Discussion Forum

Pulling a bunch of posts on the same problem together....

July 28, 2011 5:33pm EDT

Parametric Sweeps to model Progressive Die

I am trying to model the deformation of a part with a series of progressive dies. (A contact elastic-plastic deformation)

I can sweep the prescribed displacement to move a die to shape a part. (the maximum displacement of the tool is multiplied by a global sawtooth function defined by the parameter that goes from 0 to 6 in .1 increments)

I can create a parameter to change the shape of the die to progressively form the part. (a parameter changes the angle in the geometry from 45 to 15 to 0) i.e. bend the part 45 degrees then 15 degrees then flat.

However I am having a lot of difficulty putting these two parameter changes together.

The key here is I need to copy the STRESS state from from the end of one die sweep to the beginning of the next die run. So I cannot copy the deformed shape from one run to a new geometry for another run.

In a parametric sweep, with 2 parameters, displacement and angle in that order, the parameter values are:
0 135 0.1 135 0.2 135 0.3 135 0.4 135 0.5 135 0.6 135 0.7 135 0.8 135 0.9 135 1 135 1.1 135 1.2 135 1.3 135 1.4 135 1.5 135 1.6 135 1.7 135 1.8 135 1.9 135 2 165 2.1 165 2.2 165 2.3 165 2.4 165 2.5 165 2.6 165 2.7 165 2.8 165 2.9 165 3 165 3.1 165 3.2 165 3.3 165 3.4 165 3.5 165 3.6 165 3.7 165 3.8 165 3.9 165 4 180 4.1 180 4.2 180 4.3 180 4.4 180 4.5 180 4.6 180 4.7 180 4.8 180 4.9 180 5 180 5.1 180 5.2 180 5.3 180 5.4 180 5.5 180 5.6 180 5.7 180 5.8 180 5.9 180 6 180

This does not work.

Any thoughts or alternate approaches to solve this?

Thanks,

Roy Martin

pulling in yet another post on same topic

July 29, 2011 12:12pm EDT

Parametric Sweep solving inconsistancies

I am trying to model a progressive die, a contact elastic-plastic large deformation model.

One consistant problem I have is the inconsistant way the parametric sweep solves the problem.

I set up a run of the first tool geometry. So the parametric sweep only has one parameter for prescribed displacement.

parameter values are 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0

The study develops solutions for all parameter values. It does not sweep the geometry.

Next step, add a second parameter to the sweep, this one control geometry shape, and angle in this case.

in this first test case, the angle is constant for the run.

parameter values are 0 135 .1 135 .2 135 .3 135 .4 135 .5 135 .6 135 .7 135 .8 135 .9 135 1.0 135 1.1 135 1.2 135 1.3 135 1.4 135 1.5 135 1.6 135 1.7 135 1.8 135 1.9 135 2.0 135

This is where it gets weird. It will begin to loop through the first parameter for each of the 2nd paramters. And each time it loops it resets the initial conditions to zero.


So, any suggestions on how to sweep a displacement from 0 to 2 with one geometry, keep the deformed shape and stresses, then sweep a displacement from 0 to 2 again with a modified second geometry, then keep the deformed shape and stresses, then sweep for a third time with yet another modified geometry.

Or is this not possible with COMSOL.

OK,

I think I have come up with a way to model the action of a progressive die forming a part.

Model the part and ALL the tools in one geometry. Stack the tools vertically in the model.

Use prescribed deformation to move the tools into then out of the part in sequence with 1 parameter and logical conditions to time the motion of the tools.

This seems to work and allow each new tool to affect the deformed AND stressed shape of the part.

now I just need to figure out how to apply the ALE (moving mesh) physics to prevent the deformed mesh from inverting...