The Role of Machine Learning Algorithms in Materials Science: A State of Art Review on Industry 4.0
- PDF / 3,853,006 Bytes
- 21 Pages / 595.276 x 790.866 pts Page_size
- 52 Downloads / 143 Views
ORIGINAL PAPER
The Role of Machine Learning Algorithms in Materials Science: A State of Art Review on Industry 4.0 Amitava Choudhury1,2 Received: 5 March 2020 / Accepted: 30 September 2020 © CIMNE, Barcelona, Spain 2020
Abstract The 21st century has witnessed a rapid convergence of manufacturing technology, computer science and information technology. This has led to a paradigm of 4.0. The hitherto known developments in metallurgical and materials practices are largely driven by application of fundamental knowledge through experiments and experiences. However, the mounting demands of high performance products and environmental security calls for the ‘right first time’ manufacturing in contrast to the traditional trial and error approach. In this context, a priori capability, for prediction and optimization of materials, process and product variables, is becoming the enabling factor. In recent time, research in material science is increasingly embarrassing the computational techniques in development of exotic materials with greater reliability and precision. The present study is aimed at exploring the computer vision and machine learning techniques in different application areas in materials science.
1 Introduction It is appropriate to state that advances in materials science shape not only our daily lives but also promote growth. Materials are now inseparable with progress. This makes the search of new materials a contemporary and critical subject in material science. To accelerate the discovery and design process for new materials, various computational approaches have been introduced, in tandem with the experimental processes. The mechanical and physical properties of materials largely depend on the grain distribution, shape and the size of the microstructure constituents. Thus, identification, classification and quantification of the microstructural constituents is important to establish the structure property correlation of a specific material. To accelerate the manufacturing process, many industries are recently taking interest in automation by using the application of computer vision and image processing, which further enables cost-effective design of materials to achieve * Amitava Choudhury [email protected] 1
Department of Metallurgy and Materials Engineering, Indian Institute of Engineering Science and Technology, Shibpur, India
School of Computer Science, University of Petroleum and Energy Studies (UPES), Dehradun, India
2
targeted properties [1–5]. Microstructure modeling can be effective in introducing automation or feedback based control of different processes such as deformation processing and heat treatment [1]. The previously mentioned approach predicts the microstructural features, including volume fraction of phases during heat treatment of steels. Earlier, optical microscopy was the only available technique for microstructure analysis. However, in recent times, various image-processing techniques have been developed to analyze the same [5]. Due to their (microstructural) rec
Data Loading...
