AnsweredAssumed Answered

Flex command deletes split (curve) faces in multi-surface part

Question asked by Michael Palumbo on May 24, 2011
Latest reply on Jun 7, 2011 by Michael Palumbo

Like • Show 0 Likes0
Comment • 6

I drew a fuselage section as a flat multi-body/surface part and used the flex command to get the correct radius. This model is being sent into ABAQUS for analysis as a Parasolid (x-t). To help aid in the placement of rivets once in ABAQUS I am using the split curve to create lines to seed my mesh with. Rivets are placed at mesh vertices and with the split line + seed intervals I can force the mesh to have nodes in the correct locations. I am able to create all the correct rivet lines on my flat part but once I use the flex command the surface becomes single faced. I tried using the wrap command but haven't gotten it to work on anything other than the skin surface. I've attached some screen shots of the model. The third image appears to be multi-faced but it is not.... some kind of graphical error due to the close separation of the skin surface to the stringers and frame. Thanks in advance.

- Mike

Robert Stupplebeen Jun 7, 2011 9:59 AM

Correct Answer

If I were doing the analysis I would spend the time to get the CAD right before handing it off to Abaqus. Abaqus and all FEA packages are very picky with geometry being dead-on. Since you called out that the report only states a radius for the panel I believe the geometry is simply an extrude and a revolve without merge bodies.

Alternatively in Abaqus you could model the flat panel with tied constraints between the stringers in addition to the rivets and then conform it to a cylinder by utilizing a pressure load and contact. Then use an anneal step to remove any residual stresses. Then apply the loads you intended and remove the tied constraint.

Personally the first option sounds much easier to me and I'm sure this forum can help. If you choose the latter option I suggest asking for help here http://www.eng-tips.com/threadminder.cfm?pid=799.

I hope this helps.

Rob Stupplebeen

See the reply in context

No one else had this question

Outcomes

Matt Lombard May 25, 2011 1:31 AM

You should never expect any kind of results from features like Flex and Deform other than just visual results. If you want real geometry, you have to make it directly. Those features don't simulate real manufacturing processes, and usually make a mess of your geometry. They are not shortcuts to save a lot of work unless all you need is a "looks like" model. If you are going to analysis, you owe your downstream people and applications cleaner data.

Actions

Michael Palumbo @ Matt Lombard on May 25, 2011 8:08 AM

Matt to be honest I did use the flex feature as a short cut. I found that trying to create the fuselage as a curved part from the start was a difficult task (or at least for me it was). To back up a little, I am doing this model as part of a project for my masters program at school. The drawings I am referencing are out of an old FAA report and many of the drawings for the fuselage section are drawn as flat, labeled "panel curvature in right hand projection omitted for clarity." When importing the model into ABAQUS everything looks as it should and that's all I really need, a model which looks like the one in the report. I will be performing analysis of the model and comparing it to test results in the report. I know they aren't going to be exactly the same but I'm ok with that. The majority of the project will be focused on analysis in ABAQUS so I'm trying to reduce the amount of time modeling parts in SolidWork so I can focus on the FEA.

Actions

Jerry Steiger @ Michael Palumbo on Jun 6, 2011 8:51 PM

Michael,

You're probably stuck between a rock and a hard place. I'm guessing that the easiest way for you to get the rivets in the right places is to put the rivet holes in the parts in the flat and then use the Flex command, but your holes are probably not going to be round anymore, which may cause problems. Possibly you could make small holes in the flat and use those to drive larger holes in the flexed shape? Perhaps you could cut small crosses in the flat and use those to locate the holes in the flexed shape?

Jerry Steiger

Actions

Michael Palumbo @ Jerry Steiger on Jun 7, 2011 10:05 AM

I have tried doing something similar to this in the past but do to the amount of rivets here my machine simply can't the amount of holes I would need to model.

Actions

Robert Stupplebeen Jun 7, 2011 9:59 AM

Personally the first option sounds much easier to me and I'm sure this forum can help. If you choose the latter option I suggest asking for help here http://www.eng-tips.com/threadminder.cfm?pid=799.

I hope this helps.

Rob Stupplebeen

Actions

Michael Palumbo @ Robert Stupplebeen on Jun 7, 2011 10:12 AM

I ended up doing something like you suggested a couple of days ago. The tie constraint in ABAQUS worked very well in bonding all the components together. I was able to get the split lines to work after the flex command by using the spacing and converting that to an arch length. Then I drew lines extending radially from the center axis of the skin. I used these lines to create planes and used the planes as intersecting split tools. Since the number of rivets is very large I will be using sub-modeling in ABAQUS to create rivets in the areas where my strain data differs from that of the FAA report.

Actions

6 Replies

Matt Lombard May 25, 2011 1:31 AM
Mark Correct
Correct Answer
You should never expect any kind of results from features like Flex and Deform other than just visual results. If you want real geometry, you have to make it directly. Those features don't simulate real manufacturing processes, and usually make a mess of your geometry. They are not shortcuts to save a lot of work unless all you need is a "looks like" model. If you are going to analysis, you owe your downstream people and applications cleaner data.
Like • Show 0 Likes0
Actions
- Michael Palumbo @ Matt Lombard on May 25, 2011 8:08 AM
  Mark Correct
  Correct Answer
  Matt to be honest I did use the flex feature as a short cut. I found that trying to create the fuselage as a curved part from the start was a difficult task (or at least for me it was). To back up a little, I am doing this model as part of a project for my masters program at school. The drawings I am referencing are out of an old FAA report and many of the drawings for the fuselage section are drawn as flat, labeled "panel curvature in right hand projection omitted for clarity." When importing the model into ABAQUS everything looks as it should and that's all I really need, a model which looks like the one in the report. I will be performing analysis of the model and comparing it to test results in the report. I know they aren't going to be exactly the same but I'm ok with that. The majority of the project will be focused on analysis in ABAQUS so I'm trying to reduce the amount of time modeling parts in SolidWork so I can focus on the FEA.
  Like • Show 0 Likes0
  Actions
  - Jerry Steiger @ Michael Palumbo on Jun 6, 2011 8:51 PM
    Mark Correct
    Correct Answer
    Michael,
    
    You're probably stuck between a rock and a hard place. I'm guessing that the easiest way for you to get the rivets in the right places is to put the rivet holes in the parts in the flat and then use the Flex command, but your holes are probably not going to be round anymore, which may cause problems. Possibly you could make small holes in the flat and use those to drive larger holes in the flexed shape? Perhaps you could cut small crosses in the flat and use those to locate the holes in the flexed shape?
    
    Jerry Steiger
    Like • Show 0 Likes0
    Actions
    - Michael Palumbo @ Jerry Steiger on Jun 7, 2011 10:05 AM
      Mark Correct
      Correct Answer
      I have tried doing something similar to this in the past but do to the amount of rivets here my machine simply can't the amount of holes I would need to model.
      Like • Show 0 Likes0
      Actions
Robert Stupplebeen Jun 7, 2011 9:59 AM
Unmark Correct
Correct Answer
If I were doing the analysis I would spend the time to get the CAD right before handing it off to Abaqus. Abaqus and all FEA packages are very picky with geometry being dead-on. Since you called out that the report only states a radius for the panel I believe the geometry is simply an extrude and a revolve without merge bodies.

Alternatively in Abaqus you could model the flat panel with tied constraints between the stringers in addition to the rivets and then conform it to a cylinder by utilizing a pressure load and contact. Then use an anneal step to remove any residual stresses. Then apply the loads you intended and remove the tied constraint.

Personally the first option sounds much easier to me and I'm sure this forum can help. If you choose the latter option I suggest asking for help here http://www.eng-tips.com/threadminder.cfm?pid=799.

I hope this helps.

Rob Stupplebeen
Like • Show 0 Likes0
Actions
- Michael Palumbo @ Robert Stupplebeen on Jun 7, 2011 10:12 AM
  Mark Correct
  Correct Answer
  I ended up doing something like you suggested a couple of days ago. The tie constraint in ABAQUS worked very well in bonding all the components together. I was able to get the split lines to work after the flex command by using the spacing and converting that to an arch length. Then I drew lines extending radially from the center axis of the skin. I used these lines to create planes and used the planes as intersecting split tools. Since the number of rivets is very large I will be using sub-modeling in ABAQUS to create rivets in the areas where my strain data differs from that of the FAA report.
  Like • Show 0 Likes0
  Actions

Flex command deletes split (curve) faces in multi-surface part

Outcomes

Related Content

Recommended Content