• State Suggested Answer
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 113 subscribers
  • Views 4980 views
  • Users 0 members are here

RAM SS - Software automatically subdivides walls (in a model space) and produces wrong reactions

VBWGE
Offline

Hi,

I have a constant issue when RAM SS automatically subdividing continuous walls in RAM Model space. It usually happens after the whole model being run in RAM Frame module. Subdivision usually occurs at the change of thickness of the slab.

I did a quick comparison of transfer loadings on the slab from wall above:

-in the first case software automatically subdivided the wall and I exported the reactions into RAM Concept (attached)

-in the second case I manually changed the wall to be continuous (after it's being subdivided by RAM SS)

Results in the end were very different and it seems that wall has become discontinuous in the first case!

How to resolve this issue as currently I'm stuck with changing every wall in the model?

Please find files in the "secure file upload"

Thanks

  • 0
    I assume you are exporting geometry from Ram Concept to RAM Structural System. In that case, wall geometry will be split or segmented where the slab thickness changes.

    Whether the walls are segmented or not in RAM SS should not significantly affect the stiffness, so long as the continuous segments have the same thickness and material properties.

    When it comes to wall transfer loads from RAM SS back to Ram Concept, this post explains how the wall above forces are treated in an elevated slab analysis in Concept:
    https://communities.bentley.com/products/structural/structural_analysis___design/w/structural_analysis_and_design__wiki/modeling-podium-slabs

    Having the walls segmented actually seems like a good idea to me because it gives you a more refined result (so long as the properties for all the segments are correct).



  • 0 Offline in reply to Seth Guthrie

    Thanks Seth,

    Did you have a chance to look at the comparison of results I attached before? In the case when wall was divided by software into 3 segments the reactions I was getting were very suspicious as reaction of end of 1 segment was totally different to beginning of the next segment (e.g. 4.6 & 194; -29 & 40) as well end reactions for wall as a whole were totally different to if the wall would be modeled with segments. Basically, no transfer loadings from wall to supporting slab were identical.

  • 0 in reply to VBWGE
    I have not had a chance to check results. All you sent in were images, so if you want to include your files I can check them directly. Otherwise I'll look as some simple test cases tomorrow. The equilibrium, or resultant total reactions should work out whether the walls are segmented or not though.



  • 0 in reply to VBWGE

    Your case is extremely complicated with walls that are shifting from level to level, some only a small distance over, and some very long. I ran a simpler test model to demonstrate how things should work.

    Here's is the setup, the north wall is one long 20' segment setting on the slab below. The south wall is segmented into three pieces 5', 5' and 10' long.

    As such, Ram Frame analysis (which is the source of the transfer gravity forces in my test) gives a single total axial force in the north wall with no net moment, while the south walls report axial forces (and counteracting moments) individually.

    In Concept, the long wall then shows a uniform load, while the south wall has a saw-toothed shape load.

    The net axial forces are the same in on both sides.

    I feel the segmented wall is probably a more accurate portrayal of the forces really moving through the system, but it's obviously a lot harder to hand check and get comfortable with these results. If the walls above are quite stiff, then the exact distribution of these forces is not very significant, since the wall stiffness will help spread the demand out through the slab anyway (remember, in Ram Concept analysis, a wall above is like a deep upturned beam in the finite element analysis).

    So, my suggestion is to mesh long transfer walls periodically. I like to segment walls in RAM SS where joints naturally occur (e.g. where walls on other levels terminate or a major grid intersections). Then, when you look at the wall elevation, the vertical lines for the wall ends and segment locations will line up from floor to floor. This regulates the mesh into nice rectangles and often prevents meshing errors.



Communities
Support and Services
Training and Learning
Social Media