'll hit on that more. Another one is that grid forming is not grid forming is not. Grid forming is not good for me. I used to joke that I really hope all of the grids and the and the industry standards organizations that get together and they harmonize, and there was a bit of that, but we don't have a UL version and an IEC version of grid forming, but we do very much have a NESO version, an AEMO version, an ENTSO-E version, and so on. These are individual. Those are system operators, right, that you've just been talking about and they all have their own specifications. Yep. They they do reference each other quite a lot. And you can kind of follow a paper trail. This one references that one references that one who reference, you know. And so we don't have like a international standard. This is how good filming works. No. Okay, so we've got two so far. The first one is if you want to do good for me, it's not just about the inverter itself, it's the whole plan. And the second one is that grid forming standards globally are very different and perhaps shouldn't be correct. And there's even then a subset of the second one. I'm an engineer, um, where you have people asking for grid forming that aren't even on the grid, and then it's not necessarily that they have their own grid code, but a data center, for example. Uh, it's all over the news that they they are struggling to get grid connections. Many of them are saying, I don't need a grid connection, I'll bring my own power. But my own power plant will then be subjected to the volatility of my, uh, machine learning AI training models. And to cover that, I need a battery. And in that situation, a grid forming battery is very helpful, but it no longer has to adhere to any given grid code. It just has to serve the purpose of protecting the the gen sets, whatever they meet that may be, and to make sure that that data center has a high uptime. So you have even within or let's say parallel to the grid codes, you have a whole other subset of kind of customer specific requirements. Sometimes that's the second thing. I think that a lot of people get wrong. And the third thing, this is maybe the most controversial or. Okay, I like. It eyebrow raising and I bring it up because I think there's a lot of opponents out there, not necessarily only opponents of grid forming, but opponents of converter based generation. And, you know, let's get the good old synchronous machines back. Uh, they are good. They're great technology. People believe that grid forming controls are that any converter cannot be fast enough, cannot be as fast as a synchronous machine because they say it's too slow. Yeah. And so what I'm going to say about grid forming that people get wrong is that grid forming is actually slower than grid following. Okay. And it it comes in cause and effect. So what you want is you want your effect. You want the response to be very, very fast. And if you have a grid event and your grid following converter and you have a meter that has to measure that event or some higher order system has to command you and say, hey, I've detected an event. Please provide power. That command has to get supported through some controllers. The PCs starts responding. That all takes time. So the effect of grid following is very slow. And that's because it's controls are actually going so fast as it's trying to hold its setpoint. If the voltage moves or if the frequency moves, it is extremely fast. And it's tracking that so well that even when the voltage moves completely out of bounds, you still have the PCs locked into that voltage. And it does nothing to resist that inherently. Okay. So you're saying that grid forming is the the faster one to respond the right way around? The effect from grid forming is faster, okay. Because the response is slower. And so okay, it's a bit of a paradigm shift. So the grid forming converter, what it does in the transient, it doesn't track the voltage or the frequency. It says okay the voltage has deviated, but I'm holding my voltage for a certain amount of time and with some complex controls behind it. And because I'm holding my voltage, rather than tracking this other voltage, there will be an inherent physics based flow of current, which means a flow of power because you're getting two different voltages, exactly two different voltages or two different voltage angles. We can get into to more specifics, but there's a change on the grid. And because I'm doing nothing from my controls perspective, I'm actually doing something to the grid. I'm resistance. Faster. I'm not I'm not waiting for the detection. The control. And then in response, because I'm essentially just holding my position, I'm automatically responding. Exactly. Okay, great. So those are our three uh, three I'm going to say 3.5 as well because you snuck an extra one in there. Uh, so that's you did did a good job. I think people might be at this stage kind of scratching their head and going, this is this is complicated. Uh, there's there's there have been things that have been mentioned. Uh, a transient. I don't know what that means. Let's, let's let's go right back to basics. So, um, what is because you mentioned two types grid following and grid forming. How would you explain these two and how are they different verbally? You want me to explain the verbs? We can we can do we can do it verbally. Yeah. Yeah I'm going to need a whiteboard. I'm going to need a pint of beer for the. No, no I don't use that analogy for reactive power. Okay. So the the key difference I guess I already started to talk about it actually in the, in the last question, uh, with grid following, you can picture it as something that's really, as I mentioned, tracking the voltage and frequency. And it's doing that to try to hold its last set point. Whereas with grid forming, because it's a little bit of a slower response, it's providing some, some inherent resistance to it. I've done some kind of games with with colleagues where we like, we push against each other's hands, okay. And there's a different control loop where if you're pushing against a hand and you're physically touching the hand, you can feel that force. And so as soon as it changes, you inherently also change your position of your hand. Whereas if if you're just doing it visually, there's a bit of a leg, there's a control delay and you end up having some some offshoot and it's a little bit messier. And so the one where you're putting your hands together, this is grid forming. Correct? It's one way to visualize. It as it moves you kind of as you say, it's automatic resistance. There's this. The video of this episode is going to be great, by the way, because it's going to be me doing all these silly things. All of these things. Yeah, yeah. And then. But the second one is this. And so one hand moves. You're kind of, I need to move my other hand to try and sort of catch up with where that is. And that's kind of what you're saying about the grid following is having to sort of track the movements and then try and sort of adjust its position. And when you say a set point, I'm not sure everyone knows what a set point is. So like like, how does that fit in? So set point is really it's it's a command right. So it would be if, if you were a machine and I was trying to control you, I would say I'd take two steps to the left, take two steps to the right. That's your set point. And you're going to try to do it, but it's with power. It's a little bit more complex than that because it's it's a continual set point. It's not just moving your position. It's it's delivering a certain amount of force into or out of the electrical grid. Okay. Another way to look at it would be thinking about water. I think when people think about electricity, they love the analogy of water because you can't see electricity. You There's certain situations you can see electricity, but you shouldn't be able to see it. You shouldn't. Yeah. If you see it, you've got other problems. Um, but even though you can't see electricity, there's a lot of great engineers and scientists over the years who have devised all these great kind of, um, plots and graphs to help us as humans visualize it. Then those kind of were simplified. And this water analogy with, with pipes is brought up. So you could picture the grid as this massively, massively wide bathtub. Right. But it's very shallow because there's no inherent storage. Right. So there is a whole lot of area for you to take or add water. The surface area is large, but the actual volume is is low. Okay. And now you've got a grid following best connected to it and a grid forming best connected to it. The grid following best. It's not directly connected. You're basically shoveling a bit of water from from your battery reservoir into the grid, or you're trying to shovel it back out. And that takes time, and it's sloshing around and it's messy and and it needs to be commanded. Whereas the grid forming best if you hook up a siphon between the two systems, what happens is when one level changes, the other level will naturally change with it because they're really directly coupled. Okay. So that's one other way to try to visualize it a little bit. Obviously we don't have a bathtub here. Yeah. But the but the siphon kind of like will help people get that concept of that as grid forming. Like they have that almost physical link between the two. Exactly. And the, the amount of inertia you have, which would probably need explanation as well. But the magnitude, how big, how much grid forming you do is then kind of defined, let's say on the cross-sectional area of that pipe okay. So if you have a really big pipe, you can very quickly inject into that bathtub or pull from it. And if you have a small one it'll be slower. Okay. And maybe, maybe just kind of just talking about what you see in terms of your experience on, on the ground. So how many grid forming inverters are there, let's say, in the world today of of 100 projects, how many would you expect to be grid following? How much do you expect to be good for me? It varies greatly from country to country, so I think I would try to break it into a couple of key regions. So we have here in the NESO grid, they had the stability Pathfinder project that. Is that the National Energy System Operator? Correct. Thank you very much. They had a program called Stability Pathfinder. It was phase one, phase two, phase three. Phase two was pretty good for batteries. Phase three not so much. And then some subsequent tenders also have have raised some eyebrows because batteries weren't, uh, weren't elected for some of those services. But there are long term 2029 stability service tenders coming and less long duration energy storage. Um, so there's many more opportunities come in. And just a high level Those stability pathfinders. They're trying to say. We want certain services from these inverters that go beyond kind of what a normal grid following system can do. We want extra things from those grid forming inverters. So what are what are the system operators getting from those grid forming inverters. Two key things here in the UK inertia and short circuit level or fault current. How much? Well, how to say we haven't touched that level of the engineering one on one. But this term it's coming. In in name only inertia and short circuit level. Yes. In different countries the the grid operators are asking for different things. So I hadn't yet asked you or answered your first question about how many batteries are actually grid forming. Uh, the UK, there were only five awarded in the first, uh, Stability Pathfinder. Two, uh, a couple of those are online, but not all of them yet. And those are grid forming those. Those would be in grid forming. They have to be in grid forming to provide this service. Uh, you have a handful of other ones that that have entered in the market and that they plan to, to offer for, for future inertia and SCL contracts. But we're talking about. Yeah, a handful. Then you jump to Australia. Okay. And around about the same time Australia put out something called the voluntary grid forming specification. It's in the name. You don't have to do it. But if you do it, do it like this. Right okay. At the same time they also mentioned oh we're going to introduce this system strength charge. So that means if you're not providing system strength, which is another. Like a broad term to describe a few of these things. Exactly. Um, if you're not providing that technically you're kind of consuming it. And so you have to pay us for that. This charge will be locational. It might change from year to year, and it's not yet fixed. Loosely speaking. And so a lot of people who wanted to build batteries in Australia, of course, looked at that and they said, that's risk. At the end of the day, on my balance sheet. I'd want to have that. How can I get out of this? Well, you can build a grid forming battery and then this potential system strength charge, however big or small or volatile it would be, goes away. And so what was the end outcome? So let's say we have 100 inverters in Australia. How many of those would you say grid forming versus grid following currently connected. It's still fairly small, but I would say approaching 50%, especially for for new connections. There's there's some interesting statistics where you can really see that a huge batch are still grid following, but a huge batch. Our grid forming part of the reason why you don't see the change so quickly in AEMO is that they do have a very stringent connection process, and so there are still just a lot of things in the queue that were originally put forth as good following, and so they're still making it through. It's not yet 100% by any means, but I would say more than half. Okay. I think we can look up the numbers later. I think more than half is exactly right. I think your, your you're totally aligned with Modo's research on grid forming inverters, which is nice. Nice to hear. Um, and I think maybe those are kind of two examples that are good because Australia is almost like the foremost country in this, I would say, in terms of the number of good forming inverters they have. Definitely. And GB is relatively advanced for Europe, but obviously there's nowhere near Australia. Um, but it's kind of easy to see that if system strength like this broad term for system strength, uh, is starting to be taken more and more seriously, you are going to get more sort of grid forming inverters coming to the market. I'm aware that we very briefly touched on inertia and, uh, short circuit level. And I feel like it's a it's something we should just try and give some concept to. I know these things are so hard to explain because they are kind of complicated electrical engineering pieces, but I would love to give it a go. Um, so could could you just try and like, like a high level, like inertia? People will be thinking, well, look, how is a grid forming inverter? How is a grid inverter of any kind trying to trying to provide any kind of inertia because there's not a spinning thing inside it. So like how does that even work. Mhm. Yeah. Yeah I think anyone can understand some of these concepts. You don't have to be an engineer. Uh, you just need someone to talk you through it. So I'll try to do that particularly I'm going to start with inertia. I think that's something that's a little bit more tangible. You can see it kind of. Um. I'd like to ask you a question. Are you a cyclist, Ed? Yes, I am a cyclist. You are a cyclist. You like to go fast. I love to go fast. You love to go fast. What happens when you hit a hill? I slow down. You slow down. You slow down a lot. Yeah. What happens when a lorry drives past you? Genuinely slowed down even more. Just just fear. You get blown around a lot. And yeah. I get blown around. Yeah. What do you think happens to that lorry driver when when you pass him. Let's say he's just parked off. Nothing. Absolutely nothing. And why? It's because he has a lot of physical mass in this truck. Right. And so a lorry speeding along the motorway at, let's say, 50 miles an hour, right, to just to draw a parallel to the 50Hz of, of the grids here, they have so much mass that even even if they don't touch their accelerator, they will continue to to hold their speed for a given amount of time. Whereas you as a cyclist, as soon as you hit a small hill, you're slowing down. But this concept of inertia, you can really draw it towards, uh, this automotive slash cycling world. Okay. And you can think of the grid as being a lot of these massive lorries speeding along at 50mph, and they've got thick, rigid steel cables linked to them. So even if one starts to slow down, the others can kind of pick up the slack. Or if one speeds up, it won't really speed up until it speeds up this whole kind of weird. Like mad Max. Yeah. Crazy construction of all these lorries speeding down, right? Okay, so all these lorries, they're all kind of interconnected, going at 50 miles an hour. Yeah, tons of inertia. Right now we start having concerns about the environment. We want to avoid the mad Max scenario. Um, so what do we do? We start retiring some of these large lorries. Yeah. The drivers are now given bicycles. When the wind blows, they speed up when the sun is on their back. They also speed up. You know, it's getting a little bit esoteric here, but the the point is you're losing some of that inherent physical inertia. That big weight in the lorry, you're starting to take them off. It's gone. Right. Okay. And then when the hill the hill in this example is, is the grid load in demand when turning on a bunch of boilers or you have during during football finals, the tea kettles going on. You know, all these classic examples where suddenly you have a huge surge of load or of course when loads go off, that can also be a problem. Um, the speed of this system is now much more greatly impacted because of that lack of inertia. So hopefully that can give like a rough idea to to help someone visualize what inertia is and maybe why it's important we want to maintain 50Hz or 50mph, because that's what the whole system is set up to do. If you go too high or too low, generators start tripping off. Load, start tripping off. You have a blackout eventually. So if you come back to me as a, as a, as a battery with an inverter as the cyclist in this example, like how does a how do I, how can I start to replicate the weight of those lorries. Um, as someone who now has a grid forming inverter, how does that work? So batteries themselves. They are really heavy. First off, yeah. Good point. Not that that helps them with with inertia, but what they can do is they can provide a large amount of power in either direction. Not all the time, but but nearly all the time there. Of course, they're not weather dependent. They they don't need the wind or or the sun to actually provide this power. They need a little bit of range on the SoC. They need healthy power converters. They need to be designed properly. And protection settings. Coming back to was it number one the first thing that people get wrong? Yes. It's more than just the the power converter. So so batteries, they have this great ability that they can change their power when when you want them to. Okay. A wind turbine, especially economically it's going to be trying to run at its maximum output which the current weather allows. Same with solar. Um, they could potentially be controlled down. So if we have a over frequency, if the grid is speeding up, that could help to slow it down a little bit. There might be some flexibility in wind and solar, but but let's just make it really simple, right? Let's just keep it as a battery which can move in both directions and put that with a grid forming inverter. So how does that contribute inertia. So by quickly injecting a little bit of power into or out of the grid, you could maybe picture it as a small jet pack on on your bicycle. Okay. Right. You push a button when that hill has hit, you've slowed down, and then you just give a little button. Yeah. A boost from the battery. Why would you. Have a jet pack? You've got you've got an e-bike right there. E-bike example right there. You should use that. Missed opportunity for me. Yeah. Okay. So but that's kind of how it works. Right. So if you're thinking about this kind of concept of like this traveling mass, hitting a hill, slowing down, you would say like battery with grid forming because it can just put in a sort of sudden boost of energy to help get you up that hill. That's how they do it. Yeah, exactly. Okay. And in the reverse, if you're speeding up too much, let's say a large load has been disconnected. You need something to quickly slow down the grid. You need something to remove energy from the grid. And there again, a battery because it's inherently bidirectional. It's got this great ability to to go either way. Okay. We can we can maybe start to use that same analogy for fault current. Okay. Let's get let's get into fault current. Okay. So maybe just straight away. Like what. Like I don't think I people listening will know what fault current is I hope. But let's just make sure we've got the beginner's guide to it. Okay. So in an electrical system, the reason why we have switches on our outlets and why we have fuzed plugs. Or at least you do here in the UK, um, is because electricity can be dangerous, right? If it goes in the wrong place, it goes with a lot of power in the wrong place. It can cause severe bodily damage. Equipment damage. And it's generally a dangerous thing. So we need to have a way to to isolate or to disconnect electricity when we've detected an anomaly, when when it's going somewhere where it shouldn't and generally or exclusively the now I won't say exclusively generally the way to do that is looking at the current. There are some more advanced ways you can look at distance protection voltages. Um, yeah, there are, but. Let's not go into those. Let's just go looking at current. Current and why do we need fault current? If you don't have fault current, that fuze in your plug will not see enough energy flowing through it to to melt and to break the circuit. So instead of a very short high energy that then gets melted and cut off, you will just have a still high amount of energy, but continuous. And wherever you've had your short circuit, wherever you've had some equipment damage, or maybe there's a person that's closing the circuit. Current just continues to flow and continues to flow, and nothing can see that that current is an anomaly. So nothing isolates the circuit. And so your fault current is part of that isolation of the circuit. Correct. The fault current or rather the magnitude. How much fault current is the the signal for a protection relay to see? Oh, I've got a problem. I'm going to open a breaker or it's not. Like a fuze. A breaker is like a fuze, like a beefy light switch that can hold a lot of capacity, a lot of current. So you either have that as a signal for your for your circuit breaker or just for a fuze itself. It's a physical energy that that little piece of exotic metal, whatever. You have different combinations for DC versus AC, but that that bit of metal gets so hot that it just melts and opens the circuit. So fault current is needed. We need a certain amount of it to be able to isolate these faults, these these issues in the grid so that they don't continue and get worse. Now, when I say we need enough of it, we can't be melting fuzes or opening breakers for every day. Current flow for everyday power demands, right? So inherently your fault current needs to be at least a little bit above your maximum current that you might experience any time in the year or any time in a couple of years with with load growth. That being said, um, you build in some buffer this with a different protection relays. You might need different amounts of fault current for them to reliably operate. We're talking 30% above, maybe 50% above. Um, historically, synchronous machines have given us, especially if after some hundred milliseconds when when this breaking would occur, you're looking at 3 to 5 times their capacity is is what they can deliver in in short circuit level. It's a pretty chunky contribution. It's a lot. Yeah it's a lot. Now to provide that from a power converter is a little bit difficult. And maybe coming back to the example with Lori's as we retire those Lori's, we don't only lose the inertia, but we lose that significant amount of fault current from them. It gets a little bit weird because we jump to fault current. But you also have what I would call sort of I used to call reactive inertia. It's not a real term, but a voltage stiffness. Okay. Much much as the physical inertia analogy gives you a stiffness in your frequency and your speed, you're not varying your speed that much. Uh, with with voltage stiffness, you need something to really hold the voltage. And we haven't talked a lot about voltage yet. I think we've been focused more on frequency and inertia. But coming back to that analogy, that the voltage, the best example is that's the the balance of you on your bicycle. Right. You need to keep going, but you also need to keep going fairly upright. And when someone's blowing past you or based on weather based on loads, you can wobble a little bit. Let's let's come back to voltage in a second. Right. But but we just we started off with those two for system strength. And I kind of I want to do a quick recap because I know that, uh, we're getting into like technical detail here. Um, but I hope people are kind of getting a sense of one of the things that could be easily thought about power systems is that you're just flowing power from A to B, and it's almost like water going through a pipe. And in some senses that is kind of true. But in other senses, you've also got all these other parts of the system that are going on. So things like inertia to make sure that the grid keeps on moving at 50Hz or 60Hz, depending on where you're listening to this, to this podcast. But then you've also got things like fault current or it's called short circuit level in GB, which is around this ability to be able to make sure that when you have faults in your system, you are having the systems like trip off and they they broadly sit under this kind of system strength bucket. And the reason why we're spending so long talking about this is because if you want to move away from a system that's running on coal or running on gas, these kind of legacy systems, you have to have these, like these healthy indicators in your system to be able to run a power system effectively. Have I done an okay job of kind of recapping us? Exactly. Okay. You don't need me. I'm going to pack up and okay, okay. This is a very succinct, succinct, um, summary. Yeah. Then we then do voltage. I feel like we now we've started on the process. Let's let's talk a bit about voltage. Why does voltage matter? The reason why I wanted to talk about voltage is because the fault current is actually kind of a proxy for system strength in terms of voltage. Okay. So even with a pathfinder or well, forgetting about the UK for a while, going back to AEMO. Australia. Australia. They don't care about a specific amount of inertia. They don't care about a specific cell. What they want at level A short circuit level. What they wanted was a grid forming operation for system strength. Okay. It was kind of a nebulous term that just system strength. Why is why so different? Like, why do you have GB looking at inertia and short circuit level. But Australia. Looking at voltage. Sorry I cut you off but I think that's where you're going. Uh, it's partly where I was going. So NESO they had a I would say much more targeted scheme. They they had the carrot approach. Right. We'll pay you if you bring a certain service but also in a certain area. So we didn't hit on that earlier. But the Pathfinder to it was really targeted in terms of certain regions of the country are missing certain key metrics for the grid to to run healthy. And if you build a battery there or nearby with some fudge factors, we will pay you for that service. So because theirs was much more targeted and because they're paying you for it. They of course wanted to quantify it a little bit. And they and they broke it into the inertia and the short circuit level. AEMO took the stick approach. If you don't do this, you're going to get hit with a fine, a fine or a I think it's called a fee, not a fine. Okay. Um, no, it's a charge. There's a system strength charge in any case. Um, it was also somewhat locational. Uh, so you had to prove that you could reliably connect to a grid connection with a certain SCR. Another acronym, short circuit ratio. Okay, which we can maybe get into later, but it's a it's a figure that for a time was pretty good in, in terms of saying how, how strong a certain point and the grid was. And because their grid is stringy, as you said earlier, they have a lot of points that that are pretty weak. They have a low SCR, and then you have some points that are more central around the population centers, which are of course a little bit higher. Just imagine how big Australia is, how long the cables are that connect different parts. So the ones that are around lots of demand and generation have a sort of healthy grid. But if you go like further, further, further down these lines, you get to these places where the grid is sort of quote unquote, weaker. Exactly. And you know, where a lot of the the new PV plants are being built or wind farms are being built. They of course not in the city center. Further afield, also a lot of brownfield projects where you once had a coal plant, which of course is nowhere near where where the people are actually living. But now you have a great grid connection and a lot of land. And so it's great to see a lot of those projects being enacted. But then one of the the knock on effects, or one of the challenges, is that that is a weak position in the grid. So I don't want to say that Australia didn't look at the location at all, but for them it was less about I need a certain amount of fault current for a certain breaker to trip, or my grit is lacking this much inertia. It was really more about the voltage, the waveform. So it's like another component of this like broad bucket system strength. You've got inertia, you've got short circuit level or sorry fault current. Same thing more or less. And then you've also got the voltage that you're kind of putting into the bucket as well. So exactly. And that's why a lot of engineers who work with grid forming and grid codes, they talk about a voltage source behind a Thevenin impedance. Oh wow. Okay. That's catchy. We should get a t shirt made. I've already got some. Um, I do anyways, um, beside the point, uh, I mentioned that fault current is kind of a proxy for system strength, and it's a loose proxy. And in in reality, um, it's not necessarily that we don't have enough fault current in parts of the UK to for protection to reliably trip, but it's a fairly easy thing to measure how much a plant, how much fault current a plant can provide. It's a fairly easy thing to quantify, which is loosely related to how strong your voltage is, how stiff your voltage is. What I mean by stiffness. Coming back to the analogy. So you hit a hill, you slow down or you hit a valley, you speed up. A similar dynamic works for for the balance. And then especially we're getting away from lorries. So we everything is two wheeled. And so the balance is much more complex. Right. And so if you want. To be healthy you need to be like and to be. This should be obvious to people. But to be healthy on a bike, you want to be upright. For most of the time there are there are points where maybe you're going around a curve and you okay, leaning is part of the sport, but for the most part you want to be upright. Exactly. And you need some sort of force. You need a means to to rebalance yourself loads when they connect as they pull out active power through the grid. All of the little bits of grid that they go through, the impedance, some of that is reactive. And then that creates a reactive demand that then skews the voltage even more than just the active power. And so you need someone that corrects it. And unlike the frequency which is a fairly common good across grids to some extent voltage is hyper local. It's really each node can can have a unique voltage. And so what what grids need as these large lorries retire is not just the inertia that they had, but also they had this ability. You know they've got four wheels okay. They've got a lot of vertical surface which could catch the wind, but nevertheless they're very robust vehicles. And so we need something to, to hold the voltage if the loads are changing and some of the loads themselves are even reactive. So then you have even more, um, reactive and voltage changes on your network. We need still a capability to, to rebalance that voltage. And you do that with reactive power. Okay, so this is why I say that the fault current, it kind of gets at how much reactive power you can provide and or how fast. But really it's about the how much reactive power versus the change in voltage. So how stiff your voltage is. And I think those components I think you might need a degree in electrical engineering to like follow through how how they connect. But just at a very, very simple level, going back to your sort of person on a, on a bike who's potentially moving your grid, forming inverters because they have the ability to sort of flexibly change the power they're putting into the grid, and sort of how that matches with the grid's current frequency. That gives them the ability to, to, to push voltage or to pull voltage from the system. Correct. But I do have to say that a grid following converter also has that capability. Okay. But it's about again that the speed and the cause and effect and reaction times. That's really where grid forming actually provides something new or provides it in a more stable way. The big thing is that, largely speaking, reactive power is dictated by changes in voltage, active changes in frequency angle. Okay. Just just to just again, this is it's kind of it's complicated to think about some of these concepts. Not not for you I know for me, yes. But it's the reason why we're opening up the bonnet and looking inside is because if you want to move from a system of big thermal generation, and you want to have a system that has renewable generation for long periods of time, the thing that you have to get right, yes, you have to get supply and demand right. The thing that a lot of people understand, but all of these other components around voltage, inertia, short circuit level, fault current, you get to have a healthy and, um, uh, reliable grid. You have to have these things. Right. And that's why we're looking at this. That's why we're spending so much time on it. I hope people will kind of follow along why we're having this conversation, because it is a bit techie. It is also a little bit commercial now. Okay. You you can make money selling reactive power. It's it's not an energy. Um, you know, there are now reactive markets. So I know in Germany I think in the UK for a while it's not a large market. You can't earn a lot of money on it, but it is something that's, uh, that's anyways inherently within the capability of your plant to provide. It used to be defined in your grid connection as well. You used to get like a requirement for it. But before we go to active and reactive power, we have we have our three things that we've kind of put in our broad systems stability bucket. We're adding an extra one here in terms of sort of active and reactive power. Would you like to just describe at a higher level, like what those things are to people who at home, just turn on a plug and the light turns on. Sure, I can give it a try. I mean it. The active power. What they usually say. And engineering one on one. Whatever is that? That's. That's the power that does work. So that's the thing that is actually spinning a motor or a pump, whatever, in your refrigerator, in your mixer and your dishwasher, all these home appliances. Imagine all the pistons in your car all working perfectly in sync and turning the wheels. Exactly. That's that's active power. It's it's doing something that that you can see or that you can hear the reactive power. You can hear it in some situations. You don't want to hear it. But but it is needed as a means to help that active power get through the grid. So I mentioned earlier that as you pull active power through elements of the grid, some of those elements, they inherently have some reactance. It's called and that active current through the reactance creates a reactive demand that needs to be satisfied. Okay. So so you have let's let's recap. You have your your generation going through the grid that's creating reactance. It's creating reactance. And maybe I'll come with a new analogy here. Okay. So picture a circus tent okay. Over the entire country. Yeah right. It'll be pretty hard to maintain that circus tent. You need poles everywhere, right? Let's imagine every single substation is a pole. Okay. And now different substations have different voltage levels. Um, but there's a nifty thing in electrical engineering called per unit, where you, you normalize all of those voltages by their base. And so when they're at a healthy a nominal voltage, we say it's one per unit. So you can do that with frequency as well. You know if you're chugging along at 50Hz that's one per unit for the UK and much of the world. That's a healthy, healthy measure. A healthy measure. So now just to get away from all of these different voltage levels that we have in your house, it's very low. Then it goes up to to medium up to high. And then even there's extra high in some codes. So we've normalized them all. They're now all on the same level. Right. So this circus tent, it doesn't look like the the typical circus tent anymore. But it's just a straight content. It's a sun shield that you don't need here. But but that's what it is. Now, the the active power. You want to get it from A to B, you want to get it from the generation to your house. And as it's flowing over this, this tarp, it's causing some drip. Right. Yeah. It's drooping the the fabric, the material. And so along the way, you need to occasionally adjust the height of these different poles so that you do still maintain more or less this, this clean even one per unit. Yes level. And to do that you need reactive power okay. Because the reactive power will control like how big the poles are, how high the poles. Are high, how much the poles are pushing on the target. Okay. Or if they're pulling it down a little bit, I don't know if that's gonna help anyone. Actually I love I love that magic. Yeah. It's so good. It's so good. It feels like a really intuitive thing that people can, can grasp on to. Much like the lorries connected together or hitting a hill together, or a rider trying to sort of keep their balance and stay stay in voltage. Okay. I do feel like I need to apologize to some colleagues. This is a lot of pseudoscience, but it's for a good reason. And anyone who knows as much as me or more than me, please spare me. Uh, yeah. We're in. Um. We're in this funny world, right? Um, where there is this depth of electrical engineering that we could go into. But really, what we're trying to do in this episode is trying to bring some of these concepts to life for people who might not have heard about them before. Um, just to get sort of the interest beyond just the absolute basic thing that you might read in the newspaper, which is, um, not enough generation to meet demand grid problem. And that is just it's just too simple. The grid is much more complicated than that. And we're never going to get into like the full bells and whistles here. But this is just to give you a sense. Um, of riders on bikes, lorry sitting hills, tarps over the country and big bathtubs. Uh, this is what this is what you need to sort of start to thinking about if you can get a just a, just an idea of how some of this stuff works. Exactly. Okay. And maybe one more I'm going to have system stability. I'm going to add in an extra bucket, uh, something called black start. So let's say the worst thing happens, the grid goes down. Um, it used to be that, like, if you wanted to get the system back online, you'd start contacting these big spinning generators, big gas units, and right off they go. Start generating, get them all into sort of 50Hz or 60Hz, depending on where you are, and connect the system and then start to build the system back together like that. That's kind of super high level. Can your inverters, can they contribute towards that or is it not something that's for them? Absolutely. They've been contributing to that for for years, especially on island systems. Uh, batteries. One of their first main selling points is that they can do black start, and they have been doing it for a while. So, uh, an industrial complex might say. All right. I've heard there's some frequency response market. I can I can kind of finance the battery with that, but then I can secure all of my huge campus of whatever pumps and bells and whistles that produce widgets and cogs that that I use to make money. And if I lose power, I have to throw away an entire batch. And I have all of this economic damage. So I would really like a backup system to, to kind of bridge that. And so they've been doing that for for years. Um, one of my first projects when I got into battery energy storage straight out of university, was a battery black start kind of lighthouse project in Schwerin, in the northeast of Germany, and we went on to actually test it, to validate it. They disconnected, so we ramped the voltage up again. Guys in the gas turbine cross town, 14km, something like that. They say, all right, we got a healthy voltage. We're going to start turning on our coolant pumps. Uh, some some ventilation. We're going to start, uh, it's a, I don't know the English word for it, but a hydraulic system to slowly start spinning the the turbine up to speed. Before then, you can actually ignite the gas and actually bring it into generation mode. And so that is. And it worked. Relying on. Yeah. Relying on that, uh, battery and inverter combination to give the voltage to the, the gas turbine to get it started. So is that something that is a grid? I know we kind of this whole episode is being grid following, grid forming. Is that something that is both grid following and grid forming or is that just one or the other. Purely good form in that regard. The other fun thing about that first project was that in the second phase, we also energized wind turbines. You know, this was this new use case. We needed to prove that not only can we black start the grid, and then I would also introduce a new term. But I'll say that we black started ourselves and we restored the grid. Okay. And I think for, for a while it's no fault to, to you for using black start because that's what the industry said it was, generally speaking. Black start is black start. And it means you're restoring the grid. But there's coming new regulations and rules that that led us to need to kind of differentiate between those two phases. Yeah, a lot of power plants are now required to be able to black start themselves, but they're not yet required, let alone paid to actually restore external equipment. And the reason that is, is because if you're if you're a black start plant that then restores part of the grid. If you get to your neighboring power plant and you have to first energize all of its equipment and wait for it to start up, that's precious time, right? We had blackout recently here in Europe. They got everything back online. Amazingly, uh, quickly, I would say in referring to Spain. Very impressive how quickly they got it back online. And you can imagine that if every single substation you come to, you have to stop, energize the transformer, energize another transformer, get some loads spun up. You know, it takes longer. Um, so a lot of plants, especially in Germany, it's in Poland now as well. Italy, it's becoming really a common requirement that every new power plant needs to be black star capable. So for itself. Okay, so there's the black star which is like your independent part. And then the restoration, which is like the overall process on the on the grid. Exactly. Okay. Okay. This is good because I otherwise would have put my foot in that for, for the rest of the episode. But there is a clear delineation between being able to start yourself versus being able to start external stuff. Yeah, let's I think it's a fascinating part of like the that broad system, uh, strength. Let's, let's get to sort of two final bits. Right. So, um, one thing that I know people wanted to sort of talk about and understand a little bit more is the types of inverters, uh, that you have on data centers. This feels like a kind of a topic that's all over power markets at the moment. So from your work with data centers, what are you seeing? I would divide it into two main categories. So there's some that do have grid connection or are going to have grid connection relatively soon, but they now have new codes and regulations to adhere to because they they kind of brought it upon themselves. Right. They come with a totally new, uh, load profile. It's something that grid operators had never seen before, and it demanded new regulation. It could, in theory, be very choppy. Right. The loads that come off these data centers could be. Yeah, they could be sort of big. Big on and off switches effectively. Extremely choppy. And the other thing is that if the data center itself, because they've grown in magnitude, they used to be tens, maybe hundreds of megawatts. Now we're talking about gigawatts. So they're on the scale of some of the biggest power plants we have. When the biggest power plant we have trips, it's a huge issue. We lose a lot of active power, and we lose a lot of reactive support as well. When a large load trips, we also lose active power. It's in a different direction, but we lose a lot of it. So locally you're going to have a sudden spike in voltage because that that load. Coming back to that analogy with the tent over the country at one point really being pulled down, and then suddenly it's not pulled down anymore. Yeah, it's going to shoot through the roof. Uh, so you need to locally be able to manage that. Or best case, you don't trip at all. Or if you do trip, you very quickly have a battery to mimic your load. So the data center, with its more sensitive power electronics and and the IT server racks behind it. Those might go offline, but as long as a battery is there to mimic that load, then the grid doesn't see anything. Okay, so that's a big use case that that can be done with grid following the speed there. Grid following is perfectly suitable. Um, the default duration is a couple hundred milliseconds. And then when voltage comes back, most of the codes say you need to recover your load within 1000 milliseconds. But this isn't just sort of batteries providing energy to the data centers. This is batteries providing a lot of the supporting, uh, components that give you sort of good system strength. So, as you say, like they're working on voltage to make sure that that data center can connect without having problems for the local grid. They're doing in the first instance. More so I would say grid compliance support. Okay. So by by balancing out Not just that, that sudden drop of the complete data center. That's one subset of the use cases. Another one is that very choppy and fast intermittent load. You have some volatility in the load of these AI training models, but also in between their compute and their checkpoint save phase. You have these massive power gradients from 90% to to 50%. Let's say these are huge swings. And coupled with the fact that now these data centers themselves are huge, it becomes a bigger issue for for the grid. It's it's similar to, you know, back in the day, PV and wind, the grid had a pretty light hand in terms of regulation on it because there wasn't a lot of it. It couldn't really do all that much damage. And then as we got more and more wind and PV penetration, regulators realized that and they started to react and say, okay, now you have to have special voltage ride through capabilities and maybe you have to provide some some frequency response. Your requirements are getting more intense. The same has happened with the data centers. And to fill those requirements, since they're not themselves generators like PV or wind, they're they're looking to batteries to to fill that gap. So that's one I said I'd split into two. Yeah. Again, the engineer in me classifying and making subclasses, bucketing things. So that's one grid code compliance, making sure that the grid sees a smooth load and that it doesn't suddenly, abruptly drop some. I think I a grid in the grid operator in Ireland is also looking at reactive support. So once you get large enough, it's not enough to just do no harm on the active power sense, but also really, um, contributing to reactive power and voltage support. I believe Fingrid, the grid operator in Finland is also looking into that. So we're seeing it's a it's still very much evolving right. That subset. Put it away. Now, the other one I mentioned earlier, some data centers that they don't get a grid connection or it's going to be too long. And so they're just going to bring their own power. Mhm. Uh for those the use case is much different. They still have this kind of volatile load from from their machine learning training models. But now they're supplying it with a gas turbine or a bunch of gas reciprocating engines, maybe some fuel cells, things like that. And much like the grid, these equipments. They also have mechanical limits. Right. You can only abuse them so much before something's going to snap or something's going to break. Like stable running levels. Exactly. Stable running levels. They don't want to see too much torque gradients on the shaft, things like that. Coming back to the lorries and the bikes. You know. You can. You literally mean snapping. You don't mean. No. Yeah. Yeah, I literally mean snapping of. And then when that happens. Yeah. Expensive. Expensive? Exactly. So to avoid that again they're looking to to batteries. Mhm. Uh, it circles back to what I said about grid forming is not good for me. It's not grid forming. In that case you're you're not you're not required to be conformed with any grid code. But now you have a customer code or a customer use case. Right. And they say I want to make sure that, yeah, the torque gradient or the frequency variation that my generator see is limited to plus minus x. And that's where batteries, particularly grid forming batteries come in. And they can be a stronger voltage source a stronger grid. Right. We talked about system strength. They bring a lot of system strength compared to just a single generator on its own or a bunch of gas gen sets. They're relatively weak. And so now they take on the bulk of the load volatility. And they spare those those thermal machines making sure that their shafts and bearings and everything survive for years to come. Another day. There we go. And then, last question. What's a contrarian view you hold about our energy system. A contrarian view. How how contrarian do you want? I, I'm not sure whether that sort of engineers should have, like, deeply contrarian views. I feel like I feel like, uh, if, you know, you've you've listened to this, Ben, in as much as you represent the broad bucket of, like, electrical engineers, I think people are starting to realize just how dependent we are on a group of very highly qualified electrical engineers who keep stuff running that, uh, particularly in regards to places like Spain. But as we change our systems, like there is this niche, I'm going to say niche, cause I don't think anyone really understands it. But there's a niche of the skills that's really kind of quite hard to understand, but also like it's critical to getting stuff right. And I would I would be a little bit worried if they, if they had deeply contrarian views. So tread carefully. Okay. And I'm also not allowed to list again. Right. I think. No less. No less. I'm gonna have to pick one. Then I'll just. I'll just come up with one. I don't think it's necessarily contrarian, but it's maybe. Well, it does speak to your concerns a little bit, but I don't necessarily see it on everyone's radar. Okay. Which is potentially a risk. And I don't know if it's exactly contrarian. Um, but it comes back to fault current. Okay. Which is going to excite you. Right. Nice. I love it. Um, so we talked a lot about fall current and why it's needed. I already started hinting at it's a proxy for system strength, but it doesn't necessarily mean we need fault current. What we need is voltage. Stiffness. Okay. And engineers talk about small signals and large signals. So basically can you operate under daily stress. Wiggles and wiggles and hiccups and whatever. Or at end. Or can you operate under these large events, these huge faults or losses of load? Things like that. So large signal, small signal stability. You know what we're losing a lot of with these retiring lorries is is really just that small signal stability in terms of the everyday stuff. And I think a lot of grid codes, they focus right away on those. Worst case, how does your system behave if you have a direct short circuit right on your point of interconnect, and how much, how many amps of fault current you deliver or how many MVAs of SCL can deliver. Um, so there's I think sometimes too much of a focus on the large signal when it comes to grid forming. Um, and in particular, I think that's leading us to a point where we're going to have a glut of fault current. Okay. So lots of people with the ability to inject fault current at the same time. Which will potentially be too much. And the reason it's happening, there's many reasons why it's happening, I would say. So we talked about earlier these synchronous machines, they can provide multiples of their rating in fault current. 3 to 5. Times 3 to 5. Especially in the very peak, it's high and then it tails off a little bit. Okay. But when when you actually need to to open the circuit it's about 3 to 5. Let's say someone later will correct me. I'm sure we'll get a comment. Yeah, there's a range, but let's let's use let's use five. Right. That's sometimes easy math. In the olden days, if you had 100GW of load in your system, you would dispatch probably Premiere it order. You would have enough synchronous generators online to feed that load, plus a little bit of headroom. Plus something can trip. And then you get into reactive power and say, let's call it, um, let's call it 120 Giga Volt amps of power. So you take that times the factor five, you get to 600. Okay. Um, nowadays we're retiring all of those plants. We have PVA inverters. We have the wind power converters, we have the batteries, we have HVDC. Some people are even throwing synchronous condensers. Right. And they're all being given this instruction to provide fault current. Exactly. Some of them, the majority of them right now have only ever given one per unit. So just what they can do, nothing more. Mhm. Now what I think a lot of people miss is that currently whether it's Germany or Denmark, I haven't checked the numbers for, for the UK, but a lot of grids even to just get to 50% renewable energy. We at any given time of the day, we have about three times our load in power converters on the grid humming away. So that means before with the synchronous generators. Okay, maybe we had 3 to 5 times the the amount of their capacity in fault current. And nowadays just with the current PV and wind that we have and HVDC everything. You're actually already looking at kind of a factor of three. Even though they're all individually only doing one per unit. So what happens if you scale that up and you have more fault current than you need? You have to start upgrading equipment. Okay. You can run into situations. And I think there's already some European TSOs that that have encountered this. There's been a couple of presentations at various technical conferences that I like to hang out at. As you can imagine. And yeah, there's issues. So another reason why this is coming is not just because we have this massive amount of power converters coming online. The second reason is our grid is getting more and more meshed, right. It used to be that you had one big coal plant out on that side of town, and a nuclear plant way down on some other village somewhere, and they had a lot of fault current, but it was quite far away. Nowadays everything is being in the medium voltage. In the high voltage. Okay. Also extra high voltage. It's much more meshed. So the fault current that they provide has a much easier path to get to any location where there might be a fault. Yeah. So the amount of fault current, it's still less. Yes, but the effective path to the fault is getting a little bit easier for it. So to say so those are two reasons. And now the third reason is people are coming with grid forming right. And for example Germany, they're willing to pay you for grid forming inertia. Specifically I think we kind of glossed over Germany, actually. It's okay. Um, let's come back to it. Right. So Germany is now willing to pay for grid forming. Specifically, they're willing to pay for inertia, which you can only get qualified for if you're a grid forming. A lot of our customers are interested to use overload capability. Um, going above 100% of your rating, you know, 120, 130% is is not uncommon. Um, and in doing so, You want to do that for that active power, for the inertia. But the way grid form is defined. You don't get to choose where that current flows. Remember I said at the beginning of the conversation? Grid forming is slower. And that's why the the response is inherent. And we have all of this current capability in our power converter, which could support an inertia event, but it could also support a voltage event, you know, in a different axis. So that means there's overcurrent capability going above and beyond. 100% is now being built on the grid, not just only for inertia, but if there's a fault, it will also feed into the fault. So you have multiple factors where where our installed base of power converters is, it has to be massive to ever get to 50% renewables or 60, 70, 80. You know, who knows. That has to grow. The grid is getting more meshed. Those generators are more distributed, and some of them are now being built with even more than nominal capability. All of that adds to, in many situations, too much fault current. Why not change your grid requirement for fault current and just dial it down a bit? So Tennet in the Netherlands, they presented at one of these conferences. They they mentioned they tried out a couple of different things. They of course have a fairly large amount of offshore wind coming into their system. And in general, they started noticing that they're starting to get to much fault current. They tried different things. They tried sort of breaking that mesh at times. But that mesh is also quite good. So if you've built this lovely grid that can feed from any angle, and now suddenly you're trying to disconnect it so that that wasn't great. Um, for my memory as well, they talked to some of the wind turbine operators to try to to limit it. And again, it was a bit difficult. Eventually, uh, what they said is they're going to have to upgrade ratings of, of their equipment. Okay. So your contrarian contrarian view is that yes. It's great that we're getting all this folk current coming in from these inverters, but we have the knock on effect, which you maybe don't see, uh, initially, which is that for a lot of our system, we also need to make sure we do sort of this great grid upgrade at the same time to make sure that we can handle that additional fault current that's coming into the system. Exactly. Ben, I think I need to go and put my head in an icy bath of water and just try and to sort of, sort of distill some of that down. That's been a really, really, uh, thorough view for the layperson of, like, how this electrical engineering is working, how it stacks up. I hope you've been able to to sort of follow it through. It's, uh, it's a fascinating part of how this all works. And if you can start to understand this, you can really start to understand how power markets work. So, um, Ben, thank you very much for your time. Uh, it's been a wonderful episode. It was a pleasure. Thanks so much for your time. And if you have any follow up questions, you know how to reach me.