Solid Modelling - Deflection Issues

[Anonymous]

Hi,

I have had to undertake some modelling of a flat solid.  A series of angle sections welded to a flat plate.  Under a patch load I wished to see the deflections.  My worry is that the deflections see strangely small.  

Due to my worries, I though I would simply build a 10mm flat plate and apply a UDL to make sure that the modeller is conforming to typical bending formula .  

I once again found that under the same applied loading over the same span, the deflection came in at around the 1mm mark.  Seeing as common sense, and the bending formula have this coming in at around 25mm I am a little miffed.  

Unfortunately the model is too large to send.  but the applied unfactored loading is as follows:

55 kPa live 

6 kPa dead

The attached dwg is the profile and the extrusion of this must span simply with no rotational restraint over 0.8m.  

Any help would be much appreciated.  

Kind Regards,

Andrew 

[Artur.Kosakowski]

For hand calc do you include the steel plate as working together with all the beams (bi directional behavior) at the same time or do you treat it as the additional load for the angle assumed as a single supported beam? What moment off inertia is assumed - the weakest or do you include the fact that the angle modeled as connected with the beam cannot freely rotate along its longitudinal axis and is therefore much stiffer?

If this is not the explanation try to delete all the meshes and save the rtd file under a new name with no results. After being compressed the file should be of a reasonable size.

[Anonymous]

HI,

Thanks for your reply.  

I simply undertook the hand calc on a flat plate and compared the result to a flat plate made in robot.

Yes I need the steel angle to act monolithically with the steel plate and hence this is why I have turned to ROBOT.  I can model the plate supported on the angle but I wish to obtain the benefit of welding the section to the plate.  

Thanks for the tip.  I have attached the model (Ready to go, I Think?)

The dead load is supposed to be applied over the entire plate and the live only on the patch.  

Kind Regards,  

[Artur.Kosakowski]

Could you attach your hand calcs + the sketch of the static scheme you assumed so that I have something to refer to? There are no supports in the model - if you defined them in all nodes along the edges perpendicular to the beams this may  be one of the reasons too (this is not going to be fully pinned support condition at the end). 

[Anonymous]

Arthur,

I have not checked hand calcs for this situation.  

when I did not believe the results of this analysis. I undertook a different analysis of a simple 10mm thick flat plate with a 100kpa load over a 1m span.  I did this to convince myself of the analysis method robot undertakes.  

the static calculation of this for a 1m wide plate is below.  

5/384 * (100*1^4) / (210x10^6 * 8.33 x 10^-8)

this gives 74mm deflection.  

The structural Model (which I have attached)

gives me 3mm.  

There should be a support along two of the egdes of the top steel plate providing restraint in the Z direction only.  Agree, possibly I should have set the supports on one side to restrain in x and y also.  

A Quick note - I the model, the material is a set as concrete by mistake.  When changed to steel, the deflections are reduced further to below 1mm.  Sorry for not updating model. 

[Artur.Kosakowski · Accepted solution]

Andrew,

Unfortunately for FEM based programs meshing matters. You should try to create sort of cube shape elements rather than ones that have their horizontal sizes much larger than vertical ones (see the attached pictures; support conditions left as they were therefore the horizontal displacements should be ignored). On the other hand I think you don't need the should (size of the model) for your job. I would suggest model the structure as a shell with the plate being straps of panels with different thicknesses (thicker parts being the sum of the thickness of the steel sheet plus one leg of the angle) and lines of vertical panels for the other legs of the angles.

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[Anonymous]

Thanks.  

You are quite right.  The mesh was inadequate.  Hopefully my memory is up to the challange of the other extrusion

Regards,

Andrew