The symposium will focus on recent advances in algorithm development in mesoscale simulation techniques. It will be particularly centered on methods intended to bridge atomistic scale calculations with macroscale (i.e. engineering scale) techniques. In particular, Monte Carlo, cluster dynamics, dislocation dynamics, phase field, and polycrystal homogenization techniques for mechanical fields (self-consistent methods, Fast Fourier based methods, Finite element methods) applied to material science problems will constitute the main set of tools focus of the symposium. If the Multi-scale Materials Modeling approach is to be used both for virtual processing and virtual characterization with truly predictive capabilities, a strong mesoscale component needs to be developed in order to integrate the great knowledge obtained from the nanoscale science with the material response and its microstructure evolution at the macroscale. Both, lower scale calculations and larger scale finite element models have progressed steadily, with the development of advanced algorithms to solve more accurate physical models and/or improve the efficiency of existing methodologies. On the other hand, mesoscale tools have somehow lagged behind and the development of new capabilities remains crucial to obtain a coherent methodology capable of passing accurate and statistically representative information from atomistic models to engineering approaches in quasi-static as well as dynamic frameworks. This self-consistent methodology promises to guide experiments in the design of materials with the appropriate properties for the intended application. While non exhaustive, the following presents a list of topical developments for which discussions at an international symposium are becoming critical to accelerate the advances in our community: (i) Monte Carlo algorithms have been successfully applied to study many different physical phenomena. Still, advanced sampling algorithms, parallelization techniques, searching strategies and implementations in novel architectures are the subject of current research. (ii) Cluster dynamics is a mean-field approach that solves the coupled defect-diffusion equations to study the evolution of defect concentrations. Recent developments in stochastic solvers have improved the usability of the technique to explore longer time scales and more sophisticated microstructures. (iii) Dislocation Dynamics numerical bottleneck resides in the calculation of the long range interaction forces between dislocation segments. Multipole expansions have been proposed to mitigate the computational burden, but still severely limit the applicability range. Recently developed algorithms based on FFTs might help in the force calculations, mostly in anisotrpic materials. (iv) The Phase Field method has emerged as a powerful and flexible tool for quantitative modeling of the coevolution of microstructure and physical properties at the mesoscale. Simulation of multi-phase and multi-physics phenomena has progressed considerably. Direct coupling to thermodynamic databases can make many new material systems available for modeling, but remains challenging. (v) Elasto-viscoplastic Eshelby-like micromechanics models with both mean field and full field resolutions have been succesfully used to simulate the mechanical response and texture evolutions of a relatively wide spectrum of metals with cubic and non-cubic crystal symmetries. Novel algorithm developments are needed to deal with Gibbs phenomena associated with FFT approaches and the treatment of material interfaces.
Topics will include:
Advanced Monte Carlo algorithms
Cluster dynamics
Dislocation dynamics
Phase field approaches
Polycrystalline models
Hybrid atomistic-continuum methods
04月17日
2017
04月21日
2017
注册截止日期
留言