Modeling an "equivalent" plate to represent a waffle like structure

[JBlack68]

To all

I am looking for a way of modelling a waffle-like structure by a simple equivalent plate. Creating all the waffle structure using to 2D element could be done. it's just a matter of taking the time. Because the waffle pattern is quite small compare to the overall size one could end up with a lot of elements. While I am sure I could run a very large model, I like to use elements where it matters.

What makes the geo a bit more difficult is that the waffle is partially hollow to accommodate small cooling channels.

So the questions are;
1. What would be the best way of approximating a typical waffle cross section to a (uniform) plate thickness? I think at this stage I am more interested in the stiffness being right
2. with an "equivalent" plate, stress prediction would not be correct (am i right??) so is there a way of scaling the stress to estimate the stress in the waffle ie the real structure

An hint, help would be appreciated

Thanks

Regards

JXB0809

 

[rb1957]

"waffle" is a word with many meanings ... I'm picturing something like a Belgian waffle (a grillage network of stiffeners).

perhaps the best way is to make a stand-alone model of the waffle, and to compare it to a simple plate model. then you should be able to trick the material properties to give you near the right stiffness. The advantage of doing it this way is you can then take displacement results from the full model, apply them to the detail waffle and get good estimates of stresses in the waffle.

the alternative these days to say "that's took much like hard work, i'll just model the waffle in the full model and let it run a minute or two longer".

remember the day when designers had to visualise their creations and make 2D drawings based on this understanding of their structure; as opposed today when we can make 3D models and we don't have to visualise (and understand ?) near so much.



another day in paradise, or is paradise one day closer ?

 

[JBlack68]

thanks. The (sub)structure exactly looks like a Belgian waffle. Only it's made of aluminium! oh and the waffle bit (where you put your jam or whatever you like on it, the "holes", only existd on 1 side )

A stand alone FE is on the "possible approach/method" list

Thinking along the line
Ixx value of the waffle = Ixx_waffle
Iplate = bt^3/ 12 = Ixx_waffle ---> tequivalent = (12*Ixx_waffle/b)^1/3

which obviously only works for one I value

 

[IDS]

The simplest way to get direct output of accurate stresses with reasonably accurate deflections would be to model the waffle as a grillage of T beam elements. You will probably need to adjust the torsional stiffness to get good deflections, but the stress results should be OK.

Alternatively adjust the depth to give the right EI in one direction, adjust the E in the other direction (if required), which will give you good results for deflection and bending moment, you can then calculate the bending stresses externally.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[MrFEMcredible]

Get hold of a copy of R.D. Blevins "Formulas for Natural Frequency and Modeshapes" and you'll find formulas therein for 'grilled' plates equivalent stiffnesses to be used in the formulas for flat plates presented therein.