Microstructure strongly influences the mechanical properties of cast iron. By inoculating the melt with proper inoculants, foreign substrates are brought into the melt and eventually the graphite can crystallize on th...Microstructure strongly influences the mechanical properties of cast iron. By inoculating the melt with proper inoculants, foreign substrates are brought into the melt and eventually the graphite can crystallize on them. The elements and substrates that really play a role for nucleation are yet unknown. Until now there is very little knowledge about the fundamentals of nucleation, such as composition and morphology of nuclei. In this work we utilized EN-GJL-200 as a base material and examined several produced specimens. The specimens were cast with and without inoculants and quenched at different solidification states. Specimens were also examined with a high and low oxygen concentration, but the results showed that different oxygen contents have no influence on the nucleation in cast iron melts. Our research was focused on the microscopic examination and phase-field simulations. For studying the samples we applied different analytical methods, where SEM-EDS, -WDS were proved to be most effective. The simulations were conducted by using the software MICRESS, which is based on a multiphase-field model and has been coupled directly to the TCFE3 thermodynamic database from TCAB. On the basis of the experimental investigations a nucleation mechanism is proposed, which claims MnS precipitates as the preferred site for graphite nucleation. This theory is supported by the results of the phase-field simulations.展开更多
Fast prediction of microstructural responses based on realistic material topology is vital for linking process,structure,and properties.This work presents a digital framework for metallic materials using microscale fe...Fast prediction of microstructural responses based on realistic material topology is vital for linking process,structure,and properties.This work presents a digital framework for metallic materials using microscale features.We explore deep learning for two primary goals:(1)segmenting experimental images to extract microstructural topology,translated into spatial property distributions;and(2)learning mappings from digital microstructures to mechanical fields using physics-informed operator learning.Loss functions are formulated using discretized weak or strong forms,and boundary conditions-Dirichlet and periodic-are embedded in the network.Input space is reduced to focus on key features of 2D and 3D materials,and generalization to varying loads and input topologies are demonstrated.Compared to FEM and FFT solvers,our models yield errors under 1–5%for averaged quantities and are over 1000×faster during 3D inference.展开更多
文摘Microstructure strongly influences the mechanical properties of cast iron. By inoculating the melt with proper inoculants, foreign substrates are brought into the melt and eventually the graphite can crystallize on them. The elements and substrates that really play a role for nucleation are yet unknown. Until now there is very little knowledge about the fundamentals of nucleation, such as composition and morphology of nuclei. In this work we utilized EN-GJL-200 as a base material and examined several produced specimens. The specimens were cast with and without inoculants and quenched at different solidification states. Specimens were also examined with a high and low oxygen concentration, but the results showed that different oxygen contents have no influence on the nucleation in cast iron melts. Our research was focused on the microscopic examination and phase-field simulations. For studying the samples we applied different analytical methods, where SEM-EDS, -WDS were proved to be most effective. The simulations were conducted by using the software MICRESS, which is based on a multiphase-field model and has been coupled directly to the TCFE3 thermodynamic database from TCAB. On the basis of the experimental investigations a nucleation mechanism is proposed, which claims MnS precipitates as the preferred site for graphite nucleation. This theory is supported by the results of the phase-field simulations.
基金the funding support provided to develop the present work in the project Cluster of Excellence“Internet of Production”(project:390621612).
文摘Fast prediction of microstructural responses based on realistic material topology is vital for linking process,structure,and properties.This work presents a digital framework for metallic materials using microscale features.We explore deep learning for two primary goals:(1)segmenting experimental images to extract microstructural topology,translated into spatial property distributions;and(2)learning mappings from digital microstructures to mechanical fields using physics-informed operator learning.Loss functions are formulated using discretized weak or strong forms,and boundary conditions-Dirichlet and periodic-are embedded in the network.Input space is reduced to focus on key features of 2D and 3D materials,and generalization to varying loads and input topologies are demonstrated.Compared to FEM and FFT solvers,our models yield errors under 1–5%for averaged quantities and are over 1000×faster during 3D inference.
