(gentle music)
We will take input from the market and from other customer sites worldwide from the different regions and create our own critical mission profiles for those grid forming applications. The junction temperature, for example, of semiconductor will quickly jump and therefore we have to improve our semiconductor and our power modules and we do it in the way that we specify them to be able to operate with an extended and increased junction temperature.
(screen whooshing) (gentle music) (screen whooshing) (gentle music) This is the podcast for engineers, the podcast you just have to listen to if you're interested in what's going on in the semiconductor market. My name's Kelsey Markl, your host, and I'm joined today by Manuel Strugholz, who is the global system architect for Commercial and Utility Scale ESS systems. Manuel, thanks for being here. Thank you for having me here. Okay, so we're following up this episode.
This is the deep dive following Marco's episode on ESS and we're gonna really get into the details of grid forming versus grid following, but for our audience, can you just tell us what that is? So think about it. The classical energy grid had large so-called baseline power plants. For example, large coal power plants, which were basically review responsible for stabilizing the grid.
They included large rotating (indistinct) in form of their generators who would then some kind of imprint their voltage and also the frequency to the grid. And normal installations like photo type power plants are grid following, so-called grid following inverters, which were the standard until now, we're reading this information and so are synchronizing to this.
Now, we have a change and a trend towards more renewable energy sources and those are all until now so-called grid following instances and power plants who are reading this information from the grid with a change that large coal plants, for example, are shut down. This information that is imprinted on the grid that these inverter sources read is getting more and more weak, less easy to read, and which potentially leads to an unstable grid.
Okay, so you're kind of started defining already grid forming then. So why is grid forming so important now? Yeah, somehow we are going from a centralized grid to a more distributed grid with lots of distributed energy sources and this means those energy sources will have to supply the information, the synchronization information to the grid to make it more stable that other sources can use this distributed information, frequency, voltage, current ratings, and so on.
Read those from the grid and synchronize to them. But this information has to be put into the grid in a distributed manner and not in a centralized manner. And this means that grid forming is really controlling and imprinting this information to the grid and not only following this information that is already on the grid. Okay, so it seems like grid forming systems have a lot of responsibility on them. What specific attributes do they have to have in order to meet those demands? Yeah, right.
They have to do synchronization or are responsible for synchronization and angle stability, which means that if the frequency, the shifts a little bit in the frequency component in the angle, they have to counteract that. Also, they have to make sure that a frequency or a drop in a frequency and the line frequency does not get critical. They have to counteract that. They do not have to follow that shift in frequency.
They have to counteract and put their frequency in a kind of imprint to the grid and also they have to regulate the voltage. If they recognize that the voltage is dropping on the grid, they also supply more energy to the grid to raise this voltage again. Okay, so they're really, really just, I mean in the true sense of the word stabilizing everything,
Correct. That's correct. Okay. for example, different techniques, how they react and how they read this information from the grid in comparison to a grid following inverter. For example, a grid following inverter uses a PLL, which is aggravation for face lock loop at some kind of technical component who reads the data or the frequency data from the grid. So the inverter can synchronize to that. A grid forming inverter does this in an active way. They use other technical ways to achieve that.
Also, a grid forming inverter often has the ability to black start a grid. This means if there is no external stimulus to the inverter system itself, they can create the stimulus by themselves and be able to, yeah, like the way the name suggested is black start a grid. It's also often called islanding operation, which means some instances of these grid forming inverter resources can start the grid back up and then these smaller islanding grids are stable.
Other grid following resources can synchronize up to them, provide their energy, and then hierarchy per hierarchy level, the whole grid will get back up again and synchronized until you have the whole grid running. Okay, so if we for some reason lose energy source, lose an energy source from the grid, then these grid forming capabilities allow it to like restart it, give it a jolt of energy to restart it. (chuckles)
Right. That's correct. Okay, so it's sounds like grid forming also in our conversations with Marco has really become mainstream. What types of installations can we expect to see with grid forming technology? Yeah, in general, this is all the, in the range of the utility scale installations, not really in the commercial installations.
So in the local backup solutions for self consumption optimization, but really the larger installations who are directly connected to the grid at a substation level for example. Okay. The whole city. or at least parts of the city. Okay. And regarding the architecture, it's both string level power conversion and central power conversion. And therefore you have to look at different requirements. For example, a central power conversion inverter.
Our system right now is predominantly air cooled, so we have to cope with that and find solutions how to enable this grid forming capabilities through this rather inefficient, cooling way. String power conversion systems are normally and predominantly liquid cooled, which means you have to also have a look into that if you want to develop solutions for that. You mentioned these use cases have different requirements, but is there a standardization body setting standards in this area?
There are working groups worldwide, working on this topic and trying to, or not trying, there are defining standards right now, but there are no standards right now. And until the availability is there for the standards, we will take input from the market and from other customer sites worldwide from the different regions and create our own critical mission profiles for those grid forming applications. Okay. Critical mission profiles, it sounds pretty intense.
Can you explain exactly what that means? Yeah, a critical mission profile is basically the time throughout the day at which times inverter has to provide 100% load or maybe 120% load for 10 minutes, 150% load to the grid or from the grid for 10 seconds. Those are examples. And these different operational states are then combined to a so-called mission profile, which then represents a typical day in such a grid, which relies on grid forming capabilities.
Okay, so yeah, those critical mission profiles seem pretty critical. Is this what Marco and do you mean by overload conditions? Exactly. That's the point. There are certain points in the day where maybe grid tends to become unstable and therefore the grid forming inverter has to cope with that and through an overload condition, he has to put more current back into the system to get the grid stable again.
And this is more than normally the inverter is designed for and therefore the term overload because it's more than their design load, like 150% instead of 100%. Okay, is this something that we can address at the semiconductor component level? Of course we can.
We have different solutions for that, but basically we have to understand that an overload condition always means we had a lot of power cycling, we call it power cycling, where from one moment to the other, you have huge swings and output power of the semiconductor and this results in terminal cycling.
So the junction temperature, for example, of the semiconductor will quickly jump and therefore we have to improve our semiconductor and our power modules and we do it in the way that we specify them to be able to operate with an extended and increase junction temperature. For example, instead of 150 degrees, our semiconductors can now operate with 175 degrees continuously. Though this will get a lot of stress out of the customer designing the system and optimizing the cooling solution.
Speaking of the cooling solution, we are also improving the inside of the module, how the whole stack, this is consisting from the chip, the bond wires, all the substrates down to the heat sink level, that we form a very rigid solution there that thermal stresses do not result in weakening of the whole product and the power semiconductor module itself. Okay, improving the interconnections then within the module. Yeah, exactly. Like with the .XT solution.
Okay, is there anything else that we're improving on the chip level or the semiconductor level? Basically we are trying to lower the whole RTH, so the terminal resistance down to the heat sink level. This means that swings and huge power swings will not lead to huge junction temperature swings and thermal stresses inside the module.
Therefore, we improve our base plate modules both for forced air cooling base plate, indirect liquid cooling, and even with technologies like our wave solution for direct liquid cooling, which then helps to reduce the temperature swings and therefore improve the lifetime of the whole system. And I think Marco mentioned something about switching the materials, silicon to silicon carbide, for example. Is that something we're also looking at that can address some of the grid forming requirements?
Of course. Of course. Using switching, for example, from IGBTs to silicone carbide allows you to use lot higher or much higher switching frequencies. And silicone carbide has lower switching losses in comparison to IGBT, which then lowers your overall losses and therefore the raise in temperature does not even happen. Okay. You're trying to shut down the source of these problems before they even occur.
Okay, so it sounds like a lot of it comes really down to the component level for grid forming some of the requirements. Do we expect grid forming to replace grid following eventually? Absolutely, right now, grid forming capabilities come and then a premium regarding the CapEx. There is still lots of engineering work to, and good faults go into this, but we see throughout the world that and the different regions that grid forming requirements are always there.
We hear it from the market and therefore both for central power conversion systems and future string power conversion systems, grid forming capabilities and therefore overload requirements are becoming a standard. Yeah, I think this is a real societal problem that we're all addressing together how we stabilize the grid as we need more and more energy and switch to renewables. Thank you Manuel so much for being with us and explaining it to us. Thank you for having me.
And to our listeners out there, please stay tuned. (gentle music)