Discrete Dislocation Dynamics

采用离散位错动力学方法，通过分析不同加载条件下位错系综的演化，对小尺寸试样的塑性力学问题有了新的认识。主要重点将放在晶界（GB）的作用和施加的应力或应变梯度上出现的位错结构。在FOR 1650的报告期1中，已经发现GB的塑性变形-在DDD中建模为不可穿透的-只能在假设有限和应变相关晶界屈服强度的连续框架中描述。然而，目前还不清楚如何制定这样的晶界屈服强度。广泛的离散位错动力学模拟将允许得到一个更好的理解这种行为的系统变化的取向差和塑性性能（例如初始位错密度）的相关晶粒。应详细描述相邻晶粒中位错之间弹性相互作用的作用。此外，在位错网络内的位错增殖机制的取向和微观结构的依赖性将被探索。拓扑和统计措施应发展定量表征位错微观结构，而超越简单的密度措施和“结计数”。预期的结果也将允许指导塑性连续描述的发展，例如处理GBs和位错连续理论中滑移系相互作用的制定所需的关键机制。为了提高计算效率，应实现DDD代码的MPI并行化，这将允许模拟多晶样品。

The discrete dislocation dynamics method will be used to gain new insight in the plasticity of small scale specimens by analysing the evolution of dislocation ensembles under different loading conditions. The main focus will be on the role of grain boundaries (GBs) and imposed stress or strain gradients on the emerging dislocation structure. In reporting period 1 of FOR1650, it has been found that the plastic deformation of a GB -- modelled as impenetrable in DDD -- can only be described in a continuum framework assuming a finite and strain dependent grain boundary yield strength. However, it is yet unclear how such a grain boundary yield strength can be formulated. Extensive discrete dislocation dynamics simulations will allow to derive a better understanding of this behaviour by a systematic variation of misorientation and plastic properties (e.g initial dislocation density) of the involved grains. The role of the elastic interaction between dislocations in the neighbouring grains shall be characterized in detail. Furthermore the orientation and microstructure dependence of dislocation multiplication mechanisms within a dislocation network will be explored. Topological and statistical measures shall be developed for quantitatively characterizing dislocation microstructures, while going beyond simple density measures and "junction counting". The expected results will also allow guiding the development of continuum descriptions of plasticity, e.g. the handling of GBs and the key mechanism needed for the formulation of the interaction of slip systems in dislocation continuum theories. For improved computational efficiency a MPI parallelisation of the DDD code shall be realized which would allow simulating poly-crystalline samples.