...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Computer Methods in Applied Mechanics and Engineering >Parallel, physics-oriented, monolithic solvers for three-dimensional, coupled finite element models of lithium-ion cells
【24h】

Parallel, physics-oriented, monolithic solvers for three-dimensional, coupled finite element models of lithium-ion cells

机译：锂离子电池的三维耦合有限元模型平行，物理化的整体单片溶剂

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

This work is concerned with physics-oriented, monolithic solvers for the large, sparse systems of linear equations arising in three-dimensional, coupled finite element models of lithium-ion cells and related electrochemical systems. The governing equations are obtained from a thermodynamically consistent transport theory and established both for the general case with varying temperature and for the special case of an isothermal scenario. For preconditioning the respective monolithic systems of linear equations, each system is first divided into distinct blocks associated with individual physical fields, and the resulting block structure is then respected and preserved during the preconditioning process. Field-specific preconditioning components are independently selected and synthesized, resulting in greater flexibility compared to conventional preconditioning approaches. Two different concepts are developed for constructing advanced preconditioners from elementary preconditioning techniques, and the GMRES and BiCGSTAB methods are adopted as linear solvers. Three comprehensive numerical examples involving up to 7.7 million degrees of freedom are studied. In each example, the novel preconditioners exhibit much better robustness, efficiency, and weak parallel scalability than conventional ones in terms of both solver iteration counts and total solver times including preconditioner setup. Thus, the new solution approaches are the key to large-scale, high-performance computations in settings with strict demands on accuracy. (C) 2019 Elsevier B.V. All rights reserved.
机译：这项工作涉及物理化的线性方程的大型稀疏系统，在锂离子电池和相关电化学系统的三维耦合有限元模型中产生的大型稀疏的线性方程系统。控制方程是从热力学一致的运输理论获得的，并为常规情况建立了不同的温度，并且对于等温场景的特殊情况。为了预处理线性方程的各个单片系统，每个系统首先被划分为与各个物理字段相关联的不同块，然后在预处理过程期间遵守并保存所得到的块结构。与传统的预处理方法相比，独立地选择特定的预处理组分并合成，导致更大的灵活性。开发了两种不同的概念，用于构建来自基本预处理技术的先进预处理器，并采用GMRES和BICGSTAB方法作为线性溶剂。研究了三种综合数值示例，涉及高达770万自由度的综合数值。在每个示例中，在求解器迭代计数和包括预处理器设置的总求解时间方面，新颖的预处理器表现出比传统的更好的稳健性，效率和弱的并行可扩展性。因此，新的解决方案方法是具有严格对准确性要求的设置中大规模的高性能计算的关键。 （c）2019 Elsevier B.v.保留所有权利。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号