Multi-Scale Analysis and Simulation in 3D Crystal Plasticity Large Deformation Finite Element Platforms to Predicting and Designing Thermomechanical Responses of Metallic Nano-Layers

Zara Moleinia, Purdue University

Abstract

The current work centers on multi-scale approaches to simulate and predict metallic nano-layers thermomechanical response in crystal plasticity large deformation finite element platforms. Multiple parts including multi-scale modeling divided into two major scales; nanoand homogenized levels, sensitivity analysis, micro-scale simulations, precipitate hardening, and computational aid hardware and software development are researched. Copper/niobium nano-layers are designated as case studies. At the nano-scale, a size-dependent constitutive model based on entropic kinetics is developed through the thermodynamical phase space of metallic nano-systems. A deep-learning adaptive boosting technique called the single layer calibration is established to acquire associated constitutive parameters applicable to a broad scope of setups entirely different from those of calibrations. The model is validated by experimental data with solid agreements followed by the simulation of multiple cases regarding size, loading pattern, layer type, and geometrical effects specifying the inconsequential effects of layers or loading orientations, predominant influence of size over the other traits, the impacts of a constituent in a bi-crystal cases, and generalized size effects on yield and flow strengths as well as transition strain. Sensitivity analysis is performed on the size-dependent constitutive model as a diagnostic-prognostic field through the factor prioritization in order to capture the influential parameters where the size effects designed parameters, namely, m, cs, csat, and h0 are found the dominant factors on the main behavioral features. At the homogenized level, macro-homogeneity is utilized through the statistical mechanics of the microcanonical ensemble and the Clausius-Duhem inequality to link the scales through entropy flux. A homogenized crystal plasticity-based constitutive model is developed with the aim of expediting while retaining the accuracy of the computational processes for which effective constitutive functionals are realized. The high nonlinearity of the functional constants is dealt with through a metaheuristic genetic algorithm approach leading to determining the associated terms in an optimized fashion. The homogenized constitutive model results are favorably verified with nano-scale data while expediting computational processes by several orders of magnitude. The temperature effects are captured through developing a temperature-dependent constitutive model where elastic constants and effective functional parameters are determined and calibrated. The model is validated with experimental data with multiple demonstrations of temperature effects identifying the degradation of a thin nano-layer at high temperature into a thicker one at lower temperature and dramatic drops of up to 80% in flow strength at about 1000K. The work is expanded to micro-scale where a crystal plasticity constitutive model is developed with the same backbones of the model presented in the nano range. The implicit trace of size is designed in this format as the physics behind the spectrum lead. The deep-learning single layer calibration is utilized to obtain the associated constitutive parameters where validated by the experimental data. Precipitate hardening phenomenon is realized and implemented considering Orowan strengthening mechanism in nano-systems. Wherefore several cases indicating the impacts of size and volume fraction of precipitates on mechanical properties are assessed and discussed revealing the exponential increase on flow strength and hardening with respect to precipitate volume fractions, the high stress absorption by precipitates creating extreme stress gradient with the matrix, and the effect of a strengthened constituent, here Nb, on the overall behavior of these nanolamellars.

Degree

Ph.D.

Advisors

Bahr, Purdue University.

Subject Area

Analytical chemistry|Artificial intelligence|Chemistry|Mechanics

COinS