Big River 2: The Mill that Learns

Welcome to Steel Stories by US Steel. In this podcast, we go beyond the headlines, bringing you insights, ideas and stories shaping how steel is made, used and reimagined. Through conversations with industry leaders and innovators, we explore what it takes to keep forging the future of America. Welcome back to Steel Stories. I'm your host, David Kirkpatrick, and I'm very excited for our episode today with Chris Crowley, who's the Chief Information Officer of Big River Steel. Welcome, Chris. Thanks, David. Happy to be here. Thank you so much for joining us. Today we're going to be talking about the role of digital technology in steel making and particularly how Big river steel, Big River 2 mill, is fundamentally changing the way steel is made thanks to software and AI. So, Chris, really happy to have you here to explain all that to us. Well, I'm happy to be here. Looking forward to the discussion. Well, you know, let's just step back. I know that maybe you should just start by saying what is changing regarding the way software changes steel? I mean, really and truly what we're seeing, you know, is just the, the evolution of digital processes in manufacturing, right? These legacy systems that for the longest time really didn't have a digital footprint, we're seeing it get embedded into the equipment. So you're seeing the soft systems connect to these hard systems at a greater rate than at any other point in time. So with BR2, the number of sensors, the number of data points that we're getting out of that mill is more than anything that we've gotten before, even at BR1. So it's just a massive quantity of data that we're able to extract out of our processes that helps us to improve our systems across the board. Well, that's a huge change in itself. But then we layer onto that the introduction of artificial intelligence, which is in itself a revolution, am I right? Oh, absolutely. This is by far one of the most disruptive technologies that I've ever seen in my career. So AI and the potential for AI is, is something that everybody's talking about. There's trillions of dollars across the world that are being poured into AI implementation. You can look at the broader economy within the United States and the amount of money that's going into data centers as well as the chip manufacturers to support all of this AI. It's truly a world changing technology. Well, we know that to do AI, right, you need a lot of data to train it on and for it to act upon. So obviously the first part, which you described, is a prerequisite to applying the AI. But talk about, given that you now have all these sensors and you've got a super modern mill, I mean, talk to us a little bit maybe before you even answer this question about the Big River 2 plant and what makes it different? Partly, it's just really new, right? Yeah, it's definitely an age thing with Big River 2. It is one of the newest facilities in North America today. You know, we commissioned the mill last year. The technology is a few years behind that, but at the end of the day, when we went live, we have more sensors built into the lines than any other facility within uss. Right. And this is an electric arc furnace mill, which primarily is taking scrap steel and essentially recycling it with some additional materials added. So it's basically an electric operation which makes it different from a lot of more historic mills in the steel industry that would be blast furnaces and that kind of thing. So, okay, given we've got a new mill, we got all this data, we've got the opportunity to use AI. How is your job as the CIO changing? And what's. What, what are you excited about? Well, my job as the CIO is really changing in regards to how do we adopt this technology and make it work for the business. One of the biggest challenges is there's so many unknowns about AI still. You know, it's. It's something that a lot of organizations are trying to implement, they're struggling to implement, and trying to find a solution within a manufacturing environment is all that more challenging. Manufacturing has notoriously been a challenge for anybody in it, and I think that, you know, they'll agree with me. But what we're seeing really and truly is the breakdown of what had been a very standard siloed process within manufacturing. We're seeing a flattening of the digital environment across the board. We are making connections between process and business that, you know, in the past were not there. And we're really looking at taking advantage of AI agents to really just superpower our employees to be more productive. Yeah, I want to get into the topic of agents in some detail in a minute, but. So I know one of the things that you've said is that what we now are moving toward is something that can really be called a learning mill. So what does that mean? Yeah, I mean, Big River Steel was built with the learning mill in mind. And what that really means is taking advantage of the data that you're getting and building a cyclical data flow. Right. So as we're running processes, all those processes are generating data. And as we're generating that data, we want to collect that data, we want to process and analyze that data, and we want to extrapolate answers from that data and feed it back into the lines so that we can improve our processes. It's a continuous learning cycle. So the more we produce off the mill, the more our models can adapt to the changing environment that is manufacturing. And we improve our overall efficiencies, we improve our overall quality. And that's really and truly what a learning mill is. Interconnected systems across the board that can communicate without a human being in the loop. Right. Well, I'm sure that even in the past, both at electric arc furnaces and blast furnace based operations, improvements were continually being made, but they were, the pace of that was dramatically different. Am I right? And so what you're saying is that with this application of AI on top of the data, this is a kind of process, a manufacturing system that will continue getting better, more efficient, higher quality, really indefinitely. Is that kind of the hope? That is the hope, yeah. Ultimately what we want to do is we want to get to where we have the most efficient manufacturing facility in the world, really and truly, you know, with AI on top of a massive data set, we want to be able to take full advantage of the amount of data that we're producing to come up with insights into our processes that nobody's caught before. We want to use AI to automate a lot of the redundant tasks that take up a lot of our employees. Time that really isn't a huge value add to the business. At the end of the day, I've told my entire team this is probably one of the most exciting times to ever be in IT and for steel manufacturing, where Big River Steel is, I've told them that, you know, we, we are in a very unique situation because with the modern technology that we've been able to build into these plants, we're not straddled with a lot of technical debt. Because we're so new, we can really focus on just moving forward instead of having to bring ourselves out of, you know, decades worth of technology challenges and take advantage of the new modern systems. I believe that one of the things that makes the automation of Big River 2 possible and this continuous learning process so desirable and implementable is that you have designed the coils and the systems of the mill itself to be more integrated even in a physical sense. Is that correct? That's correct, yeah. So at BR2 we have the autonomous quail yards. We basically built this out to where as the coil is coming off the hot Mill. It's going into the hot mill bay. And then depending on what the end product is, if that coil has to go to the pickle line or if it has to go down to the GALV line, the system is automatically staging and moving coils around without any real human intervention whatsoever. And this runs 24 7. Nobody has to go out there and say, okay, you know, the, the GALV line now needs coil X, Y and Z. Let's go ahead and grab those out of this bay and move them down there. The cranes are automatically tied in to load those cars on the shuttle cars, and we move them up and down the yard and feed them into all of the process lines. So it's a full autonomous interconnected system that's moving coils from the finishing bays all the way across to the entries of, of the downstream units. Wow. So, in a way, the whole thing was automated to begin with by design. But in fact, once you apply this learning concept using AI, you have sort of a new level of autonomous improvement on top of something that's already intrinsically automated. So I guess I'm interested in understanding what that's going to make possible over time and just talk a little bit about what it means to have all this integration and what a learning mill will be able to do differently as it goes forward. Yeah, I mean, historically, you know, steel manufacturing is, is a, is a challenging industry. There's, there's a lot that goes into the making of steel and there's a lot of information that has to be processed. It's gotta be looked at. You've got, you know, wear and tear on your systems. You've got things that agents, you've got different setups, you've got different temperatures. All of this has to come together to make a good quality coil for our customer. And the idea is that if we can centralize that data and we can put those models in place that are learning from the data as we go, we can adapt as we're processing in line. So the PLTCM is our pickle lime tandem cold milk. Okay. That's where the coils will go through an acid bath to get rid of the surface impurities. Right. So once we go through the PLCM, if we can identify that we've got an issue with a coil, we can feed that back to the hot mill and say, okay, was this an issue that was caused at the hot mill? And then we can adjust the PLTCM as needed if we have to. So it's, it's a communication between these mills when historically those mills operated independently of each other. So historically in an older design mill, you'd sort of just take a coil, which is like a big roll of a flat steel, and just move it from one operation to the other and a whole new process would start and it wouldn't really be connected at all, but now it's really connected. So for example, so I guess if, for example, if, if the thickness was slightly incorrect or if the, you know, some other chemical quality of the steel could be improved slightly, you can, you can adjust that stuff. And some of that, I assume a lot of that is actually done autonomously now too, or am I right about that? So a lot of that data is being fed already. Right. So we, we receive production information from the level twos, which ultimately is driven by all the sensors that we have. That information does get passed on to those additional lines. But what we're trying to do is we're really trying to make more of a live ecosystem where data is flowing in and then going back out as the lines are processing to kind of improve the overall process. That's really what we're getting to. Historically. The in and out and then the processing of the data and then sending the data off as needed has been in place for a while, but with AI and the processing and compute power that we have nowadays, we're able to do that significantly faster. So it's just becoming more real time changes to the processes that used to take a long time. Do you have samples of manual things that really, maybe nobody ever really wanted to do that are going to be able to be avoided going forward as you apply AI more and software more methodically? Yeah, I mean, right now we're, you know, we're tracking automatically the scrap that's being left in our rail cars. Right. If, if we catch a rail car that's going out, that's got scrap left in it, we're sending out an automatic notification to an employee to say, hey, you know, this car needs to be brought back in, we've got to clean it out. We, we don't want to deal with that. That's a simple, you know, nobody wants to sit there and watch 98 rail cars go out so that, and make sure that they're clean. We have a system that's going to manage all of that. There's probably a number of different things throughout the organization that is going to make this significantly better. A small example is trying to find documentation. Right. Everybody hates having to go and search for documents. So what we did is we just Implemented a simple LLM chatbot here at Big River Steel. Its knowledge base is our entire document repository. So anybody in the company can ask it a question and they automatically will get the document that's associated with it. Instead of me having to go and navigate to our PTO policy, I can say, what's my PTO policy? And it provides me that answer immediately. So that's like a custom data set that was developed just for Big River Steel that isn't like out there on the Internet, but you're able to use AI to navigate it. That's correct. Yeah. I got, I got to believe that's going to enable all kinds of efficiencies and, and new opportunities. Yeah. And I mean, that's just, you know, one of the small examples. We're looking at all the business entities for this, right? The sales team, they may have trained training videos that they're doing or training sessions, just like, you know, sitting down on a teams call. And they may record it, they may go through the entire software. Well, we're looking at AI to take those, those videos and automatically create training documents from those so that we can then upload those and it can be searchable by an LLM. So anybody who was in that recording or who wasn't in that recording can easily go back and say, you know, yes, I, I know we had this training. I forgot how to do this one step. Instead of having to go back and find exactly where that training was done in the video, they can actually just ask a question of the AI bot and it'll bring up a video screenshot that shows what was actually done. Wow. You know, it's almost like the sky's the limit, right? There's there, there's like, you think you like, turn, you know, spin the bottle and find something that can be improved with AI. Almost. Right. Well, and that's, that's what, you know, that's our, that's our point, which is humans are fantastic at being creative, right? We come up with these ideas, we imagine what the future looks like. AI is just the enabling tool that we can now implement that helps us get to that, and that's the way that we're approaching this is the sky is the limit. We don't know where this technology is going to take us, but what we do want to do is we want to make sure that we're using it to better our lives versus making our lives miserable. How have the employees responded to all these innovations? Everybody at BRS has done a really good job. I don't Think I've met a single person down here who has said, I don't want to see AI down here. Everybody has embraced it. Everybody has said, I already know two or three different things that this technology could do for me. So our list of projects to build out AI solutions at BRS is just growing by the day. How are customers responding to all this? And are you implementing some of their input into the way that AI is adjusting the processes? Are they getting better quality? I mean, obviously they're getting better quality. More reliable products, I would assume, based on everything else you've said. But are you talking to them a lot about how AI and software is changing the nature of the product they're delivered and just talk a little bit about that interaction with the customers around the technology? 100%. We are definitely talking to our customers about AI and how this, this technology is kind of pushing us forward in the future. One of the things that, you know, we immediately saw some, and this isn't necessarily just directly related to AI, but one of the things that we saw was the high quality of steel that came off of the esp. That was, that was something that. Off of the, the acronym again, off of the, sorry, our endless Strip processor. That was, that was at BR2. So when we implemented that, the quality of material that was coming off of that, we got a lot of customer feedback and they were very, very excited to have product that was that high quality coming off that mill and that, that just attests to more. We build out this digital infrastructure and then the more we put AI on top of it, the better the customer experience will ultimately be. If we do it in a strategic way, if we think through what we're doing and the impact we're going to have on our customer, all we're going to be doing is delivering better quality. We're going to be delivering better efficiencies, more insight into what, where their products are in the process. You know, we've often thought about the Domino's app. When you order a pizza, right? That's our end goal. We want these people to order a coil from us and we want them to be able to see exactly where it is in the process. And then when it drops in their, their facility, wherever it may be, if it's a warehouse or if it's a plant, they're going to get a notification that says, hey, your coil has been delivered. So that's perfectly possible. And you're moving in that direction now. Yes. How far are you from basically that Domino's type experience? Now, would you say we're probably about halfway there. We've still got some work we've got to do, but. But we've definitely laid the foundation for all of what's coming from Big River Steel. I would think that would be a real competitive advantage when you can offer a customer that degree of visibility into what's going on. So congrats on that effort and I'm sure you'll get there. Anything else I should have asked you or that you had hoped to talk about in this conversation that we didn't get to? Not off the top of my mind. I think we've covered a lot, so it was great getting a chance to talk to you, David, and really appreciate the opportunity. We really did. Well, Chris, thank you so much for joining us and I hope we'll check in with you again down the road to see how this evolves because this is really state of the art steelmaking and it's very exciting to see it happening at US Steel. So congrats. Well, thank you. Yeah, we're really excited for the future of US Steel and Big River Steel and what, what this means to steel manufacturing in North America, for sure. Well, thanks and thank you all for listening and watching and please join us for the next Steel Stories. I'll see you. Steel Stories is produced by US Steel. The views expressed by our guests are their own and do not necessarily reflect those of US Steel. To learn more about our people, our capabilities, and where steel is headed Next, visit ussteal.com you can find steel Stories on Apple Podcasts, Spotify or wherever you listen. And be sure to subscribe so you don't miss what's ahead. Thanks for listening.