modeling boundary conditions for a 3D shell structure 2

[ingthere]

I have a 3D steel panel modeled with shell elements. And this panel is fixed at its two ends with bolts to solid blocks, which are the supports for the panel, and they are fixed to the ground.

Having both ends fixed, I obtain high horizontal reactions at the solid supports under gravity loading. And since the panel is fixed to the solid supports, naturally deflections and stresses at the panel is significantly affected by these boundary conditions (when I release one end of the panel in horizontal axis, deformations at the panel change significantly, which also effects maximum stresses at the connection)


Cable deflection theory says, it is normal to have high horizontal reactions, because cable elements don't have bending resistance, and all the load is carried as tensile forces, which create the horizontal reactions. However, what about these horizontal reactions for a bending resistant panel element?

Although I was trying to develop a model which reflects the reality, having fixed boundary conditions for both supports(it is a model of a test specimen, which is perfectly fixed at two ends), now I am not sure if my model is correct. Should I actually release one end in horizontal axis to obtain zero horizontal reactions?

Thanks for your help!

 

[SWComposites]

first, there is no such thing as "perfectly fixed" in reality, including in your test specimen.

second, what you need to model is the flexibility of the connection bolts, and possibility part of the supports.

third, to determine if your model is "correct" you should be comparing the model predictions to the measured test results.

 

[ingthere]

I don't have the test results yet, so at the moment I can't compare the results with reality.

I modeled the connection b/w the panel and the solid support flexible. My problem is, when I have uniform gravity loading on the panel, with these supports fixed to the ground,(solid blocks that bear the panel) the vertical global deformation of the panel is also affected.

Agreed with perfectly fixed comment, still "the flexibility" that these supports have in reality is very insignificant. That's why I count them fixed.

I also attached a detail about the solid support.

 

[ingthere]

Is it a dumb question to answer or was I not clear? I am still confused, should i consider those horizontal reactions in my design? Could you least refer an example or a book so that i can work on it ? Thanks!

 

[ishvaaag]

Normally you wouldn't want one of these things (even at the factored level) under any practical foreseeable loading enter the range of deformation that develops significant tensile action on the attachment (from loading orthogonal to the midplane of the trapecially conformed sheet). Hence you can dismiss the tensile force from the loading.

Other thing is that a single bolt through a thin sheet can't deliver fixity; hence most designers would choose to design your item as simply supported.

If you are going to develop some correlation between your tests and some numerical prediction of deflection, just factor (down) the deflection under the simply supported case and make everyone that the formulas apply only as an approximation of deflection obtained from your tests.

 

[SWComposites]

if the ends are sort of fixed, then there will be some level of horizontal reactions at the end, so yes they should be included in the FEM

 

[ingthere]

Ok thank you very much! Excellent idea to use a calibrated factor on the results of simply supported case. I will do that!

Yes, horizontal reactions will be there anyway, since also there will be some flexibility in the bolts, I will again have to calibrate the results with the tests to be able to use the numerical model for design..

Thank you for your help!