Results Analysis

Hi Kiwi!3D community

I’m an architect, and I’m developing my investigation on a bio-inspired active-bending kinetic structure, this is the result of the design:

Now that I’m in numerical analysis phase, I’m using Kiwi!3D and I have a question to make:

Which from theese options should I pick to obtain material flexural strength related to cross-section? This with the aim of prevent material limit exceeding (GFRP=240 N/mm²).

Aditionally I´ll let the file here
02_K!-Shell.gh (23.9 KB)

In this case, uniaxial bending and neglectable normal forces, you can just use P_M1 and divide it by the section modulus W = t²/6. By that, you get the stress in the outer fibers of the shell.

If the stress distribution is more complex, you can use the von Mises stresses. They return the maximum stresses in the upper part of the shell. The lower part has unfortunately be computed by yourself. (Maybe we add this in a future version.) But this is only necessary if the moments and the normal forces do not have the same sign:
sigma_1_bot = -m[0]/W + n[0]/thick
sigma_2_bot = -m[1]/W + n[1]/thick
sigma_3_bot = -m[2]/W + n[2]/thick
vMises_bot = sqrt(sigma_1_bot² + sigma_2_bot² - sigma_1_bot * sigma_2_bot+3 * sigma_3_bot²)

02_K!-Shell_mod.gh (33.9 KB)
I added this in your definition. They all yield approx. the same stress for this problem. Related to your other question you have to check your input regarding the units. The Young’s modulus of the material is interpreted as kN/m². You have either change the unit by right-clicking the component or adapt the rest of the model. Or just use the standard material component, where you can define a custom Young’s modulus matching the 240.

Hi Anna, thank you so much for your answer, it has been a great help.

I have an aditional question to make, does Kiwi take into account element stiffness? or should I input gravity force? I’m asking because the element must withstand its own weight.

This is an experiment with 0,0025 m thickness, which looks a little wavy.

Anotación 2020-05-02 161021

Element stiffness is taken into account (E,nue,t: based on Kirchhoff-Love shell theory). Gravity force/dead load have to be considered by a force component with respective load value and direction (V = rho * t).

The waviness might be a result of a too coarse refinement. This should vanish if you put a higher refinement in the long direction of the strip (either U or V in the surface refinement component). You can check this by right-clicking on AnalysisModel and checking “Show Refinement”.

Hi Anna, thank you so much for your help, it’s been very clear, for gravity force I used surface load with the paramaters that you gave me

I have an aditional question to make, can I simulate wind force to a model in Kiwi? And If I can, how can I do it?

Regards

hey, yes you can simulate a constant wind force. Just connect a ValueList to T of the load component and choose PRES_FL. This is a pressure follower load. So if the structure deforms, the loading adapts. By the input D, you can change the direction of the load and V is the amplitude.

Hi, Anna

If I want to deform a model with a point load, should I use “Pres” or “Dead” type load? and which is the difference? taking into account that my model is kinetic,

Regards.

hey,

you have three options: DEAD, PRES, PRES_FL

DEAD never changes the given direction, PRES computes the direction by the normal once but never updates it during the deformation, and PRES_FL updates the normal direction during the deformation, i.e. the load is always perpendicular. So this influences the form of the bent stripe (the deformed shapes in the picture are the same since they are fictitious and not the outcome of an analysis). But you probably want to remove the force anyway, I guess, since it is not a permanent force. If you do this, the load type does not matter for the final result after the relaxation.

Hello Anna,

I am trying to model the wind load so that it goes in the opposite direction (from top to bottom) but even though I change the direction by means of the vector command, it does not happen. What I interpret is that by default it models the suction load.

Thanks for your help.

hey, you have to set a negative value V in order to change the direction. The pressure load is always aligned to the surface normal and the direction input D is therefore not used.