Developing Smarter Steel with Artificial Intelligence Powered Software Analysis
Steel is critical to the future success of our world. As one of the only materials to be completely reusable and recyclable, it will play an essential role in building the circular economy of the future. Steel continues to evolve, becoming smarter and increasingly sustainableSteel is critical to the future success of our world. As one of the only materials to be completely reusable and recyclable, it will play an essential role in building the circular economy of the future. Steel continues to evolve, becoming smarter and increasingly sustainable.
Due to its versatility, steel has undergone enormous technological advances; in particular, Advanced High Strength Steels (AHSS) have drawn attention and are excellent candidates to meet the demands of the automotive industry and renewable energy infrastructure.
ArcelorMittal, the world’s leading steel and mining company, is focusing on creating smarter steels that are cleaner, stronger and reusable. Henrique Severiano, Laboratory Specialist, and Fernando Generoso, Product Development Specialist at ArcelorMittal in Tubarão, Brazil tell us about their material characterization with microscopy, the use of artificial intelligence (AI) and share their thoughts about the future of steel:
You are responsible for the development and evaluation of the steel quality. What does your job involve?
Our job is to use metallurgical knowledge to develop alloy design and process parameters tailored to fulfill customer needs.
That includes knowing the final application of the steel, understanding its function and what features are necessary for optimal performance.
Based on that, several quality control parameters and tests are designed to guarantee the steel is suitable for that specific use.
What challenges do you face in your daily work?
The so-called ‘second phase’ constituents in steel can be selectively colored by corrosion or thin-film formation from specific chemical etchants. This provides a feature to identify and quantify in each one of them. With the evolution of steels and the refinement of the microstructure, the usage of scanning electron microscopy (SEM) is fundamental to properly identify and quantify microconstituents. Two of the main challenges in analyzing SEM-generated microstructure images are scale and automation, as these type of images can contain a variety of artifacts and noise that cause traditional analytical techniques to fail, particularly as images become more complex. The traditional quantification by intercept method is greatly dependent on the user’s expertise and is highly unproductive; therefore, more advanced artificial intelligence (AI)-assisted software is desirable to improve and speed up the process.
How do you benefit from scanning electron microscopy and artificial intelligence (AI) in your lab?
Machine learning classifications are much more noise tolerant than their traditional counterparts. They can be used to distinguish features that have little or no difference in the SEM gray scale values, but instead have differences that are set apart only by texture. This can outperform typical software characterization that is based solely on 2D images and grayscale differences.
Typical Microstructure of a Ferritic-Bainitic Steel
Etchant: Nital reagent 2%.
Light Microscopy, 500x
Bainite – SEM, 10.000x
Typical Microstructure of Dual Phase Steel
Etchant: Nital reagente Nital 2%.
Light Microscopy, 500x
Martensita – SEM, 19.000x
Under different process conditions, the same steel can end up generating very different and complex microscale interactions of microstructural constituents, with different mechanical properties. The characterization of these phases in a large area, with high reliability and speed is of paramount importance for the development of new products, failure analysis, among other applications. Industries and research institutions that intend to play a leadership role in the development of new and more advanced steels need to improve their process simulation and product characterization capability and capacity in order to keep up with the technical evolution in steel production that is becoming increasingly complex and challenging.
We are using ZEISS ZEN Intellesis for autosegmentation and improved analysis of our second phase constituents in steel. This is changing the way we characterize materials, gaining both speed and reliability at the same time.
The dots used in the grid count are precisely what ZEISS ZEN Intellesis paints (blue) on the microstructure, thus differentiating the Bainite (blue) and Ferrite (yellow) phases.
The dots used in the grid count are precisely what ZEISS ZEN Intellesis paints (blue) on the microstructure, thus differentiating the Perlite (blue) and Ferrite (red) phases.
Perlite 48,204 % Ferrite 51,796% Time: 01min
Percentual Perlite = 47,63% Ferrite = 52,37% Time: 25min
Bainite, Martensite and Ferrite Segmentation
The dots used in counting the grid of this steel in particular using two different colors is precisely to distinguish the different phases in the same microstructure. ZEISS ZEN Intellesis in this case distinguishes the steel phases through the colors. Primary phase: Ferrite (green) and the second phases: Bainite (blue) and Martensite (red).
Martensite 22,82 % Bainite 13,39% Ferrite 63,79% Time: 02min
Percentual Martensite = 22,86% Bainite = 12,16% Ferrite = 64,98% Time: 18min
Steel is a versatile material. Do you see any future applications or trends that haven’t yet been explored?
Mankind has been using steel for about 4,500 years and its usage is more diverse than ever, from bottle caps to skyscrapers and so on. Steel consumption per capita is used worldwide as part of the economic development index. We have no doubt steel applications will keep impressing us. For the near future we can easily foresee steel being used alongside other materials such as polymers, other alloys, ceramic, etc. in the form of composites or complex structures or equipment to boost performance.
Alongside this continuous increase in applicability, the steel production process is undergoing a big transformation towards neutralizing CO2 emissions. ArcelorMittal has recently launched its first three XCarbTM initiatives as part of its journey to reach its commitment to net zero carbon emissions by 2050.