AxisVM lets you perform linear and nonlinear static, linear and nonlinear dynamic, vibration and buckling analysis. It implements an object-oriented architecture for the Finite Element Method. Note that the finite element analysis is only a tool, not a replacement for engineering judgement.
Each analysis consists of four steps:
1. Model optimization
2. Model verification
3. Performing the analysis
4. Result file generation
To reduce analysis time and memory footprint, AxisVM optimizes node order. If the total number of degrees of freedom is over 1000, it creates an internal three-dimensional graph from the model geometry and begins to partition the system of equations using the substructure method. The system is stored as a sparse matrix. The parameters of the optimized system of equations appear only at the end of this process. This process results in the smallest memory footprint and fastest calculation time but it assumes that the biggest block fits into the available memory. If it doesn’t, AxisVM stores the system as a band matrix and begins to reduce the bandwidth of the system by iterative node renumbering. If the two longest rows fit into the available memory, the system can be solved. Changes in the memory requirements for the band matrix is displayed real-time. The duration of the optimization process and the final memory footprint depends on the size of the system and the available memory.
The system of equations can be solved the most efficiently if the whole system fits into the physical memory. If the system does not fit into the physical memory but its largest block does, the running time will be moderate. If the largest block does not fit into the physical memory, the necessary disk operations can slow down the solution considerably.
The input data is verified in the first step. If an Error is found, a warning message is displayed and you can then decide whether to cancel or continue the analysis.
AxisVM displays the evolution of the solution process by two progress bars. The bar on the top displays the current step performed, while the other displays the overall progress of the analysis process.
The equilibrium equations in the direction of constrained degrees of freedom are not included in the system of equations. Therefore to obtain support reactions you must model the support conditions using support elements.
The Cholesky method is applied to the solution of linear equilibrium equations. The eigenvalue problems are solved with the Subspace Iteration method.
During the processing of the results the program sorts the results according to the original order of the nodes and prepares them to graphical display. In the following chapters we will show the setting of the parameters of the each calculation methods.