C. Przybyla, D. McDowell, Georgia Institute of Technology, Atlanta, GA
Models for prognosis applications are desired that reflect underlying microstructure. Conventional macroscopic viscoplastic constitutive models for - Ni-base superalloys that are used for component level stress analysis typically do not contain explicit dependence on the underlying microstructure. Microstructure-sensitive models should recognize the strong influence of the size and spatial distributions of the precipitate phase on various aspects of stress-strain response, including cyclic plasticity, creep, stress relaxation, and monotonic work hardening. The primary microstructure attributes that significantly affect the stress-strain response of IN100 are the grain size distribution and precipitate volume fraction and size distributions. An Artificial Neural Network (ANN) is used to inform dependence of material parameters in a Chaboche type viscoplastic model on these microstructure attributes; microstructures within the range in which the ANN was trained are employed to this end using a combination of experiments and polycrystal plasticity calculations. Such a model is applied to an example of notch root analysis.
Summary: Models for prognosis applications are desired that reflect underlying microstructure. Conventional macroscopic viscoplastic constitutive models for - Ni-base superalloys that are used for component level stress analysis typically do not contain explicit dependence on the underlying microstructure. Microstructure-sensitive models should recognize the strong influence of the size and spatial distributions of the precipitate phase on various aspects of stress-strain response, including cyclic plasticity, creep, stress relaxation, and monotonic work hardening. The primary microstructure attributes that significantly affect the stress-strain response of IN100 are the grain size distribution and precipitate volume fraction and size distributions. An Artificial Neural Network (ANN) is used to inform dependence of material parameters in a Chaboche type viscoplastic model on these microstructure attributes; microstructures within the range in which the ANN was trained are employed to this end using a combination of experiments and polycrystal plasticity calculations. Such a model is applied to an example of notch root analysis.
