AnsweredAssumed Answered

How do I fully and correctly constrain a panel modeled with 2D elements where the panel is simply supported on two ends?

Question asked by William Hensley on Feb 4, 2021
Latest reply on Feb 5, 2021 by Kiana Russel

Like • Show 0 Likes0
Comment • 8

I have a panel modeled with shell FEM elements. I want simply supported ends so I applied 3 translational constraints to a small rectangular surface at both ends. How would I go about constraining the rotational DOF. The problem I´m experiencing is that I tried to constrain the rotational constraints on the edges such that I didn´t over constrain my model and make it a clamped clamped constraint, but it doesn´t seem to matter what I do. It´s always brick walling both sides. This makes no sense to me. I even tried splitting the surface and applying all rotational contraints on a small surface near 1 edge, but that still resulted in clamped clamped. Makes no sense to me.

I´m doing a buckling analysis. I want 0% fixity at the two ends. The other two sides are free edges.

Please be gentle with me, it´s my first post here. Yes, I have searched my an answer prior but there doesn´t seem to be much info on simply supported panel buckling analysis.

I have attached some documentation. to show my work and how I´ve tried to constrain my part.

Jim Steinmeyer Feb 5, 2021 12:40 PM

Correct Answer

If the panel is bolted at both ends and completely unsupported through the middle I would change the way I would fix it. I try to get as close to "real Life" as I possibly can. If I had a plate supported by angles fixed in location like in the shot below I would use split lines on the plate at the edges of the angle leg. Depending on what you are evaluating, you can treat the angle as immovable and fix the split faces leaving the angles out of the model. I would then select the inside of the bolt holes and fix them as well. Or if the angles are welded securely but you want to account for the flex in the angle you can fix the outside leg of the angle and constrain the bolt holes to the holes in the angle using the bolted feature. It depends on what you are trying to evaluate.

Using your simplified 2D sketch and the way you ended the plate at the bolt holes I would select the edges where the bolts are and constrain translation in all 3 directions from both sides but that would allow rotation unless you want to again consider the edge of what ever the plate is bolted to and use split lines again. There will be differences in the results for all of these evaluations. Your job is to determine how the part will react in real life and what constraints can be simplified and what needs to be considered. Experience is the best guide here.

Does this help?

See the reply in context

Attachments

buckling problem.png113.6 KB

No one else had this question

Outcomes

Jim Steinmeyer Feb 4, 2021 3:50 PM

If I understand what you desire correctly I think what you want is to constrain one edge in the X, Y and Z directions but allow rotation. The other end should be constrained in the Y and Z directions but allowed to translate in the X direction. Think about how the plate would react in real life. When the middle sags the ends either stretch or they have to move toward the center where the sag is. You need to allow your model to do the same thing.

Actions

William Hensley @ Jim Steinmeyer on Feb 4, 2021 5:19 PM

Your advice worked perfectly. Now it´s acting as a simply supported beam. However, I´m still confused. why couldn´t you allow both ends to stretch, rather than move translationally in the x dir? Are you saying that for a simply supported panel, that one side has to be fully constrained in 1,2,3 and the other is forced to move in the 1 direction?

Actions

Jim Steinmeyer @ William Hensley on Feb 4, 2021 5:41 PM

The challenge with simulation is that if the item is not secured in location on the screen it can zip off into never, never land ( it could go an indeterminate distance in the x). The solver requires a known fixed point to start from.

Actions

William Hensley @ Jim Steinmeyer on Feb 5, 2021 11:57 AM

Well yes, at some physical point on a model, there does have to be a constraint for all three directions, even if there´s no net force applied in that said constraint, but what I was trying to say was different. I did work on structural members called ribs on the wing of an aircraft and usually, we would say that the fastened ends on each side of the rib were 50% fixity, or in other words, in between fixed and simply supported. Okay, so this isn´t really simply supported but it does have some characteristics of simply supported. What I´m saying is that I find it hard to picture something that is bolted down to another big structure, is going to literally move in some direction. In fact, it seems impossible because if the constraint is made with bolts on both ends, then the holes(from the panel) would have to move relative to the bolts(I made models such that the width of the panel was equal to the distance from one hole to another hole on the other side), so basically, you´d be shearing the bolts to allow for relative translational movement. I feel as if that´s what I did here. The only other way that I can see translational movement on one end of the panel is if the structure that the panel is bolted to moves as well. That would allow translational movement at the constraint holes of the panel, but 0 translational movement relative to the bolts that fasten the panel to the other structure.

Sorry it´s so long but I hope I´ve expressed my confusion in a way you can understand. I´d like to hear your thoughts.

Actions

Jim Steinmeyer @ William Hensley on Feb 5, 2021 12:40 PM

Does this help?

Actions

William Hensley @ Jim Steinmeyer on Feb 5, 2021 11:15 PM

You´re awesome Jim! Thanks for your thought-out response. I completely agree. I think that constraining the fastener holes is definitely the best bet. I can´t remember if I constrained the mating part to the rib panel or if I just constrained the bolt hole section. Definitely weren´t any welds. One thing that I could do is run two tests one that is clamped and one that is simply supported. I did that all the time. I also thought about what you said about how the ends either have to stretch or move translationally, and I can visually see how depending on the buckling shape, you could see substantial translational movement, regardless of how it happens.

Actions

Kiana Russel Feb 7, 2021 11:48 PM

Plane stress model of composite beams with interlayer slips is developed by the state space method. By virtue of the method of Fourier series expansion, the solution is obtained analytically for beams with two simply supported ends Upsers.com

Actions

Jim Steinmeyer @ Kiana Russel on Feb 5, 2021 11:00 AM

Kiana Russel wrote:

Plane stress model of composite beams with interlayer slips is developed by the state space method. By virtue of the method of Fourier series expansion, the solution is obtained analytically for beams with two simply supported ends.

I was waiting on someone far more knowledgeable than myself to answer here.

Actions

8 Replies

Jim Steinmeyer Feb 4, 2021 3:50 PM
Mark Correct
Correct Answer
If I understand what you desire correctly I think what you want is to constrain one edge in the X, Y and Z directions but allow rotation. The other end should be constrained in the Y and Z directions but allowed to translate in the X direction. Think about how the plate would react in real life. When the middle sags the ends either stretch or they have to move toward the center where the sag is. You need to allow your model to do the same thing.
Like • Show 0 Likes0
Actions
- William Hensley @ Jim Steinmeyer on Feb 4, 2021 5:19 PM
  Mark Correct
  Correct Answer
  Your advice worked perfectly. Now it´s acting as a simply supported beam. However, I´m still confused. why couldn´t you allow both ends to stretch, rather than move translationally in the x dir? Are you saying that for a simply supported panel, that one side has to be fully constrained in 1,2,3 and the other is forced to move in the 1 direction?
  Like • Show 1 Like1
  Actions
  - Jim Steinmeyer @ William Hensley on Feb 4, 2021 5:41 PM
    Mark Correct
    Correct Answer
    The challenge with simulation is that if the item is not secured in location on the screen it can zip off into never, never land ( it could go an indeterminate distance in the x). The solver requires a known fixed point to start from.
    Like • Show 0 Likes0
    Actions
    - William Hensley @ Jim Steinmeyer on Feb 5, 2021 11:57 AM
      Mark Correct
      Correct Answer
      Well yes, at some physical point on a model, there does have to be a constraint for all three directions, even if there´s no net force applied in that said constraint, but what I was trying to say was different. I did work on structural members called ribs on the wing of an aircraft and usually, we would say that the fastened ends on each side of the rib were 50% fixity, or in other words, in between fixed and simply supported. Okay, so this isn´t really simply supported but it does have some characteristics of simply supported. What I´m saying is that I find it hard to picture something that is bolted down to another big structure, is going to literally move in some direction. In fact, it seems impossible because if the constraint is made with bolts on both ends, then the holes(from the panel) would have to move relative to the bolts(I made models such that the width of the panel was equal to the distance from one hole to another hole on the other side), so basically, you´d be shearing the bolts to allow for relative translational movement. I feel as if that´s what I did here. The only other way that I can see translational movement on one end of the panel is if the structure that the panel is bolted to moves as well. That would allow translational movement at the constraint holes of the panel, but 0 translational movement relative to the bolts that fasten the panel to the other structure.
      
      Sorry it´s so long but I hope I´ve expressed my confusion in a way you can understand. I´d like to hear your thoughts.
      Like • Show 0 Likes0
      Actions
      - Jim Steinmeyer @ William Hensley on Feb 5, 2021 12:40 PM
        Unmark Correct
        Correct Answer
        If the panel is bolted at both ends and completely unsupported through the middle I would change the way I would fix it. I try to get as close to "real Life" as I possibly can. If I had a plate supported by angles fixed in location like in the shot below I would use split lines on the plate at the edges of the angle leg. Depending on what you are evaluating, you can treat the angle as immovable and fix the split faces leaving the angles out of the model. I would then select the inside of the bolt holes and fix them as well. Or if the angles are welded securely but you want to account for the flex in the angle you can fix the outside leg of the angle and constrain the bolt holes to the holes in the angle using the bolted feature. It depends on what you are trying to evaluate.
        Using your simplified 2D sketch and the way you ended the plate at the bolt holes I would select the edges where the bolts are and constrain translation in all 3 directions from both sides but that would allow rotation unless you want to again consider the edge of what ever the plate is bolted to and use split lines again. There will be differences in the results for all of these evaluations. Your job is to determine how the part will react in real life and what constraints can be simplified and what needs to be considered. Experience is the best guide here.
        
        Does this help?
        Like • Show 0 Likes0
        Actions
        William Hensley @ Jim Steinmeyer on Feb 5, 2021 11:15 PM
        Mark Correct
        Correct Answer
        You´re awesome Jim! Thanks for your thought-out response. I completely agree. I think that constraining the fastener holes is definitely the best bet. I can´t remember if I constrained the mating part to the rib panel or if I just constrained the bolt hole section. Definitely weren´t any welds. One thing that I could do is run two tests one that is clamped and one that is simply supported. I did that all the time. I also thought about what you said about how the ends either have to stretch or move translationally, and I can visually see how depending on the buckling shape, you could see substantial translational movement, regardless of how it happens.
        Like • Show 1 Like1
        Actions
Kiana Russel Feb 7, 2021 11:48 PM
Mark Correct
Correct Answer
Plane stress model of composite beams with interlayer slips is developed by the state space method. By virtue of the method of Fourier series expansion, the solution is obtained analytically for beams with two simply supported ends Upsers.com
Like • Show 1 Like1
Actions
- Jim Steinmeyer @ Kiana Russel on Feb 5, 2021 11:00 AM
  Mark Correct
  Correct Answer
  Kiana Russel wrote:
  
  Plane stress model of composite beams with interlayer slips is developed by the state space method. By virtue of the method of Fourier series expansion, the solution is obtained analytically for beams with two simply supported ends.
  I was waiting on someone far more knowledgeable than myself to answer here.
  Like • Show 0 Likes0
  Actions

How do I fully and correctly constrain a panel modeled with 2D elements where the panel is simply supported on two ends?

Attachments

Outcomes

Related Content

Recommended Content