Abstract Title
Material Parameter Estimation from Microstructure Morphologies with Machine Learning
Abstract
Machine learning plays an important role in understanding and predicting the parameters of a microstructure. Focusing specifically on Iron-Chromium alloys and using the physics-based coding program MOOSE, we are able to create accurate phase-field models of the alloys spinodal decomposition. The program allows us to control the initial parameters of our two-dimensional image, creating the perfect database to run through AI. The AI takes this subset and trains itself on predicting the initial parameters of the phase-field models. From there, models with unknown parameters can be submitted to the system, and the AI can predict the parameters, even though it has not seen that specific model. Because Iron-Chromium alloys play a key role in many industries, the applications of this research in determining the strength and endurance of steel under intense heat stress is vast.
Material Parameter Estimation from Microstructure Morphologies with Machine Learning
Machine learning plays an important role in understanding and predicting the parameters of a microstructure. Focusing specifically on Iron-Chromium alloys and using the physics-based coding program MOOSE, we are able to create accurate phase-field models of the alloys spinodal decomposition. The program allows us to control the initial parameters of our two-dimensional image, creating the perfect database to run through AI. The AI takes this subset and trains itself on predicting the initial parameters of the phase-field models. From there, models with unknown parameters can be submitted to the system, and the AI can predict the parameters, even though it has not seen that specific model. Because Iron-Chromium alloys play a key role in many industries, the applications of this research in determining the strength and endurance of steel under intense heat stress is vast.