Steel in the age of electrification and AI

Dan Bogler 00:00
I read a report from the International Energy Agency, which said that the transition to green steel globally will cost around 1.3 trillion dollars by 2050. So that's a considerable amount. Can the industry afford this? Are the margins good enough?

Alexander Fleischanderl 00:17
Now you're touching a sensitive point, but it's a very important point to talk about. Technology is only one pillar, and luckily in the transition for the steel sector, it's quite advanced. So, we have the building blocks, or many of the building blocks in the system. But there are other things that are clearly lagging. One is the energy transition. I said we need more renewable energy, so we need a factor 10 to 20 or 30 times more power given the new technological pathways.

Dan Bogler 01:01
Hello, and welcome to another episode of Moving Forward, a new podcast series from Mitsubishi Heavy Industries. Over the course of this series, we will be exploring the big technology and energy challenges of our world today, and discussing the solutions that MHI is developing to meet them. I'm Dan Bogler. I'm a former Financial Times journalist, and I've been the co-editor of Spectra, MHI's online content platform, for the past six years. Today, I'm delighted to be joined by Dr. Alexander Fleischanderl. He is the Chief Technology Officer and Head of Green Steel at Primetals, an MHI group company. Welcome, Dr. Fleischanderl. Very nice to meet you. Perhaps you could start by telling us a little bit more about yourself and about Primetals.

Alexander Fleischanderl 01:57
Yeah, Dan. Thanks for the kind introduction. I'm super excited. Thanks for having me here for this interview and this podcast. Well, I'm Alex Fleischanderl, working for Primetals Technologies in Linz in Austria. So that's where I'm based, and I'm working for the company for almost 30 years. And that's the super exciting thing that the steel industry is really fascinating me. It's my passion. So, well, who is Primetals Technologies? It's a truly global technology provider for the metals sector, mainly for the iron and steel sector. So, we are about 7,000 people around the globe, based in 23 countries, and we are truly proud of our large patent portfolio that we hold.

Dan Bogler 02:55
Well, kudos to you for still being enthusiastic about your job after 30 years. So, let's turn to steel. It's now 3,000 years since humans started to smelt metal, and steel is more indispensable than ever before. It's in boats, it's in cars, it's in buildings, it's even in wind turbines. But it also has a big environmental footprint. It produces, according to the statistics, something like 8% of global CO2 emissions. Why is steel so hard to decarbonize?

Alexander Fleischanderl 03:31
Yeah, first of all, as you said, Dan, steel is the backbone of modern society. There is nothing moving in industry, there's no industrial development without steel. On the other hand, this sector is the largest industrial emitter, and the main reason why it's a hard-to-abate industry and it's so complicated to decarbonize is that it's not only a switch in the fueling, so a transition in the energy, but also a complete redesign of the chemical pathways. So, meaning we have to reduce the iron ore. We have to get the oxygen out from the iron ore, and this can only be done -- or is mainly done today -- with coal and coke. So, the industry still realizes 70% of the capacity that is produced via coal, and that makes it so difficult in the switch of the energy and chemical system.

Dan Bogler 04:44
Yes. So, if I can rephrase it slightly: so steel is the archetypal hard-to-abate industry, isn't it? It's the one where you don't just need low-carbon sources to replace the power, but also to replace the heat in the processes, which, as you say, drives the chemistry. Electrification is presented as one of the big solutions here. Tell us a little bit about how that works, and particularly how electric arc furnaces work.

Alexander Fleischanderl 05:17
Yeah, electrification is indeed one of the pathways in the steel transition, and why it's so important is because it's available now. While we are talking about other alternatives, and I’m sure we’ll come back on this a bit later, electrification can be implemented right now, and the core pillar is definitely installing electric arc furnaces that replace the traditional blast furnace/basic oxygen furnace route by melting either scrap at the end of the life cycle, so that provides a low carbon footprint. Electrification goes far beyond the electric arc furnace. So, it's about electric heating. It's about waste heat recovery. It's about other heating systems, inductive heating for slabs and other steel applications. We can already achieve significant decarbonization by electrification, but it's not the only pathway.

Dan Bogler 06:29
Yes, and as far as I understand it, you know, correct me if I'm wrong, one of the big problems with electric arc furnaces, of course, is that they are very power hungry, and so in turn, in order to have a significant impact on emissions, we need to feed them through low-carbon electricity -- through renewables, basically. And are there enough renewables available to power all of the electric arc furnaces that the steel industry needs?

Alexander Fleischanderl 06:58
Yeah, not yet. But well, the energy transition is going on in parallel, and the issue is indeed that the electric arc furnaces are quite energy hungry, as you said. On the other side, it's also not a continuous process, so there is a batch operation and on-off with the power, and that makes it difficult, especially for renewable power grids, that we have strong fluctuations in the in the power consumption. So, this requires also innovative technologies how we can stabilize the grids to avoid any backdrop and improving the power quality at the end in the grid. So, this is really essential. So, for this, we have recently developed a technology called the Active Power Feeder that improves the power quality by a factor 10 at the minimum, but also improves the energy efficiency for the electric arc furnace itself in parallel.

Dan Bogler 07:59
Right. So, if I understand it correctly. An electric arc furnace is basically a very big and not completely stable drawer of power, and then maybe emits some power back to the grid. Is that right?

Alexander Fleischanderl 08:12
Yeah, exactly. So, well, how it works is a big, large furnace with electrodes, and you feed the charged material in there, so scrap and pig iron or direct reduced iron, and then well you ignite the electrodes and you melt the charged material down, but you have to short circuit a certain point that creates that backstroke to the grid.

Dan Bogler 08:41
And what is the attitude of the customers of Primetals, the big steelmakers around the world? Are they busily installing electric arc furnaces and replacing their old blast furnaces, or are they still hesitating? Are they worried? Are they worried about the power supply?

Alexander Fleischanderl 08:58
Yeah, indeed, it's not only the power supply, but it's indeed the role of Primetals to work closely with the customers and to design the whole ecosystem. It's not only about the electric arc furnace itself, but as we said, it's about the power system. You have to take care also about the byproducts that you are producing, so the slag, the dusts that arise from the furnace. Indeed, it's also about the flexibility for the charged materials, so for the feedstock. The scrap availability is limited; it's not equally distributed around the world. And it's about the quality of the scrap, talking about the ‘tramp’ elements, especially copper, that becomes more and more of an issue with the increasing electrification -- so more cables, more motors, more electronics, and when the scrap returns at the end of the life cycle, you have a lot of copper in there and copper is a real enemy for high-quality steel. It makes the steel especially brittle, and that's actually what nobody wants to have. We want very tensile, high tensile strength steel for the automotive and many other applications.

Dan Bogler 10:19
And you mentioned electrification is also a help or even a solution further down the mill. So, you know, after the blast furnace, after the slabs come out, when they're rolled, pulled, turned into wires, and so on. Where does electrification help there?

Alexander Fleischanderl 10:36
To maybe bring a simple example. Today liquid steel is solidified in a continuous casting machine, and the product that comes out is a steel slab that is cooled down and stored at the slab yard and then reheated again. The reheating is done with fossil fuels, so usually natural gas or other fossil gasses, and this adds a lot of carbon footprint to the steel, that is embedded. We have a really fascinating technology that combines the continuous casting with the rolling. The technology is named Arvedi ESP, so it's an endless strip production, and this avoids cooling down the slabs. It takes the hot slabs and keeps the temperature and rolls it. And there's just a little bit of electrical temperature adjustment, which is done by inductive heating, so fully electrified, and it's the only technology that is globally available that does not use any cubic meter or kilogram of fossil energy in the process. So, it's 100% electrified and can reduce the carbon footprint by more than 90% compared with conventional continuous casting and then hot strip mill afterwards.

Dan Bogler 12:06
That is a very compelling example. Obviously, if you don't have to reheat it all over again, that makes a lot of sense. Turning now beyond electrification, as you said, electrification is the pathway that's available today. But for the future, we're hoping that actually by using green hydrogen, we can go further in reducing emissions in steelmaking. Tell us a little bit about that, and in particular about your HYFOR technology.

Alexander Fleischanderl 12:31
Yeah, indeed. So, there was a big bet on hydrogen. The steel sector was a bit overhyped. Let's say we're back now to reality, but still the hydrogen is required, and it's a very viable pathway to utilize hydrogen, especially for reducing iron ores. So, for the chemical energy that is used. I'm not talking about using hydrogen for heating primarily. However, we have to recognize that green hydrogen is not available at scale and reasonable prices, so we will have to wait a bit longer. But as you mentioned, I would like to talk a bit about our HYFOR technology. So that's exactly where hydrogen comes into play. HYFOR is an abbreviation for hydrogen-based fine ore reduction, and it uses exclusively hydrogen for reducing fine iron ores. And what makes the technology really unique is the ability to use any type of iron ore -- and there's a fine grain material that is mined or directly coming from the mine without any agglomeration steps. So, this reduces, first of all, an investment for the agglomeration plant, but it also avoids the energy that is used there and reduces the carbon footprint.

Dan Bogler 14:05
So, if I understand it correctly, in the old way of working, you would get new pure, you know, iron ore from the mines, and you would have to treat it first. You'd have to agglomerate it first before you could put it into a blast furnace. But this new technique doesn't require that, and you can use any kind of quality of iron ore. Is that right?

Alexander Fleischanderl 14:26
That's right. So, usually iron ore is pelletized for high-quality pellets, but not any iron ore. Every iron ore can be pelletized, so it depends on the geology. Yeah, and this makes now every iron ore around the globe accessible for direct reduction. It's the first time there is no other technology that can achieve it. And the other USP is as we use these fine grains from the iron ore, the reduction kinetics is much better. So, the reduction gas does not have to penetrate through the pellet; it's much faster in the reduction kinetics; and it runs on lower temperature. So, it improves the energy requirement again. And the last benefit, really, and the most important one, is the extremely high yield. So, it's a closed cycle, and we achieve hopefully 99.9% of yield in the process, which is outstanding.

Dan Bogler 15:30
Right. So, for in layman's terms, all of the iron ore gets turned into steel, basically.

Alexander Fleischanderl 15:36
Exactly.

Dan Bogler 15:37
That does sound good. Tell us a little bit more. So, my understanding is that Primetals is developing a prototype plant using HYFOR for one of its Austrian customers, but also, you know, some other partners are involved.

Alexander Fleischanderl 15:53
Yeah, indeed this is correct. So, we have over the past four years operated a pilot plant -- also in Austria -- and the results were pretty impressive, which led us to the decision that we go for the next step in the de-risking of the technology. And we have decided to invest jointly with our partners to go for an industrial demonstration plant. And as the investment is high, we've joined forces and have built a strong partnership. This includes voestalpine as the leading steel producer in Austria. It includes Rio Tinto, as one of the major mining companies, and it's backed also by Mitsubishi Corporation.

Dan Bogler 16:42
Fantastic, and the idea is that you will go into effectively commercial production late next year. Is that right? In 2027.

Alexander Fleischanderl 16:52
Yeah, it's not a commercial production. It's still an industrial demonstration plant, but it's correct. So late 2027, we will start the operation. The plant is already in the installation phase, and what we are going to prove there is we have to move from an interesting and promising technology towards industrial confidence, I would say. So, meaning, we operate it for the first time continuously over a long time, and we will collect a lot of data. We will learn about the reliability of the technology, about the consumption figures, about maintenance work to be done, and to convince the market, the customers, to invest into next level direct reduction technology.

Dan Bogler 17:47
Yeah. Now, Alex, you know, thinking about both in terms of the timescale and the cost scale, you know, like any new technology, when this starts, it starts small and it starts expensive. You know, what do you think is a realistic time frame for it scaling properly and becoming, you know, genuinely affordable or cost competitive.

Alexander Fleischanderl 18:08
Yeah, we have a pretty sound business plan on that one. So, we hope that we will have de-risked it after a year of operations in 2028, and then be able to roll it out to the market. So, hopefully, we will see the first of its kind HYFOR plant in the market in 2030. It's not too long -- four years from now.

Dan Bogler 18:32
That is not too long. Stepping back a little bit, I read a report from the International Energy Agency which said that the transition to green steel globally will cost around 1.3 trillion dollars by 2050. So that's a considerable amount. Can the industry afford this? Are the margins good enough? You know, what's your view there as a supplier?

Alexander Fleischanderl 18:56
Now you're touching a sensitive point, but it's a very important point to talk about. Technology is only one pillar and luckily in the transition for the steel sector it's quite advanced. So, we have the building blocks or many of the building blocks in the system. But there are other things that are clearly lagging. One is the energy transition. I said we need more renewable energy, transitioning from coal into hydrogen or electrification. So, we need a factor 10 to 20 or 30 times more power given the new technological pathways. It's not there. It's not reasonably priced, at least not in all the regions around the globe. We have to scale the hydrogen economy. Also, carbon capture, storage and utilization is viable, at least as a bridging technology. So, we are working on that one. But as you said, it's also about the conditions to make a decision on that scale. Usually, we are talking about investments in the billions of U.S. dollar investment levels. This requires very stable and reliable political frameworks, and these are missing in many regions in the world.

Dan Bogler 20:24
Yes, and I would like to come back to that in a minute as well. But first, we've talked a lot about Primetals and decarbonization. But Primetals is a busy company and they do lots of other things. Tell us a little bit about how you're using digitalization, AI, big data to help your customers.

Alexander Fleischanderl 20:42
Yeah, indeed. So, we treat digitalization and AI as equally important to decarbonization or the, let's say, sustainability in in general, because at the end, digitalization is an enabling technology also to decarbonize. So, we talk about the twin transitions, both go together: decarbonization and digitalization. Well, what is important is that in Industry 4.0 we have collected a lot of data. Everybody was proud about the big data storage, but nobody really made use of all the big data. Maybe, they generated a few dashboards for the management, but it did not really help to transition. So, what has changed now over the past few years, is we are well aware that the big data have to be clean data, and from such a data core -- also empowered now by the much faster computing power and putting AI on top -- we can really support customers in the transition by big data. It's a complex system, however, because you have not only to manage the technology itself. It's about the materials system. It's about the energy system. It's about the steel, the product quality system. So, you have to stay within a small tolerance system when we talk about steel quality, and that's now where the big data, and especially AI systems, really support operators. We’ve embedded all these technologies recently in the central operation cockpit. Central operation cockpit is a technology that is very similar to a flight simulator for pilots.

Dan Bogler 22:41
I was about to say it sounds like you're turning a steel plant into a fighter jet.

Alexander Fleischanderl 22:45
Yeah, indeed, that that's a very good example there because it's a training center for the operators. First of all, it's unique because, well, without damaging the equipment, you can train emergency situations in the central operation cockpit. It's a digital process twin, fully embedded with everything that you have in the real plant. And the other thing is also you can train young talents, and we are lacking talent in this industry, and bring them up to date on the new technologies, the way of operation in using this central operation cockpit.

Dan Bogler 23:31
Tell me a little bit more about simulations. Now, as I understand it, a steel plant is a very complicated system, and, you know, to produce a particular customer order, so you know ArcelorMittal is producing a certain you know shipment for the car industry: they have to set it up and set it up again and set it up again. How does AI help there?

Alexander Fleischanderl 23:52
Yeah, it's, I would say when we talk about AI, I'm not talking about generic AI.

Dan Bogler 24:01
Yes, you're not talking about large language models. I understand.

Alexander Fleischanderl 24:04
You have to have definitely all the skills and the operational intelligence. We are talking about really cyber-physical systems. We are talking about chemistry. We're talking about thermal knowledge, mechanical, electrical systems. So, you have to embed everything. AI only works if you have this deep industry knowledge, and therefore we believe this can only work between the steel producers and technology providers like Primetals that can embed these AI systems in a steel plant with a deep knowledge of the process itself and the operations.

Dan Bogler 24:49
My understanding is then you can really reduce the time you need to reconfigure the plant.

Alexander Fleischanderl 24:54
That's the point. So, that's the point. We could not simulate complex production units like a blast furnace -- there are so many chemical reactions and physical systems inside there. So, we simplified a blast furnace pretty much, and even then, the calculation and simulation time took weeks for one single result. Now, with the computational power and with all this acceleration power, adding AI and the intelligence from the processes, we hold the big data at the end. That is reducing significantly the simulation time, and we can achieve results within a couple of hours. That was not possible years back.

Dan Bogler 25:45
That's impressive. Slightly different question: so, steel plants, although they're getting safer every year, they're still hot, dangerous, dirty places to work. Is there a way that you can eventually, through a combination of AI and big data and robotics fully automate them? And is that indeed an objective? Do the customers want their plants to be fully automated?

Alexander Fleischanderl 26:09
Yeah, that's a good point. Yeah, so I already touched upon the point that we have a lack of skills.

Dan Bogler 26:17
Yes.

Alexander Fleischanderl 26:17
On the one hand. On the other hand, we have to carefully think about all the safety aspects in a steel plant. It's a hot dirty place with big machinery, and it's super helpful to get humans away from this operation where you expose them to heat or any risk. That's the one side. So, we are quickly moving towards autonomous plant operation. We are not there. This will for sure take a couple of years more. But I would say almost every day we embed another digital assistant that is supporting, moving step by step towards autonomous plant operation. But where we are at the moment, is that we can operate already now one production unit, like a rolling mill or a continuous casting mill, with one operator in the pulpit. That used to take a couple more people, and what it means is still the operator has the decision power, but he is, or she is, well supported and guided by explainable AI that is recommending how a decision should be made. The final decision is still with the operator, and it should be. It's not that AI is replacing the humans, but it's a collaborative way where AI and digitalization in general supports the decision making, achieving better decisions for better results and a better steel product at the end.

Dan Bogler 27:55
That's great. Let me come back briefly to what you mentioned earlier, which is policy support for the industry and for the energy transition in the industry has been inconsistent, you said, or variable. Where has it been better? Where has it been worse? Is there an area, dare I say, Europe, leading this?

Alexander Fleischanderl 28:15
Yeah. So I mean, Europe has taken a leadership role with the callout of the Green Deal in December 2019. So, already six years have almost passed, and I think implementing the ETS scheme, the CBAM, the Industrial Acceleration Act, and all the other instruments is pretty helpful. But, unfortunately, it was not all well thought out. It was one piece after the other, which sometimes happens if you're a first mover, right? That you have eventually to adjust and adapt the systems. But what makes it difficult for the steel producers then to invest? So, we are really talking about large investments. As we said, these assets are designed to operate for the next four decades, and so they really rely on a very stable framework and support from the government. And government's role is really to set this framework and also to provide then support, financial support, to de-risk the new technologies and really to facilitate green steel that when produced is also competitive. So that's still a core issue.

Dan Bogler 29:39
I was going to say if I'm a European steelmaker and I've invested heavily in producing green steel, and then my customers start buying cheaper imports from China or Korea, that's not tenable.

Alexander Fleischanderl 29:50
Indeed, I mean CBAM helps there. So, that's the Carbon Border Adjustment Mechanism, pricing carbon properly. Avoid carbon leakage at the end -- ‘gray’ steel from other regions coming to coming to Europe and to balance the conditions. But I would say we have really now to avoid that the first movers are penalized. They should be rewarded. They have invested heavily. Now we are changing a little bit the rules in in Europe, and that's difficult now. But maybe just taking one or two other regions, I think also India -- and this is a bit of surprise, a positive surprise -- has now really set a clear strategy on how they are going to decarbonize. It's a bit slower, will take a bit longer, but they have a clear strategy and have defined the rules for green steel, so for the carbon that is embedded in a levelized green steel definition. And also in China: China is moving fast in the transitioning of the energy sector, but also in the steel sector. So, we see movement. I would say the train is moving in the right direction. The transition is irreversible. It will go on. It's just a matter of speed.

Dan Bogler 31:22
That's very encouraging to hear. Thank you, Alex. Finally, just a little bit personally. You know, you said at the beginning that you are passionate about this industry, despite having worked in it for many, many years now. What is it, or is there one particular thing you could point to, that you say this was a moment, or this was a result that made me feel I'm making a difference; Primetals is making a difference.

Alexander Fleischanderl 31:46
Yeah, well, I live always in the here and now and looking forward. That's it. That's my passion. So, but they have many, many moments for sure. Also, in the past. But let's take an actual example, and that's the HYFOR smelter project. The industrial demonstrator that we're implementing. So, I think it's not something we have on slides. It’s really something tangible, and we believe this will have a huge impact on the industry, on society, and if this really becomes true, as I said, that we see the first plant, and after 2030, I think I personally, but the whole company can be really proud that we have contributed to the society with such a huge impact.

Dan Bogler 32:39
Right, and I think you don't just sit in an office, right? I think I've seen a Primetals video where you're climbing around in the plant, you know, right in, you know, as far as I can see in front of the furnace. Is that right?

Alexander Fleischanderl 32:52
Yeah, I'm still pretty much hands-on. Yeah.

Dan Bogler 32:55
Well, we're dead on time. Thank you so much for your time. It was a joy to speak to you, and we learned a lot about Primetals, about the steel industry, about the future of hopefully affordable, high-quality green steel. Thank you, Dr. Fleischanderl.

Alexander Fleischanderl 33:13
Likewise, thanks, Dan.