Parallelized Meso-Scale Kinetic Monte Carlo Simulations for SOFC Characterization
Parallelized Meso-Scale Kinetic Monte Carlo Simulations for SOFC Characterization
复制标题
SOFC 表征的并行细观动力学蒙特卡罗模拟
DOI：
10.1149/05701.2437ecst
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
N. Watanabe
中科院分区：
文献类型：
--
作者：
T. Tada;N. Watanabe
我们提出了一种用于SOFC表征的并行中尺度动力学蒙特卡罗代码。该代码包含五个特性：1)单元胞最小输入以构建更大的块结构；2)根据三维形态堆积块；3)基本事件的符号定义；4)在open- mpi混合并行化中使用域分解并行实现；5)封闭、周期和开放条件的边界条件。通过计算氧离子在含有4%氧空位的立方ZrO2中的自扩散系数，证实了并行化kMC程序的可靠性。在32核计算中测试了并行化效率，并证实了良好的并行性。
We propose a parallelized meso-scale kinetic Monte Carlo code for SOFC characterization. Five properties included in this code are 1) minimum input of the unit cell to build larger block structures, 2) piling blocks according to the three dimensional morphology, 3) symbolic definition of elementary events, 4) parallel implementation with the domain decomposition in open-MPI hybrid parallelization, and 5) boundary conditions for the closed, periodic, and open conditions. The reliability of the parallelized kMC program was confirmed from the calculated self-diffusion coefficients of oxygen ions in the cubic ZrO2 with 4% oxygen vacancies. The parallelization efficiencies were also tested up to 32 cores calculations, and good parallelization was confirmed.