STAAD TANK ANALYSIS - STIFFNESS

[Jacst3]

I'm analyzing a large, multi-cell concrete tank in STAAD. I've "cut" the tank into a quadrant to make the model small enough to analyze. I feel I need to place "fixed-but" supports at the wall edges that would have been connected to adjacent walls. I think I need to provide a spring stiffnes for all degrees of freedom to make the model think it is still connected to other elements. Has anyone ever calcualted this spring stiffness before? I've used this eqation: K = (Ec/48)*(tw/h)^3 which I took for a uniform cantilever wall from ACI 350.3. I'm not getting correct results in my model. Any thoughts?

Thanks,

Jason

 

[JoshPlumSE]

There is a Bentley: STAAD Products forum that may be a better spot for this. It's not that I think this an inappropriate topic for this forum. But, rather that I feel you might get a better response from the folks who monitor that forum.

 

[cessna98j]

I would think if you're trying to simulate the adjoining tank wall as a boundary condition, it's a bit more complicated than just the cantilever equation you speak of. A cylindrical tank will have hoop stresses for example that wouldn't be accounted for using that stiffness equation.

You could try a smaller model of similar proportions and develop an empirical stiffness that you could apply to the larger model. You may have to do some iterations but I think this approach sounds reasonable.

 

[Gumpmaster]

I think a symmetrical support boundary condition would be best for this type of application. I've asked Bentley about it, and no dice. Does RISA have that type of boundary condition?

I know high end FEM programs (ansys and such) use a symmetrical boundary condition.


Below are responses the responses I got from Bentley

Dear User,

Further to the response below, one of my senior colleagues had the following suggestion:

Actually, if the Y-Z plane is the plane of symmetry, the symmetrical supports are
FIXED BUT FY FZ MX

An anti-symmetrical model with anti-symmetrical loads can be defined as well. Then Symmetrical plus anti-symmetrical becomes the response of the structure in the plus X half of the complete structure; and Symmetrical minus anti-symmetrical becomes the response of the structure in the minus X half of the complete structure.


And from a previous email.

Dear User,

In case, the structure is symmetric (and the loadings are symmetrical as well), you can indeed model only half of the structure with the line of symmetry being adequately supported (supports being FIXED BUT and free in the Y direction and restrained in the other directions).

However, these methods are appropriate if the algorithms are old and slow. But STAAD has the advantage of possessing the advanced analysis engine and hence, from the analysis point of view, it would not be a problem to model the full structure.

From the modeling point of view as well, if you model half of the symmetrical structure, you can easily model the other half using the “Mirror” facility within a very short interval.

I hope this helps.

 

[westheimer1234]

just model the whole concrete tank..

how would you change the support condition for different load combination..

 

[JoshPlumSE]

One significant point, is that you can only use symmetry to reduce the size of your structure if BOTH your loading and your structure are symmetric. So that's something to think about.

That being said replacing half the structure with a symmetric Boundary condition is very easy in just about any program (RISA or STAAD) as long as the symmetry is about one of the global axes. If the symmetry is about an inclined axis then it would get a bit more complicated. But, that's just because you have to define an "inclined support".

 

[Jacst3]

Hey - thanks to all for your help. When I tried to model the whole struture my machine kept crashing so I opted to "break" it into symmetrical parts. I've added "fixed-but" supports to the symmetrical break lines as the Bently response suggested and it seems to work. However, it seems that my "weird" result were more due to the the elastic mat foudation supports that I'm using. I'm getting very large shear stesses in my conc. walls at the junction of four walls at the top when all tanks are full of liquid. when I model the structure on a fixed base at all the nodes I get great results, but they can't be completely true. I need to evaluate further - and if the high shears are true then I can remedy with a "brace beam" at the top of the wall.