I have a use case where I want to simulate a real 3-point-bending test, performed on a specific specimen. On the real test a preload is applied, then displacement and load are set to zero with a tare function, and finally a specified displacement is applied. Since the displacement was set to zero at the preload point, the reported displacment is relative to the point at which preload was achieved.
I have attempted to simulate this using two static steps with nonlinear geometry. The first step applies a preload, and the second step applies a displacement. However, this does not have the intended result. If the preload causes a displacement of 1mm relative to the top of the sample, and I call for a 3mm displacement step afterward, the end result is a final displacement of 3mm only, rather than the 4mm I expect.
I have an idea of how I might solve this with outside scripting, since both steps are static. Instead of running a single .pmx file with two static steps, I would run two .pmx files, each with a single static step. The first simulation would contain only a force step with the desired preload. Once the analysis has run I would take the final displacement of the anvil under the preload and use it as a parameter in the second simulation, which would be identical except for using a displacement step instead of a force step. The displacement value in the second simulation would be the sum of the actual displacement from the force step and the desired relative displacement.
This plan seems to me like it would work, but I would like to know if anyone else has a simpler or more direct way of modeling the preload-displacement combination.
I mainly work with amplitudes more than steps to sequence loads and displacements. That’s why I’m not sure to understand what you mean.
¿Do you mean subsequent steps do not take previous deformations under consideration? I think they do. Isn’t the fix parameter take care of that?
I only mean basic case when you have first step with some load and second with another load while the first one is deactivated. Then the second step doesn’t use the deformation from the first step even when Nlgeom is enabled unless there’s plasticity. That’s why you can e.g. precede the load control step with displacement control one to establish contact without affecting the results. Or analyze independent load cases.
However, if you consider other points from the post linked above, you could keep/redefine the first load in the second step or use the FIXED parameter to freeze the deformation from the previous step.
@FEAnalyst
Thanks . Now I realize I didn’t understand your response because I also didn’t understand the initial question.
ryantheengineer is trying to apply a superposed displacement BC into a predeformed part. That’s not possible as you say but there could be an easy workaround.
Three point bending test has two supports and one punch. Initial preload can be imposed on the supports upright direction and later test displacement BC imposed on the punch downwards.
EDITED: A pretension secction in the Puch itself should also work. No need to touch the suporting area.