QUASI STATIC RADIOSS

Eser Dogan

Hello,

I want to make quasi-static nonlinear analysis in Radioss. My model does not have any contacts but I take time step error.

Can you help me ?

Eser Dogan

This my input file, How can I fix it ?

 

Unable to find an attachment - read this blog

Eser Dogan

Can you help me

 

Simon Križnik

Hi,

 

• the loading function should be ramped up from zero
• use Tools>Engine file assistant to define simulation parameters
• ctrl+F find IMPL/QSTAT = implicit quasistatic

Unable to find an attachment - read this blog

Eser Dogan

Hello, thank you very much for your response.

If I have any loading when time is zero, what can I do ?

Eser Dogan

Also, Quasi-static solver takes 15 min to solve it; however, explicit solver takes 1 min.

What is the reason of this difference ?

Simon Križnik

Hi,

 

glad to help.

 

In the explicit quasistatic analysis it is recommended:

• The loading should be gradually ramped up from zero over a reasonable period (at least 0.3 s). Instantaneous loading (non-zero loading at time zero) causes propagation of a stress wave through the model similar to impact load.
• The kinetic energy should be bellow 5% internal strain energy throughout the simulation to ensure a quasistatic process without inertial effects.
• Vibrations should be damped using for instance /DAMP Rayleigh damping.

 

The implicit quasistatic analysis with instantaneous loading has convergence difficulties as the equilibrium solution at initial load is far from the initial position. As a workaround try:

• create two engine files with the first engine ending after the preload is applied using the ramping function and the second simulation starts in the equilibrium condition
• perform pretension analysis and export deformed shape and stress fields using STATE file. Then initialize the stresses and deformed position in the original solver deck. However, this approach is not tested as I am away from my Hyperworkstation.

Implicit and explicit solutions can be compared only if the explicit solution satisfies above recommendations regarding slower loading and damping- both will increase explicit solution CPU  time. The implicit solution could be further optimized for better convergence. Note also that Radioss implicit is only meant to complement explicit (capability to switch during simulation)- otherwise use NLSTAT non-linear quasi-static analysis in Optistruct.