Some information (different application) that involves transferring static analysis results
to the dynamic stage.
The following provides some guidance, notes and necessary input features for modelling
the preloading phase of a bolted structure and the subsequent phase where the structure is
impacted by a separate object at a prescribed initial velocity (for example, an SPH bird-strike).
Much of this has been drawn from the very nice shared notes by Jim Day in the following,
regarding several preload approaches:
http://ftp.lstc.com/anonymous/outgoing/jday/bolt_preload3.pdf
http://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/preload.pdf
2. Implicit Dynamic Relaxation (ImpDR) - IDRFLG=5/6
When IDRFLG=5 or 6, LS-DYNA performs an implicit analysis for the preload phase of
the simulation. Parameters for controlling the implicit preload solution are defined using
appropriate *CONTROL_IMPLICIT keywords to specify solver type, implicit time step,
etc. When using this option, one must specify DRTERM to indicate the termination "time"
of the implicit preload analysis. When DRTERM is reached, the implicit preload phase
terminates and LS-DYNA begins the next phase of the analysis according to IMFLAG in
*CONTROL_IMPLICIT_GENERAL. For example, if it is desired to run an implicit pre-
load phase and switch to the explicit solver for the subsequent transient phase, IDRFLG
should be set to 5 and IMFLAG should be set to 0.
IDRFLG=6 also performs an implicit analysis as with IDRFLG=5 (all parts) but only for
the part subset specified with DRPSET. This option was added at the request of the Aero-
space Working Group to allow implicit initialization of the jet engine bird strike models.
The initial time step size for implicit analysis (DT0) can have an effect on the converged
values of the preload. However, the smaller the time step, the greater the CPU.
Static analysis (IMASS=0) is well-suited to inducing preload. However, no rigid body
modes can be present for a static analysis. One has the option of dynamic implicit
(IMASS=1) combined with an extended loading period. Invoking dynamics can also in
general ease convergence. The step size has units of time if dynamics is invoked.
Note that for implicit dynamics, some additional numerical damping can be introduced via
the gamma and beta values: GAMMA=0.6 and BETA=0.4 have been used to good effect,
to eliminate high amplitude dynamic oscillation.
By default (IAUTO=0), the time step is set to be constant. Using the time step adjusting
option (IAUTO=1), the code automatically adjusts the time step size based on ease or
difficulty in achieving convergence.
4. Transient Implicit/Explicit Single Switch
As an alternative to using Implicit Dynamic Relaxation (ImpDR), in some cases the preload
can be established in the early part of the regular transient simulation.
Use one input deck where switching between implicit and explicit solver is determined by
a curve. The abscissa of the curve is time and the ordinate is set to 1.0 for implicit and to 0.0
for explicit (curve is a step function). This switching is activated by setting IMFLAG at
*CONTROL_IMPLICIT_GENERAL to -|curve ID|. Switching from one analysis to the
other is seamless and has no CPU or I/O overhead.
The initial time step size for implicit analysis (DT0) can have an effect on the converged
values of the preload. However, the smaller the time step, the greater the CPU.
Static analysis (IMASS=0) is well-suited to inducing preload. However, no rigid body
Modes can be present for a static analysis. One has the option of dynamic implicit
(IMASS=1) combined with an extended loading period. Invoking dynamics can also in
general ease convergence. The step size has units of time if dynamics is invoked.
Note that for implicit dynamics, some additional numerical damping can be introduced via
the gamma and beta values: GAMMA=0.6 and BETA=0.4 have been used to good effect,
to eliminate high amplitude dynamic oscillation.
to the dynamic stage.
The following provides some guidance, notes and necessary input features for modelling
the preloading phase of a bolted structure and the subsequent phase where the structure is
impacted by a separate object at a prescribed initial velocity (for example, an SPH bird-strike).
Much of this has been drawn from the very nice shared notes by Jim Day in the following,
regarding several preload approaches:
http://ftp.lstc.com/anonymous/outgoing/jday/bolt_preload3.pdf
http://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/preload.pdf
2. Implicit Dynamic Relaxation (ImpDR) - IDRFLG=5/6
When IDRFLG=5 or 6, LS-DYNA performs an implicit analysis for the preload phase of
the simulation. Parameters for controlling the implicit preload solution are defined using
appropriate *CONTROL_IMPLICIT keywords to specify solver type, implicit time step,
etc. When using this option, one must specify DRTERM to indicate the termination "time"
of the implicit preload analysis. When DRTERM is reached, the implicit preload phase
terminates and LS-DYNA begins the next phase of the analysis according to IMFLAG in
*CONTROL_IMPLICIT_GENERAL. For example, if it is desired to run an implicit pre-
load phase and switch to the explicit solver for the subsequent transient phase, IDRFLG
should be set to 5 and IMFLAG should be set to 0.
IDRFLG=6 also performs an implicit analysis as with IDRFLG=5 (all parts) but only for
the part subset specified with DRPSET. This option was added at the request of the Aero-
space Working Group to allow implicit initialization of the jet engine bird strike models.
The initial time step size for implicit analysis (DT0) can have an effect on the converged
values of the preload. However, the smaller the time step, the greater the CPU.
Static analysis (IMASS=0) is well-suited to inducing preload. However, no rigid body
modes can be present for a static analysis. One has the option of dynamic implicit
(IMASS=1) combined with an extended loading period. Invoking dynamics can also in
general ease convergence. The step size has units of time if dynamics is invoked.
Note that for implicit dynamics, some additional numerical damping can be introduced via
the gamma and beta values: GAMMA=0.6 and BETA=0.4 have been used to good effect,
to eliminate high amplitude dynamic oscillation.
By default (IAUTO=0), the time step is set to be constant. Using the time step adjusting
option (IAUTO=1), the code automatically adjusts the time step size based on ease or
difficulty in achieving convergence.
4. Transient Implicit/Explicit Single Switch
As an alternative to using Implicit Dynamic Relaxation (ImpDR), in some cases the preload
can be established in the early part of the regular transient simulation.
Use one input deck where switching between implicit and explicit solver is determined by
a curve. The abscissa of the curve is time and the ordinate is set to 1.0 for implicit and to 0.0
for explicit (curve is a step function). This switching is activated by setting IMFLAG at
*CONTROL_IMPLICIT_GENERAL to -|curve ID|. Switching from one analysis to the
other is seamless and has no CPU or I/O overhead.
The initial time step size for implicit analysis (DT0) can have an effect on the converged
values of the preload. However, the smaller the time step, the greater the CPU.
Static analysis (IMASS=0) is well-suited to inducing preload. However, no rigid body
Modes can be present for a static analysis. One has the option of dynamic implicit
(IMASS=1) combined with an extended loading period. Invoking dynamics can also in
general ease convergence. The step size has units of time if dynamics is invoked.
Note that for implicit dynamics, some additional numerical damping can be introduced via
the gamma and beta values: GAMMA=0.6 and BETA=0.4 have been used to good effect,
to eliminate high amplitude dynamic oscillation.
By default (IAUTO=0), the time step is set to be constant. Using the time step adjusting
option (IAUTO=1), the code automatically adjusts the time step size based on ease or
difficulty in achieving convergence.
option (IAUTO=1), the code automatically adjusts the time step size based on ease or
difficulty in achieving convergence.