US Steel seeks to forge the future of high technology in factories new and old

When US Steel took full ownership of Big River last year, it also acquired the plant’s artificial intelligence know-how and was a signal of the 120-year-old manufacturing giant’s commitment. advance technology in its factories. But implementing the kind of technology used in Big River at the steelmaker’s other mills, some of which are more than 100 years old, has proven a difficult task, according to the company’s chief information officer.

Competitors like Nucor Corp.

have an advantage over US Steel in part because they mostly operate newer plants with more cost-effective electric arc furnaces and newer technology, similar to Big River, said KeyBanc Capital Markets analyst Philip Gibbs. US Steel has struggled to be profitable in recent years, he said, in part because its equipment is old and expensive to operate and maintain.

When US Steel announced its acquisition of Big River, some saw it as a Hail Mary pass to regain its footing, Mr. Gibbs said. The strategy worked in the short term as demand for steel increased for some time after the pandemic, he said, but added that it remains to be seen how this will play out in the long term as the market tightens.

US Steel said the Big River acquisition has already more than paid off and the company posted its highest first- and second-quarter profits this year before falling demand created headwinds for the third quarter. .

While the Big River deal didn’t revolutionize the way people work at former US Steel facilities, it did serve as inspiration for what’s possible, according to Kevin Burns, process excellence engineer at the plant. Steelmaker’s Gary Works in Gary, Ind.

Cutting-edge technology was part of the main idea when Big River was built, according to Cody Hore, a third-generation steelworker and one of the plant’s first employees. “The next thing you know is that US Steel knocked on our door,” he said. “They wanted to get their foot in the door with rising technology.”

“When someone first suggested I try to automate one of our cranes, I thought, ‘No, no, that’s never going to happen,’” Burns said. “Then you go there and realize, ‘Oh, it’s not an impossible thing. You can do it. You can do it safely. Oh, that’s a very good idea. Maybe we should.

Once North America’s largest steel mill, the 110-year-old Gary Works on the shores of Lake Michigan features equipment that was first put into operation in the 1950s and 1960s, though it was updated over the years.

Gary Works, still one of the largest steel mills in the United States, employs 4,000 people and has an annual crude steel capacity of 7.5 million net tons. Big River has 3.3 million net tons of annual crude steel capacity with approximately 750 employees, which in part reflects its greater efficiency.

Big River uses state-of-the-art technology to make core steel mill functions, such as cooling hot steel coils, more efficient. If the coils are too close together, they take longer to cool, which is why Big River’s automated machine learning crane is so important. The temperature in parts of the mill can reach 150 degrees during the summer, so keeping things cool can be a challenge. Big River Steel recently installed a slushie machine to help keep employees cool.

Big River also uses cameras to power machine learning algorithms that can detect flaws in reel slabs or determine if someone is creating a safety hazard by getting too close to certain machines.

Big River has built some of its algorithms in-house and implemented some through a third party. But algorithms deployed at Big River cannot be plug and play at other plants, according to Christian Holliday, senior director of Digital Studio and Big River Steel Integration at US Steel. Each factory usually needs to build and train its own models according to the unique environment. Older factories may face more challenges because they don’t always produce the necessary data, he said.

That means retrofitting older factories involves installing data-generating devices such as sensors and cameras on existing equipment, according to US Steel chief information officer Steve Bugajski.

Finding the best place to place a sensor on a piece of equipment can be a challenge because it needs to be protected from steam and heat, according to Burns. New equipment often comes with built-in sensors that are safely located inside the machine, something that can’t be done when retrofitting older equipment, Burns said.

Also, older equipment tends to be heavier and bulkier and produce less vibration, so sensors designed to detect vibration might need to be more sensitive, Burns said. In some cases, sensor installations need to be done during a planned equipment outage, which occurs on average once a year, according to Burns.

Older factories such as Gary Works may also lack wireless networking capability to connect sensors, Bugajski said. Bandwidth is notoriously poor in environments such as steel mills that are full of concrete and steel. Newer, more advanced plants can also struggle with network capacity, but connectivity planning is often factored into the design phase, according to Forrester analyst Paul Miller.

The deployment of new technologies also creates challenges in teaching new skills to workers and helping them adapt to change.

US Steel recently began offering digital training to non-IT employees, including machine operators who spend their time in the field in the plant. Mr Holliday said US Steel is on track to meet its goal of having 100 employees trained as “digital agents” by the end of 2022.

“Not everyone embraces change, but we really try to work with people who are,” he said.

Mr Burns of Gary Works said when he entered the steel industry 22 years ago decisions were usually made based on who was the loudest in the room.

“Now we have all of this data, so you can go back and look at what happened and make good decisions based on the data you have at hand instead of relying on the guy who’s 30, 40 years here, and what he remembered about the event,” Mr. Burns said.

Despite the challenges, Gary Works deploys machine learning algorithms designed to make operations more efficient.

An area called the hot strip mill, where steel slabs are turned into coils, can create bottlenecks in production, according to process innovation engineer Todd Hardesty. Now a machine learning system can analyze why certain slabs took longer than they should have. Problems such as slower human operators or faulty equipment can be solved, he said. The factory is extending the time management system to other bottlenecks, according to Hardesty.

Gary Works is also working to create a digital twin, or a live virtual representation of what certain key equipment is doing at any given time, with the ultimate goal of predicting finish times and optimizing output, according to Veeraiah Katta, architect applicable for the United States. Steel.

US Steel hopes to accelerate the creation of such analytical models by creating a single repository for all of its data. The project, a so-called data lake, has been underway since 2018 and will eventually help factories build complex models faster.

Despite the challenges, US Steel remains committed to its modernization efforts, according to Mr. Bugajski, the company’s CIO.

The Big River acquisition did not provide a single model for modernizing former US Steel plants. But Mr Burns of Gary Works said it had helped move the company in the right direction.

“It really takes you from like, ‘I don’t know,’ to ‘Oh, wow, they do it all the time,'” he said.

Write to Isabelle Bousquette at isabelle.bousquette@wsj.com