AnsweredAssumed Answered

Displacement on a Press Plate Issue in Nonlinear

Question asked by Megan Derwich on Jul 22, 2014
Latest reply on Jul 25, 2014 by Jerry Steiger

Like • Show 0 Likes0
Comment • 8

Hi,

I am really new at that nonlinear simulations and I am having some issue with making a press plate move down on to my part. My project is taking a lid to a cup and applying it to a cup using a press plate (similar to a top down compression test). I currently have it cut in half, so I can have a higher mesh, because of this I have semmetry on all the 3 of the part in the assembly. When I have my mesh on draft quality, my displacement on the press plate would push the lid down and on to the cup, However, when I changed my mesh from draft quality to normal mesh quality the displacement on my press plate goes from moving 17 mm plus to moving 1.2 mm. This is the video I have with draft quality mesh lid application https://www.youtube.com/watch?v=8NxoawqFrPw&feature=youtu.be

Here are my current setting, fixtures, and warning that I have on my parts.

On Flate Faces -2 Translations: x-0mm, y--15mm, z-0mm

On Flate Faces - 3 Translations: x-0mm, z-0mm

On Flate Face - 4 Translations: x- 0mm, y-0mm, z- 0mm

The warning I get when I am running the simulation is Max. allowable iteration increased by 5, and PCDLSSO102: Assembled Stiffness.

Can I get some help with why my displacement changes and does not push my press plate into my lid in the non-draft quality mesh.

Thank you.

Nicholas Luyster Jul 22, 2014 6:04 PM

Correct Answer

Hi Megan,

Can you simplify things down a bit more by making sure the lid and the press and the cup are all initially touching. It makes it harder to converge when things aren't initially touching. Take a look at this video at 3:30 to see how you can make the bodies initially touching.

Looking at your geometry, you'll probably want to go with option #2 like Shaun said. The ruffles on the lid will prevent you from option 3.

All the best,

Nick Luyster

SolidWorks Simulation Training

See the reply in context

No one else had this question

Outcomes

Shaun Densberger Jul 22, 2014 3:12 PM

Could you post the model itself? At the moment I can't think of a reason why the order of the elements would cause this effect (draft quality elements have linear interpolation functions, whereas normal quality elements have quadratic); very interesting.

I also want to point out that you can reduce this problem down in size even further with symmetry. Currently you have a 1/2 symmetry model, but from what I've seen in the Youtube video, you could also do:

Quarter Symmetry
Circular Symmetry
2D Axisymmetric

Ideally, we'd like to do #3, as the 2D idealizations are by far the fastest models to simulate. However, the rib features on the lid technically violate the assumptions of a 2D axisymmetric model. If you feel that these features have minimal impact on the results, you could suppress them, thereby allowing you to run a 2D simulation.

That being said, I see no reason why you can't do #2. All you need to do is cut your geometry up such that you have some angular section of it that is repeated cylindrically. Again, based on what I see in the Youtube video, this would probably be from the center of one rib to the center of an adjacent rib.

The simulation time is roughly N^2 where N is the number of DOF associated with your model. So a 1/2 symmetry model will take 1/4th the time; a 1/4 symmetry model will take 1/16th the time. The circular and 2D cases wil be even faster.

Actions

Nicholas Luyster Jul 22, 2014 6:04 PM

Hi Megan,

Looking at your geometry, you'll probably want to go with option #2 like Shaun said. The ruffles on the lid will prevent you from option 3.

All the best,

Nick Luyster

SolidWorks Simulation Training

Actions

Jared Conway Jul 22, 2014 11:38 PM

What's the goal of the analysis?

Starting with everything touching is a good idea.

I think you "got lucky" with the draft solution. High quality is the way to go.

Actions

Megan Derwich Jul 23, 2014 11:25 AM

Jared- The goal of this analysis is to find the resultant forces on the top of the press plate and have the results be similar to displacement and the force of the actual real cup that I am testing. It is to be able to take new products and be able to accurately simulate forces for lid application and different lid designs.

Nicholas and Shaun- Last night I ran my simulation with a normal mesh, it got to 94.4% when it hit the warning of max iterations increased by 5. After looking at the results the total Y displacement was 5.9mm even though it was set to move the press plate 7mm. The lid did not fully set on the cup due to the change in displacement (this is still working with the half of the cup).

I thought that taking the off the on flat faces fixtures would prevent the warnings that I am having from coming up early in the simulation run. My current fixtures now are just the reference gerometry on the top of the press place, Translation: x-0mm, y-0mm and z--7mm. I then made a refence geometry on the top of the lid and made the translation to the top plane x-0mm and y-0mm. My idea was that the press plate would not move in the x, y, plane and the only be able to move the lid in the z direction.

Currently I am trying to get my model to accept circular symmetry. here is the picture of the current model.

The only issue is that I can not select the circular symmetry option.

All the properties have stayed the same as well as the component contacts of an incompatable mesh.

Actions

Shaun Densberger @ Megan Derwich on Jul 23, 2014 2:04 PM

Last night I ran my simulation with a normal mesh, it got to 94.4% when it hit the warning of max iterations increased by 5. After looking at the results the total Y displacement was 5.9mm even though it was set to move the press plate 7mm. The lid did not fully set on the cup due to the change in displacement (this is still working with the half of the cup).

I think that this is just because the simulation didn't finish, so all you're seeing is the results at pseudo time step t < t_final.

I think that it'd be worthwhile to remove the die that pushing the lid down from simulation, and instead apply the enforced displacement on the portion of the lid that the die was in contact with (I doubt that this will add a measurable about of stiffness to the model). I'd also adjust your model such that you need as little enforced displacement as possible (ie parts in initial contact as Nicholas suggested and determining how much displacement is needed for the lip of the lid to snap over the cup). I'd also make sure that you have Arc-Length control enabled (if you already don't) under the Advanced Options in the Solution tab. All of these should help improve the numerical stability of the problem, which is most likely the cause of your current issues.

Currently I am trying to get my model to accept circular symmetry. here is the picture of the current model.

Hmmmm. I'm not sure why the circular symmetry constraint is locked out, but in this case you could just apply a normal constraint on the exposed surfaces.

Actions

Jared Conway @ Shaun Densberger on Jul 24, 2014 3:36 AM

I don't think circular symmetry is supported in nonlinear. And in this case normal symmetry is the better selection. See kb for explanation.

Since you are still learning, I'd recommend building a really simple model to do some testing on and that you can post.

I suspect a combo of setup issues and challenging contact and nonlinear problem that will require some massaging to work properly. Are you working with your var on this? Or reaching out for training help? That might be more efficient if you can't post the model.

Actions

Megan Derwich Jul 25, 2014 10:01 AM

So after multiple trials and remaking the orginal assembly, it seems to work. I took off my mesh controls because they kept causing errors on the on normal mesh. Attached is the video of the fully run simulation with extremely close to the min. forces needed to apply the lid to the actual real life part. https://www.youtube.com/watch?v=IbpHzv58TfM and the final results

I would like to thank everyone for helping me with this! The forum has been extremely helpful and the responce to my problem has been quciker than the paid SolidWorks help.

Actions

Jerry Steiger @ Megan Derwich on Jul 25, 2014 4:53 PM

Megan,

Cool!

Jerry S.

Actions

8 Replies

Shaun Densberger Jul 22, 2014 3:12 PM
Mark Correct
Correct Answer
Could you post the model itself? At the moment I can't think of a reason why the order of the elements would cause this effect (draft quality elements have linear interpolation functions, whereas normal quality elements have quadratic); very interesting.

I also want to point out that you can reduce this problem down in size even further with symmetry. Currently you have a 1/2 symmetry model, but from what I've seen in the Youtube video, you could also do:
1. Quarter Symmetry
2. Circular Symmetry
3. 2D Axisymmetric
Ideally, we'd like to do #3, as the 2D idealizations are by far the fastest models to simulate. However, the rib features on the lid technically violate the assumptions of a 2D axisymmetric model. If you feel that these features have minimal impact on the results, you could suppress them, thereby allowing you to run a 2D simulation.

That being said, I see no reason why you can't do #2. All you need to do is cut your geometry up such that you have some angular section of it that is repeated cylindrically. Again, based on what I see in the Youtube video, this would probably be from the center of one rib to the center of an adjacent rib.

The simulation time is roughly N^2 where N is the number of DOF associated with your model. So a 1/2 symmetry model will take 1/4th the time; a 1/4 symmetry model will take 1/16th the time. The circular and 2D cases wil be even faster.
Like • Show 1 Like1
Actions
Nicholas Luyster Jul 22, 2014 6:04 PM
Unmark Correct
Correct Answer
Hi Megan,

Can you simplify things down a bit more by making sure the lid and the press and the cup are all initially touching. It makes it harder to converge when things aren't initially touching. Take a look at this video at 3:30 to see how you can make the bodies initially touching.

Looking at your geometry, you'll probably want to go with option #2 like Shaun said. The ruffles on the lid will prevent you from option 3.

All the best,

Nick Luyster
SolidWorks Simulation Training
Like • Show 0 Likes0
Actions
Jared Conway Jul 22, 2014 11:38 PM
Mark Correct
Correct Answer
What's the goal of the analysis?

Starting with everything touching is a good idea.

I think you "got lucky" with the draft solution. High quality is the way to go.
Like • Show 0 Likes0
Actions
Megan Derwich Jul 23, 2014 11:25 AM
Mark Correct
Correct Answer
Jared- The goal of this analysis is to find the resultant forces on the top of the press plate and have the results be similar to displacement and the force of the actual real cup that I am testing. It is to be able to take new products and be able to accurately simulate forces for lid application and different lid designs.

Nicholas and Shaun- Last night I ran my simulation with a normal mesh, it got to 94.4% when it hit the warning of max iterations increased by 5. After looking at the results the total Y displacement was 5.9mm even though it was set to move the press plate 7mm. The lid did not fully set on the cup due to the change in displacement (this is still working with the half of the cup).

I thought that taking the off the on flat faces fixtures would prevent the warnings that I am having from coming up early in the simulation run. My current fixtures now are just the reference gerometry on the top of the press place, Translation: x-0mm, y-0mm and z--7mm. I then made a refence geometry on the top of the lid and made the translation to the top plane x-0mm and y-0mm. My idea was that the press plate would not move in the x, y, plane and the only be able to move the lid in the z direction.

Currently I am trying to get my model to accept circular symmetry. here is the picture of the current model.

The only issue is that I can not select the circular symmetry option.
All the properties have stayed the same as well as the component contacts of an incompatable mesh.
Like • Show 0 Likes0
Actions
- Shaun Densberger @ Megan Derwich on Jul 23, 2014 2:04 PM
  Mark Correct
  Correct Answer
  Last night I ran my simulation with a normal mesh, it got to 94.4% when it hit the warning of max iterations increased by 5. After looking at the results the total Y displacement was 5.9mm even though it was set to move the press plate 7mm. The lid did not fully set on the cup due to the change in displacement (this is still working with the half of the cup).
  
  I think that this is just because the simulation didn't finish, so all you're seeing is the results at pseudo time step t < t_final.
  
  I thought that taking the off the on flat faces fixtures would prevent the warnings that I am having from coming up early in the simulation run. My current fixtures now are just the reference gerometry on the top of the press place, Translation: x-0mm, y-0mm and z--7mm. I then made a refence geometry on the top of the lid and made the translation to the top plane x-0mm and y-0mm. My idea was that the press plate would not move in the x, y, plane and the only be able to move the lid in the z direction.
  
  I think that it'd be worthwhile to remove the die that pushing the lid down from simulation, and instead apply the enforced displacement on the portion of the lid that the die was in contact with (I doubt that this will add a measurable about of stiffness to the model). I'd also adjust your model such that you need as little enforced displacement as possible (ie parts in initial contact as Nicholas suggested and determining how much displacement is needed for the lip of the lid to snap over the cup). I'd also make sure that you have Arc-Length control enabled (if you already don't) under the Advanced Options in the Solution tab. All of these should help improve the numerical stability of the problem, which is most likely the cause of your current issues.
  
  Currently I am trying to get my model to accept circular symmetry. here is the picture of the current model.
  
  Hmmmm. I'm not sure why the circular symmetry constraint is locked out, but in this case you could just apply a normal constraint on the exposed surfaces.
  Like • Show 1 Like1
  Actions
  - Jared Conway @ Shaun Densberger on Jul 24, 2014 3:36 AM
    Mark Correct
    Correct Answer
    I don't think circular symmetry is supported in nonlinear. And in this case normal symmetry is the better selection. See kb for explanation.
    
    Since you are still learning, I'd recommend building a really simple model to do some testing on and that you can post.
    
    I suspect a combo of setup issues and challenging contact and nonlinear problem that will require some massaging to work properly. Are you working with your var on this? Or reaching out for training help? That might be more efficient if you can't post the model.
    Like • Show 0 Likes0
    Actions
Megan Derwich Jul 25, 2014 10:01 AM
Mark Correct
Correct Answer
So after multiple trials and remaking the orginal assembly, it seems to work. I took off my mesh controls because they kept causing errors on the on normal mesh. Attached is the video of the fully run simulation with extremely close to the min. forces needed to apply the lid to the actual real life part. https://www.youtube.com/watch?v=IbpHzv58TfM and the final results

I would like to thank everyone for helping me with this! The forum has been extremely helpful and the responce to my problem has been quciker than the paid SolidWorks help.
Like • Show 1 Like1
Actions
- Jerry Steiger @ Megan Derwich on Jul 25, 2014 4:53 PM
  Mark Correct
  Correct Answer
  Megan,
  
  Cool!
  
  Jerry S.
  Like • Show 0 Likes0
  Actions

Displacement on a Press Plate Issue in Nonlinear

Outcomes

Related Content

Recommended Content