Learning Forum

[Tonmoy]

Hi everyone!

I am using explicit dynamic relaxation (DR) to initialize my LS-Dyna model under gravity load. My model includes some soil elements and a foundation pad placed on the top of it. I applied fixed boundary condition at the outer surface of the soil and applied the gravity load just on the pad but I have some questions about it.

First, why the fringe component of the z-stress shows positive values for most of the soil elements? Shouldn't they be something close to zero?

Second, how should I decide if I need to opt for the implicit or the explicit DR if my model terminates normally for both of them?

Thank you for your help.

[Alex R.]

Hello,

First, why the fringe component of the z-stress shows positive values for most of the soil elements? Shouldn't they be something close to zero?

 Please note that based on this scale the elements in red can be somewhere between -2.624+4 to 2.674E+3, it cannot be evaluated if the fringe values are close to zero or not. I think changing the scale will help here.

Second, how should I decide if I need to opt for the implicit or the explicit DR if my model terminates normally for both of them?

Does the model start oscillating after switching to the trainset phase of the simulation? Can you verify the load, such as gravity forces or reaction forces? If the model does not oscillate after the DR phase and the reaction forces are agreeing with hand calculations the DR is most likely successful.

Thank you,

Alex

[Tonmoy]

Thanks Alex for your suggestions. I can now understand the confusion on the stress values: I rather tried to read the Z-stress of the surface elements and found that they are some positive values and the value depends on their location. The one with just beside the foundation pad has this large positive stresses.

The comments on DR make sense. In my model, both the reaction forces and stresses looked fine, but I noticed that with implicit DR, the total KE from glstat increased more quickly compared to explicit DR. However, as Dr. Basu mentioned, I’d rather switch to explicit gravity initialization instead of using DR.

Thank you again for your time and help to respond to my queries. I'm marking the question as answered.

 

- Tonmoy.

[Ushnish Basu]

Firstly, I don't think PML elements should be used in a model undergoing dynamic relaxation. More pertinently, I would recommend doing gravity initialization via explicit analysis with a ramp load and damping.

[Tonmoy]

Thank you Dr. Basu for your help. I have fixed the model by replacing those keyword cards and now doing an explicit gravity initialization.

 

- Tonmoy.