Tweet
Hi,
I am working on the simulation of donning a prosthetic liner on the residual limb. The method I used can be described briefly as follows.
1) The residual limb model was reconstructed based on the MRI/CT data. At the beginning, an initial liner shape was designed by offsetting the skin surface outwards by some millimeters. So the system has bones, patella tendon, soft tissue, and liner. All the boundaries shared by the corresponding pairs. To avoid using contact, I did the following three steps.
2) I used multigeneration method for all the materials. The whole process could be completed in 3 steps. For the bones, patella tendon and the soft tissue, the material mechanical properties were 0 from t=0s to t=2s. After t=2s, the real mechanical properties would be assigned to the corresponding parts. For the liner, the material mechanical property was 0 from t=1, and then the real mechanical property would be assigned to it after t=1s.
3) The first step could be called liner design: I used a prescribed displacement boundary condition to move all the vertices of the liner inner surface (i.e., the skin surface) inwards around 10mm (not identical for all the vertices because some smoothing work was done for the shape of liner to make it more beautiful, so the displacement for each skin node was calculated according to the liner shape we want). When t=1s, I switched on the liner mechanical property. So currently the shape of the liner is its original shape, so the liner design was finished.
4) The second step could be called liner expansion, just as in the real world we expand a liner to prepare for putting it on the residual limb. In this step, I still used the prescribed displacement boundary condition to make the displacements of all nodes of the skin surface 0 relative to their initial positions (use absolute displacement). So now the liner was expanded to the very same shape as the residual limb.
5) The third step could be called liner don, just as in the real world the liner interacts with the soft tissue and reaches a final shape that can keep a balance these two parts. In this step, I switched on the mechanical properties of the soft tissue and the bones.
My question is: Step1 and Step2 could run successfully, but Step3 would crash just after t=2s. I tried many methods to tune this. In my understanding, because the liner was stretched, it would apply very large stress or force on the skin surface when t=2s. So we need to do something to balance the stress or force. I tried applying nodal forces (obtained by computing the reaction forces of each node of the skin at the end of step2), applying pressure (obtained by computing the stress of each face of the skin at the end of step2), applying traction (obtained by computing the traction of each face of the skin at the end of step2) at t=2s, and then decreased the values to 0 linearly. But none of them worked. The computation could not go any further after t=2s.
Is there any problem on my method or setup? Can you give some suggestion to make it work?
Thanks.
Sinba
I am working on the simulation of donning a prosthetic liner on the residual limb. The method I used can be described briefly as follows.
1) The residual limb model was reconstructed based on the MRI/CT data. At the beginning, an initial liner shape was designed by offsetting the skin surface outwards by some millimeters. So the system has bones, patella tendon, soft tissue, and liner. All the boundaries shared by the corresponding pairs. To avoid using contact, I did the following three steps.
2) I used multigeneration method for all the materials. The whole process could be completed in 3 steps. For the bones, patella tendon and the soft tissue, the material mechanical properties were 0 from t=0s to t=2s. After t=2s, the real mechanical properties would be assigned to the corresponding parts. For the liner, the material mechanical property was 0 from t=1, and then the real mechanical property would be assigned to it after t=1s.
3) The first step could be called liner design: I used a prescribed displacement boundary condition to move all the vertices of the liner inner surface (i.e., the skin surface) inwards around 10mm (not identical for all the vertices because some smoothing work was done for the shape of liner to make it more beautiful, so the displacement for each skin node was calculated according to the liner shape we want). When t=1s, I switched on the liner mechanical property. So currently the shape of the liner is its original shape, so the liner design was finished.
4) The second step could be called liner expansion, just as in the real world we expand a liner to prepare for putting it on the residual limb. In this step, I still used the prescribed displacement boundary condition to make the displacements of all nodes of the skin surface 0 relative to their initial positions (use absolute displacement). So now the liner was expanded to the very same shape as the residual limb.
5) The third step could be called liner don, just as in the real world the liner interacts with the soft tissue and reaches a final shape that can keep a balance these two parts. In this step, I switched on the mechanical properties of the soft tissue and the bones.
My question is: Step1 and Step2 could run successfully, but Step3 would crash just after t=2s. I tried many methods to tune this. In my understanding, because the liner was stretched, it would apply very large stress or force on the skin surface when t=2s. So we need to do something to balance the stress or force. I tried applying nodal forces (obtained by computing the reaction forces of each node of the skin at the end of step2), applying pressure (obtained by computing the stress of each face of the skin at the end of step2), applying traction (obtained by computing the traction of each face of the skin at the end of step2) at t=2s, and then decreased the values to 0 linearly. But none of them worked. The computation could not go any further after t=2s.
Is there any problem on my method or setup? Can you give some suggestion to make it work?
Thanks.
Sinba
Comment