ABAQUS and PYTHON continuous time history solving

[a126z]

Experts, I would like to ask is there a way to solve the problem according to the time step in ABAQUS? That is, load once and solve once, but do not end the solution, the second time continue to load on the results of the first operation, continue to solve, and repeat this process. Can this be done quickly using PYTHON code?

 

[FEA way]

That’s basically the idea behind multistep simulations. You just need general static (or dynamic) steps with nonlinearities. You could also use the restart feature that allows you to continue the analysis from a selected point (including its end).

 

[a126z]

Thank you very much

 

[a126z]

I would like to ask if there are n analysis steps in ABAQUA, it is possible to analyze one step at a time and output only the displacement result of the corresponding step, but the final ODB file is only one, which contains these n analysis steps.

 

[FEA way]

You could use the restart capability to add more steps to a multistep analysis one by one. Then it's possible to combine output databases.

 

[a126z]

Thank you for your advice. I restarted the model, but I realized that this method slows down the calculation, is there a way to fix this?

 

[FEA way]

Wouldn’t a single multistep analysis be sufficient for you ? Then you have one run, you can postprocess the results of individual steps and those results will be stored in a single database. But if, for some reason, you want to solve only for a single step at a time then restart seems to be the best option.

 

[a126z]

Since what I want is so that I can load and output the results at the same time, and then continue to load based on the results afterwards to realize the loop, I'm not sure if a single multi-step analysis can do it?

 

[FEA way]

If you want to modify the load after solving one step (based on its results) then restart will be the way to go.

 

[a126z]

However, using the reboot method will slow down the operation speed. I don't understand why it takes longer time to calculate in the loop even though I have used the method of overwriting the previous ODB result file.

 

[FEA way]

Maybe you could use Isight to automate your analyses. But still, they will have to be submitted multiple times leading to some slow-down.

 

[a126z]

But why can this process be done efficiently inside ANSYS by writing APDL command flows and recording macro files? I think ABAQUS also has such a way to do it quickly.