While the development of materials development systems using information science and computers, including artificial intelligence (AI), has been actively promoted in Europe and the United States, the concept of "Materials Integration" was proposed in the Cabinet Office SIP Phase 1 "Structural Materials" and Phase 2 "Materials Revolution" to accelerate the development of materials by connecting the four elements of materials engineering shown in Figure 1: process, structure, properties, and performance, using computers and information science methods, including AI. In the 5th Science and Technology Basic Plan, its promotion is stated as an integrated material development system that supports Society 5.0. Among these, the system developed by the industry-academia-government all-Japan system centered on NIMS and the University of Tokyo, mainly for metallic structural materials, is called the MInt system, which is an acronym for Materials Integration by Network Technology.
In general, materials development proceeds by clarifying the linkage between process, structure, properties, and performance, as shown in Fig. 1, and the MInt system can predict the entire process, structure, properties, and performance by freely connecting various prediction calculation tools called modules based on AI, materials engineering theory, and empirical rules.
Furthermore, as shown in Fig. 2, on the contrary, it is possible to optimize the process and chemical components based on the desired performance using AI, making it a unique system in the world.
Fig. 3 is an example of a panel display showing the flow of calculations in the MInt system. One of the key features of the MInt system is that the necessary calculations can be performed automatically by connecting the modules to be combined through an intuitive interface on the screen and visually constructing the flow of calculations called the workflow. This is one of the main features of the MInt system.
By using the MInt system, we can greatly speed up the verification of ideas, which has traditionally been done mainly through experiments. As an example of a specific result, we can now estimate with high accuracy the lifetime (rupture lifetime due to creep phenomenon) of heat-resistant high Cr steel welded components used in thermal power plants when they are used under high temperature and high pressure. This makes it possible to replace thousands of hours of experiments with just a few hours of calculations. In addition, when combined with AI-based optimization techniques, the MInt system can efficiently search for optimal material and process conditions based on desired performance. In the case of research and development in SIP, it has begun to find optimal conditions for materials and processes that have been overlooked in the past. In this way, the MInt system will speed up the verification of ideas and enable exhaustive and efficient optimization by AI, which is expected to strengthen the competitiveness of Japan's advanced materials industry.