Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Help finding appropriate Young's mod and Poisson's ratio values for contact model
Posted 01.05.2017, 10:13 GMT-4 5 Replies
Please login with a confirmed email address before reporting spam
I'm trying to model the pressure distribution of an experiment I conducted where a fixed silicone rubber hemisphere was pressed against a glass plate by moving the platform the plate was mounted on into the hemisphere using fixed distances (0.5 -> 2 mm). This was conducted to determine the surface area of contact for specific forces.
Because the silicone hemisphere was moulded in the lab (and assuming it isn't hyperelastic), I need a way to find the appropriate values for the hemisphere's Young's modulus and Poisson's ratio. However, this has been more difficult (to guess) than I thought it would be.
At 2 mm of compression, the resulting force was about 8.5 N. However, using the model I have with [Young's mod = 0.008 GPa, Poisson's ratio = 0.47], a total force of 220.7 N is calculated. By increasing the Young's modulus, this total force decreases but the model struggles to solve. Increasing the Poisson's ratio allows the model to solve at higher Young's mod values but the solutions don't always seem correct. e.g. with [Young's mod = 0.0005 GPa, Poisson's ratio = 0.4999] the total force = -18.718 N.
(the surface plots for the von mises stress of each of these examples are attached)
Therefore the help I need is in finding appropriate Young's mod and Poisson's ratio values for the model resulting in an 8.5 N total force for a 2 mm displacement.
(My model is attached)
P.s. the silicone rubber is 'Material 4' in the model and its properties are changed in the Parameters under Global Definitions.
Because the silicone hemisphere was moulded in the lab (and assuming it isn't hyperelastic), I need a way to find the appropriate values for the hemisphere's Young's modulus and Poisson's ratio. However, this has been more difficult (to guess) than I thought it would be.
At 2 mm of compression, the resulting force was about 8.5 N. However, using the model I have with [Young's mod = 0.008 GPa, Poisson's ratio = 0.47], a total force of 220.7 N is calculated. By increasing the Young's modulus, this total force decreases but the model struggles to solve. Increasing the Poisson's ratio allows the model to solve at higher Young's mod values but the solutions don't always seem correct. e.g. with [Young's mod = 0.0005 GPa, Poisson's ratio = 0.4999] the total force = -18.718 N.
(the surface plots for the von mises stress of each of these examples are attached)
Therefore the help I need is in finding appropriate Young's mod and Poisson's ratio values for the model resulting in an 8.5 N total force for a 2 mm displacement.
(My model is attached)
P.s. the silicone rubber is 'Material 4' in the model and its properties are changed in the Parameters under Global Definitions.
Attachments:
5 Replies Last Post 03.05.2017, 11:16 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Hello Steven,
I took a quick look around your file and noticed two things:
- your parameters are called "Young" and "Poisson", but material 4 uses Y and P.
- you have not turned on Almost Incompressible Material, although you have a Poisson ratio very close to .5.
There may be other things but those two stood out for me.
Best,
Jeff
I took a quick look around your file and noticed two things:
- your parameters are called "Young" and "Poisson", but material 4 uses Y and P.
- you have not turned on Almost Incompressible Material, although you have a Poisson ratio very close to .5.
There may be other things but those two stood out for me.
Best,
Jeff
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Thanks for your reply, Jeff
Y and P are variables with the expression Young and Poisson respectively because I wanted to include these in my table.
As for using the nearly incompressible material option, this causes even less convergence and stranger solutions such as the one attached
Y and P are variables with the expression Young and Poisson respectively because I wanted to include these in my table.
As for using the nearly incompressible material option, this causes even less convergence and stranger solutions such as the one attached
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
You're right, I totally missed the fact that you used variables to turn Young into Y and Poisson into P.
A couple of additional comments:
- At nu=.47 not turning on the almost incompressible material switch is probably safe, but at .499 I'd be cautious.
- If the experimental data suggests that the total force is lower, wouldn't that suggest that the Young's modulus is lower than the one you've tried? The softer the silicon, the less force it takes to get it to deform a given amount, no?
- It looks like in your file the contact zone is only a small number of mesh elements. You could try refining the mesh.
- The two materials have Young's moduli that are several orders of magnitude apart. The accuracy could be suffering from ill-conditioning, see www.comsol.com/community/forums/general/thread/141001 and the blog it points to.
- You are computed Force by "manually" integrating the pressure over surface 7. I would have used the preimplemented Total contact force, z component (solid.T_totz).
Jeff
A couple of additional comments:
- At nu=.47 not turning on the almost incompressible material switch is probably safe, but at .499 I'd be cautious.
- If the experimental data suggests that the total force is lower, wouldn't that suggest that the Young's modulus is lower than the one you've tried? The softer the silicon, the less force it takes to get it to deform a given amount, no?
- It looks like in your file the contact zone is only a small number of mesh elements. You could try refining the mesh.
- The two materials have Young's moduli that are several orders of magnitude apart. The accuracy could be suffering from ill-conditioning, see www.comsol.com/community/forums/general/thread/141001 and the blog it points to.
- You are computed Force by "manually" integrating the pressure over surface 7. I would have used the preimplemented Total contact force, z component (solid.T_totz).
Jeff
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Thanks again for your replies, Jeff!
I definitely meant to write decreasing when it came to the young's mod to have lower forces. I've been doing this so I'm not sure why I accidentally wrote increase.
I think you're right about the ill-conditioning. When I changed the glass to acrylic, my model solved without any problems. I've tried changing the mesh size previously but it hasn't helped much. I just tried solving the model with an extremely fine mesh and very tight tolerance in the stationary solver but I'm still having this problem with glass. Is there anything else I can try to help my model converge correctly?
Shouldn't the integrated pressure across the area be the same as the inbuilt contact force? I've found that they're slightly different but I don't understand why.
I definitely meant to write decreasing when it came to the young's mod to have lower forces. I've been doing this so I'm not sure why I accidentally wrote increase.
I think you're right about the ill-conditioning. When I changed the glass to acrylic, my model solved without any problems. I've tried changing the mesh size previously but it hasn't helped much. I just tried solving the model with an extremely fine mesh and very tight tolerance in the stationary solver but I'm still having this problem with glass. Is there anything else I can try to help my model converge correctly?
Shouldn't the integrated pressure across the area be the same as the inbuilt contact force? I've found that they're slightly different but I don't understand why.
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Updated:
8 years ago
Hi Steven,
If we go with the hypothesis that the discrepancy between your experimental data and model results as well as your difficulty in getting the model to converge at all in some cases are coming from the large difference in Young's moduli between the two materials, then a logical thing to try would be to artificially decrease the Young's modulus of glass in your model. As long as it remains, say, a couple of orders of magnitude larger than that of the spherical cap, that won't affect the solution significantly but it would make a big difference in terms of the suspected ill-conditioning.
Best,
Jeff
PS: I always go with the pre-implemented expressions when they exist: someone's already done the thinking for me and I don't trust my ability to invent a better wheel.
If we go with the hypothesis that the discrepancy between your experimental data and model results as well as your difficulty in getting the model to converge at all in some cases are coming from the large difference in Young's moduli between the two materials, then a logical thing to try would be to artificially decrease the Young's modulus of glass in your model. As long as it remains, say, a couple of orders of magnitude larger than that of the spherical cap, that won't affect the solution significantly but it would make a big difference in terms of the suspected ill-conditioning.
Best,
Jeff
PS: I always go with the pre-implemented expressions when they exist: someone's already done the thinking for me and I don't trust my ability to invent a better wheel.