rm447
Bioengineer
- Sep 19, 2007
- 21
Hi! Has any body ever had to deal with a similar problem as follows?
I have two stl files, one is of an object in an unstressed state, and one is of the same object after some deformations were applied to it through manipulation by some unknown forces. I need to find the resulting stress field so I need to be able to put these deformations in the form of nodal displacements to include in my ABAQUS input file, and then find the stresses using my material model. How would one go about such a problem of reversely finding displacements using the initial and final shapes?
Here is what I though might work but it didn't. May be some one can better this approach or suggest a different one?
I used a graphical analysis software that can do a 3D analysis to find how different the two geometries are. It automatically pairs up two points on the objects to find the deviation between them (I already had registered and aligned the two objects using their principal axes). I picked some of these points that had higher deviations (both positive and negative). Since the information was not given at vertices, I then used Patran to find nodes nearest these points. Then I applied the deviation as B.C.'s to these nodes. I thought that if I apply enough nodal displacements it can deform that area in a smooth way to result in the post-manipulation shape. But the problem is that this was not the case. Because I am only moving some of the nodes, I am only getting some spikes at those nodes and the rest of the surface remains pretty much intact.
Does any one have any ideas? I will appreciate any input! Thanks.
I have two stl files, one is of an object in an unstressed state, and one is of the same object after some deformations were applied to it through manipulation by some unknown forces. I need to find the resulting stress field so I need to be able to put these deformations in the form of nodal displacements to include in my ABAQUS input file, and then find the stresses using my material model. How would one go about such a problem of reversely finding displacements using the initial and final shapes?
Here is what I though might work but it didn't. May be some one can better this approach or suggest a different one?
I used a graphical analysis software that can do a 3D analysis to find how different the two geometries are. It automatically pairs up two points on the objects to find the deviation between them (I already had registered and aligned the two objects using their principal axes). I picked some of these points that had higher deviations (both positive and negative). Since the information was not given at vertices, I then used Patran to find nodes nearest these points. Then I applied the deviation as B.C.'s to these nodes. I thought that if I apply enough nodal displacements it can deform that area in a smooth way to result in the post-manipulation shape. But the problem is that this was not the case. Because I am only moving some of the nodes, I am only getting some spikes at those nodes and the rest of the surface remains pretty much intact.
Does any one have any ideas? I will appreciate any input! Thanks.