Import Initial State from Static Analysis for Implicit Dynamic Analaysis

[aurum]

My earthquake analyses steps are as follows:

1. I have a dam-reservoir-foundation system analysed for static loads (gravity, hydrostatic etc.). This is performed in a static, general step.

2. I open a separate cae file which is exactly the model definition from step and import the .odb from the step 1 as the initial state for subsequent dynamic analysis (seismic load).

3. Using the GUI/python interpreter, I created a Predefined field as the initial state by selecting dam, foundation and reservoir instances each using the .odb file from the step 1.

Q1.Is this the right way?

Q2.I tried verifying if the initial state has been imported properly by doing another static analysis step instead of dynamic analysis with no loads by reasoning that the results from this step should be identical to the results from step 1. But they are not (both quantities and directions of displacements (upward displacement instead down))

Help me point out the errors!

 

[NRP99]

Why not use in same CAE file dynamic step after static step so that the results are carried forward? If its nonlinear analysis it should carry forward the "state" of component to proceeding step.

 

[FEA way]

It would be easier to perform a multi-step analysis (unless you want to modify the assembly in the second step) - static followed by the dynamic procedure. In addition, did you consider running a response spectrum analysis ? This is often done when evaluating a structure's response to earthquakes. It's a linear dynamics approach involving some approximations but in many cases full non-linear dynamics simulation is not necessary.

If you run a static analysis with no load then Abaqus will solve for static equilibrium and it will be something like a springback analysis after metal forming - the structure will recover from elastic deformation and you will be left with residual stresses.

 

[aurum]

I was not entirely sure multi-step analysis would carry forward the results from the static steps to subsequent dynamic steps. A lot of my earlier simulations were done that way and out of blue, I had this doubt if should I import results or should I do multi-step. Thanks for confirming.

Response spectrum analysis was done and I need to do quite some non-linear seismic analysis for the work. And that makes sense about the structure behaving like in a spring back analysis and why the static analysis with no load had results in the opposite direction to the initial state. Great. I get it now.

Thanks, @NRP99 and @FEA way.