Nonlinear - Solution

The Solution tab in the Nonlinear dialog box sets solution related options. The following are the options you can set on this tab.

Stepping Options

Time information is associated with the definition of time curves for loads and boundary conditions. Time is a pseudo variable for static problems without creep, visco-elasticity, or thermal loading using time-dependent results from a transient thermal study.

Start time Starting solution time. Not used by the Arc-Length method.
Restart Click this check box to restart from the last successful solution step. Available only if restart data exist with option Save data for restarting the analysis activated in the previous run.
You can change any load parameters (Time curve dialog box) when restarting the analysis.
You can change fixture conditions from free to fixed degrees of freedom and vice versa when activating the Restart option.
For example, to change a fixture condition from fixed to free when restarting the analysis:
  • Under Translations (Fixture PropertyManager), type 1 in the direction which you apply the fixture.
  • Under Variation with Time, select Curve and click Edit. In the Time curve dialog box, enter the curve data:
    X (Time, sec) Y Value Fixture condition
    0 0 fixed
    1 0 fixed
    1.05 Off (restart) change from fixed to free
    2 Off free

To change a fixture condition from free to fixed when restarting the analysis:

  • Under Translations, type 1 in the direction which you apply the fixture.
  • Under Variation with Time, select Curve and click Edit. In the Time curve dialog box, enter the curve data:
    X (Time, sec) Y Value Fixture condition
    0 Off free
    1 Off free
    1.05 0 (restart) change from free to fixed
    2 0 fixed

    During the first run (remember to select Save data for restarting the analysis for the first run ( Start time = 0 < t < End time = 1 sec) the solver ignores this fixture and the selected entity where the fixture is applied is free to move. Activating the Restart option and rerunning the analysis (Start time = 1 sec < t < End time = 2 sec) , the solver applies the fixture and the selected entity is restrained from moving in the specified direction.

End time Ending solution time. Not used by the Arc-Length method.
Save data for restarting the analysis Check this flag before running the study for possible restart. The software takes some time and disk space to save the data required for a proper restart. If you clear this check box, you have to start from the beginning.
Remeshing the study deletes all restart information.
Time increment Sets the procedure for incrementing time at each solution step for the Force and Displacement control methods. For the Arc-Length control method, the program uses this value to estimate an arc-length increment.

Automatic (autostepping)

When checked, the program determines an increment internally for each solution step to improve chances of convergence. This option supports all control techniques. If the flag is checked, the following entries are used:

Initial time increment

The program uses this increment as the initial guess for the time increment.

Min

Minimum time step. The default is 1e-8 seconds.

For nonlinear dynamic studies only, if the specified minimum increment is equal or greater than the initial time increment, the program resets the minimum time step to 10% of the initial time increment.

Max

Maximum time step. Default is the End time for the Force and Displacement control methods.

For nonlinear dynamic studies only, the specified maximum time step is not used. The program resets the maximum value to the initial time increment.

No. of Adjustments

Maximum number of time step adjustments for each solution step.

Fixed

Fixed time step increment. Default is to run 10 steps.

Geometry nonlinearity options

Use large displacement formulation Uses the large displacement formulation.
Update load direction with deflection If checked, the direction of the applied load (normal uniform pressure or normal force) is updated for every solution step with deflection.
Examples
(a) Applied normal load on undeformed geometry.
(b) Flag for Update load direction with deflection is unchecked. Original load direction is maintained on deformed geometry.
(c) Flag for Update load direction with deflection is checked. The load direction is updated, and remains normal to the deformed geometry for each solution step.
When you apply a torque, the program calculates the force and the moment arm that creates the torque, and applies the force to the nodes. These forces retain their initial directions throughout the solution and as a result may cause high unexpected stresses to develop.
Large strain option (For plasticity material models only). Uses the large strain formulation.
Keep bolt pre-stress

When this option is cleared, the bolt's length at zero stress state L0 is determined based on the length of bolt at the start of analysis Lst, which corresponds to the un-deformed geometry state of the components attached through the bolt connector. The bolt's length at zero stress state is calculated from:

L0 = Lst / (1+(P/A*E))

As the nonlinear analysis progresses, the bolt's length Lstep at each analysis step adapts itself to the deformed geometry of the attached components as they deform due to the applied loads. The bolt's final stress at the end of the nonlinear analysis differs from the user defined pre-load stress. The bolt's axial load at each analysis step is calculated from:

Pstep = A* E* (Lstep - L0) / L0

When this option is checked, the program first runs an analysis with the user defined pre-load P as initial condition without any external loads. The deformation of the parts connected through the bolt is calculated and is used to determine the bolt's length at zero stress state L0. Let’s define Lf as the deformed length of bolt which corresponds to the settlement of the connecting parts due to pre-stress. The bolt’s length at zero stress is then calculated from:

L0= Lf / (1+(P/A*E))

For the second step of the analysis, all aplied loads are included. The bolt's axial load at each analysis step is calculated from:

Pstep = A* E* (Lstep - L0) / L0

During the analysis, if (a) Lstep <= L0 then the bolt is loose, and if (b) Lstep > L0, the bolt is under tension and keeping parts together.

Notation:
  • P: User defined axial pre-load
  • Pstep: Axial load of bolt at current analysis step
  • A: Bolt section area
  • E: Bolt material modulus of elasticity
  • L0: Original length of bolt at zero stress state
  • Lst: Length of bolt at start of analysis (corresponds to the un-deformed geometry state of the components attached through the bolt)
  • Lf: Deformed length of bolt after settlement of connecting parts due to pre-stress (Keep bolt-prestress selected)
  • Lstep: Deformed length of bolt at current analysis step

Solver

Sets the solver to be used in performing nonlinear analysis.

Automatic solver selection The program selects the most robust between two solvers depending on the size of the model and available RAM:

Intel Direct Sparse

For small and medium size models with slim geometry. The Intel Direct Sparse solver requires more RAM than the FFEPlus Iterative Solver.

FFEPlus

For medium size models with bulky geometry, and large models.

Direct sparse Use the direct sparse solver. This solver have more chances of convergence for highly nonlinear problems.
FFEPlus Use the FFEPlus iterative solver. This solver is less demanding on memory. It maybe faster for large problems.
Large Problem Direct Sparse The Large Problem Direct Sparse solver, by leveraging enhanced memory-allocation algorithms, can handle cases where the solution is going out of core.

Incompatible bonding options

Sets the solver to be used in performing nonlinear analysis.

Simplified Applies the node-based bonding contact.
More accurate (slower) Applies the surface-based bonding contact, which results in longer solution time than the node-based contact formulation.
The program stops the analysis if:
  • The number of step-size adjustments in any step exceeds the maximum number of step adjustments.
  • The step increment required for convergence becomes smaller than the minimum step increment.