The keywords required to launch a RADIOSS implicit analysis are defined in the Engine file and begin with /IMPL (see RADIOSS Engine Input for more details). The only exception is /EIG, which is defined in RADIOSS Starter for running modal analysis. Most RADIOSS output terms have the same meaning for implicit and explicit analysis.
For static problems, a time step is used. This has no physical meaning and only serves to define the increments for loading and displacements; that is the cycle number here will correspond to the increment number.
Therefore, there isn't a big change in the Engine input file. Review the output, like /PRINT (for example: /PRINT/-1 will print out at each increment) and some output file frequencies, as large time step is used.
|
The basic keywords for linear analyses are:
1. | Linear Static: |
*/IMPL/LINEAR: Static linear computation
/IMPL/SOLVER/n: Linear solver method, where n indicates the type of solver
/IMPL/PRINT/LINEAR/Nprint: Printout frequency for linear resolution
The printout frequency keyword is used mainly for iterative solver (or mixed). When used with a direct solver, the relative residual will be printed.
/IMPL/LINEAR/INTER: Contacts can be taken into account in a linear analysis through a simple two-step linear approach: A linear first step that evaluates penetrations and generates interface reaction springs followed by a linear second step.
2. | Linear Quasi-static: |
*/IMPL/QSTAT: Quasi-static analysis
/IMPL/QSTAT/DTSCAL: Scaling factor of adding inertial stiffness
When /IMPL/QSTAT/DTSCAL is used with /IMPL/LINEAR/INTER, the scaling factor is only applied to the second step.
3. | Euler Buckling: |
*/IMPL/BUCKL: Euler buckling modes
*/IMPL/LINEAR: Static linear computation
Conforming to RADIOSS Bulk input, the input format for buckling analysis changed in HW10.0. It becomes /IMPL/BUCKL/1 and Lanczos solver (BCSLIB-EXT) is used for buckling mode extraction. Buckling from restart or with pre-stresses is also available (/IMPL/BUCKL/2).
Animation output files (*Annn) are special in the case of a buckling analysis. Each file contains one buckling mode (where the critical loading is presented by the time value). If a linear computation precedes a buckling analysis, then the first two animation files (*A001, *A002) contain results of linear structure computation and the buckling modes are available starting from the third animation file (*A003).
Iterations are needed in an implicit nonlinear analysis. The simulation is typically divided into several increments (steps) to manage the nonlinearity that will be represented by the time step control.
A nonlinear solver (Newton-based method) is used to find the balanced forces on the updated geometry (large displacement, large strain) at each step.
Two types of control parameters should be defined:
• | Nonlinear solver (using /IMPL/NONLIN/n) |
• | Steps (using /IMPL/DT/n, /IMPL/DTINI and /IMPL/DT/STOP) |
The basic keywords for nonlinear analysis are:
* /IMPL/NONLIN/n: Nonlinear solver method, where, n indicates the type of nonlinear solver
/IMPL/SOLVER/n: Linear solver method, where, n indicates the type of linear solver
/IMPL/PRINT/NONLIN/Nprint: Printout frequency for nonlinear iterations
* /IMPL/DT/n: Time step control, where, n indicates the time step control method
* /IMPL/DTINI: Initial time step
/IMPL/DT/STOP: Min-max time step values
/IMPL/QSTAT: Quasi-static implicit
/IMPL/SPRBACK: Spring-back computation