Ok let’s imagine I have a flat bar and I want to bend it on the thin direction as step 1 and then on step 2 I want to see how the material would behave if I twisted it at that spot. I know this may sound weird, but could I some how reset all displacement values to 0 and then twist the structure starting at the pre-deformed position? This would give me an idea of how the twist would individually affect the shape of the part given its new shape that is still under load from the initial step. I don’t want to run them independently, I want to see the effect of the twist given the bend as the starting point.
If you include plasticity, you can run it in two steps - the second one will use the deformed step from the end of the first one. However, if you are only interested in the deformed shape (not material state) from the first step then export the deformed mesh and import it for a new analysis (like in nonlinear buckling analyses with imperfections from linear buckling).
Abaqus/Explicit has the *ANNEAL procedure that performs virtual annealing - sets all stresses, plastic strains and velocities to zero. And you can also choose whether to update the reference configuration when importing the model (deformed mesh and material state) between the analyses.
2 Likes
Are you talking about post-processing the results only?
1 Like
I think its mostly a post processing “thing” like FEAnalyst suggest but more of a displacement annealing keeping the streeses the same, can I see what step two did to the form of the part? So we take the new shape, set displacements to 0 in the new shape, keep internal streases, apply new displacements and calculate a new final shape. This would tell me if the new motion has correctly deformed the part per my espectation and then I could still investigate the internal stresses.
The stress annealing step sounds really useful for brazing steps for example between parts of dissimilar metals. When the liquid metal solidifies but stay close to melting for a temp recipe amount of time, it will result is very interesting and useful internal stresses. There are very many applications I can think of where this would be useful. But I digress.
I think it would be as simple (lol) as substracting the first step’s displacement vector/matrix from the second step’s vector. Sounds simple if we think of it as the math representation. But how, I dont know. I think I could write a python program maybe to process the frd file that way? I dont know how it would be done as a formal part of prepomax with a button. I dont even know what one would call it…discrete step displacement vs continuous or commulative displacement?
This could be added to the equation-based field outputs. But currently, mixing/selection steps is not integrated.