Tension Only Bracing

[andrewdroth]

Hi all,

I'm have a static model of a tank support structure. I'd like to make the cross bracing only active when it is in tension, and ignored when in compression.

I have used gap elements and also used truss elements where I turned off the bracing that went into compression. I get different results in the two methods.

What would be the most acurate way to have only members that are in tension active?

[John_Holtz]

Hi Andrew,

The two methods should give the same result. Are you sure the two models are identical? The two areas to check in the gap model are:

1. The type of gap element is set to "Tension without gap".

2. The gap element stiffness must equal A*E/L of the truss elements, where A = cross section area, E = modulus of elasticity, and L = the length of the brace.

Let us know what you find out.

[andrewdroth]

I just tried this again and am getting very different results.

I have set the k value using the formula you and the help menu provided. The gap stiffness is considering the same area that my truss model is using. The gap elements are set to "Tension Without Gap".

I used my gap element DS to determine which cross brace members were in tension. I then turned off all the cross brace that was not in tension in my truss member DS. I am getting very differnet axial forces.

I then copied the truss member DS to make sure I was using identical models. Changed the truss members to gap elements and ran it again. I am still getting very different axial forces (1121 lb in the Truss model, and 8914 lb in the gap model.)

These numbers should be identical, correct?

[John_Holtz]

Strange indeed! I do not have an explanation.

Are the other results like stress and displacement different also? If they are, then the problem remains a mystery. If the results are not different, then the problem is related to reporting the axial force in the truss or gap elements.

Please create an archive of the model and provide post it to the forum. If you are not familiar with creating archives, or if the archive is too large to post, please see the thread Create, Post, or Provide an Archive of your model.

[andrewdroth]

I think we figured it out.

Because the truss members can take compressive load the axial force I'm seeing is a net result of tension and compressive forces. Whereas the gap member can only take tension, therefore there is no net effect.

Does that sound plausible?

[John_Holtz]

If you are happy, then I am happy . But I do not understand your reasoning.

In one of the posts, you described running the DS (Design Scenario for those that didn't know) with gap elements, and then "turning off" the trusses corresponding to gap elements with 0 load in another DS. So mathematically the two models are identical, and the results should be identical.

How did you "turn off" the truss elements that were in compression? Did you change them to a different part number and suppress that part? If by turning them off you mean that you hid them in the Results environment, then your explanation makes more sense. (That possibility just came to my mind!) If this is the case, then the two models are not identical and will not give identical answers.

I assume that there were no warning messages during either analysis, especially messages about the max/min stiffness ratio being larger than 1E10. Any warning message could indicate that the stiffness matrix is "unstable" which can cause inaccuracies in the solution.

[andrewdroth]

Yes, I turned them off by swithing them to a suppressed part.

One of the truss elements went into compression in the results when it was tension as a gap element.

I don't think they are exactly identical because the truss members can take both compressive and tension loading.

But I'm hapy for know.