This dialog presents individual phases of calculations:
- Model reduction (renumbering of nodes and elements).
- Definition of the stiffness matrix for individual structure elements.
- Matrix decomposition K = L * U
- Problem solving for successive load cases.
In the sparse method, the matrix decomposition performs so that only non-zero elements of the stiffness matrix are stored. This opposes the frontal and skyline methods, in which all matrix elements are stored from the diagonal to the last non-zero element, including all zero elements.
The sparse method involves solving the linear equation system K * x = b or the equation system of eigenvalue problems Kφ - λBφ = 0. The direct sparse solvers (SPDS) are efficient computational techniques based on decomposition of the matrix K = L * U with a considerably less number of matrix elements different from zero. Frontal or skyline is used for regular methods. The matrix decomposition is performed with the use of the Gauss elimination method.
The Sparse direct solver is a substructure-by-substructure (superelement) approach involving deep, step-by-step multilevel nesting. SparseM is a multifrontal solver which may be applied both with the nested dissection method (NDM) and the minimal degree algorithm (MDA). (see also Parameters of the Sparse M iterative solver).
See also: