Plate with thickness variation

[Me09]

Hi Guys,
I want to model a plate with thickness variation. the nominal thickness is 2 mm but after measurement found the thickness is vary from 1.85 to 2.15 mm. My question is how I can model the plate? if I make the plate as a part with nominal thickness how I can apply the thickness variation?

 

[Me09]

so if we have a plate and thickness is not regular how we can define the dimension? how we can define the thickness 2 ± 0.15 mm?

 

[rstupplebeen]

More details would be very helpful. What is the end goal of your analysis? DO you have a full CMM scan of the part? Could you model the plate at these 2 limits. I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[Me09]

Thanks Rob,
the problem is about modelling of a cantilever which is a part of machine. the machine sensitive to the frequency of the cantilever and I have found that the thickness of the cantilever is not the same. I only can measure the cantilever thickness from the edges and can not measure it in 3D to model exact shape. I want to model the cantilever when the thickness is not curtained. Do you think it is possible to do that? can I define an standard deviation for the geometry when I model the part or during the process in abaqus?

 

[shingouz]

Hi,
I don't think it is possible to define a deviation from nominal geometry while preprocessing, but why don't you study the effect of thickness by modeling the plate with min thickness and max thickness and see the differences?

 

[Me09]

Thanks shingouz

I am not sure if we model the minimum and maximum geometry can get good result as the geometry randomly changing along the part so we need a logic behind this to explain about the effect of random geometry changing.

I have heard using Ansys we are able to define the geometry deviation. Can we do the same thing in Abaqus?
Is there any other idea to model a simple part with thickness of 2 ± 0.15 mm?

 

[n3l3]

Ok, did not get too much into your problem, but this ideas came into my mind, quite probable that you considered'em before:

>Do a parametric analysis (either "parametric" in abaqus or just repeat the simulation for thicknesses e=2.15,2.10,2.05 ... and do a postanalysis in your excel or so sheet.

>Use 3d elements instead of shell.

good luck
n3l3

 

[rstupplebeen]

I would assume a linear change in thickness from 1 known point to the next then throw a factor of safety on it. You could also drive the thicknesses with Python or Isight and basically do a monte carlo simulation on thicknesses. I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/