Steel Times International digital August 2019

INNOVATIONS

INDUSTRY NEWS ROUND-UP

GALVANISING TECHNOLOGIES

TESTING & ANALYSIS

Another bumper section of 22 pages devoted to new contracts and orders

A round-up of the latest global steel industry news

John Cockerill examines galvanising technology for AHSS steel grades

ArcelorMittal discusses automated surface inspection

www.steeltimesint.com Digital Edition - August 2019 - No.12

WILL COKE BE NEEDED IN 2030? WE REVIEW EUROCOKE 2019

From John Cockerill to CMI. From CMI to John Cockerill.

The encounter between traditional values and modern trends has helped us to offer the bestadapted response to modern clients’ aspiration.

Inspired by the visionary and entrepreneurial personality of its founder, CMI once again becomes John Cockerill.

Since 1817, the strong commitment to a culture of creative thinkers helped us to provide innovative and profitable answers to the needs of our clients.

Resolutely oriented towards the future, innovation forms an inherent part of our engineering.

While growth is substantial, the challenge is to focus on generating sustainable progress. This is what we have done for the past 200 years.

Design | Engineering | Technical Assistance & Training | After Sales

johncockerill.com/industry

INDUSTRY METALS Reheating Furnaces Pickling Acid Regeneration Cold Rolling Strip Processing Automation

CONTENTS DIGITAL EDITION - AUGUST 2019

INNOVATIONS

INDUSTRY NEWS ROUND-UP

GALVANISING TECHNOLOGIES

TESTING & ANALYSIS

Another bumper section of 22 pages devoted to new contracts and orders

A round-up of the latest global steel industry news

John Cockerill examines galvanising technology for AHSS steel grades

ArcelorMittal discusses automated surface inspection

Picture courtesy of: Friedrich Kocks GmbH & Co Kg

www.steeltimesint.com Digital Edition - August 2019 - No.12

2 Leader

by Matthew Moggridge, editor.

52 Galvanising lines Galvanising technologies for AHSS

4 News round-up The latest global steel news.

59 Digital manufacturing Through-process optimisation

8 Innovations The latest new products and contract news.

66 Digital manufacturing Securing operational technology in metal manufacturing

34 Eurocoke Summit 2019 Will coke be needed in 2030?

72 Reputation management Keeping up appearances

40 Automated surface inspection Performance optimisation of automated surface inspection

76 History Mining ore on the English Weald

WILL COKE BE NEEDED IN 2030? WE REVIEW EUROCOKE 2019

EDITORIAL Editor Matthew Moggridge Tel: +44 (0) 1737 855151 matthewmoggridge@quartzltd.com Consultant Editor Dr. Tim Smith PhD, CEng, MIM Production Editor Annie Baker Advertisement Production Martin Lawrence SALES International Sales Manager Paul Rossage paulrossage@quartzltd.com Tel: +44 (0) 1737 855116 Sales Director Ken Clark kenclark@quartzltd.com Tel: +44 (0) 1737 855117 Managing Director Steve Diprose stevediprose@quartzltd.com Tel: +44 (0) 1737 855164 Chief Executive Officer Paul Michael SUBSCRIPTION Elizabeth Barford Tel +44 (0) 1737 855028 Fax +44 (0) 1737 855034 Email subscriptions@quartzltd.com Steel Times International is published eight times a year and is available on subscription. Annual subscription: UK £178.00 Other countries: £254.00 2 years subscription: UK £320.00 Other countries: £457.00 ) Single copy (inc postage): £40.00 Email: steel@quartzltd.com Published by: Quartz Business Media Ltd, Quartz House, 20 Clarendon Road, Redhill, Surrey, RH1 1QX, England.

Tel: +44 (0)1737 855000 Fax: +44 (0)1737 855034 www.steeltimesint.com Steel Times International (USPS No: 020-958) is published monthly except Feb, May, July, Dec by Quartz Business Media Ltd and distributed in the US

by DSW, 75 Aberdeen Road, Emigsville, PA 17318-0437. Periodicals postage paid at Emigsville, PA. POSTMASTER send address changes to Steel Times International c/o PO Box 437, Emigsville, PA 17318-0437.

Printed in England by: Pensord, Tram Road, Pontlanfraith, Blackwood, Gwent NP12 2YA, UK ©Quartz Business Media Ltd 2019

ISSN0143-7798

www.steeltimesint.com

1

2

LEADER

You simply can’t miss Future Steel Forum 2019...

Matthew Moggridge Editor matthewmoggridge@quartzltd.com

Future Steel Forum 2019 takes place over 25-26 September at the Sofitel Hotel in the Hungarian capital, Budapest.

The Future Steel Forum has been a pioneering conference. When it was launched in 2017, I can’t think of any organisation running a similar event in the steel space.

Recently, however, other organisations have announced their own events.

Notably, the US-based Association for Iron & Steel Technology (AIST) offered its members a ‘Digital Transformation Forum for the Steel Industry’ and it was good to see some familiar faces on their schedule of presentations.

This year the Future Steel Forum offers delegates an exciting programme, including, for the first time, a Chinese steelmaker, the Rockcheck Steel Group.

Delegates will hear from the company’s chief executive officer, Catherine Zhang, who will offer her perspective on the Chinese approach to digitalisation and steelmaking.

From the USA, Dr. Lane Thames from Tripwire Inc’s Vulnerability and Exposure Research Team, will discuss Fog

Computing, the Cloud and cybersecurity.

Stephen Pratt, founder and CEO of San Francisco-based Noodle AI, a company involved in the development of Big River Steel’s Osceola ‘learning mill’ in Arkansas, will look at artificial intelligence, as will Patrick Henz from Primetals USA. Among many notable speakers from Europe, Costanzo Pietrosanti, chair of the Smart Factory Focus Group at ESTEP, will look at industrial cybersecurity. There will be papers from Primetals Technologies, PSI, SMS group, Fives, Tenova and Danieli, not to forget a panel discussion on Industry 4.0 and the Workforce following a presentation from Professor Farrokh Mistree of the University of Oklahoma entitled Frontier Issues for Succeeding in a Digitalised World. Professor Mistree will be on the panel alongside Liberty House Group’s Eric Vitse, IBM’s José Favilla and Fives’ Kriistian Van Teutum. And if that’s not enough, there will be a presentation from POSCO on the ‘smartisation’ of ironmaking.

To register for the conference and to see the full two-day programme, simply visit www.futuresteelforum.com

Offering The Most Suitable Dimensional Measuring Solutions Regardless of Shape, Size, Speed or Temperature, We Measure It All !

Static, Oscillating, Rotating, Full Profile Capture, We Do It All !

www.zumbach.com • sales@zumbach.ch

Temposonics® Magnetostrictive Linear Position Sensors

SOLUTIONS FOR HIGH TEMPERATURE APPLICATIONS

SOLUTIONS FOR HIGH TEMPERATURE APPLICATIONS

ABSOLUTE POSITION SENSORS WITH INDEPENDENT DETACHED ELECTRONICS

Provides accurate measurement

Varied operating temperatures to +105 °F) Detached Electronics Absoluteup Position Sensors°C with(+221 Independent Accurate Measurement | Varied Operating Temperatures up to +105 °C (+221 °F)

Model RD4

Model ET

Model GBS

Contact applications support at 1-800-633-7609 / sensorsinfo@mts.com MTS Systems Corporation, Sensors Division • 3001 Sheldon Drive Cary, NC • www.mtssensors.com/industrial

mtssensors.com | info.us@mtssensors.com | 800 633 7609

http://bit.ly/2GqhbXo

4

• Domestic sponge iron producers in Iran produced 2.5Mt of the product during the first month of 2019, according to figures released

by IMIDRO (Iranian Mines and Mining Industries Development and Renovation Organisation). The export value of Iran’s minerals and mining industries’ exports stood at US$9.2 billion over the past year. Projects worth an estimated US$2.17 billion have already been announced.

Source: Tehran Times, 1 June 2019.

• While rumours were rife about whether or not Sanjeev Gupta (pictured above) was likely to make a move on British Steel, the reality was something else. On 2 June, the website This Is Money. com suggested that Mr Gupta and his green steel vision was the perfect fit, and many industry observers agreed, but in mid-August a report appeared on the BBC website stating that the investment arm of Turkey’s armed forces pension fund had emerged as the preferred bidder for the steelmaker. Ataer Holding is the name of the Turkish armed forces pension fund and it owns almost half of Turkish steelmaker Erdemir.

Digital Edition - August 2019

NEWS ROUND-UP

• India has been removed from the USA’s preferential trade programme as the country ‘has not assured the United States that it will provide equitable and reasonable access to its markets’ to quote the Donald.

The USA’s Generalised System of Preferences gives developing countries easier access to the US market, according to a report by CNN. Source: CNN, 2 June 2019.

• China’s Baowu Steel Group is the second-largest steelmaker in the world, but could it ever knock ArcelorMittal off the top spot? Its merger with arch rival Magang (Group) Holding will put it within striking distance, according to online reports.

According to Baowu, the merger is ‘a major step toward reaching the goal of 100 million tons a year’.

Source: Asia-Nikkei.com, 3 June 2019.

• A global recession could become a reality and it’s all the fault of The Donald, according to online reports. The US president’s trade war with China is behind shrinking factory activity in Asia and the situation could spread as higher trade tariffs take their toll on global commerce.

Source: Free Malaysia Today.com, 3 June 2019.

• Trouble might be brewing in the Bellari district of the State of Karnataka in Southern India, following a decision to sell 3,666 acres of land to the Indian steelmaker JSW. Kota Srinivas (pictured above), a politician from the Bhartiya Janata Party (BJP) – who was elected unopposed to the Karnataka Legislative Council last summer – has urged the Government to think again. He said there was no urgency to convert the lease agreement over the land into a sale deed.

STI will keep you posted.

Source: Deccan Herald, 2 June 2019.

• Herbert Eibensteiner has been appointed chairman of the management board of Austrian steelmaker voestalpine. He takes over from Wolfgang Eder who has been elected to the supervisory board.

Eibensteiner started his career at the company as a plant engineer in 1989.

Source: voestalpine, 3 July 2019.

For more steel industry news and features, visit www.steeltimesint.com

• ArcelorMittal, the world’s biggest steelmaker, is reducing production in Europe in three factories (Dunkirk, France; Bremen and Eisenhuttenstadt, Germany) due to demand conditions in the market. It will also extend the maintenance shut-down period of its Asturias facility in Spain in Q4 2019. The company expects production to recover in all these factories once demand conditions improve.

Source: The Corner.eu, 3 June 2019.

NEWS ROUND-UP

• ArcelorMittal’s bid to acquire Essar Steel has been approved, but the resolution plan submitted by the world’s biggest steelmaker – to safeguard the rights of operational creditors and other financial creditors – has been changed. The National Company Law Appellate Tribunal has said the proceeds of the sale were to be shared among creditors.

Source: The Hindu, 4 July 2019

• And talking of Essar Steel, ‘private sector Essar Steel’ has signed a deal to sell 1Mt of hot-rolled steel to the Indian arm of South Korean steelmaker POSCO for US$730 million. Over the past five years, POSCO in India has bought roughly 2.5Mt of hotrolled steel to use as substrate for value-added products made at the company’s Maharashtra plant in Western India.

Source: Argus Media, 5 July 2019.

5

• Villagers in Joda in the Keonjhar district of Orissa, India, are protesting outside the main gates of Tata Steel’s iron ore mining facility because, they claim, the steel giant has ignored their demands. The taking over of village roads, jungles and even cremation ground has caused villagers numerous problems, according to a local news website.

Source: Orissa Post.com, 5 July 2019.

• Taiwanese steelmaker Chung Hung Steel Corp has downplayed the effectiveness of punitive US levies on steel imports from Vietnam, claiming they have had limited effect on the company and that Vietnam would remain its largest overseas market this year. The company claims that the US imposes 3% anti-dumping tariffs on steel imports from Taiwan, meaning there is no need to export goods through Vietnam to reach the USA. Source: Taipei Times, 5 July 2019.

• Three steel concerns owned

by Steel Authority of India Ltd (SAIL) have been put out to tender through something known as an ‘expression of interest’. The three ‘steel units’ are Salem Steel, the Durgapur Alloy Steel Plant and Viveswaraya Iron & Steel, based in Bhadravati in Karnataka. All three have been on the Indian Government’s divestment agenda since 2016.

Source: Economic Times, 5 July 2019.

• Provincial administrations in South Korea are stirring the pot through suspending both POSCO and Hyundai Steel operations because of illegal emissions of air pollutants. Hyundai in South Chungcheong Province was told to cease work for 10 days from 15 July and POSCO’s Pohang and Gwangyang works have been told the same. It has been reported that ‘punishment is inevitable’.

Source: Korea Times, 6 July 2019.

• Following the opening of its additive manufacturing (AM) TechCentre Hub in Singapore, German steelmaker ThyssenKrupp has published a white paper examining the growth opportunities behind AM, claiming that it could create three to four million jobs by 2030 in the 10 countries that make up the Association of Southeast Asian Nations.

Source: 3d Printing Media Network, 5 July 2019.

Digital Edition - August 2019

6

NEWS ROUND-UP

• India’s Union Budget is pushing so-called ‘big ticket’ infrastructure projects, which in turn should boost demand for steel. Finance Minister Nirmala Sitharaman has announced steps to ‘scale up’ infrastructure developments such as road, rail and logistics upgrades with a view to achieving greater connectivity. The digital economy and job creation were also high on the budget agenda.

Source: The Hindu Business Line, 7 July 2019.

• Tata Steel’s Colorsteels service centre in the UK has been given a boost with a new hub in Glasgow. A 1,000 sq metre site has been set up to service fabricators in the region who make flashings, fabrications and gutter systems using Colorcoat prefinished steel products.

Source: Pressroom.today, 8 July 2019.

• Indian steelmaker JSW has been given the goahead to acquire Asian Colour Coated Ispat (ACCIL).

Source: News Today Net, 7 July 2019.

• Questions are being asked about whether or not POSCO will build a steel plant in India – and it is thought unlikely after the company withdrew from the Odisha project following public resistance to the proposed plant and regulatory hurdles. However, it transpires that the Indian Government has been asking the South Korean steelmaker to look again, and company executives have met with Indian officials, it is claimed.

Source: News Article Insiders, 8 July 2019.

Digital Edition - August 2019

• A report by Reuters claims that ThyssenKrupp is still planning to float its elevator business, following rumours in the media that talks on selling the unit would start in the autumn. It is thought that Kone of Finland might be a potential buyer.

Source: Reuters, 8 July 2019.

• Indian steelmaker JSW’s crude steel output grew 3% to 4.24Mt in Q2 2019, up from last year’s 4.12Mt. During Q2 2019, the company’s production of flat-rolled products grew 1% to 2.91Mt. Output of long-rolled products stood at 1.05Mt, up 12% from the same period in 2018.

Source: Economic Times, 8 July 2019.

• Steel magnate Lakshmi Mittal, the driving force behind ArcelorMittal, the biggest steelmaker in the world, and the 11th richest man in the world, according to the Sunday Times Richlist, was a supporter of Boris Johnson’s leadership campaign. Mittal has also supported the Labour Party – to the tune of £125,000 – when Tony Blair was Prime Minister.

Source: The Times, 1 August 2019.

• South Korean steelmaker Hyundai Steel claims it will reduce air pollutant emissions by over 50% by 2021. The company plans to operate a sinter gas treatment system (SGTS), which it has been using at its plants in Dangjin, South Chungcheong Province.

Source: Korea Times, 9 July 2019. www.steeltimesint.com

NEWS ROUND-UP

• The Russians have approached the World Trade Organisation to complain about US anti-dumping measures on imports of hot-rolled and flat-rolled carbon-quality steel from Russia. An online report from Reuters states that the Russian complaint ‘listed eight separate US announcements between 1999 and February 2019’. The US has 60 days to settle.

Source: Reuters, 9 July 2019.

• Workers at Riva Stahl in Trier and neighbouring Horath in Germany have gone out on strike for four weeks in protest against poor wages.

It is claimed that some workers at the wire works are on pay that is one third of the sector’s union-agreed rates. Others are on below 12 Euros per hour. The workers voted to strike by a large majority on 11 June.

Source: World Socialist Web Site, 9 July 2019.

www.steeltimesint.com

• Iran’s annual steel production capacity will be boosted by 10Mt/yr when plans to establish a steel mill in the Persian Gulf Special Economic Zone come to fruition. IMIDRO (Iranian Mines & Mining Industries Development & Renovation Organisation) signed on the dotted last month. The project will take five years and cost $5 billion and create 10,000 jobs.

Source: Financial Tribune, 10 July 2019.

• US automotive producers sold 8.3 million passenger cars, trucks and sports utility vehicles during H1 2019, down 2% on the same period the previous year (8.5 million units). The downward trend is likely to put pressure on steelmakers.

Source: Argus Media, 10 July 2019.

• A fire that broke out at MMK’s Mill 2500 in Russia on 8 July is having a negative effect on hot rolled coil production. The mill has declared force majeure on some contracts and given smaller orders to the company’s shop number 10, which is normally reserved for the production of large coils. Around 125 to 150kt of hot rolled coil will be affected.

Source: Argus Media, 10 July 2019.

• Tata Steel is not planning another joint venture in Europe, but will make its European operations sustainable by encouraging them to be cash-positive, according to CEO and managing director TV Narendran.

Source: Business Standard, 11 July 2019.

7

• Voestalpine Eifeler Coatings has opened a new coatings facility in Cleveland, Tennessee, USA. The plant will supply PVD coatings to manufacturers in Tennessee, North Carolina, South Carolina, Georgia, Alabama, Mississippi, Arkansas and Florida.

Source: The Fabricator.com, 11 July 2019.

• Two leading Indian steelmakers, Tata Steel and JSW, are both focused on water sustainability. Tata plans to continue investing in sewage treatment plant and creating new rain water harvesting structures, while JSW is creating water recycling infrastructure and developing technology to lower fresh water intake.

Source: Economic Times, 12 July 2019.

• Cosimo Massaro, a crane operator at an ex-ILVA ArcelorMittal steelworks in Taranto, was killed when he was blown into the sea during a gale.

Divers at a depth of six metres recovered his body following a three-day search. Eight people are under investigation.

Source: ANSA.it, 15 July 2019.

Digital Edition - August 2019

00 8

INNOVATIONS

Primetals acquire ABP Induction Systems Mitsubishi Heavy Industries (MHI) and Primetals Technologies is to acquire ABP Induction Systems (ABP), a global manufacturer and servicer of induction furnaces and heating systems currently owned by private equity company CM Acquisitions of Chicago, USA. According to Primetals, ABP offers a variety of best-in-class products and comprehensive services to blue-chip customers, including leading automotive OEMs and suppliers, industrial manufacturers, independent foundries as well as steel plant manufacturers and steel producers.

MHI and Primetals Technologies will jointly take ABP’s shares. Future business activities will be conducted in close co-operation with and under the leadership of Primetals Technologies.

The completion of the acquisition of ABP is planned to close around the end of August.

ABP provides state-of-the-art equipment for ferrous and non-ferrous metal casting, forging and steel making.

The company specialises in induction melting, holding and pouring furnaces as well as induction heaters and its business is built upon a large and glob-

al customer base with more than 1,600 active units worldwide. “ABP also has a core competence in the service business and provides comprehensive aftermarket solutions to customers through the entire product lifecycle,” said Primetals.

The company has service centres strategically located close to major industrial areas in Germany, the USA, China, India, Mexico, Russia, South Africa, Sweden and Thailand.

ABP exclusively provides special induction heaters to Primetals Technologies for endless strip production.

Satoru Iijima, chairman of the board and CEO of Primetals Technologies, said that ABP’s induction heaters are one of the most crucial elements for endless strip production. “With ABP becoming one of MHI’s group companies and the further close ties that will bring, we can develop and provide customers with even more advanced technologies,” he said, adding that the acquisition will enable Primetals Technologies to combine ABP’s competence in induction heating and related activities with its know-how as a worldwide engineering, plant-building, lifecycle services and digitalisation partner for

the metals industry. “ABP´s well-experienced portfolio and its know-how will certainly complement our wide range of customer plants, namely mini mills and long rolling plants, especially in emerging markets, as well as in endless strip production,” he said.

Till Schreiter, CEO of ABP, commented: “ABP’s state-of-the-art induction products and technology-driven culture will fit well with both shareholders.

Through a closer tie-up with MHI and Primetals Technologies, ABP can pursue further growth potentials, which will also lead to a contribution to them”.

With MHI and Primetals Technologies, ABP has access to their resources worldwide, which will improve ABP’s global market presence, provide opportunities to develop new business sectors, and drive digitalisation.

“This will assure long-term stability for our facilities, employees and customers”.

ABP will be a group company of MHI under the ownership of Mitsubishi Heavy Industries America, Inc., headquartered in Houston, Texas, and Primetals Technologies USA LLC, Alpharetta, Georgia.

IFM 7 Twin Power from ABP, capacity 13.4 tons, rated power 6MW

Focus on Primetals Technologies

For further information, log on to www.primetals.com

00 10

INNOVATIONS

Primetals’ 150th process optimisation order Primetals Technologies has announced its 150th order for the process optimisation of a continuous caster. The company claims that advanced automation packages and cyber-physical models are in high demand at present, especially in the USA and China. Automation and digitalisation of continuous casters, says Primetals, are increasingly being installed on third-party plants, and steelmakers are showing ‘growing interest’ in ‘digital twins’.

“Nowadays, advanced automation packages and cyber-physical models – so-called ‘digital twins’ – are ‘state-ofthe-art’ and form the basis for ongoing optimisation of production processes,” claims Primetals. The company’s head of the Linz, Austria-based electrical and automation division of Primetals, Dr. Peter Juza, commented: “In keeping with these demands, we have been continually developing our solutions in recent years, and can now offer our customers the most comprehensive portfolio for the automation, digitalisation and optimisation of continuous casters. The number of orders in this sector more than doubled during the last fiscal year,” he said.

According to Juza, “Our offer of a subscription license for a system for optimising continuous casting processes is arousing great interest in the world's steel industry.”

Primetals claims that new releases, upgrades and updates are installed after consultation with the customer, keeping the customer’s application up-to-date and allowing operators to benefit from ongoing developments, improvements and inventions.

Primetals says that the modularity of the system allows subscriptions to individual functions and models to be taken out or cancelled as required. Newly developed features and functions that become available over the course of time can also be implemented on request. Juza said that the increasing number of customer inquiries ‘convinces us that we have the right model for the future’.

In December last year, a US-based steel producer was the first company in the world to purchase a subscription license from Primetals Technologies for its continuous caster. Juza commented: ”The increasing number of customer inquiries convinces us that we have the

right model for the future.” ”Primetals Technologies combines many years of experience in the digitalisation of steel production processes with profound metallurgic know-how – which is a great advantage for our customers,” Juza said.

The company is particularly proud of its metallurgic and digitalisation knowhow, citing its Model Suite of modeling packages, which, it is claimed, is based on state-of-the-art technology for dynamic secondary cooling and soft reduction. It takes into account thermodynamic effects, such as shrinkage and phase transitions, thus making a significant contribution toward directly improving quality during the solidification phase. Plant operators, says Primetals, can have a digital twin of their continuous caster and the continuous casting process, enabling them to optimise production processes and the production of new grades of steel in the simulation and replay mode. What Primetals describes as the ‘Process Intelligent Cockpit’ process can be used for training continuous caster operators.

Focus on Primetals Technologies

FIVES SOLUTIONS, ENGINEERED FOR YOUR SUCCESS

Mastering technology to help steelmakers achieve performance excellence: — Strategic consulting and process optimization services — Ultra-low NOx reheating furnaces — Advanced cold rolling workshops for stainless and silicon steels — Complete strip processing lines to produce AHSS and UHSS, including: • Wet FlashCooling®, providing the cooling rate up to 1,000°C/s • CELES EcoTransFlux™, transverse flux induction heating up to 1,200°C — Eyeron™, real time quality qualification system — NeoKoil® Smart line, automatic line management Over 1 600 industrial installations worldwide, more than 150 years of process know-how and experience.

SEE FOR YOURSELF AT FUTURE STEEL FORUM SEPTEMBER 25-26, 2019 BUDAPEST, HUNGARY

Contact us: + 33 1 45 18 65 35 (France); + 1 234 217 9070 (USA) steel@fivesgroup.com; www.fivesgroup.com

00 12

INNOVATIONS

Steel Dynamics order state-of-the-art equipment

Steel Dynamics Inc of the USA has awarded SMS Group an order for the supply of a complete steel production line. Commissioning is scheduled for mid-2021.

The order is for equipment running from the steelworks to the CSP plant and from the cold rolling mill to the galvanising line at the company’s Sinton facility in Texas.

The 2.7Mt Sinton steel plant, claims SMS group, is setting new standards.

“In addition to the mechanical equipment from the liquid phase up to strip processing, the scope of supply…comprises X-Pact electrical and automation systems as well as technical support during installation and commissioning.

The Sinton works will be equipped with two direct-current electric arc furnaces (EAF) with a 190 ton capacity and two twin ladle furnaces (LF) as well as a double vacuum tank degasser. It has an annual liquid steel capacity of over 3Mt /yr.

According to SMS Group, the CSP plant will enable SDI to produce thin slabs with thicknesses of up to 5’2” Digital Edition - August 2019

(130mm) and slab widths of up to 84” (2,134mm).

The caster is a single-strand curved mold plant (VLB – vertical liquid bending) with a metallurgical length of over 82’ (25 metres) and a yield of 7.5 tons/ minute. Casting speeds of up to 19.7’ per minute (or six metres per minute) are attainable.

“With this new plant for SDI, SMS group will again make a statement for cost-efficient production paired with higher throughput and better quality,” said Cosimo Cecere, head of sales and project management, CSP Plants, SMS group.

A five-stand pickling line/tandem cold mill will follow the CSP plant. The process section of the pickling line/ tandem cold mill will be equipped with the latest turbulence technology and a 600-kN leveling unit. With the aid of a payoff reel upstream, the pickling line and tandem cold mill can be operated in parallel independent of each other.

This permits the hot strip to be pickled and oiled or directly guided to the tandem cold mill. The annual pickling

capacity will be 1.0Mt/yr.

The five-stand, six-high tandem cold mill will have a wide roll gap setting range to ensure excellent cold strip tolerances and strip flatness. The strip width is 78" (1,981mm) and will be able to roll the cold strip down to 0.08" (0.20mm).

Finally, a carousel reel will coil the rolled cold strip. In coupled mode with the continuous pickling line, the annual capacity will be 0.77Mt. The skin-pass mill for post-treatment of hot and cold strip with priority at cold strip skin-passing, is designed for an annual capacity of 0.4Mt. For cold and hot strip galvanising, the plant will be equipped with a continuous galvanising line including a horizontal Drever furnace heated by a direct-fired and a radiant-tube zone.

The line will be prepared for the later installation of an ultra-fast gas cooling section to allow for future production of high-strength steel. The scope of supply includes a DUMA-BANDZINK air knife system that will set the zinc layer thickness to ensure high surface quality.

INNOVATIONS

For further information, log on to www.sms-group.com

A change system with two zinc pots will permit the strips to be coated with a conventional zinc layer or an aluminium-zinc alloy.

For post-treatment, the line will be equipped with a four-high skin-pass mill stand, a tension leveller and two horizontal shuttle-roll coaters, as well as with an oiling machine in the exit section. The galvanising line will be able to process strips with a thickness

of up to 0.16" (4mm) and a width of up to 76" (1,930mm). It will have an annual capacity of 0.5Mt. The technological element of the plant revolves around the X-Pact electrical and automation systems, which assure that productivity, production flexibility and final product quality will meet actual and future requirements.

“The applied concept and equipment provide a solid foundation for the dig-

13

italisation of the plants and processes and for future-oriented extensions and applications. Diagnosis and visualisation concepts will be used in the entire plant. They assure and sustain effective and preventive maintenance activities,” explained SMS Group.

SDI will be selling the products in the rapidly expanding high-strength steel tubes and multi-phase steel markets for automotive and structural steels.

SMS group believes that with the new equipment, SDI will become the leader for this new type of minimill producing hot strip.

“In particular, it is the new, extremely high-performance continuous caster in combination with the proven thermo-mechanical rolling process that will allow the production of special steels in unprecedented dimensions,” said SMS group.

Mark D. Millett, CEO of Steel Dynamics commented: “The leading position in technology and the longstanding partnership between SDI and SMS group were the decision-making factors for awarding the order to SMS group”.

Burkhard Dahmen, CEO of SMS group, said that Steel Dynamics and SMS group have successfully completed numerous highlight projects in the past.

“In constructing the new plant complex in Sinton, Texas, we continue our fruitful and trusting partnership,” he said, adding that the new project will once again illustrate the position of SMS group as leading partner in the world of metals.

Superior high-temperature tags and labels for track and trace applications.

Polyonics® offers thermal transfer printable polyimide and aluminum bar code tag and label materials so you can accurately control inventories from when the materials leave the furnace through to delivery to your customers.

Keeping you on track www.steeltimesint.com

polyonics.com/HotMetal 603.352.1415 | info@polyonics.com

Personal Chemistry

Digital Edition - August 2019

00 14

INNOVATIONS

Danieli revamp Metra extrusion press

Danieli Breda (DB) is to revamp the largest extrusion press installed for Metra at Rodengo Saiano in Italy.

Originally supplied by DB as a 55MN short-stroke press, the line will be updated to operate with greater force at 75 MN, processing billets up to 2,100mm long.

The latest order follows on from a successful revamp at the plant to increase the extrusion force from 31

to 40 MN. In addition to the planned latest revamp, the line will be converted from the current back-loading design to a front-loading operation.

According to Danieli Breda, an extremely aggressive project follow-up and smart engineering solution to overcome a tight time schedule are crucial.

All the presses installed at the

Rodengo Saiano facility have been supplied or revamped by DB. There has been a long-standing partnership between DB and the Metra Group, beginning in the 1980s and the result of mutual confidence, engineering flexibility and a strong commitment from both sides.

For further information, log on to www.danieli.com

Best quality from Danieli Centro Maskin’s HiGrind techno Italian plant builder Danieli claims that two new billet grinders will ensure over 100kt/yr of throughput for Chinese steelmaker, Shagang Group, in full skin-pass mode. The equipment, which has not been delivered yet, can process 140- and 150-mm billets up to 16m long and will meet very high requirements in terms of surface roughness Digital Edition - August 2019

and overall quality.

In fact, the two grinders, including all handling equipment, hydraulics and secondary filtering equipment, are being manufactured at Danieli China in Changsu.

Shagang Group is a special steel manufacturer operating seven mills.

The first Danieli Centro Maskin grinder

was supplied to the business 10 years ago. The new installation, which should be in place by Spring 2020, will allow the company to increase the quality of its products in line with market demand.

According to Danieli, the HiGrind process is the most advanced system for controlling tolerance of the grinding

INNOVATIONS

15

Focus on Danieli

ology depth and the overall grinding quality, as well as ensuring high output, yield recovery and overall operational cost reduction.

For further information, log on to www.danieli.com www.steeltimesint.com

Digital Edition - August 2019

00 16

INNOVATIONS

Doosan to wow visitors at IMHX 2019 exhibition

Doosan claims that its stand at IMHX 2019 logistics exhibition (24-27 September, NEC, Birmingham, UK) is set to be one of the largest at the show, and that each zone – spanning manufacturing, construction, agriculture, transport, warehousing and aviation – will display Doosan’s best-in-class solutions specific to the sector. The machines on show will be kitted-out with attachments appropriate to typical industry applications, offering visitors the chance to see the optimum handling solutions for enhancing performance in each industry.

The very latest counterbalance and warehouse truck models will be available to see, from the new feature-rich B15R-7 three-wheel and B45X-7 fourwheel series of electric forklift trucks, right up to Doosan’s heavy-lifting diesel and gas models – including its giant 25 tonne capacity counterbalance forklift, the DV250S-7. However, there will be more than just counterbalance trucks to see.

In advance of the launch of its new range of warehouse trucks, Doosan will be offering a glimpse of what’s to come by debuting three new models in Digital Edition - August 2019

the walkie and stacker powered pallet truck range – the BPL18/20S-7 with lateral drive, the BPM20S-7 with middle drive, and the BDL13S-7(H) double stacker. “All three pedestrian operated trucks reflect Doosan’s renowned design characteristics of compact functional form, high maneuverability, durability and rugged, reliable performance,” the company claims.

Doosan’s new 1,500kg capacity electric pallet truck – the BPSC15E-7 lateral drive, will also be on show at IMHX. It features a built-in charger, long-life batteries and a compact chassis for extra maneuverability.

In the Future Zone, visitors to the Doosan stand will be given the opportunity to see the future of materials handling, with a large-screen video display highlighting some of the technologies Doosan is working on to drive productivity in the warehouse and in the yard. Automation and lithium-ion power are just two aspects of a journey of discovery that will highlight the scale and scope of Doosan as a $21 billion global engineering business, involved in everything from power plants, engines

and construction equipment to consumer goods.

In fact, one of Doosan’s lithium-ion powered trucks will be on display.

While electric power is gaining wider acceptance, diesel will remain a significant power source for heavy handling tasks for some time to come. The company will be showing, for the first time in the UK, its new Euro Stage 5 engine – designed

by Doosan to comply with stringent EU emission standards.

As a tough all-rounder, durability for outside duty is assured with a rugged rear-drive axle, IP65-rated Curtis controller and IP20 PAL brushless AC motors – offering maximum resilience to water and dust. It has a tight turning circle, smooth and responsive acceleration, and a conveniently placed direction switch on the hydraulic control lever for switching direction of travel quickly and easily.

Doosan’s new range of four-wheel electric counterbalance forklift trucks – the B45X-7 Series – will also be on display. The new model is claimed to offer greater durability, enhanced productivity, increased visibility and an extensive range of safety features as standard.

Brian Grady, sales and marketing director at Doosan Industrial Vehicle UK, commented: “This year at IMHX, we hope to demonstrate that Doosan has an established and rapidly expanding portfolio of materials handling equipment.” in just about every industrial sector.”

For further information, log on to www.doosan.com

Finite Network Method Precise simulation of EAF high current systems

Exclusively available at BSE !

Principles.

Concept.

Advantages.

● Detailed simulation of all electro-magnetic properties of EAF high current system.

Finite Network Method (FNM) makes realistic calculation feasible for the first time:

● Exact dimensioning of components and ultimate power input optimisation.

● Accurate calculation with a precision of less than 1 %; all other methods have error ranges >10 %.

● Local current density.

● Optimum layout of masts and roller guides for best power input.

● Simulation also for peripheral and metallic structures where eddy currents are induced. ● Precise unsymmetry calculation as the basis for good operational results. ● Optimisation of existing equipment and optimal design of new equipment.

Inhomogeneous current density distribution in the high current cables depicted in a cross-section.

Badische Stahl-Engineering GmbH Robert-Koch-Straße 13 D-77694 Kehl/Germany Phone (+49) 78 51/877- 0 Fax (+49) 78 51/877-133 eMail info@bse-kehl.de www.bse-kehl.de

● Exact short circuit impedance and unsymmetry; short circuit tests (dip tests) are not necessary any more. ● Measure the real arc voltages in real-time. ● Real static and dynamic forces and momenta acting on the mechanical system. ● Local magnetic field strength and shielding effects. ● Induced eddy currents.

● Optimal design of high current conductors and symmetrizing loops. ● Prevention of hot spots in the furnace. ● Balanced power input and improved regulating performance. ● Optimal material sizes, optimal cooling.

● Effects of ferromagnetic material like steel.

Simulation of graphite electrodes with clearly visible proximity- and skin-effect.

Very inhomogeneous current density distribution at the external delta closure of the transformer.

00 18

INNOVATIONS

Slovenian steel company embraces Industry 4.0 PSI Metals has been commissioned by the SIJ – the Slovenian Steel Group and its steel company SIJ Metal Ravne – to handle the implementation of the PSImetals modules Production, Quality, Order Dressing and Order Scheduling and Line Scheduling. PSImetals 5.18 will replace two legacy systems in the melt shop and downstream area.

Together with PSI, SIJ Metal Ravne is replacing its legacy manufacturing systems and creating a new digital manufacturing environment. The scope of the project includes sales order-based production planning, production and technological process control and quality management control as well as certification and production of stockyard management.

Vladimir Arshinov, CIO of the SIJ Group, said the project reflected the digital manufacturing strategy of the business. “It will allow us to complete

Digital Edition - August 2019

construction of a fully integrated digital manufacturing architecture on SIJ Metal Ravne: from CRM system to the ERP system including a new production management system which incorporates the technological management system of the melt shop." He said the company was considering further rolling out digital manufacturing architecture to other companies in the group.

SIJ Group expects the new solution to improve operational KPIs such as customer order performance, production cycle and stocks level, and to enable its second biggest steel plant, SIJ Metal Ravne, to enter new markets by fulfilling the highest certification demands.

SIJ’s Group IT team will participate in the project from the very beginning, starting with a know-how transfer by PSI experts in the nearby PSI office in Graz, Austria. This, it is claimed, will empower the SIJ Group IT team to

perform future configuration work and application maintenance. The project will start in September 2019 and is expected to go into operation within 18 months.

SIJ Group is the largest vertically integrated Slovenian steel producer and one of the largest producers of stainless and special steels in Europe. SIJ Metal Ravne from the SIJ Group is the third largest tool steel producer in the European Union.

Based on its own software products, the PSI Group develops and integrates complete solutions for optimising the flow of energy and materials for utilities (energy networks, energy trading, public transport) and industry (mining, metals production, automotive, mechanical engineering, logistics).

For further information, log on to www.psi.de

IMAGINE 4D PANTHER

® MASTERS OF

ULTIMATE PERFECTION Imagine you have an extraordinary surface inspection system for long products – powered by Automation W & R

KNOW-HOW FOR TOMORROW

4D PANTHER is a german trademark registred by KOCKS

www.kocks.de

00 20

INNOVATIONS

Zumbach’s BENDCHECK keeps straightness under control Manufacturers of steel tubes and bars are increasingly confronted with the fact that their products must meet higher quality requirements. Tight specifications for precise outer diameters have been established in the market for many years, but demands for product straightness are now increasing, says Zumbach of Switzerland.

In the past, claims Zumbach, a few off-line samples were used to analyse the entire production. Important process information was lost and a complete quality control across the entire production could not be guaranteed.

With Zumbach’s new BENDCHECK system, straightness, diameter and ‘ovality’ can be measured directly in the production process. With more measuring points, the head and tail bend of the product can be determined in addition to the straightness. Each pipe or bar is, therefore, completely recorded, measured and logged. Deviations are indicated in real time, so that the production process can be optimised quickly and profitably and unsatisfacto-

ry products can be sorted.

A customer-friendly display and control unit visualises all important and relevant data at a glance. Measured values, statistics, trend charts and process data can be displayed clearly and conveniently, the company claims.

As part of the fourth industrial revolution – otherwise known as Industry 4.0 – in-line measuring systems are taking over increasingly more important roles. The seamless acquisition and processing of measured data enables exchange with higher-level systems via communication protocols such as OPC UA and, therefore, making an important contribution to the digitalisation of production processes.

Zumbach’s BENDCHECK system is claimed to offer an industrial solution that records and makes available all the important data of a modern production.

For further information, log on to www.zumbach.com

ABB ArcSave for lead ABB’s ArcSave electromagnetic stirrer will be delivered in March 2020 and installed on an energy-efficient, 55-ton electric arc furnace (EAF) at the new voestalpine BÖHLER Edelstahl steel plant in Kapfenberg, Austria, helping optimise annual output of 205kt of high-performance steels from mid-2021.

According to ABB, ArcSave requires no contact with the bottom of the EAF, and enhances stirring during the melting of large scrap items, reducing stratification via forced convection. This improves EAF operation by homogenising temperature distribution and chemical composition, while speeding scrap and ferroalloy melting compared with natural convection alone.

ABB claims that ArcSave also contrib-

New rolling mill wid Kelk’s ACCUBAND C775-FF is claimed to provide steel mill operators with continuous, stereoscopic, non-contact material width measurements in cold rolling mills and process lines. The width gauge includes four CCD cameras, arranged in fixed pairs, to scan steel strips at a rate of 1,000 scans per second in order to accurately locate the position of the edge, regardless of pass line variations. The edge data is then combined with the camera separation distance, obtained during calibration, resulting in accurate strip width and centre line deviation, the company claims.

The compact design of the ACCUBAND C775-FF allows the width gauge to be installed in tight spaces while still providing one of the most accurate measurements in the industry (±0.4 mm at 2 sigma). In addition, integrated Digital Edition - August 2019

INNOVATIONS

21

ding Austrian steelmaker

utes to lowering environmental impact by reducing electricity usage, process additions such as alloys and lime, and consumables such as electrodes.

Anders Lehman, vice president of ABB Metallurgy Products, said that ArcSave was created to help manufacturers improve productivity while

prioritising safety, energy efficiency and product quality. “As the leading provider of electromagnetic stirring solutions, we are extremely proud to supply one of the world’s most advanced special steel plants,” Lehman said.

Plant construction SMS group is responsible for the engineering, process

technology and start-up of the fully automated melting system. SMS group specified ArcSave at the core of the Kapfenberg plant, which unites costand energy-efficiency with a high degree of automation and digitalisation.

ArcSave’s advanced, efficient technology is claimed to align perfectly with voestalpine’s ambition to be at the cutting edge of specialist steel production.

For further information log on to www.abb.com/metals

dth gauge from Kelk LED front light illumination provides long life, reliable performance, and minimal maintenance, says Kelk.

“For decades, KELK has deployed optical system technologies for hot rolling mills, with hundreds of installations around the globe,” said Hiro Kitagawa, vice president, sales and marketing, KELK. “Our design and engineering teams have now leveraged this intellectual property to develop highly-accurate and advanced sensors for cold rolling mills and processing lines.”

The new ACCUBAND C775-FF joins KELK’s current model ACCUBAND C765-F cold width gage and ACCUSPEED laser velocimeters, to provide additional solutions for cold rolling mills and process lines. Movable cameras on model C765-F automatically position themselves over the edges of the strip, based on the nominal width input, to www.steeltimesint.com

provide accurate width measurements, whereas cameras on the new model C775-FF are pre-positioned, making the system more compact, to fit into tight spaces, and easy to maintain, by eliminating moving mechanical parts.

In addition to cold strip mills and process lines (including galvanising and

pickling lines), other specific applications for the ACCUBAND C775-FF include tension levellers, side trimmers (entry and exit), tandem mill exits, temper mills, and slitters/rewinders.

For further information log on to www.kelk.com

Digital Edition - August 2019

00 22

INNOVATIONS

ExxonMobil introduces a special oil f

Digital Edition - August 2019

INNOVATIONS

for steel mills

23

ExxonMobil has introduced Mobil SHC Elite, a synthetic gear and bearing circulating oil for steel mills. The lubricant’s advanced formulation can help protect equipment operating at extreme temperatures, while increasing energy efficiency and extending oil drain intervals.

According to the company, extensive testing shows that Mobil SHC Elite can deliver double the oil life of traditional synthetic products, and up to 12 times the oil life of mineral products.

It can also help protect equipment,operating at temperatures up to 150°C in intermittent service, meaning it is ideally suited for a range of steel mill applications, such as helping to protect hot mill gear drives, blast furnace distributor gearboxes and furnace tilt gearboxes.

ExxonMobil claims that the advanced synthetic lubricant is specifically formulated to deliver excellent wear protection and oxidation resistance without any of the compatibility challenges associated with glycol-based products used in high temperature applications.

In addition, Mobil SHC Elite has demonstrated an enhanced torque ratio, enabling it to deliver a 3.6% energy efficiency improvement when compared with conventional mineral oils, says ExxonMobil.

“Protecting core processing equipment from high in-service temperatures is among the key challenges facing steel mill operators,” said Emre Noyan, industrial marketing manager at ExxonMobil. “Mobil SHC Elite’s extended oil life, wide temperature range performance and energy efficiency improvement can help operators increase uptime and cut costs – giving them a competitive edge.”

Mobil SHC Elite has received approval for its proven performance from leading gearbox manufacturers including Siemens, whose FLENDER gear units depend on effective lubrication to ensure reliability in intense operating environments.

For further information, log on to www.mobil.com/industrial

www.steeltimesint.com

Digital Edition - August 2019

00 24

INNOVATIONS

Auto separator system makes life easier, says Georg

Heinrich Georg claims that setting up the separator shafts on slitting lines has, up to now, been extremely time-consuming.

The company claims that it was necessary, for every slitting programme, to remove the conventional separator shafts, fit the necessary separator disks and space rings manually, and then replace the shaft. The result was unproductive downtimes during separator-shaft changing, high labour costs

Russian steelmaker Severstal has placed an order with the French plant builder Fives for two new high-capacity walking beam slab reheating furnaces.

The order is part of a major reconstruction of its No2 rolling plant at the Cherepovets Steel Mill in the north west of Russia.

The new Stein Digit@l furnace – each with a production capacity up to 440 tons/hr hour – will replace two of the steelmaker’s four existing furnaces serving the hot strip mill 2000 of rolling plant No2.

According to Fives, the Stein Digit@l furnace is ‘a crown jewel within the range of Fives’ technologies for reheating and heat treatment’.

Fives claims that its Stein Digit@l furnaces offer a consistent flame shape Digital Edition - August 2019

and significant space requirements for the changing table.

But the problem appears to be solvable with Georg’s patented automatic separator system, which moves each individual separator disk automatically to the right position, relying on a procedure that takes seconds and is claimed to ‘significantly enhance the productivity of slitting lines’.

According to Georg, at programme

changes, the shaft control system receives the new position data from the higherlevel distributed control system (DCS) and positions the separator disks accordingly.

Georg says that the system is set up for the various programmes by means of a user-friendly man-machine interface on the main operating panel or on a separate control desk. For maintenance purposes, the shafts can be replaced in less than

Two Fives furnaces for Russian steelmaker

INNOVATIONS

five minutes, it is claimed, using a simple cassette change. The separator cassette can be inspected away from the production line.

The individual disks are moved by electric motors to their position on the shaft and are locked there pneumatically.

The drive system is located inside the shaft for protection, but also because of

the compact design. Georg says that only minimal modifications are necessary even for retrofitting into an existing line.

Antonio Garcia, head of Georg’s strip lines division, says the automatic separator system is important for boosting the efficiency of the slitting lines. “Push-button setting of the separator disks has proven its value in a number of European mills,

and heating quality, ultra-low NOx, substantial fuel saving, minimised scale loss and decarburisation, and reduced OPEX due to lower maintenance and fuel consumption.

The two furnaces for Severstal will be able to reheat a wide range of products within a charging temperature mix from half cold 20°C to half hot 500°C, in line with the hot rolling mill requirements at Severstal.

“One of the unique features of the furnace is the use of AdvanTek side firing burners with digital control. This creates so-called “micro zones”, each 2.8 metres long, which enable the furnace temperature to be controlled in a very precise manner, guaranteeing high heating efficiency,” says Dieudonné Hounliasso, sales engineer at Fives

Stein, the Fives subsidiary that developed the furnace.

The combustion system is controlled

by Virtuo-R, Fives’ proprietary Level 2 software tool for optimum thermal control. It is claimed that Virtuo-R helps to minimise heating defects, optimise operation time required for heating and substantially reduce gas consumption and scale losses. As a result, the Stein Digit@l furnace meets the most stringent environmental requirements by providing NOx emissions of less than 60 ppm for dry waste gases.

“Today, this equipment offers the best return on investment on the market for demanding slab reheating applications,” summarises Kristiaan Van Teutem, sales director of Fives Stein.

According to Vadim Germanov, chief

www.steeltimesint.com

25

and significantly increased the productivity of those lines. I am confident that our presence at the Steel Summit 2019 will also enable us to convince our American customers,” he said.

For further information, log on to www.georg.com

executive officer of Severstal’s Russian Steel Division, the new reheating furnaces will ensure homogeneous heating of slabs which is imperative to produce high value-added products.

“Moreover, we will be able to reduce gas consumption and increase the capacity of the hot strip mill 2000 by 140,000 tons per year, which will bring a financial benefit of around €20 million per year,” he said.

Fives has recently supplied slab reheating furnaces to JSW’s Dolvi Works in India, ATI Flat Rolled Products in the USA, Colakoğlu in Turkey and OMK, Vyksa Steel Works in Russia.

For further information, log on to www.fivesgroup.com Digital Edition - August 2019

00 26

INNOVATIONS

US Steel KoĹĄice places order with Tenova

For further informa

to www.tenova.com

Digital Edition - August 2019

INNOVATIONS

LOI Thermprocess

ation, log on

m

www.steeltimesint.com

27

Slovakian steelmaker US Steel Košice has placed a order for the supply of an annealing and coating line (ACL) for dynamo steel strips to Tenova LOI Thermprocess, a leading global supplier of heat treatment lines and furnaces.

According to Tenova, the plant bears the internal name of Dynamo-Line no.

4 and will meet the highest requirements for the production of non-grain oriented electrical steel. The contract includes the engineering, to a large extent the turnkey delivery of all equipment as well as the supervision of the assembly and commissioning, including training.

Tenova, part of the Techint Group, designed and delivered the entire line and was chosen by US Steel Kosice for several reasons: The multiplicity of references for comparable facilities, the short realisation time and the mathematical model were of crucial importance.

Techint group companies – Tenova LOI Thermprocess and Tenova Italimpianti, supplied the complete annealing and coating line (ACL) including the furnace, the entire strip handling and chemical processing with drying oven following the coil coating, the associated electrical, instrumentation and control technology including the automation system.

US Steel Košice is a wholly owned subsidiary of US Steel Group headquartered in Pittsburgh in the USA. It is a fully integrated metallurgical plant and one of the largest industrial companies in the Slovak Republic.

In addition to the production of thin sheet, ultra-fine sheet and galvanised strip, non-grain-oriented electrical steel sheets have been produced for more than 30 years. The company is undergoing extensive modernisation in order to improve product quality. One of the measures is the dismantling of two outdated annealing and coating lines and at the same time the construction of the new annealing and coating plant.

Between 2000-2019, Tenova received a total of 60 orders for heat treatment systems for electrical steel, 34 of which are completely new installations.

Digital Edition - August 2019

00 28

INNOVATIONS

on s u c Fo ova Ten

Chinese order Consteel from Tenova Chinese steelmaker Sichuan Guanghan Desheng Iron & Steel has awarded Techint company Tenova the contract to supply one of its Consteel electric arc furnaces, representing the 14th state-of-the-art EAF Consteel Evolution projects realised in China over the past 18 months.

According to Tenova, the shift from BF-BOF (integrated steelmaking) towards EAF technology, sustained by the reforms of the Chinese government, has led to the demand for advanced equipment and technologies that increase production and provide environmentally friendly solutions.

“Once again our technologies were recognised by the client with whom, from the first technical discussion, the co-operation was fruitful and satisfying,” said Mario Marcozzi, sales Digital Edition - August 2019

director, Tenova Metals. “Our previous reference plays a fundamental role in building this environment of trust. This is in line with Tenova’s top priority: to be a reliable partner and to always improve our service.”

Tenova claims that its Consteel Evolution offers high quality standards as well as a more environmentally friendly approach to production compared to traditional steelmaking methods. The company boasts 75 references worldwide and claims that the Consteel system is considered the best available proven technology in terms of balance/ mix between innovation, reliability and sustainability.

For further information, log on to www.consteel.com

INNOVATIONS

29

Tenova highlighed sustainability and digitalisation at METEC 2019 Italian plant builder Tenova, a Techint company, was using its presence at METEC 2019 as ‘a window onto a green and more efficient way of producing metals’.

Roberto Pancaldi, CEO of Tenova Metals, commented: “Our motto for this edition is ‘the smart way to metals, which refers to the path that Tenova has been tracing over the last years towards a more sustainable and digitalised steel production. This approach is underpinned by a consistent strategy aimed at changing the business model along the entire value chain. An approach we want to put forward and continue to develop in tight collaboration with our clients, through tai-

www.steeltimesint.com

lor-made solutions,” he said.

In Düsseldorf, Tenova was talking about Pomini Digital Texturing and its Self Regenerative Burner, the former being ‘a unique texturing machine that can be discovered through augmented reality’ and the latter a ‘smart burner able to optimise the combustion process through the application of digital technology’.

Visitors to the Tenova stand took advantage of the company’s expertise in the so-called Tenova Arena where ‘core technologies’ and ‘cutting edge solutions’ were explained.

Where sustainable steelmaking is concerned, Tenova used METEC to promote a new technology aimed at

replacing carbon partially or totally with hydrogen as the reducing agent in the direct reduction (DRI) process. The technology has already been applied to German steelmaker Salzgitter’s SALCOS low CO2 steelmaking project as well as HYBRIT, a fossil-free steelmaking process developed by Hybrit Development AB, a joint venture between Swedish steelmaker SSAB, industrial minerals company LKAB, and power generator Vattenfall.

Lastly, Tenova presented the technology behind its twin-chamber melting furnace (TGF).

For further information, log on to www.tenova.com

Digital Edition - August 2019

00 30

INNOVATIONS

Fall prevention is crucial, says Plibrico USA-based Plibrico, an industry-leading single-source supplier of aluminosilicate and high alumina monolithic refractories used in the processing of steel and aluminium, says that most equipment used for thermal processing stands well over 10 feet tall and has the capacity to hold or process over 60 tons of molten metal. During refractory installation, repair and maintenance of such large equipment, refractory professionals often find themselves raised atop platforms, scaffolding, decking and work stations. Due to the fact that refractory employees regularly work at elevated heights, it is crucial to keep them safe from fall-related injuries, as well as to ensure the job site is free of safety violations. To accomplish this goal, it is essential to understand the hazards of falls and know the Occupational Safety and Health Administration (OSHA) rules.

According to OSHA, in 2017, almost 42% of all construction worker-related deaths were attributed to falls.

Thousands more were injured. Fall Protection infractions (OSHA 29 CFR 1926.501) also topped OSHA's 2018 list of the Top 10 safety violations for the eighth consecutive year.

Incidents involving falls frequently involve a variety of factors; however, a common thread running through most is the absence of fall protection

Digital Edition - August 2019

equipment. Even if you're Nik Wallenda, the high wire aerialist of the famed Flying Wallendas family, OSHA requires protection when working on refractories at heights of six feet above a lower level: • Handrails, guardrails and toeboards: serve as barriers between the employee and an open edge. Midrails or screens need to be installed between the top of the guardrail and the walking or working surface to prevent falls.

• Personal Fall Arrest Systems: provide employees with an individual form of fall protection. For example, a body harness connected to a lanyard or retractable line secured to a fixed anchor. These types of systems are designed to go into action before contact with any lower level.

• Personal Fall Restraint Systems: prevent employees from reaching the edge where a fall hazard is likely to occur. It tethers a worker in a manner that will not allow a fall of any distance. This system is comprised of a body belt or body harness, an anchorage, connectors, and other necessary equipment.

As a second line of defence or where fall prevention systems are not practical, for instance roof work, a warning line system consisting of ropes, wires, or chains is an approved solution if it is at least six feet from open edges

around all sides of the work area.

Fixed barriers can also be installed to prevent employee access to dangerous areas.

To address any hazardous areas that may have floor openings, colour-coded covers should be used and marked with the word "Hole". Covers should be secured tight to prevent workers from falling through floors or elevated areas.

OSHA clearly states employer requirements. OSHA mandates that employers train workers on how to use personal fall protection equipment and how to work in hazardous situations. Employers must also assess the workplace to determine if walking or working surfaces have the necessary strength and structural integrity to safely support workers.

Before any work begins, conduct a hazard assessment to develop a comprehensive fall protection plan, to manage hazards and focus employee attention on prevention. Falls cause deaths and numerous serious injuries each year, many of which are preventable. Maintain the highest safety standards on your job site by installing or using fall protection systems – not all of us can be as sure footed as Nik Wallenda.

For further information, log on to www.plibrico.com

INNOVATIONS

www.steeltimesint.com

31

Digital Edition - August 2019

00 32

INNOVATIONS

Praise for SSAB’s Strenx and Hardox For half a century, trailer manufacturer Acoplados Salto has been known for the quality of its products. Seven years ago, it started to use SSAB’s steel for the trailers, and the move, claims SSAB, has been a huge success.

“Acoplados Salto uses hot-rolled and cold-rolled Strenx performance steel and Hardox wear plates in the trailers,” says Sergio Andriano, plant manager at Acoplados Salto. “As a result of using SSAB’s lightweight steel we have gained new clients who are enthusiastic about the material. We have increased our sales by 30%.”

Among the products the company makes there are semitrailers, curtain side trailers, hopper wagons and flatbed trailers. Compared with a few years ago, customers can carry much heavier loads, which increases productivity and efficiency – the lighter the trailer is, the more products it can carry.

“The lightweight material from

SSAB has enabled us to reduce the weight of our units by 1,800 to 2,000 kilograms (4,000 to 4,400 pounds), depending on the product. This has translated into a greater load capacity for our clients,” Andriano says.

Juan Manuel Pierce, head of the design department at Acoplados Salto, has been part of developing the new lightweight products. He says one of the main advantages of using SSAB’s materials is their reduced weight and high level of safety. Even though the material is lightweight, it is robust and does not crack when in contact with hard surfaces or harsh conditions, making them safer for heavier loads.

“The use of Strenx 700MC and 700CR as well as Hardox 450 has helped us achieve a product that is highly resistant to wear and tear while at the same time reducing thickness,” Pierce says. “The best properties of the SSAB steel enable us to improve the design of our products.”

For further information, log on to www.ssab.com

Digital Edition - August 2019

Since December 2017, Acoplados Salto has been a member of the My Inner Strenx programme.

“Through My Inner Strenx we have integrated new knowledge about the SSAB materials to constantly improve our production,” Pierce says. “The help and support we receive from SSAB includes technical advice on how to use the material, soldering, welding, cutting it and bending it.”

Andriano also sees another advantage to being a member of My Inner Strenx.

“Having the My Inner Strenx logo on our units represents quality products. It’s something the client values, knowing that the unit is made with SSAB lightweight material. They know it is quality,” Andriano says.

INNOVATIONS

www.steeltimesint.com

33

Digital Edition - August 2019

34

EUROCOKE SUMMIT 2019

Will coke be needed in 2030?

Amsterdam was the venue for the EUROCOKE Summit 2019, organised by Smithers Apex. The event offered delegates a comprehensive conference programme and exhibition and was preceded by a site visit of Tata Steel Ijmuiden and a pre-conference workshop on blending for cokemaking and PCI. A wide range of issues related to all aspects of the cokemaking industry was discussed IN his opening comments to delegates at the Eurocoke Summit 2019, held in a Novotel hotel in Amsterdam, Jorge Caldeira, a consultant working with Coalbiz, told delegates that interest in the conference this year had been at an all-time high with 275 delegates registered.

Caldiera then invited Tata Steel Ijmuiden’s Johan Van Boggelen, technology manager, Hlsarna Pilot Plant, to approach the lectern.

In many ways, Mr Van Boggelen was risking life and limb as he told delegates that the Hlsarna smelting reduction ironmaking process was a coal-based technology that did not require coke. By using biomass and scrap, a reduction in CO2 of over 50% can be achieved.

According to Van Boggelen, the target with Hlsarna is to be 'sustainable in every way by 2050. He admitted it was an ambitous target, but said that Tata Steel was determined to work towards it.

Hlsarna now has the support of the Dutch Government and is currently focused on the operational and equipment aspects of the process, testing equipment endurance and demonstrating long-term process stability.

Van Boggelen described Hlsarna as a Digital Edition - August 2019

‘breakthrough technology’ that offered significant environmental and economic benefits to become the ‘enabler for a sustainable global steel industry’.

Elevated prices – again Van Boggelen was followed by Eurocoke stalwart Jim Truman, principal analyst, metallurgical coal for Wood Mackenzie.

Truman was there to present a global metallurgical coal market outlook and concluded that the industry can expect ‘another year of elevated prices’. He said that Chinese macro steel demand indicators have been weak, but said that the government there was moving to stabilise the property market and boost infrastructure spending, which could mean demand growth in the region of 3%.

However, he argued that import restrictions will continue when coal demand looks weak, and said that a widespread ban on imports was not on the cards because China needs high-quality imported coals. In fact, the shortage of premium coking coals will keep domestic prices high in H1.

Truman said that India would have another year of ‘good growth’ with additional import demand in the region

of 4Mt. PCI, he said, would be especially strong and private companies will drive demand for imported coal in the near term.

According to Truman, Australian supply has struggled, but has potential during 2019, although progress is slow. The US will probably move forward with new projects, he predicted, adding that exports have grown to around the 50Mt level, but are constrained by rail capacity. New Orleans could serve as an outlet for more tonnage as it did in 2010/2011 following flooding in Queensland, Australia.

Doubling production Corsa Coal, a growth-oriented premium quality metallurgical coal producer based in Canonsburg, Pennsylvannia, sold an estimated 2.0 to 2.4Mt in 2019 and has witnessed 190% sales growth since 2016. The company’s vice president for international markets, Frederick Cushmore, said the company hopes to double ‘company-produced’ tonnes over the next three years and has customers in the USA, Asia and South America.

The company has opened three mines in the last 18 months – Acosta, Horning and Schrock Run – and has two other mines www.steeltimesint.com

EUROCOKE SUMMIT 2019

1

2

3

4

1. Jorge Caldeira of Coalbiz 2. Johan Van Boggelen of Tata Steel 3. Jim Truman of Wood Mackenzie 4. Jason Halko of Teck Metal

www.steeltimesint.com

ready to begin development.

Corsa’s Fred Cushmore said that steel production is set to grow over the next decade as urbanisation steps up a gear.

Coking coal demand is forecast to increase steadily, he added, and explained how demand will be met by a slow expansion of production globally. Furthermore, producers and investors are monitoring demand and the profitability of coal producers – and the latter are demanding an immediate return. He said that, despite profitability, it will be very difficult to raise capital and that supply growth will be very limited. Prices will remain stable for hard coking coal, he added.

Jason Fannin, senior vice president, Contura Energy looked at the impact of environmental restrictions in China on the seaborne metallurgical coal trade. Contura Energy is a Tennessee-based coal supplier with affiliate mining operations across major coal basins in Pennsylvania, Virginia and West Virginia and customers across the globe.

Interesting facts and figures Fannin’s presentation was full of interesting facts, like when a tonne of steel is exported from China it displaces a tonne of steel produced outside of China. He said it was tough to tell whether the displaced steel was from an integrated plant or an EAF-based producer. Fannin said that iron production in China drives coking coal consumption and that the country’s environmental push has knock-down effects for coking coal. And speaking of integrated mills, he believes that combining new technology with Carbon Capture and Storage (CCS) may be the only way to have a real impact on emissions. He spoke of a kinder, gentler steel industry going forward and claimed that the Hybrit initiative was not as far progressed as Tata Steel’s Hlsarna.

ArcelorMittal, he said, was adopting a multi-faceted approach to emissions.

Moving from China to India and South East Asia and an excellent presentation by Christopher Urzaa, marketing manager for coal at the Jellinbah Group. At present, he said, India produces 106Mt/yr of steel but it hopes to be producing 246Mt by 2030. There are, however, hindrances, including land, which is a perennial issue because every farmer has a say in India’s democracy. He said that iron ore, while high quality, is hampered by corruption and permitting. Throw in the fact that there

35

are 22 languages ‘enshrined in the Indian Constitution’, not including English and you can see that it isn’t all plain sailing.

Urzaa claimed that India will overtake China in terms of population by the year 2025 and that China’s current population is 1.397 billion compared to India’s 1.309 billion (in 2015). The global population is roughly 7.383 billion and one in six people on the planet is Indian, compared to one in three people being Chinese. India’s population will peak at 1.68 billion around 2060. Comparing average GDP, in India in 2017 the figure was US$6,427 (36% of the world average) while in China it was US$15,309 (in 2017) which is 86% of the world average. India expects GDP growth to be at 6% for three years. Where urbanisation is concerned, the 2015 figure for India was 32.8% and it is expected to be 40.1% by 2030. The figures for China are 55.5% and 70.6% respectively. As for steel use per person, in India it’s 75.3kg while in China it is 544.5kg (2017 figures). Out of India’s 2015 population figure of 1,309 billion, 498 million are under 20 years of age. By 2030, says Urzaa, the figure will be 1.513 billion and 475 million respectively.

Urzaa also looked at Indian steel companies and their projected tonnage going forward. JSW will be producing around 45Mt/yr (board approved) by 2025 and aspires to produce 60Mt/yr by 2030.

Tata Steel is currently producing 15Mt/yr and is expected to produce 22.5Mt/yr by 2020 and 28Mt/yr by 2025. India’s public mills are forecast to produce 23.5Mt by 2023, says Urzaa.

In Vietnam, Urzaa focused on Formosa Plastics Vietnam, which operates a 7Mt/yr integrated mill. Plans are in the pipeline to increase its capacity to 22Mt/yr. Hoa Phat, also in Vietnam, is currently producing 2.8Mt/yr of steel of which approximately 1.1Mt/yr is pig iron. By 2020, it is forecast to produce 8Mt/yr of steel.

Up until 2014, China was importing around 50% of laterite nickel ore from Indonesia, until the Indonesians initiated a ban on raw ore exports. As a result, China has set up plants utilising RKEF (rotary kiln electric furnaces) technology to smelt nickel laterite ore in order to produce nickel pig iron (NPI). Some small producers use very small (45m3 to 83m3) blast furnaces to produce NPI. Urzaa mentioned the Morowali Industrial Park, which is located in the Central Sulawesi Province of Indonesia and is Digital Edition - August 2019

36

EUROCOKE SUMMIT 2019

the main benefits of ‘smart solutions’ are: faster reaction times, better control over operations and maintenance, increased flexibility, increased availability of equipment and plant and the ability to ‘put the machine in the palm of your hand’. He then examined camera-based positioning systems, arguing that they offer simplicity and precision and can replace complicated instrumentation. Furthermore, they are retro-fitted on existing coke oven machinery and are the basis for implementation of maintenance activities.

dominated by Tsingshan Steel and PT Dexin, soon to be commissioned. PT Dexin is a joint venture with Tsingshan Steel (43%) and Jiansu Delong Holdings (45%). Dexin produces 3.5Mt/yr of steel and 1.3Mt/yr of coke; it uses 1.6Mt/yr of coking coal and 0.6Mt/yr of PCI. Looking at India and South East Asia in terms of coking coal demand, India imported 49Mt in 2018 and is forecast to import 97Mt in 2030. In 2018 India imported 8Mt of PCI (Pulverised Coal Injection). By comparision, South East Asia imported just 8Mt of coking coal, rising to 26Mt in 2030. Where PCI is concerned, last year South East Asia imported 2Mt, rising to 7Mt in 2030. The region hopes to be importing 26Mt of coking coal in 2030.

Ternium’s blending philosophies After lunch there were three interesting presentations focused on plant operations, materials and new technologies. Ternium’s Ana Patricia Jaccard examined the blending philosophies of Ternium in Argentina and Brazil, arguing that integrated steel mills require high quality control when it comes to raw materials. She said that met coke plays a key role in blast furnace performance and has a significant impact on the cost of the final product due to the costs of imported coals that make up the coal blends carried out at the coking plants.

Defining coal blend profiles, therefore, was a very important aspect of the business strategy.

Jaccard explained that both Ternium Argentina and Ternium Brazil have their own independent blending philosophies, which vary depending on ‘their processes and particularities’. Both, she said, have reduced coal blend costs. She said that while both plants made notable savings, but costs rise and for this reason she argued that there must be fast and transparent communication between members of the raw materials committees; respect for the technologies and particularities of each plant; investment in the training and development of specialists and in research and development and, lastly, leadership support and involvement.

Mathias Hoffman of Paul Wurth, part of the SMS group, discussed smart solutions for cokemaking machines, examining smart operation and maintenance utilising Paul Wurth’s XpertCloud technologies, including AIXpert for data processing and new business models. For Hoffman, Digital Edition - August 2019

5

6

7

5. Oana Niculita, John M Henderson 6. Brian Ricketts from Eurocoal 7. Ana Patricia Jaccard from Ternium

Predicting coke strength Jason Halko, Teck Metal’s global manager of technical marketing presented a paper on the effect of mineral type and size on coke quality. The research was conducted in order to help develop better CSR prediction models and find a less expensive way of obtaining coking properties other than through exploration. It was also to estimate the net present value of potential reserves.

Halko claimed that current models used to predict CSR (coke strength after reaction) are not always accurate and went into a lot of background work and past research on the subject, which found (in 1994) that mineral type is an important influence on CSR. Studies conducted between 2013 and 2015 found that in situ coal ‘may consist of minerals with different size fractions.

One of the most interesting (and relevant) papers at the conference was Oana Niculita’s Manless Coke Oven Machines.

Ms Niculita, head of sales and marketing at UK-based John M Henderson Machines, a regular EuroCoke attendee, looked at the human aspect of coke making, the health and safety challenges; and the pre-requisites, requirements, components and scope of the works for manless coke making. She offered up case studies of relevant projects and assessed the overall impact of manless operations.

In one case study of a South Korean coke plant, Niculita highlighted savings after automation of $3.3 million/year and manpower being reduced from 40 to 14 people, reducing costs by $3.12 million/ year. The operation cycle of the facility was reduced from 14.36 minutes to 11.38 minutes, a saving of $1.47 million/year.

The future of coal-based research Brian Ricketts, secretary-general of Eurocoal (the European Association for coal and lignite) looked at the future of coal-related www.steeltimesint.com

38

EUROCOKE SUMMIT 2019

EUROCOKE SUMMIT 2019 – FACTS AND FIGURES • More than 38% of global electricity came from coal in 2017 (21% in the EU).

Source: Brian Ricketts, Eurocoal.

• Steel use per person in India is 75.3kg per person (2017).

Source: Christopher Urzaa, Jellinbah Group.

• Global seaborne demand for metallurgical coal will rise from 310Mt in 2018 to 410Mt by 2040.

Source: Jim Truman, Wood Mackenzie.

• The Hlsarna smelting reduction ironmaking process can achieve at least a 20% reduction in CO2 emission intensity without Carbon Capture & Storage (over 80% with CCS).

Source: Johan van Boggelen, Tata Steel.

• 325 former coke oven workers have formed a group litigation to seek compensation for cancers and respiratory disease linked to breathing in fumes from coke ovens between the 1940s and 1980s.

research in the EU in the light of EU 2050 policy objectives.

According to Ricketts, over 38% of global electricity came from coal in 2017 with natural gas in second place (23.2%) and hydro power third (15.9%). He said that coal use was forecast to be flat to 2040 and that coal’s share in EU power generation between 1990 and 2017 dropped from 40% to just over 20%.

Source: UK High Court.

• Workers at coking plants and people who live near have a high risk of exposure to coke-oven emissions and were found to be at an excess risk of mortality from cancer of the lungs, bronchus, trachea and prostate.

Source: JMH (John M Henderson).

• Coal is used by 74% of steel manufacturers globally.

Source: BetterCoal.

• The USA accounted for 38% of 2018 seaborne coking coal supply growth.

Source: Xcoal Energy & Resources.

• Columbia is the only Latin American country with reserves of excellent quality metallurgical coal. Only 25% of the total territory is already explored.

Source: Juan Manuel Sanchez, Carbocoque.

• By 2030, India will have a mixture of large slot and stamp charged batteries.

Source: Neil Bristow, H&W Worldwide.

EU Emissions Trading System Ricketts argued that that one of the biggest threats to coal use in Europe was the EU Emissions Trading System (ETS), claiming that its Market Stability Reserve will mean 1.2 billion fewer allowances by 2023. He also talked about the Clean Energy for All Europeans initiative, which is basically a new energy ‘rule book’ designed to facilitate the transition away from fossil fuels towards cleaner energy in line with the EU’s Paris Agreement commitments for reducing greenhouse gases, or GHGs. An emissions performance standard of 550gCO2/kWh excludes all coal and lignitefired power plants from capacity markets and was campaigned against by Eurocoal.

Eurocoal is an active partner in The Coal Platform’s technology working group.

The Coal Platform is an adjunct to the European Commission’s ‘Clean Energy for All Europeans’ initiative and is spearheaded

by CoalTech 2051, a two-pronged strategic research agenda for coal R&D focused on research activities that support EU policy objectives as well as ‘research activities that promote the necessary global response’ to the ongoing use of coal outside of the EU.

Clean coal technologies for lower emissions were highlighted. The 750MW Lunen power plant in Germany, for example, is claimed to be one of the most efficient coal-fired facilities in the world while the Lagisza Power Plant in Poland featurers a 460MWe Amec Foster Wheeler supercritical, vertical tube CFB boiler.

Ricketts said that power plant efficiencies of up to 50% will be possible in future and there will be potential to reduce global emissions by 3 GtCO2/yr. He said that Carbon Capture and Storage (CCS) works and will become cheaper going forward and highlighted the CCS installation at the Boundary Dam coal power plant in Canada, in operation since 2014.

According to Ricketts, the plant captures around 1Mt of CO2 annually. He said that CCS plants are capable of capturing up to 97% of the CO2 in flue gases and second generation plants will be 67% less expensive per ton of CO2.

Bettercoal A similar theme was developed by AnneClaire Howard, executive director of Bettercoal, an organisation established by a group of major coal buyers to promote the continuous improvement of sustainability performance in their coal supply chain. She said that between 2016 and early 2019, Bettercoal assessed the performance of 20 coal companies in nine countries. Thirty mine sites were visited and the assessment covered close to 600Mt of thermal and met coal. In 2018, Bettercoal members purchased over 70% of their coal from Bettercoal-assessed suppliers. www.steeltimesint.com

EUROCOKE SUMMIT 2019

39

Rising environmental challenges Another EuroCoke stalwart, Neil Bristow of H&W Worldwide Consulting, is the go-to person for anybody who wants a well-rounded snapshot of the global coking coal market. And Bristow never fails to deliver, this time giving an interesting account of the challenges facing the industry to 2030 and asking the million dollar question: Will there still be a need for coke in 2030? Well, it’s not all bad news, but there’s a fair bit of it around. There are rising environmental challenges, the green lobby regard coal as bad and they fail to distinguish between coal types. Then there’s economics. Can aging batteries go on forever? And what about the costs of integrated steelmaking and the sustainability, quality and costs of raw materials? And that’s not all: What about staffing issues? Does anybody still want to be an engineer and if not, can artificial intelligence come to the rescue?

Coal kills?

He ran through some quotes from ‘the green world’ highlighting the attitude towards coal, stuff like ‘coal kills’ and ‘we will ban coal and close down this polluting and toxic industry.” Even the Economist, said Bristow, once described coal as ‘environmental enemy number one’.

Bristow pointed to the continual push towards zero-carbon steelmaking, asking the big questions, such as ‘is there a replacement for steel?’ and ‘how can steel recyclability be optimised?’ Can we ever get to 100% recycling of steel? Is steel’s image irreversibly tarnished?

He then sung the praises of steel as the key material for economic growth in global economies and how new steels are constantly under development. Steel, he said, was a highly versatile, economic and flexible material.

LEADERS IN CUSTOMISED HANDLING SOLUTIONS

The future of steelmaking?

He posed the question: Does the electric arc furnace represent the steelmaking of the future? And was the EAF a threat to blast furnace steelmaking? Will the use of more scrap metal kill off the need for coke? And then some more answers: it can be used in EAF and BOF steelmaking, emerging markets are ‘scrap short’ and the key to scrap availability is sustained, and particularly high quality, steel production and, of course, quality scrap. Scrap, he said, was not a ‘coke killer’, certainly not in the short term. Environmental concerns, however, will increase to 2030.

Bristow compared India and China, claiming that large coal mining in the latter would be well in decline by 2030 and that coal in India was predominantly lower quality and of the thermal variety. And while today China is a major coke exporter and India a major importer, by 2030 India could be a potential exporter.

Takeaway thoughts He offered some ‘takeaway thoughts’ in the form of more questions: What does a world of plentiful, cheap energy look like?’ What happens to the met coal market if thermal coal is banned?’ ‘Will NIMBY-ism mean the end of open-cut mining?’

Will 2030 witness fully automated mines run by artificial intelligence and what will be the cost and staffing implications?

Answers on a postcard to the usual address! �

+353 47 80500

combilift.com

40

AUTOMATED SURFACE INSPECTION

Performance optimisation of automate Automated Surface Inspection Systems (ASIS) to aid detection and classification of surface defects, especially at the hot rolled stage, can be used to avoid defect crisis and reduce costs in downstream processes. When building the classifier, combined training and putting gross defects before a pseudo defect will improve the classifier model’s performance. For the training set tested, the optimal sample size appears to be 250 samples per class. By Z Chen*, J Li**, R J. Pavlosky**, L Galey2** L Zhang*** AUTOMATED surface inspection has been researched and developed during the past few decades. In its earliest phases, surface inspection systems [1] employed the measurement of reflected light from a point detector or a set of photodetector arrays. In recent years, automated surface inspection has been significantly improved and commercialised. Automated Surface Inspection Systems (ASIS) have been applied to many different processes in the steelmaking process, including hot rolling [2], picking [3], tinning, and hot dip galvanising [4], for surface quality control and assurance.

There are three main ASIS suppliers in the market: Cognex/AMETEK, Primetals Technologies (former Siemens VAI) and ISRA-Parsytec [5]. All three types of ASIS are camera-based inspection systems that include four functions: image acquisition, defect detection /segmentation, feature extraction, and defect classification [6].

These ASIS have both defect detection and classification abilities, and the classification results can be used in near real time to assist in surface quality control and assurance. Each ASIS supplier applies

different technologies to detect and classify defects. To filter image noise and detect defects, each may have applied different detection algorithms. For defect classification, they have all used supervised pattern recognition techniques, but different algorithms are used in different ASIS as the main classification tools [5].

From May 2011 [7], ArcelorMittal Cleveland Hot Strip Mill (HSM) installed and deployed a Cognex/AMETEK ASIS, which has been optimised to meet the growing demands of providing hot band to meet critical surface quality requirements.

The Cognex/Ametek system uses a C5 Decision Tree (DT) algorithm as the main classification tool [8].

ASIS functions and classification tuning ASIS enables the online detection, localisation, and classification of surface defects while the strip is running. ASIS includes four processes: - image acquisition,- defect detection / segmentation, - feature extraction and defect classification. (Fig 1).

Image acquisition, one of the key

processes, is the process of translation from optical signals to a stored digital image by the camera sensor and electronics. If the image has not been satisfactorily acquired, the intended tasks may not be achievable, even with the aid of image enhancement.

Defect detection process includes two steps: defect identification and processing, and defect segmentation. Defect identification is to find and process defects from normal strip while the defect segmentation process segments the defects as single blobs and merges related defects as a complete defect. Feature extraction is the process of deriving mathematical measurements from the detected defects.

During defect classification, the classifier will assign the defect to a category based on the feature vectors provided by the feature extraction process. The whole surface defect inspection process is a closely forward-linked chain. The performance of defect classification will depend on the performance of all the previous processes in the chain. The performance of any process will affect the final performance of an ASIS.

Fig 1.Block diagram of ASIS process

*University of Illinois at Urbana-Champaign, zchen61@illinois.edu; **ArcelorMittal Cleveland; ***ArcelorMittal Global R&D centre – East Chicago liwei.zhang@arcelormittal.com Digital Edition - August 2019

www.steeltimesint.com

AUTOMATED SURFACE INSPECTION

ed surface inspection Fig 2. ASIS installation location at Cleveland HSM

Fig 3. Flowchart of general classifier tuning procedure.

Cognex/AMETEK ASIS at Cleveland HSM Fig 2 shows the Cognex/AMETEK ASIS set-up at Cleveland HSM [7]. The top surface inspection unit is installed at the end of the seven-stand finishing mill and the bottom side surface inspection unit is before the down coilers.

The ASIS is used to detect and classify surface defects that adversely affect the surface quality of hot rolled strip. The outputs of ASIS classification for the defects are fed into ASIS data for surface quality control and assurance in near real-time.

These defects include, but are not limited to, roll marks, drag scales, scales, furnace tears, laminations, scabs, and slivers.

Therefore, the classifier plays an essential role to achieve efficient surface quality control and assurance in near real-time.

Cognex/AMETEK ASIS deploys C5.0type decision tree classifiers as the main classification tool. The C5.0 classifier has been widely used in industry because of its advantages, such as being extremely fast www.steeltimesint.com

in classifying unknown records, capable of handling both continuous and discrete attributes, and being robust to the effect of outliers [9].

Classification tuning The classification performance for a certain defect is evaluated by two parameters: accuracy rate and confidence rate. The accuracy rate defines the percentage of the defect correctly classified by the classifier.

The confidence rate defines the percentage

Fig 4. Decision tree classifier example

41

of true sample numbers in the classified samples for a certain defect. A good classifier needs to have as high an accuracy and confidence rate as possible. Therefore, a classifier needs to classify all the real gross and important defects and reduce false positive ones. Confusions between gross and real defects including important and less important ones are normally allowed.

Fig 3 shows the flowchart of a general classifier tuning procedure. The first step is to prepare a training set, which is the defect library that contains different defect samples. Cognex/AMETEK’s Classifier Manager software is used to collect defect samples from an online database and build a defect library. Its Defect Library Manager software is then used to refine the defect library. After choosing classifier parameters, Classifier Manager is used to build a Decision Tree (DT) classifier. Finally, the classifier is tested both offline and online. Based on the test results, the whole procedure will be repeated several times until a satisfactory classifier is created.

Decision tree algorithm Decision tree (DT), is based on the ‘divide and conquer’ strategy [10], using one of the inductive learning algorithms that generates classification tree using the training data/ samples. An example of DT classifier is shown in Fig 4. Leaves of a tree are class names, and each node represents a feature-based test with branches as possible outcomes. To classify a sample, we start at the root of the tree, evaluate the test, and take the branch appropriate to the outcome. The process continues until a leaf is encountered, at which time the sample is asserted whether it belongs to the class named by the leaf. The tree is expanded until every sample is classified into one of the leaves.

Decision tree classifiers are sometimes more interpretable than other classifiers, such as neural networks and support vector machines, because they combine simple questions about the data in an understandable way [10]. Decision tree approach has substantial advantages for surface quality classification problems because of its flexibility and ability to handle non-linear relations between features and classes at a faster speed. Hence it can easily achieve near real-time classification and improve the classification accuracy to a great extent.

Digital Edition - August 2019

42

AUTOMATED SURFACE INSPECTION

Defect types Total samples

25094

SkidTear

SkidTearD

Scab

Lamn

Other

28 10 23 2 25031

Table 1 Reference coil defect samples

Defect typesTotal samples

2853

SkidTear

SkidTearD

630 300

Scab

Lamn

755 225

Other

913

Table 2 Test library defect samples

The generalised method for constructing a decision tree for an arbitrary collection (S) of samples can be summarised as follows: • If S is empty or only contains samples of one class, then the simplest DT is a leaf labeled with that class.

• Otherwise, let T be any test on a sample that will produce possible outcomes of {01,02,...,0k} Each sample in S will give one of these outcomes for T. Therefore, T portions S into subsets {S1,S2,...,Sk} containing all the samples with the outcome 0i.

The same method is applied recursively to each subset Si of S to build a tree.

C4.5/C5.0 Classification algorithm C4.5 is an algorithm used to generate a decision tree developed by Ross Quinlan [11] , and C5.0 is the latest commercialised version of C4.5. Being multi-threaded, C5.0 is faster, and more memory efficient than C4.5, but the selecting criteria used to build the tree is still the same.

C4.5 builds decision trees using the concept of information entropy.

Suppose there are n classes, denoted as {C1,C2,...,Cn}. The entropy at a given node t is: See equation (1) where St is the subset of samples in node t, |St| represents the number of samples in St, and freq(Cj ,St ) is the frequency of class Cj at node t. Entropy (St ) measures the amount of information needed to identify the class of a sample in St. Considering a parent node p with a subset Sp, after the applied test, T it is partitioned into k subsets {S1,S2,...,Sn}, the information gain of test T is defined as: See equation (2)

Digital Edition - August 2019

Info_gain(T) measures the information gained by Sp portioning in accordance with test T. By analogy with the definition of Info_gain(T), we have: See equation (3) This represents the potential information generated by dividing Sp into k subsets, whereas the information gain measures the information relevant to the classification that arises from the same division. Then, gain ratio is the ratio between info_info(T), which expresses the portion of information generated by the split that is used for classification: • Gain criterion selects a test T to maximise the Info_gain(T).

• Gain ratio criterion selects a test T to maximise the Info_ratio(T).

Although giving quite satisfactory results, the gain criterion has a serious deficiency, which is to have a strong bias in favour of tests with many outcomes [11]. However,

since Cognex/AMETEK ASIS uses binary tests, this deficiency is not a problem. In general, the gain ratio criterion is more robust and typically gives a consistently better result than gain criterion [12].

The recursive partitioning method of constructing decision trees described above will continue to subdivide the set of training samples until each subset contains samples of a single class, or until no test offers any improvement. The result is often a very complex tree that over fits outliers, mislabelled, noisy data resulting in the inference of more structures than being justified by the training samples/ set [11]. During past decades, many methods for simplifying decision trees have been proposed. Among them, pruning techniques are probably the most popular.

C4.5 uses error-based pruning [11] to discard unnecessary subtrees and replaces them with leaves.

Model overfitting The errors committed by a classification model are usually divided into two categories: training error and generalisation error. Training error is the number of misclassification errors committed on the training set, whereas the generalisation error is the expected error of the classification model on unseen data, ie test set. A good classification model must have both a low training error and a low generalisation error. Besides the training set, there is another independent dataset called test set, which will be used to test the performance of the resulting DT classifier from the training set. The error of the training set (percentage of the correctly

Equation 1

Equation 2

Equation 3

www.steeltimesint.com

Innovative OES Reliable aluminum analysis Elemental analysis is essential to the quality of your aluminum products. We offer solutions that fit your needs and your budget. For pure and alloyed aluminum and for inclusion analysis, our instruments deliver unmatched speed, precision and accuracy. Versatile, non-destructive XRF and XRD. Industry-leading OES. Highly efficient automated instruments including sample preparation. A network of local support. You’ll have the instruments and expertise to achieve superior product quality, reduce costs and maximize uptime.

Thermo Scientific™ ARL iSpark™ OES

Thermo Scientific™ ARL easySpark™ OES

Thermo Scientific™ ARL™ SMS-2500 automation

Discover more at thermofisher.com/elemental © 2019 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.

Thermo Scientific™ ARL™ 9900 XRF-XRD

44

AUTOMATED SURFACE INSPECTION

Classifier Performance Training methods

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

Combined training Separate training

93.65

78.67 6.16

74.6

75.81 21.61

Table 3 Performance of classifiers

Classifier Performance Training methods

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

Gross first

93.65

78.67 6.16

Pseudo first

82.54

67.53 6.41

Notes: pseudo defects mean unimportant defects, which do not affect surface quality. 10 test coil set is used for classifier performance test. The gross defect classification performance was evaluated.

Table 4 Performance of classifiers

classified samples in the training set) is called training error, while the error of the test set is called test error. The test error is generalisation error since the model has not seen the test data.

When a tree is too simple, both the training error and the test error are large, resulting in ‘under fitting’. Under fitting occurs when the model has not yet learned the true structure of the training set. As a result, the classification error is large both on the training set and previously unseen data. When the number of nodes increases, the tree grows larger and both errors decrease. However, if the tree becomes too large, its test error starts to increase while its training error continuously decreases. This phenomenon is called ‘model overfitting’. The reason is that when the tree is too complex, some nodes that accidentally fit the outliers, or noisy data, may not generalise well to the test set and will downgrade the performance of a tree.

For the Cognex/AMETEK system, besides the DT algorithm itself with the pruning and boosting capabilities to minimise the under fitting and overfitting problems, under fitting can be solved by increasing the training set size, to increase the tree complexity. Overfitting can be a more important problem. It may be caused by various reasons such as the presence of noise, lack of representative samples, and most importantly, the complexity of the model [13].

Classifier performance based on DT theory When building a DT classifier, the effects of its algorithm parameters on Digital Edition - August 2019

the classification performance need to be optimised, but it is not the focus of this paper which mainly focuses on the effects of classification performance using a different training set (with the same algorithm parameters). The training set has significant effects on the classifier’s performance because small variations in the training data can result in very different looking trees [14].

One of the requirements of the C4.5 DT model is sufficient data. The amount of data required is affected by factors such as the number of features and classes and the complexity of the classification model; as these increase, more data will be needed to construct a reliable model [11]. It has been proven that the model accuracy tends to increase at a decreasing rate with increases in the sample size [15]. However, if the sample size is too large, this might also hurt the model’s performance. Oates et al. found that training set size has a direct impact on tree size. Even if C4.5 is pruning unnecessary subtrees, overfitting is, in fact, occurring and it becomes worse as the sample size increases. Research shows that there is a nearly linear relationship between training set size and tree size, even after accuracy has ceased to increase [16].

Besides the sample size, the number of classes and features will also have effects on the performance. If there are too many classes, then the numbers of training samples become smaller quickly in a tree with many levels/branches. This problem is called ‘error-prone’ [17]. For a given sample size used in training a classifier, there exists an optimal feature size. This result is true for both two-class problems and multi-class

problems. In tasks where more features than the ‘optimal’ are available, decision tree quality is affected by the redundant and irrelevant features [18].

For Cognex/AMETEK ASIS, there are two types of defect library: training set defect library and test set defect library. The training set defect library is the training set and it will affect the classification performance in many ways. Specific factors in the training set, which might affect a final DT classifier performance are: 1) Sub-class separate training vs combined class training Suppose we have two defect classes; A and B, which are separated into different subclasses (A1, A2, A3, B1, B2) based on size, darkness or severity. Separate training means that a classifier is trained for five different classes (A1, A2, A3, B1, B2), but only two classes (A and B) are considered while evaluating the performance.

Combined training means that a classifier is trained for two classes (A and B). For the same dataset, separate training may have several drawbacks: first, it will have bigger entropy due to more classes, and more information is needed to correctly classify the dataset; second, a sample in one subclass may also fill into another subclass (a dark sample may also be a small sample).

2) Sequence of classes when training classifier Based on the DT algorithm, there should be no difference with different class sequences put in the classifier training. The practical study will test if there is improved classification performance with Cognex/ AMETEK DT classifier regarding the class sequence.

3) Total class categories As previously mentioned, more data will be needed to construct a reliable model with more class categories [11], which means more time and effort need to be devoted to sampling collecting. For a specific training set, if there are too many classes, the numbers of training samples become smaller quickly in a tree with many levels/ branches [17], which may lead to overfitting.

4) Data balancing for each class/ category A dataset is considered ‘imbalanced’ if one class (the majority class) vastly outnumbers the others (minority classes) in the training www.steeltimesint.com

CONNECTING THE INTERNATIONAL STEEL INDUSTRY CTIVES

MAGAZINE

FORUM STEEL FUTURE try – 2018 steel indus the 4.0 and Industry 50 now! go to page

ORT ENCE REP

CONFER

scope the micro n under rence Digitalisatio Innovaction confe li’s at Danie

TUBE, PIPE

AND WIRE ROD

tube, ts in the developmen markets The latest pipe and steel wire rod

PERSPE

etaltech r of LTI-M Edgar Rayne questions our answers

DIRECTORY

mesint.com

www.steelti

mesint.com www.steelti - Vol.42 No2 March 2018

Order your copy priced from £78 or FREE for paid subscribers

17 ORY 20

– Vol.42 No.2

dd 1

march.in

WEB

DIRECT STI BUYERS

GUIDE.indd

17 ORY 20

17 14:15:13

06/09/20

1

WEEKLY E-NEWSLETTER

Packed with information on the steel industry and continually updated with news for steel professionals, steeltimesint.com also features special articles and interviews with leading industry figures.

@SteelTimesInt

STI_Full_page_Portfolio_A4_Ad.indd 1

18 11:02:10

15/03/20

T L DIREC

March 2018

STI Cover

@S

ATIONA

IONAL –

INTERNAT

CS 4.0

S ETHI

BUSINES

RY

DUST TEEL IN

TERN IMES IN STEEL T

The Steel Times International Directory is the essential guide to steel manufacturers, producers, suppliers of plant equipment and services to the steel industry.

STEEL TIMES

Steel Times International is the key publication for the steel market, reporting on iron and steel making issues from all corners of the globe.

A round-up of the top news stories is also sent to more than 11,000 industry professionals each week. You can register online to receive the weekly newsletter and keep up-to-date with the latest news from across the steel manufacturing industry.

www.steeltimesint.com/e-newsletter

WWW.STEELTIMESINT.COM 22/01/2019 10:10

46

AUTOMATED SURFACE INSPECTION

Classifier Performance Classifier

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

Classifier A

93.65

78.67 6.16

Classifier B

85.71

60.67 19.80

Classifier C

71.43

68.18 6.58

Note: 10 coil test set is used for classifier performance test. The gross defect classification performance was evaluated.

Table 5 Performance of classifiers

Classifier Performance Ratio Gross: Pseudo

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

1:12

65.88

1:10

66.55

99.22 3.22 99.54 3.89

1:8

70.26

99.20 2.70

1:6

73.66

99.17 3.05

1:5

76.75

98.35 2.49

1:4

80.10

97.67 2.49

1:3

85.15

96.83 1.86

Note: The new sub-test set (1940 gross samples and 913 pseudo samples) is used for classifier performance test. The gross defect classification performance was evaluated.

Table 6 Performance of classifiers with low gross defects

data. One of the weaknesses of decision trees is dealing with imbalanced datasets [19] . This is because when learning under highly imbalanced training data, classifying all instances as the majority class will result in high accuracy of classification.

Therefore, a training dataset consisting of an importunately high number of examples from one class will result in a classifier that is biased towards this majority class [20].

5) Feature number vs sample number The accuracy of the model increases more slowly as the training set size increases in training [15]. However, if the training set size is too large, this might also hurt the model’s performance. What’s more, for a given sample size used in training a classifier, there exists an optimal feature size. If more features than the ‘optimal’ are available, the decision tree quality is known to be affected by the redundant and irrelevant features [18]. Then for a given feature size, there should also be an optimum sample size.

Practical study Software tools in Cognex/AMETEK ASIS are used to build and refine a training set. The final training set includes 65 subclasses, 21 classes, and a total of 6648 samples.

Two test sets were prepared, the first one consists of 10 verified coil inspection data under the normal and good production conditions, shown in Table 1, and will be referred to as 10 coil test set. The second Digital Edition - August 2019

test set consists of the previously collected samples (not used in the training), as shown in Table 2. These samples are a good test set to be used for studies, where large numbers of gross defects are needed. The classifier model is created in the Cognex/AMETEK ASIS environment.

Several classifiers were built using different training sets under the same DT algorithm parameters with all 132 available features in Cognex/AMETEK. The C5.0 DT algorithm builds a classifier to maximise the overall accuracy rate for all defect classes in the training set, and normally the performance of a classifier is evaluated using the accuracy rate of all classes. In this paper, the classifier performances on the gross/important defects (SkidTear, SkidTearD, Scab, and Lamn) are used for evaluation and comparison.

Results & Discussion Two classifiers are studied Separate Training and Combined Training. The first classifier was built using the separate training option with 65 different classes since all subclasses were treated as independent classes; the second classifier was built under the combined training option with 21 classes.

The classification results of these two classifiers are shown in Table 3.

The test results show that the classifier with the combined training has better performance (better accuracy and confidence rates, and lower unclassified

rate) than the separate training. This is consistent with the theoretical study.

Defect class sequences In the Cognex/AMETEK classifier building environment, users can choose the sequence of defect classes when training a classifier. Typically, the defect classes should be in the order of difficulty to classify. To study the effect of the sequence while training a classifier, two classifiers with reversed sequence were trained: the first classifier uses the sequence from pseudo to gross defects while the second classifier uses the sequence from gross to pseudo defects. The classifier test results are shown in Table 4.

The test results show that unclassified rate is about the same for two different classifiers. However, the classifier with gross defects first in the sequence has better defect classification accuracy and confidence rates. Therefore, whilst training a classifier in Cognex/AMETEK ASIS environment, it’s better to put gross defects first, then pseudo defects.

Total class categories Based on the discussion from the previous section, the number of total classes in the training set has effects on a classifier’s performance. In this section, three classifiers with different defect class categories are studied using the same training set (65 subclasses, and a total of 6648 samples) as for the previous study: • Classifier A: This classifier combines 65 subclasses into 21 classes.

• Classifier B: This classifier combines 65 subclasses into 30 classes. Compared to Classifier A, the gross defect classes are the same, but more pseudo defect classes are added based on their shapes, sizes or brightness differences. For example, Friction class in Classifier A is separated into FrictionBlock, Friction Diagonal, FrictionVshape etc.

• Classifier C: This classifier combines 65 subclasses into 27 classes. Compared to Classifier A, the pseudo defect classes are the same, but more gross defect classes are added based on their shapes, sizes or brightness differences. For example, Tear class in classifier A is separated into TearLarge, TearMedium, and TearSmall.

These three classifiers use the same defect samples and classifier parameters, the only difference is the number of defect classes www.steeltimesint.com

CONNECTING THE INTERNATIONAL STEEL INDUSTRY

Steel Times International Directory The Directory is the essential guide to steel manufacturers, producers, suppliers of plant equipment and services to the steel industry. It provides comprehensive listings, product information and key contact

Order your copy now from just £95*

details.

For more information on advertising or how to be listed contact: Esme Horn esmehorn@quartzltd.com Tel: +44 (0) 1737 855 136

@SteelTimesInt

STI_Buyers_Guide_Half_Vertical.indd 2

WWW.STEELTIMESINT.COM 25/02/2019 11:05

48

AUTOMATED SURFACE INSPECTION

Classifier Performance Ratio Gross: Pseudo

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

10:1

99.74

72.20 0.46

8:1

99.33

78.69 0.46

6:1

99.43

79.42 0.46

5:1

99.48

79.69 0.60

4:1

99.28

81.47 0.53

1:1

96.19

91.70 0.67

Note: The new sub-test set (1940 gross samples and 913 pseudo samples) is used for classifier performance test. The gross defect classification performance was evaluated.

Table 7 Performance of classifiers with low pseudo defects Classifier Performance Ratio Gross: Pseudo

Accuracy rate

Confidence rate

Unclassified rate

%

%

%

250

95.74

90.80 0.81

180

95.62

89.88 0.73

120

94.73

91.02 0.96

60

92.25

87.83 0.92

30

90.30

75.77 1.38

Note: The second test set is used for the classifier performance test. The gross defect classification performance was evaluated.

Table 8 Performance of classifiers with different training set size

to be classified. As a result, comparing the performance of these three classifiers will reveal the effects of total class categories for the specific training set on classifier performance. A 10 coil test set is used to test the performance and the test results are shown in Table 5.

The test results show that Classifier A has the best performance overall. Comparing the performance of Classifiers A and B, when more pseudo defect classes are added to the training set, the accuracy and confidence rates of the gross/ important defects drop. The unclassified rate is increased and most of the newly unclassified samples are pseudo defects.

Compared with Classifier A, the results of Classifier C show that when adding more gross defect classes, the accuracy and confidence rates of gross/important defect classification drop significantly, although the unclassified rate was not seriously affected (increase about 0.4%). Therefore, we can conclude that adding more classes based on their shapes, sizes and brightness differences will downgrade the performance of a classifier.

Data balancing between classes To study the effects of the data balancing between defect classes on the classifier’s performance, six classes are used (SkidTear, SkidTearD, Scab, Lamn, Stain, and ScalePits). The second test set (samples not used in the classifier training), instead of online coil inspection data, is used for Digital Edition - August 2019

performance testing since the online coil inspection data does not contain sufficiently gross defect samples for this study. SkidTear, SkidTearD, Scab, and Lamn are categorised as gross defects while the rest are categorised as pseudo defects. Note that ScalePits are important defects. However, the Cognex/AMTEK system is not able to distinguish them well from stains or pseudo dots (unimportant defects). Therefore, we put the ScalePits defects into the pseudo category in this study.

Two different studies were performed: • Low gross defect and high pseudo defect cases: each gross defect class has 50 samples. The samples in each pseudo defect class vary according to different ratios.

• High gross defect and low pseudo defect cases: each gross defect class has 20 samples, while the samples in each gross defect class vary. The reason 20 samples per class is chosen is that there are not enough gross defect samples to allow a larger number under 10:1 ratio for gross defects over pseudo defects.

Random defect samples of the six classes were selected from the second test set to form a new sub-test set, with 1940 gross samples and 913 pseudo samples. The test results are shown in the Tables 6 and 7.

The results of Table 6 show that, when there are more pseudo defects, the classifier’s performance increases as the pseudo defect sample number decreases (ie the dataset is transferring from unbalanced

to balanced). When learning under pseudo defect dominated unbalanced training data, the C5.0 DT algorithm tends to bias toward the Pseudo classes (majority classes). Therefore, we see the accuracy rate of gross defect is increased when the pseudo samples are decreased, while the confidence rate acts in reverse.

Similarly, when there are much more gross defect samples than pseudo defect samples, the C5.0 DT algorithm tends to classify pseudo defect samples as gross defects, which may lead to high accuracy, but low confidence rates for gross defects for the resulting classifier. Results in Table 7 show that when the training set is highly unbalanced and dominated by gross defects, the resulting classifier can achieve higher accuracy rate, but lower confidence rate on gross defects.

Both the theoretical study and experimental results show that a balanced training set such as 1:1 ratio for gross and pseudo defects, is better for the DT algorithm.

Feature size and training set size It is shown in the literature that for a given sample size, there exists an optimal feature size and for a given feature size, there is an optimal sample size[15] [18]. Classifiers studied in this section use 132 total features. Six classes (SkidTear, SkidTearD, Scab, Lamn, Stain, and Scale) are selected, and the second test set is used for classifier performance tests. The sample number for all classes is kept the same intentionally to form a balanced training set. Different training dataset sizes (30, 60, 120, 180, 250 samples per class) are used and the test results are shown in Table 8.

The results in Table 8 show that both accuracy and confidence rates are lower when the training set is smaller than the feature size. The accuracy and confidence rates are increased with a decreasing rate while training dataset size is increased, which matches the DT theory analysis results. The literature review also shows that, if training set size is larger than the ‘optimum’ value, the classifier’s performance will decrease due to overfitting. Although more studies are needed to find the optimal sample size, the results in Table 8 show that the 250 samples per class seem to be a good choice for this particular study.

The studies show that when building the classifier, combined training and putting www.steeltimesint.com

SENSORS Full Capacity Supplier for Rolling Mills

ience counts !

... where exper

Temperature Monitoring and Profiling Solutions in the Steel Reheat Industry ‘Thru-Process’ Monitoring solutions for Slab, Billet & Bloom Reheat Furnaces

Load Cells

• Optimise furnace programs • Save energy and increase production • Obtain optimal drop out temperatures • Minimise scale build up • Prevent hot roller wear & tear

Width Gages Tensiometers Plate Outline Gages Laser Speed/Length Measuring System

Comprehensive • Passes through furnace with Slab • Get an accurate Slab temperature profile • Measurement at up to 20 points • Live 2 way radio communications

Weighing Applications Fluid Pressure Transmitters

Safe • Safe system installation without production

Hot Metal Detectors Absolute Displacement Transducers

delays

• Reliable protection of data logger up to 1300°C

Easy • Optimise your process accurately • Validate your furnace mathematical

For details call +1-416-445-5850, or kelk.sales@vpgsensors.com

model.

kelk.com

Visit us for more information: ©2017 Vishay Precision Group, Inc.

PhoenixTM Ltd UK sales@phoenixtm.com

www.phoenixtm.com

PhoenixTM GmbH Germany info@phoenixtm.de

PhoenixTM LLC USA info@phoenixtm.com

50

AUTOMATED SURFACE INSPECTION

gross defects before a pseudo defect will improve the classifier model’s performance.

When building the training set, unnecessary classes (such as separate classes based on size, brightness etc) and unbalanced datasets need to be avoided. The studies also show that an optimal feature size/ training size relation exists, and the optimal sample size appears to be 250 samples per class with the training set tested.

The classification performance is the key to maximising the benefit of using ASIS. Although the study and test results are from Cognex/AMETEK ASIS, the best practices can be applied to any ASIS, where a Decision Tree classifier is deployed.

Acknowledgement This paper was presented at AIST 2018 7-10 May in Philadelphia, PA, USA. This work was supported by ArcelorMittal Global R&D and ArcelorMittal Cleveland.

Special thanks are given to Cleveland Hot Strip Mill (HSM) management and global RD-EC management for their great support and encouragement of this work. Also, special thanks to Joseph Bentsman from the University of Illinois at Urbana-Champaign and B G Thomas from Colorado School of Mines for ideas and suggestions

‘Automatic Surface Inspection for Galvanized Products at IMSA MEX,’ in AISTech, Cleveland, OH, 2006, 2006 [5] L Zhang, H Long and J Chapko, ‘Tuning and Deployment of a Surface Inspection System,’ Iron & Steel Technology, no January, pp 38-46, 2010 [6] L Zhang, J Franklin, T Pisanelli, R Pavlosky and H Long, ‘Optimization of Surface Inspection System Performance,’ in AISTech, Pittsburgh, PA, USA, 2008 [7] R J Pavlosky, L Galey, M B Assar and L Zhang, ‘Practical Applications of Surface Inspection System at the Hot Mill,’ in AISTech, Cleveland, OH, USA, 2015 [8] C Corp , ‘SmartView ICN System Setup and Reference’ [9] S Miertschin, ‘‘Decision Tree’ [PowerPoint slides],’ University of Houston [10] R Safavian and D Landgrebe, ‘A survey of Decision Tree Classifier Methodology,’ IEEE Transactions, vol 21, no 3, pp 660-674, 1991 [11] J R Quinlan, C4 5: programs for machine learning, Elsevier, 2014 [12] J R Quilan, ‘Decision Trees and Multi-valued Attributes,’ in Machine Intelligence 11, New York, NY, USA, Oxford University Press, Inc , 1988, pp 305318 [13] P -N Tan, M Steinbach and V Kumar, Introduction to Data Mining, Addison-Wesley

Longman, Inc , 2006, pp 172-186 [14] J R Quinlan, ‘Induction of decision trees,’ Machine learning, vol 1, no 1, pp 81-106, 1986 [15] J Morgan, R Daugherty, A Hilchie and B Carey, ‘Sample size and modeling accuracy of decision tree based data mining tools,’ Academy of Information and Management Science Journal, vol 6, no 2, pp 71-99, 2003 [16] T Oates and D Jensen, ‘The Effects of Training Set Size on Decision Tree Complexity,’ in The Fourteenth International Conference on Machine Learning, San Francisco, CA, USA, 1997 [17] A Nordman, ‘‘Classification: Decision Trees’ [Lecture 3] ‘TNM033: Data Mining’ [18] H Almuallim and T G Dietterich, ‘Learning Boolean concepts in the presence of many irrelevant features,’ Artificial Intelligence, vol 69, no 1, pp 279-305, 1994 [19] H He and E A Garcia, ‘Learning from Imbalanced Data,’ IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, vol 21, no 9, pp 1263-1284, 2009 [20] N Japkowicz, ‘The class imbalance problem: Significance and strategies,’ in Proceedings of the International Conference on Artificial Intelligence, 2000

References [1] E J Klimas, F Popovich and R J Pavlosky, ‘The practical application of automatic inspection instrumentation to cold rolled steel strip,’ SAE Technical Paper, 1984 [2] M Baileul, ‘Dynamic Surface Inspection at the Hot-Strip Mill,’ in AISTech, Charlotte, NC, USA, 2005 [3] T MacDougall and E Dillon, ‘Implementation of Surface Inspection Technology on Noisy Surfaces,’ in AISTech, Indianapolis, IN, USA, 2007 [4] M Gonzalez, A Trevino, M Bailleul and F Nlome-Nze,

Superior high-temperature tags and labels for track and trace applications.

Polyonics® offers thermal transfer printable polyimide and aluminum bar code tag and label materials so you can accurately control inventories from when the materials leave the furnace through to delivery to your customers.

Keeping you on track Digital Edition - August 2019

polyonics.com/HotMetal 603.352.1415 | info@polyonics.com

Personal Chemistry

www.steeltimesint.com

CONNECTING THE INTERNATIONAL STEEL INDUSTRY CTIVES

MAGAZINE

FORUM STEEL FUTURE try – 2018 steel indus the 4.0 and Industry 50 now! go to page

ORT ENCE REP

CONFER

scope the micro n under rence Digitalisatio Innovaction confe li’s at Danie

TUBE, PIPE

AND WIRE ROD

tube, ts in the developmen markets The latest pipe and steel wire rod

PERSPE

etaltech r of LTI-M Edgar Rayne questions our answers

DIRECTORY

mesint.com

www.steelti

mesint.com www.steelti - Vol.42 No2 March 2018

Order your copy priced from £78 or FREE for paid subscribers

17 ORY 20

– Vol.42 No.2

dd 1

march.in

WEB

DIRECT STI BUYERS

GUIDE.indd

17 ORY 20

17 14:15:13

06/09/20

1

WEEKLY E-NEWSLETTER

Packed with information on the steel industry and continually updated with news for steel professionals, steeltimesint.com also features special articles and interviews with leading industry figures.

@SteelTimesInt

STI_Full_page_Portfolio_A4_Ad.indd 1

18 11:02:10

15/03/20

T L DIREC

March 2018

STI Cover

@S

ATIONA

IONAL –

INTERNAT

CS 4.0

S ETHI

BUSINES

RY

DUST TEEL IN

TERN IMES IN STEEL T

The Steel Times International Directory is the essential guide to steel manufacturers, producers, suppliers of plant equipment and services to the steel industry.

STEEL TIMES

Steel Times International is the key publication for the steel market, reporting on iron and steel making issues from all corners of the globe.

A round-up of the top news stories is also sent to more than 11,000 industry professionals each week. You can register online to receive the weekly newsletter and keep up-to-date with the latest news from across the steel manufacturing industry.

www.steeltimesint.com/e-newsletter

WWW.STEELTIMESINT.COM 22/01/2019 10:10

52

GALVANISING LINES

Galvanising technologies for AHSS Steel grades requiring galvanising have dramatically changed in the last 20 years, going from simple carbon grades for construction to ultra-high tensile strength dual phase grades and extra soft and deep drawing qualities for the automotive industry. These new steels have required key modifications and improvements in galvanising lines. For instance, furnaces must accommodate new heat cycles, some with annealing temperatures over 900°C. Fast cooling to various temperatures is sometimes followed by aging sections in the line, and surface oxidation must be controlled to achieve good zinc wettability.

Plant operating constraints, design changes, areas of investigation and thoughts about even newer steels are discussed. By Michel Dubois* and Brice Van Houtte*

DRIVEN by the wish for sustainable development and minimising greenhouse gases, the automotive industry has pushed for increased use of thinner, higher strength, yet ductile steels to reduce the weight of the body in white and improve crash resistance. The development and increased use of advanced high strength steels (AHSS) has required the development and implementation of new technologies for galvanising lines in two main areas: • Strip heat cycles with higher temperatures and faster cooling rates • Surface control to ensure a good zinc wettability Additionally, some process steps had to be improved and upgraded with more powerful devices such as shears, bridle power, skin pass mills and recoilers.

Heat cycles and technology development For traditional carbon grades (CQ) simple heat cycles consisted of heating up to 720°C, minimal soaking and then forced cooling down to the zinc pot temperature.

For AHSS grades, however, the cycles are much more complex and varied, as illustrated below: • Peak strip temperatures have dramatically increased up to 920°C to ensure full ‘austenitisation’ as required by fully martensitic grades • After an initial slow cool, fast cooling sometimes reaches around 100°C/ sec using gas cooling or 500°C/sec using

non-oxidising liquid hydrocarbons to reach temperatures between 450°C and 250°C (both for 1mm strip) • A reheating process under a H2 atmosphere before the Zn pot may be required • Overaging for up to 120 seconds may be required for bainitic transformations • A rapid cool after coating may be added to limit either the time before martensite transformation and/or aging.

Fig. 1 shows four typical heat cycles that are used today. Because of the different cycles a single furnace is not capable of producing all of them without significant changes in the strip pass line, an operation

which requires a long stoppage time, and unlikely to be acceptable.

The galvanising heat cycle is, unfortunately, far from ideal with regard to quenchability. After fast cooling there is an unavoidable soaking time of 10 - 20 seconds at ~450°C, depending on line speed. This is the location of the bainitic nose in CCT diagrams and the soaking time corresponds to the strip length through the hot bridle, snout, Zn pot, strip wiping and cooling areas down to about 400°C in the tower. This means that the use of the classical CCT diagram is not valid and it is important that metallurgical designers allow for this fact in their steel composition/ process design calculations. For example,

Fig 1. Heat cycle one for dual phase (DP), martensitic (MS), twinning induced plasticity (TRIP) and quench & partitioning (Q-P) steels

* Michel Dubois and Brice Van Houtte work for John Cockerill in Seraing, Belgium.

Email michel.dubois@cmigroupe.com Digital Edition - August 2019

www.steeltimesint.com

GALVANISING LINES

53

Fig 3. Time to reach annealing temperature for 1mm thickness with two tube Fig 2. CGL cooling curve through a CCT diagrams, austenitisation 800°C

Fig.2 shows the cooling profile computed by dedicated software for a dual phase (DP) grade superimposed on the CCT diagram.

Technologies for heating To reach the full ‘austenitisation’ temperature, classical radiant tube technology has reached its limits, as the classical refractory metal heating elements such as Inconel 601 and refractory stainless steels are limited to about 950 - 980°C. Fig. 3 indicates the time required to reach a defined peak temperature for two different zone temperatures. It is evident that when the target is 920°C the use of classical radiant tubes significantly reduces line productivity, as the line speed has to be reduced compared to the use of modern high Cr alloys used up to 1,100°C. Some technological innovations have been developed to circumvent this problem, for instance, double P radiant tubes (see Fig. 4) enable higher average operating temperatures, thanks to the high internal circulation rate of the gases.

For horizontal furnaces that are less sensitive to strip breakages, ceramic tubes (typically made of SiC (see Fig.5)) are proposed and can operate at 1,100°C and above, but they are only available as straight tubes. Finally, the technology used in stainless bright annealing plants (Cr-Mo resistance elements operating at 1,200°C) can also be used (see Fig. 6), however this can increase problems with strip breakage and short circuits, electrical insulation and furnace lining materials. An option could be to have a dedicated section for their implementation, located after a classical radiant tube heating. www.steeltimesint.com

temperatures

Instead of radiative heating, induction with transverse flux technology can be used to overcome the reduction in steel permeability above 700°C - the Curie point.

Various patents and technical solutions including dedicated software are proposed by suppliers to avoid the differential heating between edges and centre.

Technologies for cooling Some steel grades require rapid cooling through quenching to achieve the desired properties. However, for hot-dip coating lines, in order to maintain good wettability and coating adhesion, the steel surface cannot be significantly oxidised. Water cooling, therefore, the most rapid method, is not applicable, and this impacts the maximum tensile strength that can be obtained with a defined steel composition.

Technological development has been focused on the optimisation of classical gas cooling, working on the nozzle shape, gas flow, H2 content and nozzle-to-strip distance. Cooling rates of 100-140°C/sec between 700 and 450°C for 1mm strip thickness are possible with high H2 content cooling gas.

High cooling rates using gas do have consequences. The electrical consumption of the blower is very large with required motor power reaching 1MW. This is due to the fact that the electrical power required varies with the 4th power of the heat transfer co-efficient, so doubling the cooling rate requires 16 times more power. Another issue is the possible H2 pick-up by steel that may induce hydrogen embrittlement after forming. If the risk of H2 pick-up is high, then residence time in

the cooling section should be minimised.

Also, furnace partitioning is recommended to minimise H2 content after cooling.

In order to address the above, an innovative solution using a non-oxidising liquid hydrocarbon called ‘Ultra Dry Cooling’ has been developed by John Cockerill. Cooling rates in the range of 400°C/sec on 1mm strip can be reached as well as final temperatures as low as 250280°C, such values being required for some stop-quench processes.

Some plants utilise a reheating inductor situated before the pot to enable the strip to reach the pot temperature after a low overaging (typically 250 - 400°C) treatment.

Classical longitudinal induction furnaces are being used, but the technological improvement has been limited to the use of HNX atmospheres. Gas tightness and thermal expansion in operation are critical.

Finally, a cooler operating on the liquid coating with variable blowing rates can be added after the pot, and before the classical APC cooling tower, to minimise the time before Ms temperature (or aging if martensite is formed before the pot) while avoiding damage of the liquid layer which is very sensitive to gas impingement (see Fig.8).

Heat cycle challenges The need for consistency and reproducibility of mechanical properties is paramount in an industrial plant. The example of steel grades requiring an accurate stop quench process such as quench and partitioning (Q&P) is especially of concern, given that steel composition and upstream processing may vary slightly between batches, and Digital Edition - August 2019

54

GALVANISING LINES

Fig 4. Example of double P radiant tubes with high Ni-Cr alloy for 1,100°C

Fig 5. Silicon Carbide Tube introduced in a furnace

Fig 6. (left) Cr-Mo resistance heating elements

Fig 7. Performance obtained with CMI’s Ultra Dry cooling

that Ac1 and Ac3 temperatures depend on heating rate and composition. Intelligent metallurgical models are expected to help define the set points and targets for each production batch.

Additionally, there is a need to control the temperature uniformity along and across the strip at +/-5°C, not an easy task considering the variations in steel surface emissivity that are known to be affected by surface cleanliness and oxidation. Systems including measurement and actuators to control the transverse uniformity are quite rare and/or difficult to operate industrially.

A key question is…”What will be the most economic way to produce the grades that require very different heat cycles?”

Dedicated lines, multiple pass lines selection with or without twisting towers are some options. The examples of DP and Q&P grades emphasises the point: one requires overaging, the other not. The situation is much more critical in continuous galvanising lines than in continuous Digital Edition - August 2019

annealing lines due to the constraint on the entry strip temperature; 450°C, for example, leads to fast aging of martensite. Other areas of concern are the possible need for better flatness control at the wiping nozzle for those grades that have the phase transformation before the pot, and the question of crossbow correction with the existing pot roll design.

The use of Zn-Al+Mg bath composition offers the chance to reduce the pot temperature by typically 40°C.This may provide some benefits since the soaking temperature before the pot will be significantly below the bainitic nose.

Managing surface wettability It is well known that during annealing in industrial furnace atmospheres, Cr, Si and Mn will oxidise leading to a surface oxide layer that inhibits reactive wetting between Fe and Zn. Three key ways have been proposed to circumvent this: • Precoating of steel before annealing:

Ni or Cu plating was proposed some years ago, but is not yet used industrially due to the cost and the fact that those elements can diffuse into the steel, impacting the Fe-Zn reaction. The question of optimum deposit thickness is also still to be clarified.

• Internal oxidation during annealing: as the dew point increases the oxideforming potential for the alloying elements that can be oxidised underneath the surface also rises, particularly for B, Al and Mn and Si. The Wagner theory explains the phenomena. It has limitations, however, because it is based on the competition of the oxidant diffusion to the strip core and the alloying elements towards the surface. Si for example, when present at a concentration higher than about 75% of that of Mn cannot be oxidised internally and will make a covering SiO2 layer.

Fig. 9 shows the predictions obtained by AESOP, a computer model, for two different dew points for a CP1180 grade. CO is being www.steeltimesint.com

GALVANISING LINES

55

Fig 8. Example of pre-cooler for GI

Fig 9a. Prediction by AESOP with 0°C Dew point Fig 9b Prediction by AESOP with -35°C Dew point

Fig 10. (right) Strip oxidation principle

generated from carbon in the furnace and a significant amount of water is consumed.

The surface becomes carbon-depleted so it is less easily transformed into austenite during annealing below 910°C, and also loses its hardenability. Furthermore, a high CO content in the furnace can lead to carbon deposits and to safety issues close to the entry seal areas due to the toxicity of that gas. The question of possible carburisation of the radiant tubes with its potential detrimental effect is under debate.

• Full Fe oxidation: this method is well known and was the way followed by the old Sendzimir process. However, due to the increase of concentration of the alloying elements, the Fe oxide thickness must reach 50 - 200nm before entering a fully reducing atmosphere. This requires full oxidation between 650 and 750°C. The next step of the process reduces that oxide with H2 while avoiding the migration of the alloying elements to the surface during the residence time at high temperature. www.steeltimesint.com

TECHNOLOGIES FOR CONTROL OF SURFACE WETTABILITY John Cockerill has developed two main technologies: one dedicated to direct fired furnaces (DFF) in which a slight air excess in the last zones oxidises the strip surface.

Fig. 10 shows the principle. If efficient, it requires a good control of the gas quality to ensure the right air excess. A concept upgrade, also dedicated to DFFs consists of the implementation of an additional zone dedicated to control accurately the oxidation across the strip width. The DFF is then always running in its classical gas excess mode, oxidation being controlled only at its end in a dedicated section.

Alternatives are available for full radiant tube furnaces that are known to be more energy efficient and less sensitive to the gas calorific value. Two main proposals exist, one consists of a full oxidation zone inside a full radiant tube furnace operating with a full reducing atmosphere, whereas the second proposal, developed and patented

by John Cockerill, offers a three-section furnace, the first being close to neutral for Fe, the second making the oxide with dedicated control for each side, the third having in charge the reduction of the Fe oxide while keeping some oxidant potential for the alloying elements (see Fig.11 from left to right).

Challenges on surface wettability Numerous laboratories have investigated the question of wettability; and the need for high dew point and/or full surface oxidation is recognised. However, customers require good coating adhesion and it is known that good wetting and poor adhesion are not always linked. In addition, oxidation at grain boundaries is not recommended since it may initiate cracks during deformation. Some laboratories have also investigated the possibility of operating at a very low dew point with a H2O/H2 ratio so low that the oxidation of Mn, Si, etc. does not occur. If this were Digital Edition - August 2019

56

GALVANISING LINES

Fig 11. Oxidation control in radiant tube furnace patented by CMI

Fig 13. Spring effect on recoiler leading to poor grip

Fig 12. 316L sink roll after 6 runs of 4 days in AlSi

possible, it means that SiO2-based furnace linings need to be replaced by Al2O3 thanks to its higher resistance to reducing agents.

Unfortunately, operating at -100°C dew point seems impossible industrially since experiments have shown that there is no efficient means to reduce the dew point to that value.

Alternatively, running with high H2 as in stainless steel practice is not a solution due to the problem of H2 pick-up during annealing and the impermeability of a Zn layer to the H2 diffusion.

Another issue is grade transition in industrial schedules. It is known, for example, that the oxide thickness increases much faster on CQ grades than on AHSS and, therefore, the furnace atmosphere requires adjustment at grade changes.

This may impact line scheduling, as with annealing cycles. The question of Zn reactivity in terms of its dissolution, dross formation and the galvannealing reaction with reduced Fe oxide in a `fresh and pure` state is presently not clear. Finally, there is the potential for oxides to stick on furnace Digital Edition - August 2019

rolls, resulting in dents on the strip. It is known that high Mn grades require specific roll coatings, but extensive experience is lacking.

New coatings and Zn pot area The last decade has seen the appearance of numerous Al-Zn+Mg coatings dedicated to construction and automotive products, the latter usually containing less Mg. There has also been a rebirth of aluminising in connection with the hot stamping process to make high strength safety components (up to 2,000 MPa). The question of the benefits of AlZnMg coatings for AHSS is, however, still open in relation to the expected lower pot temperature that may impact the heat cycle.

Technology development has been in two main areas: • Specific wiping devices to minimise the oxidation of high AlMg content. The Arcelor patent WO 2010130895 describes the device and is industrially implemented.

• Due to the aggressiveness of liquid AlSi, pot rollers are particularly vulnerable,

as illustrated in Fig.12. Various solutions are proposed by different suppliers. Some have proposed specific ceramic bearings and/or coating for pot rolls. Replacing the classical 316L materials with high carbon cast steels can provide longer roll life; however, even with those innovations, the pot roll life time is only three to four days.

One solution might be a full ceramic sink roll, but experience including cost efficiency is limited.

It is worth noting that it has not been necessary to innovate in induction heating technology for galvannealing AHSS, even though its magnetic permeability is lower than ferrite, since the longitudinal flux technology can operate with up to 50-60% of austenite at the pot. A special design of the induction heater is, however, necessary to heat high austenite steel strip and for steels with higher amounts of nonmagnetic phase, the question may need to be addressed also, despite a mature technology.

If not there may be concerns at the pot equipment in relation to crossbow correction. It is known that corrector roll www.steeltimesint.com

GALVANISING LINES

Fig 14. Patent EP2010690B1

57

Fig 15. Oxidation control in RTF furnace

Fig 16. Device against oxidation at wiping Patent WO 2010130895

316L Sink roll after 6 runs of 4 days in AlSi

Fig 17. number of driven rolls required for exit bridle of Tension leveller for various Tension to YS Ratio in Leveller

Fig Spring effect on recoiler leading on poor grip

penetration increases with strip thickness and hardness and the corrector roll shaft design may need to be reconsidered due to the higher level of stress that it will encounter especially at high tower tensions with thick and wide sheet.

Strip conveying for AHSS The robustness of the strip conveying process is a key issue for any continuous lines. Handling high strength steels do not require specific technologies except the entry welder, either a laser type or conventional one, but with post annealing. www.steeltimesint.com

However, some improvements may be required to be able to manage the higher material strength. For instance: • Shears and side trimmer including compactors and scrap chopper must be powerful enough • Bridles must be verified due to loss of wrapping angle • Skin pass rolling force and tension must be increased: a load of 1,200-1,500t is recommended, including 450mm work roll diameter • Tension leveller requires tensions that are usually much higher than the initial

design. Considering a minimum specific tension of 20% of the yield strength (YS), a DP1000 grade will require a value close to 40t for a 1,500mm wide, 2mm thick strip for only 0.2% elongation. This implies a significant increase in cassette stiffness as well as the number of rolls in the entry and exit bridles. For instance a 1000 MPa grade would require 5 or 6 rolls compared to 3 or 4 for a 300 MPa grade.

• Recoiler with belt wrapper must be designed to ensure that the very stiffness strip will wrap properly and grip the mandrel.

Digital Edition - August 2019

58

GUNNING ROBOTS

FOR IMPROVED

HOT REPAIR of EAF, Ladle, RH

GALVANISING LINES

Fig. 13 shows the case of a stiff sheet that does not wrap due to spring back.

What’s next?

The market has always been driven by overall cost reduction, environmental issues and end use by the final customer. Three product areas at laboratory development stage are: • Austenitic-martensite microstructures, the target being improved formability without reduction of strength • Medium Mn steels with up to 6-8% Mn, and possibly with other additions • Fully bainitic structures known for their excellent elongation-stress performance Conclusions Steel grades requiring galvanising have dramatically changed in the last 20 years, going from simple carbon grades for construction to complex ultra-high tensile strength (1,100-1,220 MPa) DP grades. As a result innovations and improvements were required on the process. Some are related to the heat cycle which requires a certain level of austenitisation, followed by a quenching and eventual aging to let the microstructure develop.

Others are related to the control of the strip surface to have a good zinc wettability and coating adhesion.

However, challenges still exist, especially relating to mass production. Product consistency and reproducibility are essential for success and profitability. Campaign management including transitions and line schedules are also key issues.

The future will probably be just as challenging as it has been in the past, and developments will always find their roots in total cost reduction. The position of AHSS must also be included in the Industry 4.0 concept which consists of intelligent factories with computer models to control processes, maintenance and quality, allowing steel manufacturers to respond quickly to a changing demand and/or environment.

However, innovation and solutions will come from people, thanks to their passion for technology, creativity, knowledge and experience. �

References 1. Improved Properties of Galvannealed Dual Phase 980 Matthew M. McCosby and Eduardo A. Silva Galvanisers Meeting 2012 2. Computer simulation of the selective oxidation of AHSS, D. Liu and T. Cormode –M Kim & J Kim 3. Development of 100% ceramic pot roll for CGL E Oyawa, S Hamayoshi Galvatech 2011

VELCO GmbH Haberstraße 40 42551 Velbert (Germany) info@velco.de • www.velco.de Tel +49 (20 51) 20 87-0

PROCESS CONTROL

59

Through-process optimisation

This article presents an overview of the newly developed through-process optimisation solution. The holistic approach and the combination of metallurgical and operational know-how with an intelligent IT-system allows steel producers to improve their overall efficiency, achieving higher levels of quality and developing and maintaining their know-how basis. By Jan Friedemann Plaul*, Wolfgang Oberaigner*, Yuyou Zhai**, Thomas Pfatschbacher*, Manfred Kuegel*

THE progress of digitalisation in many industries is creating new opportunities for the improvement of overall efficiency and quality in steel production. At the same time, end-customers are demanding steel products at higher quality levels. They look for tailor-made steel-grade solutions, short development times for new steel grades, and the manufacturer’s ability to swiftly respond to quality deviations that

have resulted in rejections. The number of customers requesting zero-defect products is constantly growing. As a consequence of this trend, steel producers need to target even higher quality levels, increased process stability, greater process flexibility, and high production efficiency, in order to be successful at manufacturing challenging products for demanding customers.

This means that steel producers need to

become even more capable and to use a system that ensures accurate and stable control of all process parameters. Producers require quick and complete access to quality- and process-relevant data, a deep understanding of how a change in process parameters will affect the properties of their products, and the know-how to develop products quickly and successfully. To meet these requirements, Primetals Technologies

* Primetals Technologies Austria GmbH **Primetals Technologies China Ltd. www.steeltimesint.com

Digital Edition - August 2019

60

PROCESS CONTROL

THROUGH-PROCESS QUALITY CONTROL

THROUGH-PROCESS KNOW-HOW

• Comprehensive data recording with full product genealogy • Quality control system functionality • Deviation & root cause analysis THROUGH• Corrective & compensational PROCESS actions OPTIMISATION • Automatic product grading • KPI evaluation and visualisation • Know-how rules editor • Statistical process control and intelligent analysis • Interface to data mining platforms

• Know-how expert service • Know-how rules generation • Definition and improvement of key performance indicators • Product development • Quality management • Know-how modules, e.g. metallurgy,operations, quality • Data based know-how generation & analysis • Trainings • Audits & consulting

Fig 1. Overview of through-process optimisation

Hot rolling/plate rolling

Steel melt shop

Cold rolling

Cold band processing Annealing line

converter Ladle furnace

RH plant

Slab caster

Hot strip mill

Picking line tandem cold mill

EAF (EBT)

Galvanizing

Chemical composition Cleanliness

Internal soundness

Homogeneity/segregation

Austenite formation Recrystallazation Precipitation formation Ferrite grain structure Control of r-, n-values Fig 2. Metallurgical know-how along the entire steel production route

Metallurgical know-how best technological properties

• Chemistry • Process parameters • Metallurgical modelling • Metallurgical know-how base • Defects/defect classification (“defect catalogue”)

• Thickness, flatness, profile, roughness • Root cause and remedial actions

Operational know-how stable production

• Operational practice • Campaigning • Preventive-maintenance practices • Consumables, energy efficiency

Material-development know-how

Sustainable high quality products

• Material-grade & process development • Simulators & systematic simulator use • Know-how base, and partnership with suppliers and steel companies

Quality management & process know-how Sustainable strategy & culture

• Quality-management consulting • Feasibility & improvement studies • Process & organisation consulting

Fig. 3 Through-process know-how modules

THROUGH-PROCESS KNOW-HOW

• Surface quality

Quality know-how stable quality properties

Properties of final product

Batch annealing

has developed and introduced the ThroughProcess Optimisation (TPO) solution, which targets the smart, digital interconnection of various process units and the accumulation of know-how along the entire steelproduction chain.

Introduction Today’s steel producers have to perform in a highly challenging market environment that is marked by declining raw material quality along with price volatility, increasing pressure from stricter environmental legislation, and fierce market competition due to excess capacities all around the world.

These challenges force many steel producers to thrive into the market of high-value and high-margin steel products like automotive grades. When moving from commodity grades and mass products to high-quality and advanced high-strength products, steel producers have to address the following issues: • Stable production with highest quality standard, minimisation of rejection rate: because of the tighter production tolerances and lower process reliability under such conditions, a low rejection rate becomes a highly important prerequisite for a profitable production.

• Extension and intensification of quality and research activities: in order to reduce the rejection rate while maintaining cost effective production, it becomes favourable to focus R&D and product development efforts and to specialise in certain product areas.

• Comprehensive know-how build-up, documentation and protection measures, which are difficult to deal with without adequate IT-system support.

• Become a learning company that is among the fastest to learn and improve, in all terms of production, automation, technological and internal processes.

In order to reduce the rejection rate and to provide support for product and production development, Primetals Technologies introduced its new knowhow based Through-Process Optimisation Solution (Fig. 1) [1,2,3], which consists of two parts: 1.

A new, intelligent through-process quality control (TPQC) system, and 2.

The through-process know-how (TPKH) packages 1. Through-process know-how 1.1 Metallurgical know-how for quality control,

Digital Edition - August 2019

www.steeltimesint.com

PROCESS CONTROL

61

Level 4 (Entreprise Resource Planning System) Level 3 (Manufacturing Execution System, APS, ODS) Through-Process Quality Control (TPQC-System) L2

L2

L2

L2

L2

L2

L2

L2

L1

L1

L1

L1

L1

L1

L1

L1

HM De-S

LD (BOF) converter

Ladle furnace

RH-OB plant

Slab caster

Hot strip mill

Pickling and tandem mill

Galvanizing

Product yard management

Coil storage

Smart Sensors/Robotics/Tech. Packages/Quality Inspection Systems/CMS Maintenance Asset Technology

Fig 4. Integration of TPQC with automation and IT-systems

Slab 1

Slab 2

Finished coil

Hot coils

Fig 5. Genealogy based data tracking across multiple processing units

HM De-S

LD (BOF) converter

Process parameter(s) threshold upper limit

Process parameter(s) threshold lower limit

Ladle furnace

RH-OB plant

HM De-S Process parameter(s) collection

Process parameter(s) evaluation

Fig 6. Quality control function along the production route

www.steeltimesint.com

Slab caster

Hot strip mill

Galvanizing

rules generation and KPI definition For production of high value added steel products many hundreds of properties, parameters and conditions have to fit together and be aligned to achieve good product quality and a cost efficient and stable production (Fig. 2). To develop rules which are controlling the process and quality parameters of the product, a deep understanding of all metallurgical reactions and complex physical processes is mandatory. In a first step metallurgical experts are defining for each process unit a ‘basic set of rules,’ which will be implemented in the TPQC system by using a special rule editor. Beside the development of formulas for the rules, the experts are also implementing a set of possible route causes in case of deviation as well a set of corrective actions for the individual process unit.

In a second step a set of compensational actions is defined for each type of deviation, which can be applied for downstream process units. The intelligent design of the rules enables an automatic evaluation of possible route causes by the TPQC system.

It is also possible to define a different set of rules for individual steel grades or groups.

To measure the efficiency and the process stability for each process unit, key performance indicators (KPIs) are developed by experts and implemented in TPQC. Together with the plant management the target setting for each KPI is agreed. During regular review meetings the performance of the individual KPIs are verified and action plans are developed and agreed to Digital Edition - August 2019

62

PROCESS CONTROL

List of possible root-causes Most likely root-cause on top

time TPQC keeps track of most frequent root-causes and allows for trend analysis with respect to continual improvement

Manual root-cause confirmation increased root-cause awareness root-cause documentation

Documentation about most likely root-cause to sort possible root-causes by their documented frequency

Fig 7. Semi-automatic root-cause analysis support

improve the performance. The KPI tracking can, therefore, be an important part of the continual improvement process (CIP) of a steel plant.

1.2 Expert service For the implementation of TPO and in order to resolve specific problems, Primetals Technologies supports its customers with its own experts and with external consultants. These specialists are highly experienced and cover various disciplines, acting as consultants for a wide range of topics; for example, plant operation, quality management, maintenance, or endcustomer qualification (Fig. 3). Primetals Technologies also provides training sessions for steel producers who decide to implement TPO. In close co-operation with the respective customer, the team of experts devises and incorporates an optimal set of rules into the customer’s TPQC system.

These rules will ensure better control of the steel-production process, optimise productquality consistency, and improve general plant operation.

2. Through-process quality control system The basis of the TPO solution is the through- process quality control (TPQC) system, which creates a central database by receiving quality- and process-relevant production data from all production units via the Level 1 or Level 2 automation systems. Additionally, laboratory measurements and data from all types of sensors and equipment are centrally stored.

Digital Edition - August 2019

Target temp 1620° C

1320° C

Blast furnace

Hot metal

LD (BOF) converter

HD de-S

DeS

Target temp 1680° C

DeC BOF #2/3

1620° C

1585°C

1560°C

Ladle furnace

RH-OB plant

Slab caster

RH-OB

CCM reduction of casting speed within limits to avoid a strand-break

LF

Immediate compensational Temperature action: reblow at BOF too low TPQC identifies possible root-causes Corrective action Check/change material related of new material in use

Input material (scrap, hot metal)

analysis, temp, weight)

Input of additions, cooling, heating agents (weight) as calculated L2/offgas model working/in use New material quality in use

Fig 8. Root-cause analysis within the melt-shop

The TPQC creates an information-rich genealogy of each individual product that is processed, and makes it possible to retrieve process data of all production steps for every part of the product. This allows users of TPQC to track quality issues in very little time and analyse them by reviewing process data for all relevant production steps, which is key for fast troubleshooting and claim management. Fig. 4 shows the TPQC System embedded into the automation and IT environment.

The essential functionality of TPQC is to ensure desired product properties and increase quality levels by monitoring all quality-relevant process parameters along the full production route at defined quality

gates. Therefore, communication between the production-management system (PMS) and individual automation systems of each production unit is required to evaluate distinct steps of the production process.

2.1 Product genealogy The TPQC automation system is collecting all relevant process and production data provided by the various plant units of the whole process chain from ironmaking through hot rolling, cold rolling, annealing and galvanising (Fig. 5). In addition to pure process data, which is collected from various processing unit, the genealogy information interconnects this data across all involved processing units. For example, www.steeltimesint.com

PROCESS CONTROL

63

improvement (Fig. 7). In order to support quality engineers and operators in respect of root-cause analysis across multiple production processes, TPQC provides a semi-automatic approach:

Fig 9. Quality assistance

the genealogy keeps track of elongation factors, head/tail changes, upside/downchanges as well as cutting and welding operations that are possible between different processing units of the production chain.

The genealogy information supports claim management, since no important information would be lost any more. Even after years it will be possible to deal with a customer claim since heat number, slab number and coil numbers, even daughter coils and sheets, can be identified and any relevant deviation or defect before delivery may have been logged in the long-term archive of TPQC. 2.2 Quality Control System Functionality Quality conformance checks are carried out by means of a specific rule system on the production data, and the results will be shown to operators and quality engineers, respectively, depending on the kind of quality issue and location in the plant. By means of the rule editor TPQC offers a flexible way to make quality checks, which paves the way for future adaptions and extensions without having to change any part of the system’s source code. The rule system of TPQC is also an important pillar for Primetals Technologies Through-Process Know-How described above.

Fig. 6 shows the principle of this process/ quality checking inside TPQC. A green traffic light indicates that all values (of a production unit) are inside the specified bounds. A yellow should express the detection of minor problems, requiring www.steeltimesint.com

some further checks. A red light should indicate more serious issues, which in most cases need a more complex remedial action and final decision plan to resolve the problem, which can even include a dismissal of the product in the worst case. What can be defined as minor and major has to be developed from the process experience and must be finally specified as expert rules.

2.3 Deviation and Root-cause analysis Root-cause analysis (RCA) [4] is a method of problem solving used for identifying the root-causes of faults or problems, well known from literature and required by the ISO/TS standards for automotive production [5]. In general, RCA is often addressed by several methods and has to be applied iteratively as a tool of continuous

1.

In case of a detected deviation, a list of highly probable root-causes is shown to the operators or quality engineers depending on the plant location and organisational responsibility.

2.

Each root-cause comes with a specific description for root-cause verification, in order to eventually remove any doubt in cases in which more than just one root-cause might be possible.

3.

After identification of the actual root-cause, the responsible person places a check-mark and thereby confirms the rootcause for the given quality deviation.

4.

The system keeps track of the root-causes for any detected quality deviation and calculates a root-cause statistic.

The documentation of root-causes for identified quality deviations is mandatory in case of some quality standards like ISO/TS 16949 [2]. In addition, a root-cause statistic can be calculated for an arbitrary time period and thereby enables the calculation of trends over time. Hence, the root-cause statistic feature is a valuable aid to provide convincing evidence on the effectiveness of the quality management with respect to continual improvements. In particular, it is a useful tool for the plant manager, enabling a strict monitoring of deviation frequencies, in order to identify the most

Fig 10. Various surface defects scattered across a hot rolled coil

Digital Edition - August 2019

64

PROCESS CONTROL

Fig 11. KPI visualisation

frequent and most costly deviations. These identified deviations may then be addressed by additional problem solving methods and corrective actions (see below), if adequate.

2.4 Corrective and compensational actions A corrective action can be defined as a set of actions to eliminate the cause of a quality deviation under specific conditions.

Unfortunately, the production environment is subject to a large number of influences like any kind of external disturbances, raw material unknowns as well as the human factor and other non-deterministic events.

Hence, it is seldom possible to eliminate the cause of a quality issue permanently.

For this reason, the definition of corrective action has to be broadened in so far as a certain corrective action for a given rootcause may eliminate a quality deviation absolutely and permanently only under the specific condition that applied previously, although not all of them can be determined or measured and archived.

A compensational action is defined as an action to repair an already affected semi-finished product, e.g. by cutting out defective sections from a strip or by means of surface scarfing of a slab as indicated in the example below. Because of the fact that the system provides guidance to operators and quality engineers, the tool can also be seen as a learning tool. It is a fact that people who are actively involved in root-cause analysis for some time will become more aware regarding quality related influences, which allows them to take preventive measures even before a quality related incident happens, that is without having to wait for TPQC rootcause suggestions at any rate and for any incident, respectively.

Thus, TPQC can be used as a pure Digital Edition - August 2019

conformance tool for quality checks and root-cause analysis support, but the real benefit of this system is realised when it is also used as a continual learning tool in order to improve operator and quality engineer skill levels. The next easy example in Fig. 8 deals with quality issues in the melt-shop.

This Fig. 8 illustrates the example of a produced heat, which is planned to arrive subsequently at a continuous casting machine.

At the end of the BOF process, the L2 system indicates a temperature of the liquid steel that violates the grade specific requirements. In this case, TPQC evaluates the temperature by means of process specific rules and records a quality issue.

In addition, the TPQC system immediately suggests a re-blow at the BOF and if the operator confirms this, a message is sent to the operator at the continuous casting machine to reduce the casting speed. By reducing the casting speed, TPQC tries to compensate for the additional processing time caused by the re-blow at the BOF,

which otherwise would increase the risk of a sequence break or slab quality issues.

In addition, the system indicates again a list of likely root causes for this problem.

In the example case, a new material had been used for the material additions. The operators, for example, had not managed to update the L2 data base with the new material data in time, which led to an inaccurate calculation of the L2 model for the BOF process and finally caused the temperature deviation.

Fig. 9 shows a screenshot of how TPQC indicates root-causes for recorded quality issues to quality engineers. The list on the left hand side of the screen shows all semifinished products for a given day (in this case filtered for heats). The selection of a heat with recorded quality issues (indicated by the call sign) shows all recorded quality alarms/issues. In this case there are two quality alarms. By selecting a quality alarm, the instructions for verification as well as root causes are shown in the text window at the bottom of the screen.

Even though the given example shows a root-cause analysis for a single process unit, this kind of root-cause analysis is not restricted to single processes or plant areas (melt-shop, hot strip mill, cold mill etc.). By means of carefully specified rules for root-cause analysis, TPQC can also indicate root-causes across process or plant boundaries. In this way, the system supports through-process root-cause analysis, which is especially useful in case of certain surface defects that can be tracked back to the liquid phase. Surface defect rules as well as surface grading will be treated in section 2.5 below.

Fig 12. SPC standard diagram

www.steeltimesint.com

PROCESS CONTROL

1. Problem Definition Why do some grades have a high number of surface defects per strip

2. Data Understanding Are these missing values, constant values, outliers

3. Models Applying machine learning methods to isolate root causes

65

4. Derive actions Create tasks and implement new rules. KPIs and SPC charts

rolling time RHF furnace gas pickling speed

Fig 13. Example for data mining project and definition of actions

2.5 Automatic product grading TPQC incorporates a rule-based module for coil-grading. The coil-grading module pre-processes surface defect maps.

Afterwards, the grading-module performs an in-depth evaluation of the defect map, in order to determine instantly whether a coil surface quality matches the pre-defined requirements of the end-customer, and what action to take if this is not the case.

2.6 Key Performance Indicator evaluation and visualisation The centralised collection of data enables the generation of key performance indicators (KPIs), which convey information about technical and business-related achievements and illustrate what progress has been made. TPQC implements various types of graphical human-machine interfaces to support staff members from the quality and production departments, as well as top-management executives, in monitoring and benchmarking production conditions with respect to specific targets that are in alignment with the KPIs. 2.7 Statistic-process control (SPC) SPC [6] is a reliable and proven tool to provide statistical evidence that a production process stays within its predefined operational range and, therefore, behaves in a controlled way.

There are significant advantages of having this statistical process control applied to quality assurance processes, since sooner or later statistical significant deviations of process values will have an impact on product quality. In general, SPC aims on sampling data which is meant to be stable www.steeltimesint.com

over a long period of time in order to allow for a reliable detection of undesired process dynamic. TPQC can follow up single process criteria but also combined calculated performance indicators and KPIs, and offers SPC charts for selectable material/product related measurements or process data.

Fig. 12 shows a typical SPC control chart with a list of subgroups. Each subgroup which is exceeding the control limit is indicated by red or blue according to a violation of the upper or lower control limit.

2.8 Machine learning capabilities The extraordinary large amount of structured data in the TPQC system is perfectly suited for analysis with data mining and machine learning algorithms.

TPQC offers a direct interface to transfer data effortlessly to leading data mining platforms. All acquired data is automatically assigned to the correct products along the process chain (steel making heat, slab, hot rolled strip, cold rolled strip, galvanised strip) by the genealogy function in the TPQC. Data mining and machine learning in combination with TPQC offers the following benefits: • Fast analysis for data quality and potential problems of measurements, manual inputs, etc.

• Quick analysis of multi-variate problems such as variation of mechanical properties by visualising the raw data from TPQC • Root cause analysis of problems using through process data and applying advanced data mining methods • Predictions using machine learning methods for end-of-line properties of products or process stability

Closing the loop by creating new rules, KPIs or SPC charts in TPQC based on the data mining results. The data mining functionality of TPQC supports quality and process engineers in optimising the product quality and stabilising the production process. It is an extremely useful tool to increase the productivity of technologists and engineers throughout the process chain. �

References [1] Kurka, G.; Hohenbichler, G.: TPQC – Through-Process Quality Control, CISA 2016, Beijing [2] Kurka, G.; Hohenbichler, G.; Pfatschbacher, T.; Pichler, L.: Know-how based root cause analyses tool to ensure high product quality and process; ESTAD 2017, Vienna [3] Ringhofer, M.; Wimmer, G.; Plaul, J.F.; Tatschl-Unterberger, E.; Herzog, K.: Digitalsierung in der Stahlindustrie, Stahl und Eisen 137, 2017, Nr. 5 [4] Wilson, Paul F.; Dell, Larry D.; Anderson, Gaylord F. (1993). Root Cause Analysis: A Tool for Total Quality Management. Milwaukee, Wisconsin: ASQ Quality Press. pp. 8–17. ISBN 0-87389-1635 [5] ISO/TS 16949(E), “Technical Specification: Quality Management Systems – Particular Requirements for the Application of ISO [6] Statistical process control – Reference manual, second edition, DaimlerChrysler Corporation, Ford Motor Company, and General Motors Corporation, 2005 Digital Edition - August 2019

66

DIGITAL MANUFACTURING

Securing operational technology in the

Eitan Goldstein* explores the real-world cyber challenges metal manufacturers face in securing operational t THE metal manufacturing industry, like many others, faces increased cyber attacks.

While there are numerous theories as to why this is, what isn’t disputed is that the threat is real and organisational defences are falling. Back in 2014, a cyber attack against a German steel mill saw attackers use emails to steal login details that allowed them to gain access to the mill’s control systems with devastating consequences.

ThyssenKrupp, one of the world’s largest steel makers, confirmed it was victim to a cyber industrial espionage attack in 2016.

More recently, Norsk Hydro battled to contain a ransomware attack, highlighting how an attack against IT infrastructure impacts operational technology (OT) systems and ultimately the organisation's ability to function.

An independent study, conducted by Ponemon Institute on behalf of Tenable,

found that 90% of organisations had experienced at least one damaging cyberattack over the past two years and 62% had two or more. The study — based on a survey of 701 cybersecurity and IT professionals whose organisations fall into the OT sector (described as industries dependent on industrial control systems (ICS) and other operational technology) — further reveals that these attacks resulted in data breaches and/or significant disruption and downtime to business operations, plants and operational equipment.

The threat is real and the study found organisations expect it to worsen. Survey respondents expressed concerns about third parties misusing or sharing confidential information and OT attacks resulting in downtime to plant and/or operational equipment increasing in the year ahead.

Perhaps of greater concern is that so-

called ‘bad actors’ may not just be looking to cause disruption, but could be intent on physical harm. In 2017 a new malware was discovered that was designed not only to damage systems, but to potentially maim humans. The malware, named Triton, and first discovered in a Saudi Arabian petrochemical plant, actively targeted safety instrumented systems.

Digital transformation at what cost?

Industry 4.0 has revolutionised manufacturing plants. Business leaders order the implementation of new technological tools that benefit the bottom line through enhancing efficiency and output. The result is that the data side of the business — traditionally the realm of IT — and the operational technology (OT) side — used to manage industrial systems (ICS) — have merged. This convergence is

* Senior director of business development, Tenable Digital Edition - August 2019

www.steeltimesint.com

DIGITAL MANUFACTURING

67

e metal manufacturing environment

technology busting open previously siloed departments and exposing the organisation to new threats. As has been documented by a number of recent breaches, OT attackers will often use credentials gained in IT environments to pivot into and attack OT infrastructure.

Caught in the middle are the IT and security professionals tasked with realising the ambition, securely.

In the first four months of 2019 the US Industrial Control System-Computer Emergency Response Team (ICS-CERT) issued 74 alerts for vulnerabilities in industrial control systems (ICS). ABB, AVEVA, Mitsubishi, Omron, Rockwell, Schneider Electric, Siemens and Yokogawa were some of the system manufacturers impacted by these vulnerabilities. This includes a remote code execution in AVEVA’s InduSoft Web Studio disclosed by Tenable Research in February. In the same four-month period, 2,817 information www.steeltimesint.com

technology vulnerabilities were also recorded on the website CVE Details.

Given that both OT and IT vulnerabilities, if exploited, pose a very real threat to metal manufacturing technologies and the environments they’re utilised within, that’s a total of 2,891 new vulnerabilities in just the last four months for the IT and/or security team to grapple with.

As the security programme lead of a large oil and gas company in the US recently told me: “Business has made the decision to implement new OT and we’ve got to catch up and secure it. The business arguments are too compelling. It’s about enabling business leaders to do what they want to do.”

Visibility is crucial Staff responsible for OT security cannot afford to be blinkered and focus only on OT vulnerabilities. IT and OT convergence means that both OT and IT vulnerabilities

can be exploited to attack critical infrastructure as already witnessed. Clear, and complete, visibility of both IT and OT systems viewed together through a single pane of glass is the only way to gain a holistic view on risks.

For many within the metal manufacturing sector, as is equally true of many others, full visibility of the network including OT infrastructure isn’t straightforward.

Referring back to the Ponemon study — using a five-point scale of strongly agree to strongly disagree — only 20% of OT sector respondents agree or strongly agree that they have sufficient visibility into their organisation’s attack surface. There are many reasons for this — a company could have numerous sites around the country or around the world, or its technology inventory could still exist in silos due to organisational mergers and acquisitions spanning many years, among many other factors. Regardless of the cause, it is Digital Edition - August 2019

68

DIGITAL MANUFACTURING

essential for metal manufacturers to take the necessary steps to develop a clear view of their networks, as a complete hardware and software inventory is fundamental to all security frameworks and compliance requirements including the CIS Controls, NIST Framework for Improving Critical Infrastructure Cybersecurity and NERC CIP. Once true network visibility is achieved, the next steps are to determine what is essential to the organisation’s ability to function and then to ascertain how vulnerable the operation is to attack. The only effective way to accomplish this is with a unified strategy that spans the IT and OT sides of the organisation. In my experience working with manufacturing organisations of all sizes, those with separate, siloed OT security programmes — operating with different tools, different KPIs and different policies than their IT security programmes — will not stay safe for long in today’s threat landscape. Traditional ways of securing systems through Excel

Digital Edition - August 2019

spreadsheets or tribal knowledge are, quite simply, insufficient for securing organisations against the modern cyber threat landscape.

Where do we go from here?

An awareness among IT and OT professionals of the threat landscape is essential if organisations are going to reduce their cyber risk. But it is not only those on the ground level who need awareness of the risks facing OT environments. It’s imperative for business executives, the C-suite and even the board of directors to understand the full scope of the cyber threats their organisation faces. Thankfully it would appear that this message has been heard. The Ponemon study found increasing communication with the C-suite and board of directors about cybersecurity threats facing the organisation is the number one priority going forward.

It’s important for metal manufacturers to also consider the physical safety

implications of cyber threats. Physical safety is the number one priority for organisations that rely on OT technology — both the safety of the workforce and of those who live and work in the vicinity.

Given the real threat to life, manufacturing organisations cannot afford to have a well-intended security update backfire, nor can they permit a threat actor to breach their OT systems. OT cybersecurity becomes a balancing act between achieving continuous uptime and effectively securing essential systems.

Effectively securing connected OT and IT environments is a work in progress, not something that will be fixed overnight.

As digital transformation continues to result in the convergence of OT/IT environments, I’m seeing organisations that rely on OT technology beginning to acknowledge the challenges. From there, we can all work together on solving the cybersecurity issues facing metal manufacturers. �

www.steeltimesint.com

CONNECTING THE INTERNATIONAL STEEL INDUSTRY

Steel Times International Directory The Directory is the essential guide to steel manufacturers, producers, suppliers of plant equipment and services to the steel industry. It provides comprehensive listings, product information and key contact

Order your copy now from just £95*

details.

For more information on advertising or how to be listed contact: Esme Horn esmehorn@quartzltd.com Tel: +44 (0) 1737 855 136

@SteelTimesInt

STI_Buyers_Guide_Half_Vertical.indd 2

WWW.STEELTIMESINT.COM 25/02/2019 11:05

ED E IT AT M G S LI LE CE E A D SP

INDUSTRY 4.0 AND THE STEELMAKING PROCESS 25-26 SEPTEMBER 2019 BUDAPEST • HUNGARY If you want to know what’s happening in the world of ultra-high technology and the production of steel, then look no further than the only steel conference in the world dedicated 100% to Industry 4.0 and how it – and its related technologies – can aid and optimise the steel manufacturing process.

DELEGATE REGISTRATION The conference fee includes a 2-day conference programme, refreshments, a networking lunch and conference proceedings. Register online to subscribe to the Future Steel Forum membership package which will include a subscription to Steel Times International, the Steel Times International Directory, relevant news alerts and admission to the Future Steel Forum, 25-26 September 2019, Sofitel Hotel, Budapest, Hungary.

Delegate registration fees for this conference are as follows:

Join us in Budapest, Hungary on 25-26 September 2019 to discuss, debate and discover the latest hard hitting topics that are facing steel manufacturers in this digital age. Visit the website to view the full conference programme www.FutureSteelForum.com

Standard rate

£990

Group bookings (5+ delegates)

£792 (per delegate)

Please note the conference fees will also be subject to value added tax where applicable.

SPEAKERS INCLUDE:

Official Media Partner

@Future_Steel

Join our Future Steel Forum Group

1 866 2019

Organised by:

Dr Svend Lassen, Tata Steel Europe

FSF_DPS_Delegate_Ad.indd 2-3

Dr Marcus J. Neuer, VDEh-Betrebsforschungsintitut

Diego Diaz, ArcelorMittal

Stephen Pratt, Noodle AI

Cyril Peillon, Fives Group

Farrokh Mistree, University of Oklahoma

Dr MilisavljeviceSyed, University of Liverpool

Dr Nils Naujok, PwC Strategy&

www.FutureSteelForum.com 22/07/2019 10:50

ED E IT AT M G S LI LE CE E A D SP

INDUSTRY 4.0 AND THE STEELMAKING PROCESS 25-26 SEPTEMBER 2019 BUDAPEST • HUNGARY If you want to know what’s happening in the world of ultra-high technology and the production of steel, then look no further than the only steel conference in the world dedicated 100% to Industry 4.0 and how it – and its related technologies – can aid and optimise the steel manufacturing process.

DELEGATE REGISTRATION The conference fee includes a 2-day conference programme, refreshments, a networking lunch and conference proceedings. Register online to subscribe to the Future Steel Forum membership package which will include a subscription to Steel Times International, the Steel Times International Directory, relevant news alerts and admission to the Future Steel Forum, 25-26 September 2019, Sofitel Hotel, Budapest, Hungary.

Delegate registration fees for this conference are as follows:

Join us in Budapest, Hungary on 25-26 September 2019 to discuss, debate and discover the latest hard hitting topics that are facing steel manufacturers in this digital age. Visit the website to view the full conference programme www.FutureSteelForum.com

Standard rate

£990

Group bookings (5+ delegates)

£792 (per delegate)

Please note the conference fees will also be subject to value added tax where applicable.

SPEAKERS INCLUDE:

Official Media Partner

@Future_Steel

Join our Future Steel Forum Group

1 866 2019

Organised by:

Dr Svend Lassen, Tata Steel Europe

FSF_DPS_Delegate_Ad.indd 2-3

Dr Marcus J. Neuer, VDEh-Betrebsforschungsintitut

Diego Diaz, ArcelorMittal

Stephen Pratt, Noodle AI

Cyril Peillon, Fives Group

Farrokh Mistree, University of Oklahoma

Dr MilisavljeviceSyed, University of Liverpool

Dr Nils Naujok, PwC Strategy&

www.FutureSteelForum.com 22/07/2019 10:50

72

REPUTATION MANAGEMENT

Keeping up appearances Clear communication as part of a careful reputation management strategy has not always been seen as a steel industry strength, but in today’s rapidly changing world dominated by social media it could not be more important.

Digital Edition - August 2019

www.steeltimesint.com

REPUTATION MANAGEMENT

www.steeltimesint.com

73

Digital Edition - August 2019

74

REPUTATION MANAGEMENT

WITH steel in the headlines so much recently, general interest in the industry is at its highest in recent memory. This provides an unrivalled opportunity to remind a notably receptive global audience of the vital role that steel plays in their lives.

While the often-conservative nature of engineers can on the one hand prove a challenge, on the other hand it means that in general the steel industry avoids two of the biggest no noes of reputation management. Engineers, like doctors, are reared on critical appraisal of data and intensive peer review. The steel industry is not given to bluff and bluster, but calmly and intelligently reviewing problems and considering all possible solutions. The industry also tends to steer away from the hyperinflated language so prevalent in the FMCG markets. How many of us have had to suffer train or plane delays and cancellations only to be told that the errant company has a “passion” for us the customer. Really!

The World Steel Association (worldsteel) carries out regular research into the global reputation of the steel industry across various audiences including customers, governments, NGOs, the media and the general public. Reassuringly, the results of this research show us that both steel Digital Edition - August 2019

as a material and the steel industry itself are largely favourably viewed. Global audiences enthusiastically agree that steel is inextricably linked with economic growth and development and know that it is infinitely recyclable without any loss of properties. Indeed, we have seen that there is a strong correlation between how much people know about steel and how favourably inclined they are towards it.

Arguing the case for steel forthrightly is more important than ever as requirements from wider society become more exacting, and competitive industries improve their products and become better at communicating their benefits.

The diminishing social license to operate of the plastics, coal and oil industries is a harsh warning of what can happen when negative stories start to dominate the agenda. However, the up side to the attention is the opportunity it provides. And only a brief look at the worldsteel website shows how seriously the industry takes its responsibilities.

The number one priority for the industry is safety and health, which must not be compromised in the face of other business issues and where any suggestion to the contrary must be immediately stopped.

Steel Safety Day was set up in 2014 by

worldsteel to reinforce awareness of the industry’s five most common causes of Lost Time Injuries (LTIs), defined as an injury that prevents an employee from returning to work for their next scheduled shift. On this day worldsteel encourages members and non-members alike to carry out safety audits at their facilities focusing on that year’s theme. Every year, Steel Safety Day audits are carried out on or close to 28 April so as to be aligned with the International Labour Organisation’s World Day for Safety and Health at Work.

One of the most encouraging achievements in recent years is the steep decline in the LTI Frequency Rate, which was 5.4 LTIs per million hours worked in 2007 and has since declined steeply to 0.97 in 2017. While this is welcome, there is still much more work www.steeltimesint.com

REPUTATION MANAGEMENT

to do. Our reputation in this area requires that we are not seen to be complacent or self-congratulatory, but that we are always striving so that injuries and occupational illnesses are not considered an unavoidable part of an historically dangerous industry, but as something to be eliminated.

Climate change is one of the other urgent challenges for the steel industry. The 2015 Paris Agreement stipulates that to prevent global temperatures from rising above 1.5°C the planet must be carbon neutral by the middle of this century. We estimate that on average the production of a tonne of steel produces 1.83 tonnes of carbon, meaning that as the world produced 1,808.6 million tonnes of steel, carbon emissions from the industry stood at 3,308.4 million tonnes, between 7% and 9% of the global total for that year.

Many steelmakers have made major announcements on

breakthrough technology investment and planning using 2050 as the common target date. But there is also a lot that can be done outside of breakthrough technology in the intervening time and this is where worldsteel’s newly-launched step-up programme will prove most valuable.

The programme covers raw materials, energy input, yield and process reliability.

The idea is that worldsteel will share leading practices from among the best performing of its member company sites with the weaker sites so that the efficiency levels of the steel industry are brought up to the level of the top performers with a resultant decrease in industry carbon emissions. The programme will be tested across nine sites in 2019 and then rolled out much more widely in the following years. Worldsteel is inviting all of its members to take part in this programme and expects a high level of participation across all regions. As such, we hope that this will develop into our flagship climate change programme.

On the topic of environmental sustainability more generally, in 2017

More information about all of worldsteel’s programmes can be found at www.worldsteel.org www.steeltimesint.com

75

worldsteel launched a recognition programme, Sustainability Champions, to encourage steel companies to increase their efforts, set higher standards and make further progress on the wider range of subjects that can be included under the sustainability umbrella. To be recognised, worldsteel member companies must submit comprehensive data on indicators including the amount of revenue spent on research and development, the amount of working days dedicated to employee education and training, and Life Cycle Inventory data that should cover more than 50% of the company’s crude steel production and be no more than five years old. Last year worldsteel recognised six of its member companies (ArcelorMittal, JSW Steel, Tata Steel Europe, Tata Steel Limited, Tenaris and Ternium) as Sustainability Champions with several more only just missing out.

Clearly, effective communications with a wide range of stakeholders, not least steel industry customers and investors, are a crucial part of a successful steelmaker’s reputation management. In addition to the primacy of safety and health and the industry’s response to increasingly stringent environmental regulations, wider supply chain transparency and the forces of rapid digitisation and automation are issues that the steel industry must be seen to be continuously engaging with in new and inventive ways to enhance its reputation as a leading industry. The programmes mentioned here are those that deal specifically with corporate reputation, but there are other worldsteel programmes with a more specific focus that will continue to develop to maintain steel’s reputation in important customer sectors where alternative materials exist; examples include constructsteel and WorldAutoSteel, which are more technical, and which elaborate on the benefits of using steel in the construction and automotive markets.

All of this will involve managing reputation in different ways as communications channels continue to evolve, particularly in the direction of digital media, which the industry needs to embrace in all its markets to ensure it can compete for share of voice. The inherent properties of steel are an advantage in this task, but the industry cannot ignore the risk of complacency. As mentioned before, the precipitous decline in the reputation of the plastics industry is instructive. �

Digital Edition - August 2019

76

HISTORY

Iron ore mine pits revealed during surface stripping

Shelly Cyrena limestone with a layer of

Mining ore on the Weald A recent archaeological excavation in advance of a housing project near Horam in the Sussex Weald of Southern England has revealed a close concentration of several hundred shaft pits for mining ore.

By Tim Smith*.

Although no dating evidence has been recovered, the shaft pits are thought to have been dug to supply nearby Waldron Furnace located 2km to NW. Documentation shows Waldron furnace operated for over 100 years (1560 to 1758) and among the many products it cast were included the most demanding, cannon.

The dominant ore on the Weald is Siderite, (iron carbonate) assaying at around 40% Fe content. It should be remembered that far lower grade ores than this were still in use in the UK into the 1980s, for example the ironstones of Northamptonshire averaging 30% Fe fed furnaces at Corby (STI April 2018 p76). The Oolitic ironstones of Rutland averaged 19-40%Fe (STI Jan/Feb 2019 p48) and were sent to Scunthorpe in Lincolnshire, while Lincolnshire Frodington ores averaged 33% Fe. The Cleveland carbonate ores of NE Yorkshire, which dominated domestic ore production, averaged just 27% Fe. Siderate ore in the Weald occurs in discrete bands, typically ranging in thickness from 13 to 450mm and occurring at the junctions of layers of Wadhurst clay and Ashburnham sand. Over millennia, up to seven bands of ore have been formed, in places to a depth of 20m. Depending on the local geology, pits vary in depth from 12m to 2m with top diameters from 1.8m to 3.5m. Some pits have vertical sides, others taper towards the bottom, while some widen slightly towards the bottom, but the unstable nature of the iron bearing strata means none expand at the base to form bell pits as was common for prehistoric flint extraction in chalk. Carbon 14 dating shows pits were dug from at least the 12th Century onwards to serve the simple hand blown bloomery furnaces, while others date from the early 16th Century when the blast furnace reached the Weald.

Excavation of some of the pits at Horam

Nodule of weathered iron pyrites

reveal a self-fluxing ore with a layer of siderite adhering to Cyrena (now renamed Neomiodon) limestone consisting of bivalve shells laid down in brackish water.

The earliest ore extraction in the Iron Age, some 500BC in Britain, was probably from stream beds and banks where fast flowing streams cut deep ghylls through the strata.

Indeed, most bloomery sites found in the Weald are alongside such streams.

Industrialising production Later, the Romans came and industrialised the production of iron, exporting billets to the Continent and for local use such as building palaces and farmsteads. The Romans needed so much ore that they mined by quarrying rather than extracting from pits. Then came the Saxons in early medieval England who seemed to have reverted to pit mining. The Norman invasion of 1066 extended the industry, demanding such products as armour, weapons, horse and ox shoes, but output per furnace was still measured in kilograms of a solid mass of spongy iron. As these bloomery furnaces grew larger, bellows were driven from a water wheel rather than by hand. It was not so much any increase in temperature of these larger furnaces that led to the accidental production of molten iron, but rather an increase in the retention period of the sponge iron formed higher up the furnace shaft which raised the carbon content of the iron and hence caused the melting point to fall from 1550°C for pure iron to around 1130°C for an iron containing 4.3% carbon.

In 1490, the first true blast furnace in England arrived on the Weald in the form of the Walloon technology brought from the Low Countries of present day France and Belgium.

Ore was extracted by both pit and quarry excavations, the latter sometimes also serving to provide marl, a lime rich clay

siderite ore

Nodule of haematite ‘kidney’ ore

which was spread on fields to fertilise them.

Documentation shows that pit mining continued with various leases from land owners allowing the extraction of ‘mine’, as the ore was called, and how the pits were to be filled in afterwards ensuring that top soil was reinstated at the surface.

Iron Masters demanded ore from the deeper layers where the silica content was generally lower reducing the amount of slag produced and hence fuel, which was charcoal, consumed.

Occasionally other ores were used.

There is evidence that some small Roman furnaces, probably built to supply iron during the construction of a farmstead or palace, used iron pyrites nodules which form in underlying chalk. Normally, the high sulphur content of such an ore (45%) would make the iron hot-short, but analysis has shown that nodules found in surface soils have weathered to such an extent that all the sulphur has been removed leaving a rich oxide ore of 69% Fe content – even richer than today’s imported haematite ores. Even more extraordinarily, we occasionally find nodules of haematite ‘kidney’ ore, on the Weald, which is normally associated with the past volcanic activity of upland Britain. We conjecture that these have been brought to the area by melt water following one of the several glacial periods that covered much of Britain, but no further south than present day London.�

*Secretary Wealden Iron Research Group www.wealdeniron.org.uk Digital Edition - August 2019

www.steeltimesint.com