Console from Prismatic shell

Odgovori
dani
Postovi: 4
Pridružio se: Čet Mar 24, 2011 4:08 pm

Console from Prismatic shell

Post od dani »

Hello, I need to make some calculation about existing visor of stadium.
The construction is from thin plates - prismatic shell (see attached files).
I have some questions?
1) The inclined surfaces doesn't make one surface, so I've made them from triangular surface. Is that right?
2) After preliminary calculation, Tower5.5've showed the deflection at the end of console as 420mm, but all we expected it is 200mm (this was from calculation made by hand). Where could be the problem or mismatch?

I am uploading the "Tower" file, so if anyone could help or suggest something.
Ops, I cannot give a link, because of your rules, so if somebody can help i'll send him mail :)

Thanks in advance.
You can write in Serbian, i'll understand.
Pozdravi ot Sofia

P.S.
Sorry for writing in English, but my Serbian is not very good. I can understand you, but cannot write.
I suppose this is like you speak Bulgarian :)
b.milan
Site Admin
Postovi: 2141
Pridružio se: Čet Jun 01, 2006 5:53 pm
Lokacija: Beograd
Kontakt:

Post od b.milan »

Hello Dani,
You can send e-mail directly to Radimpex (info@radimpex.rs) or You can upload the file to megauplad.com and write down the code (not whole link, just the part of the link, containing the code)

For example, instead of: http://www.megaupload.com/?d=I68ZV75P

Write down: [megaupload.com/?d=I68ZV75P]

Regards,
Milan
RADIMPEX SOFTWARE TEAM
dani
Postovi: 4
Pridružio se: Čet Mar 24, 2011 4:08 pm

Post od dani »

I've uploaded files on Megaupload.

1st file: Prismatic shell elements
megaupload /?d=WDTZK1QI

2nd file: Beam with variable section
megaupload /?d=RVFG3S3X

3td file: Drawing - Section
megaupload /?d=IHRUS0RP

4th file: Model in Tower
megaupload /?d=YID4YMRQ

10x for you support

Regards
b.milan
Site Admin
Postovi: 2141
Pridružio se: Čet Jun 01, 2006 5:53 pm
Lokacija: Beograd
Kontakt:

Post od b.milan »

Model is quite complex and You performed modelling extraordinary good.
I think that is the only way for modeling such irregular surface.

My calculation for Exploatation load case gave me results different than You reported.

I've got 0.2m deflection (calculated in Tower 5 and Tower 6)
Mesh density was 0.6 m and total number of nodes were ~7300
Is it ok ?

Slika



My oppinion is that modeling such construction with variable section beams is not advisable. Behavior of actual structure is too complex and there will be signifficant differences in models causing difference in final results.
For example - differences in position of loading, differences in places of supporting, torsional stiffnes of lateral beams is not taken into account in beam model... etc.
RADIMPEX SOFTWARE TEAM
dani
Postovi: 4
Pridružio se: Čet Mar 24, 2011 4:08 pm

Post od dani »

thanks
The results are the same.
The deflection 400mm was before improving the model. (I've changed the thickness of plate - from 120mm at the top to 280mm at the end.)

But I haven another question?
Is this 200mm are real?
I know (they've told me) that the deflection from Tower we need to multiply by 3, so the real deflection could be 200 x 3 = 600mm.
That's strange. Is this true, or my information isn't right.

Regards.

P.S.
In our design standards is said that deflections must be "delta<Lk/75"
But these deflections, aren't the same as in Tower. We must multuply by 3 "Tower deflection" to reach them
In Eurocode the deflections are limited by L/250
b.milan
Site Admin
Postovi: 2141
Pridružio se: Čet Jun 01, 2006 5:53 pm
Lokacija: Beograd
Kontakt:

Post od b.milan »

I don't know who told You about such thing called "Tower deflection".
With all due respect, this is an ordinary nonsense.

You have to know that finite elements method is fundamentally based and extremely dependent on exact calculation of deformation.

Let me explain:

Finite elements forms stiffness matrix. This matrix is representation of geometric and material properties of the whole structure (plates, beams, supports). Stiffness matrix represents system of linear equations.
Unknowns in this system of linear equations are deformations of nodes (3 displacements and 3 rotations for each node)
External forces (load) acts as constants vector in this system of equations.

The solution of this system of linear equations are node deformations in global coordinate system.

Those node deformations are one and only source for obtaining forces in elements. So, if there are wrong deformations, there is no theoretical chance for forces to become good. It is the simple truth that the most exact influence in finite elements method is deformation. Everything else is derived from it.

It is very easy to perform some simple checks and compare Tower results of some well known representative elementary examples from literature (or compare it with some other software).

I do not understand Your mentioning of maximal deflection values. Those limitations are not addressed to software for calculation. Those limitations are addressed to engineer who must create a design within those limitations.

Of course, beside this "elastic" deformation, there is another kind of deformation - deflection of reinforced concrete girder.
That 'non-elastic' deformation is usually larger than elastic deformation (but no one could say in advance - how much - 5%, 100% or 500%)
This kind of deformation takes into account creeping and cracking of reinforced concrete girder.
This kind of calculation in program Tower could be performed after dimensioning and reinforcement adopting.
RADIMPEX SOFTWARE TEAM
dani
Postovi: 4
Pridružio se: Čet Mar 24, 2011 4:08 pm

Post od dani »

thank you for all.
Now I`m sure that the results are all right.
This explanation is missing in our "ManuaL" and nobody cannot explain like you.

Thank you again.
If I have another questions, i"ll ask again :)

Regards
Odgovori