86Since the three-dimensional additive manufacturing is a new process, it is possible to manufacture with new compositional alloy systems that are not suitable for conventional forging or casting and may exhibit better properties than existing materials. However, it is very inefficient to search for new alloy systems only by experimental methods because of the huge cost in time and expense. In this study, we have proposed a model that can easily and quickly predict precipitation phenomena, which significantly affect high-temperature properties, by means of energetics. This model evaluates the total free energies of various microstructures that can appear in the process from supersaturated solid solution to equilibrium state and predicts the precipitation process of alloys based on the assumption of the steepest energy descent path. This model requires only a small number of input parameters for the calculation, and the prediction can be done with a relatively simple calculation of energy addition and subtraction only. Therefore, the composition, temperature, and time dependence of the precipitation process in new materials with unknown properties as well as practical structural materials in multiphase and multicomponent systems can be calculated in a short time using a personal computer. And, by combining an optimal solution search method based on information engineering, we attempted to search for heat-resistant nickel alloys with novel compositions suitable for additive manufacturing from a wide compositional space of multi-component systems [1]. The concept of Materials Integration aims to computationally link processing, structure, properties, and performance to accelerate materials research and development [1,2]. To realize this concept, we have developed a system called MInt [3]. In MInt, various prediction models based on computational simulations, theoretical or empirical formulas, and machine learning algorithms can be implemented as modules. These modules can be interconnected to design workflows that comprehensively predict material properties and performance from processing condition through microstructure evolution. Once forward predictions are established, inverse problems can be solved using AI optimizers. We applied MInt to address the inverse problem of designing the aging process of a Ni-Al two-phase alloy, a model material for Ni-based superalloys. We constructed a computational workflow to predict high-temperature strength based on the aging process [3]. By incorporating Monte Carlo Tree Search (MCTS), one of the AI optimizers, into the workflow, we explored aging patterns that result in higher strength [4]. The AI algorithm identified 110 aging patterns superior to conventional isothermal aging. Based on these AI-discovered patterns, we proposed a novel two-stage aging concept. P5-11Exploration of New Alloy Compositions Using Prediction Model of Precipitation by Energetics Yoshiaki Toda, Sae Dieb, Keitaro Sodeyama, and Masahiko DemuraCenter for Basic Research on Materials, National Institute for Materials Science (NIMS) [1] S. Dieb, Y. Toda, K. Sodeyama, M. Demura, Sci. Tech. Adv. Mater.: Method, 3 (1), 2278321 (2023). P5-12Inverse Design Based on the Concept of Materials Integration Masahiko Demura Research Network and Facility Services Division, National Institute for Materials Science (NIMS) [1] M. Demura & T. Koseki, Mater. Trans., 61, 2041 (2020).[2] M. Demura, ISIJ International, 64, 503 (2024).[3] S. Minamoto et al., Mater. Trans., 61, 2067 (2020).[4] T. Osada, et al., Mater. & Design, 226, 111631 (2023).[5] V. Nandal et al., Sci. Reports, 13, 12660 (2023).
元のページ ../index.html#86