New to non-linear analyses - need help with contact surfaces

[Alan_H]

Newby question here regarding a non-linear FEA.

I'm currently working on a non-linear simulation of a A304 SS 1/4" screw anchored into a 1/8" steel plate.  The screw has a 1-1/2" long aluminum (A6061) spacer between the head of the screw and the steel plate.  I'm attempting to determine what exterior diameter of the spacer needs to be to prevent buckling of the spacer (ID of the spacer is 0.266") when a 200 lb external load is applied to the spacer 1" away from the plate perpindicular to the screw / spacer

I'm wanting to start with a 1/16" thick spacer wall and will manually increase the spacer wall thickness until I achieve acceptible stresses.  The methods I've used to model this are:

1)  Modeled the spacer and the steel plate and utilized the bolted connection (bolt without nut) to

          generate the screw (beam preload of 5 lbs).

     Created a frictional surface-to-surface contact b/w the spacer and the plate.

2)  Modeled the spacer, screw, and the steel plate.

     Created frictional surface-to-surfaces contacts b/w the spacer and the plate and b/w the spacer and the screw.

     Created a bonded surface contact b/w the screw and the plate

To assist in the visualization, I've attached a screen capture of #2 above:

I know the problem lies in the use of the surface-to-surface contacts since those settings are the areas I'm least familar with. 

Any pointers would be greatly appreciated.

Thanks, Alan

[zhuangs]

Hi Alan,

If possible, please provide your model for better investigation, so I can help you set up the contact pairs.

-Shoubing

[S.LI]

As you said, surface-to-surface contact is pretty complicated, no matter what FEA package you use.

You really need some luck to make all settings work together perfectly.

My suggestion is to avoid surf-to-surf contact as possible as you can.

It will simplify your model and make it focusing on the problem you are really interested in.

---

Alan_H wrote:

Newby question here regarding a non-linear FEA.

I'm currently working on a non-linear simulation of a A304 SS 1/4" screw anchored into a 1/8" steel plate.  The screw has a 1-1/2" long aluminum (A6061) spacer between the head of the screw and the steel plate.  I'm attempting to determine what exterior diameter of the spacer needs to be to prevent buckling of the spacer (ID of the spacer is 0.266") when a 200 lb external load is applied to the spacer 1" away from the plate perpindicular to the screw / spacer

I'm wanting to start with a 1/16" thick spacer wall and will manually increase the spacer wall thickness until I achieve acceptible stresses.  The methods I've used to model this are:

1)  Modeled the spacer and the steel plate and utilized the bolted connection (bolt without nut) to

          generate the screw (beam preload of 5 lbs).

     Created a frictional surface-to-surface contact b/w the spacer and the plate.

2)  Modeled the spacer, screw, and the steel plate.

     Created frictional surface-to-surfaces contacts b/w the spacer and the plate and b/w the spacer and the screw.

     Created a bonded surface contact b/w the screw and the plate

To assist in the visualization, I've attached a screen capture of #2 above:

I know the problem lies in the use of the surface-to-surface contacts since those settings are the areas I'm least familar with. 

Any pointers would be greatly appreciated.

Thanks, Alan

---

[zhuangs · Accepted solution]

Hi Alan and S.Li,

Though surface-to-surface contact is complicated and there are lots of contact settings, you feel easy once you know some procedures of how to set up the contact pairs.

First, we need to set up the contact pairs.  Please search "Contact Pairs" in 'Help" and find the methods to generate contact pairs.

Second, we need to set up the contact setting.  But note that the surface-to-surface contact has default contact settings.  Once the contact pairs are generated, the default contact settings are automatically used.  Thus, the user especially the new user does not have to set up the contact setting.   Note that the Autodesk Simulation continuously improves the default settings to make surface-to-surface contact easy-to-use for the customers.

Third, If you want to know more about surface-to-surface contact, you can search "Surface-to-Surface Contacts" in "Help".  The section has topics:  Tips, Main Screen, Options, Advanced Controls.  These topics show lots of details.

We have many application engineers and developers with lots of experience using surface-to-surface contact and are eager to provide technical help and discussion to our customers.

-Shoubing

---

@S.LI wrote:

As you said, surface-to-surface contact is pretty complicated, no matter what FEA package you use.

You really need some luck to make all settings work together perfectly.

My suggestion is to avoid surf-to-surf contact as possible as you can.

It will simplify your model and make it focusing on the problem you are really interested in.

---

@Alan_H wrote:

Newby question here regarding a non-linear FEA.

I'm currently working on a non-linear simulation of a A304 SS 1/4" screw anchored into a 1/8" steel plate.  The screw has a 1-1/2" long aluminum (A6061) spacer between the head of the screw and the steel plate.  I'm attempting to determine what exterior diameter of the spacer needs to be to prevent buckling of the spacer (ID of the spacer is 0.266") when a 200 lb external load is applied to the spacer 1" away from the plate perpindicular to the screw / spacer

I'm wanting to start with a 1/16" thick spacer wall and will manually increase the spacer wall thickness until I achieve acceptible stresses.  The methods I've used to model this are:

1)  Modeled the spacer and the steel plate and utilized the bolted connection (bolt without nut) to

          generate the screw (beam preload of 5 lbs).

     Created a frictional surface-to-surface contact b/w the spacer and the plate.

2)  Modeled the spacer, screw, and the steel plate.

     Created frictional surface-to-surfaces contacts b/w the spacer and the plate and b/w the spacer and the screw.

     Created a bonded surface contact b/w the screw and the plate

To assist in the visualization, I've attached a screen capture of #2 above:

I know the problem lies in the use of the surface-to-surface contacts since those settings are the areas I'm least familar with. 

Any pointers would be greatly appreciated.

Thanks, Alan

---

---

[JAlanHughes]

Ok, figured this out.  I found on another post a reference to the training webinars under subscription and watched those.  BIG help there.

Looking back at my original model and the final model that solved the only thing I can definitely say that I changed was the use of symmetry and fully modeling the screw (including the thread) to keep the file size down some.  Part of the solution may actually have been in the modeling of the thread.  Since only the peaks of the thread contact the internal walls of the spacer the number of contact elements was greatly reduced.

As I mentioned above, I did have to model the screw up completely (couldn't ever get anything to work that made since from a results standpoint when using the bolted connection feature of Algor.

Oh, I also installed the latest service patch.  Not sure if that helped with the solution but it definitely corrected an issue with the solver generating an error and stopping about every hour or so.  That error was:

        forrtl: severe (24): end-of-file during read

Alan

[zhuangs]

Great.  Good to know you figured the problem out.

About the error 'forrtl: severe (24): end-of-file during read', I am wondering what kind of operation system you are using, Vista?

-Shoubing

[JAlanHughes]

OS is Windows XP.  We haven't upgraded to Vista yet.

[bjorn_fallqvist]

I remember seeing that error quite a few times. Not at all as frequently now, but it still happens every now and then. Annoying, but it can be helped by using restart analysis, although this is not an ideal solution.

[S.LI]

Hi, Alan,

Please share your log file so that we can know when and how the error happened.

Thanks.

---

@JAlanHughes wrote:

Ok, figured this out.  I found on another post a reference to the training webinars under subscription and watched those.  BIG help there.

Looking back at my original model and the final model that solved the only thing I can definitely say that I changed was the use of symmetry and fully modeling the screw (including the thread) to keep the file size down some.  Part of the solution may actually have been in the modeling of the thread.  Since only the peaks of the thread contact the internal walls of the spacer the number of contact elements was greatly reduced.

As I mentioned above, I did have to model the screw up completely (couldn't ever get anything to work that made since from a results standpoint when using the bolted connection feature of Algor.

Oh, I also installed the latest service patch.  Not sure if that helped with the solution but it definitely corrected an issue with the solver generating an error and stopping about every hour or so.  That error was:

        forrtl: severe (24): end-of-file during read

Alan

---

[Alan_H]

S. Li,

Sorry, I've already overwritten the log files and I haven't had the error again since installing the 2011 service pack.

Alan