摘要
为提高有限元分析效率、减少存储空间消耗,对刚度矩阵的压缩存储算法进行了研究.研究了"广义相邻节点对"与刚度矩阵中非零子矩阵的关系,确定了刚度矩阵中非零子矩阵的分布规律;提出了一种新的刚度矩阵压缩存储方法—"改进的CSR存储方法",给出了基于压缩存储的刚度矩阵的生成过程以及线性方程组迭代解法方法,并将提出的算法应用于三维体积成形有限元分析软件.有限元分析实例表明,该算法可以有效地减少存储空间,提高计算效率.
To improve the efficiency and reduce the storage space of finite element analysis,compression and storage algorithm of 3D-FEM stiffness matrix is studied.The relationship between ＂generalized adjacent double nodes＂ and the non-zero sub-matrix in stiffness matrix is researched for getting distribution of non-zero sub-matrix in stiffness matrix.A new algorithm of stiffness matrix of compressed storage-＂improved CSR storage method＂ is proposed.Based on the algorithm,the generation process of stiffness matrix is given and iterative solution of linear equations method is proposed to improve the efficiency of solving linear equations.The algorithm is applied to the three-dimensional bulk forming finite element analysis software and the numerical results show that the algorithm can effectively decrease the storage space and improve the computation efficiency.
出处
《材料科学与工艺》
EI
CAS
CSCD
北大核心
2012年第2期96-100,107,共6页
Materials Science and Technology
基金
国家自然科学基金资助课题(50875155)
教育部"长江学者和创新团队发展计划"创新团队资助项目(IRT0931)
