The term “Industry 4.0” was coined by the German government in 2013 to describe strategies for using technology to improve manufacturing productivity, including artificial intelligence (AI), robotics, and the industrial internet or “Internet of Things” (IoT). The term has since gone global, drawing the attention of the steel industry and being widely used in press releases and company reports. According to S&P Global Platts estimates, steel mills have invested anywhere from tens of thousands of dollars to tens of millions of dollars to improve data systems and make them “leaner,” more automated, accurate, cost-effective, and more efficient. Security, in short, is smarter. Data and algorithms will adjust inventory, speed up billing cycles, predict customer demand, select better raw materials, and allow production lines to “collaborate” with other equipment. “Big Data” using cloud storage and involving the Internet of Things has already entered our lives, with an estimated 12 billion devices, including autonomous vehicles and smart TVs, connected in the past five years.
Robotic process automation (RPA), which allows repetitive tasks to be performed faster and cheaper in steel mills, is being adopted more quickly in mines than in steel mills, even though its use remains relatively limited across the mining industry, perhaps because of the apparent need to keep human resources in check. Partial work.
According to World Steel Dynamics, from a global perspective, the level of employment in the steel industry has dropped by about 50% from 1972 to 2012 due to consolidation and improved production. The employment of 2 million people and the indirect employment of 4 million people further decreased. However, World Steel said employment opportunities for skilled workers in the industry remained good.
Yandex Data Factory, which is currently cooperating with more than a dozen steel, metal, and mining companies, especially cooperates with related companies in areas such as optimizing raw material consumption to reduce smelting costs and ensuring the quality of finished products. The amount of raw materials is more accurate. AI typically analyzes up to seven years of previous smelting data values to determine which types and quantities of ferroalloys or other raw materials are used to achieve the best cost per smelting that is unimaginable to operators.
AI can reduce the amount of raw material used, which could be crucial for refining gold, zinc, and copper, which use expensive cyanide, which accounts for 20%-40% of processing costs, Yandex Data Factory said. While AI can be used in any aspect of any company, it is best used on the production floor, said Jan Zavalina, CEO of Yandex Data Factory. Maybe you have six months to measure the results and learn what algorithms work with, say, reducing slab defects, he said. If it takes just over three months for an AI project to show results, you could be wrong. Indeed, corporate pressures are fast-emerging and tangible, especially in the areas of energy conservation, raw material optimization, and product quality. The biggest problem, he says, is expecting the data to explain improvements made, or provide training, and the data used must be trusted to produce good results.
Raw material quality assessment
One of the recent cases studied by Yandex Data Factory involves the optimization of the use of ferroalloys at the Russian Magnitogorsk Iron and Steel Company (referred to as RMS) with an annual production capacity of 12 million tons of crude steel. RMS and Yandex Data Factory have analyzed more than 200,000 smelting historical data in the past few years to study how to reduce the use of ferroalloys and auxiliary materials in the oxygen converter plant while ensuring product quality. According to Yandex Data Factory, the result has been an average 5% drop in ferroalloy consumption, which equates to annual savings of $4.3 million. Sergei Sulimov, a former official in charge of finance and economics of Russia Masteel, said, “We believe in the feasibility of using mathematical models for big data analysis, as well as the rapid development of Internet of Things technology, and will be in the next 3-5 years. It will reduce the cost of industrial enterprises by 5%-10%.
Real-time data is key
Markus Malinen, vice president of Europe, Middle East, Africa, and Russia at Quintiq, a subsidiary of the Dassault Group, said that big data itself is not new to the steel industry. It has been around since the beginning of computers. Using, what is newer is the increasing complexity of inputs and sensors used today, as well as optimization and self-learning capabilities. He highlighted the increasing need for steel companies to use real-time measurements to adjust production plans and plan production operator behavior to improve accuracy, ensure quality control, and remain competitive. He said the use and adoption of AI in the general industry has been low, however, it does have customers and interest and implementation have been growing. Quality has been the driver for the initial adoption of AI, but environmental concerns and market pressures have also been contributing factors.
Markus Marinen says, “One of the main areas where AI can be used is in reducing scrap generation. Each steel grade you produce has its own chemical composition and heat requirements. Your production organizer can combine several orders of the same grade. Proceed to production. Due to limitations of production equipment and materials, there is a risk that you will produce slabs that do not match the original target. Probes can be used to analyze steel production and continuous testing to precisely reduce scrap generation. If requirements are exceeded, the system will An alert of an excess requirement can be sent, or the location of the current order can be searched in real-time to see if another order is available for that steel, so production can keep running.”
Equipment manufacturers lead the way
Ron Ashburn, executive manager of the American Institute of Iron and Steel Technology (AIST), said, “As steel mills move forward with digitization, we will see equipment manufacturers leading this evolution in the steel industry.” Proactively roll out unique technologies for self-adapting and learning factories. For example, Danieli Automation announced earlier this year that it had acquired Italy’s Tel robot Laboratories to enhance its robotics capabilities. Tenova recently announced a partnership with Microsoft, whose Azure cloud platform provides Industry 4.0 solutions for the steel industry. According to Ron Ashburn, the development of digitization, big data, and Industry 4.0 in the steel industry is rapidly gaining momentum. He said that in the long run, all factories must be digitized to survive, otherwise, they will eventually be eliminated.
Access to innovation funding
Andrea Bassanino, the partner of Ernst & Young Advisory Services and head of the strategy for the Mediterranean region, pointed out at the steel industry conference held in Italy in May this year that European steel companies can obtain AI projects from the EU under the “Innovation Plan 2020”. , a fund totaling 80 billion euros for all industries in the period 2014-2020, which also has considerable funding and tax incentives in Italy. However, given that Italy’s domestic steel industry can hardly find a balance between production and consumption levels, he urged Italian steel companies to undertake Industry 4.0, which could only be financed after a thorough analysis of expected outcomes. Plus, there is interest in AI projects all over the world, with many companies receiving little government funding to get a boost. Since the beginning of this year, news of steel companies starting to implement AI projects has been increasing.
New AI is being implemented
Steel companies that are clearly leading in the field of AI include South Korea’s Pohang Iron and Steel Company, which claims that it has become the world’s first steel company to introduce AI production processes, and strives to turn itself into a smart steel factory manufacturing company. According to Noodl e.ai, an American smart technology company, Dahe Iron and Steel Company of the United States is also known as “the world’s first smart steel production enterprise”. Since January this year, POSCO has been using AI big data for deep learning to control coating weight, replacing previous manual control. After a two-month pilot test, with the support of the Department of System Management Engineering at Sungkyunkwan University, South Korea, the new system was introduced to the No. 3 continuous galvanizing line at Gwangyang Plant. By using artificial intelligence to precisely control a continuous galvanizing line, a core technology in the production of automotive steel sheets, Posco has been able to significantly reduce coating weight deviations. This is an automatic control technology, by combining the coating weight production model of artificial intelligence technology with the control model of optimization technology, it can predict the coating weight in real time and accurately meet the target coating weight.
In addition to reducing operator workload, the new system is reducing production costs that previously resulted from underskilled workers wasting expensive zinc coating. In the case of manual control of the coating, the coating weight deviation was 7 grams per square meter, but with the artificial intelligence-based system, the deviation was only 0.5 grams, POSCO said. The company also said it is one of 20 steel companies in the world that can produce so-called world-class advanced automotive coated steel, and that it sold 9 million tonnes of automotive steel sheet in 2016, accounting for 10 percent of the market. POSCO plans to further apply the automatic coating weight control solution to the company’s other continuous galvanizing lines at home and abroad, and actively introduce artificial intelligence technology in other steel production processes while building smart factories.
The Big River Steel Company in Arkansas, USA, is described as a flexible mini-process steel plant. The steel plant started production in late 2016, with a total investment of 1.3 billion US dollars, integrating sensor feedback, machine learning, and automation in the whole set of equipment. In March 2017, the company signed an agreement with Noodl e.ai to provide cloud-based supercomputing services, optimize equipment maintenance plans, production line scheduling, logistics operations, and environmental protection, according to the company’s website and the American Institute of Iron and Steel Technology.
Big River Steel CEO David Stickler said in a release that the smart system will help the plant shift its focus to equipment maintenance. Additionally, laminates were mentioned in the AI survey as an area that could be helped, saving time and money when maintenance is required. He said, “We look forward to challenging industry norms by mining the treasure trove of data, and Big River Steel has a unique position in data collection and analysis.” It is said that Noodle e. The development frontier of the megatrend.
Speaking at steel mills during the ASTI 2017 Annual Conference and Expo, David Stickler noted, “We have a large number of sensors and measurement systems, and we look for correlation in every aspect of our operations, whether it’s our scrap portfolio or breakthrough detection. We are a cloud-based company, we do not have a large system or a large IT department, which saves tens of millions of dollars. We also believe that by running on the cloud platform, we can improve network security. ”
In June 2017, Primetals Technologies and Baosteel Group Corporation signed a contract for the provision of the “Dynamic Width Control” technology package. The technology package will be installed on Baosteel’s 1580 hot rolling mill in Shanghai as part of Baosteel’s “smart workshop” pilot project, which is also one of the “Made in China 2025” government projects, to improve the width performance of the 1580 hot rolling mill, thus Reduce waste loss and improve cost control. The new technology package will be available during regular production until the end of 2017. The reduction and elimination of width deviations is a challenge for hot-rolled coil production, Primetals Technologies said in a statement. The hot rolled coil width is usually controlled by the vertical stand in the roughing mill and the relationship between tension and reduction in the finishing mill. For the latter part, the dynamic width control system of the finishing mill of the new technology package controls the width through the tension control of the finishing mill.
India’s Tata Steel is considering using AI to figure out how to spot surface defects on strip steel for automotive use. To this end, the company opened a laboratory in Amsterdam Science Park, where it is working with the artificial intelligence company Scyfer to improve the steel inspection process.
The Austrian Voestalpine Group recently announced that it plans to invest 100 million euros in two future-oriented projects in its Donawitz steel plant in line with the principles of Industry 4.0. casting lines, and a research center dedicated to the development of lighter, stronger steel grades. According to Bloomberg, the new technology means that Voestalpine will soon be able to produce 500,000 tons of some type of wire a year with just 14 employees.
Robots walk into steel mills and collect scrap
At the same time, electric robots will also be used in continuous casting production in steel mills. At the end of 2016, the robot installed by Primetals Technologies on the LiquiRob continuous casting platform of Deutsche Edel Stahl werke company replaced the manual ladle oxygen blowing, which has greatly improved the safety of workers’ operation and the success rate of oxygen blowing for the first time. This allows the use of larger and more efficient oxygen lances.
Through the introduction of hydraulic demolition robots, such as the Aqua Cutter 410V robot launched by Aquaj et Systems AB in June this year, scrap steel collection is very convenient. It uses water jets to remove concrete from steel bars at renovation, repair, and demolition sites. A valuable feature of the hydro-demolition robot is that the rebar is cleaned and descaled during the process, while alternative methods, such as pneumatic tools, could damage the rebar or create microcracks in the remaining concrete, the company said. In addition, Kobe Steel has introduced the Arcman 4 robot digital control welding system, which greatly reduces the welding cycle of medium and thick plates.
In terms of the robot industry as a whole, China is the world’s largest consumer of robots. In 2016, China sold 89,000 robots, a year-on-year increase of 26.6%. In the same year, global robot sales were 290,000 units, a year-on-year increase of 14%. China’s robotics industry is seen as having huge market potential, with sales expected to exceed 100,000 units this year, partly due to government policies supporting the smart robot industry, and some industrial sector upgrades. In China, robots were first used for welding in the automotive industry in the 1990s, and then for loading and unloading. Currently, China is working on the use of robots in the power industry and in the assembly of consumer electronic devices.