Learning Forum

[dv0003]

Hello,

I want to model a fan blade with shell elements. The original fan blade varies in thickness in all directions: height, length, depth.

I know the "thickness" feature in Workbench and I used it until now. But with this feature I can define the thickness only depending of one coordinate, e.g. the x-axis.

Is there a possibility to define the thickness of a shell as a function of more than one direction? E.g. as a table where the value depends of x and y coordinate, e.g. "thickness = f(x,y)"? Or is there a way to define a thickness for each individual element?

Thank you very much!

Best regards,

David

[Sheldon Imaoka]

Hi David Having thickness in more than 1 direction is not possible natively in Mechanical, but it is possible via APDL commands (e.g., inserting a 'Commands (APDL)' object under the geometry in the Tree Outline). The procedure is described in Section 13.2.4. "Specifying a Shell Thickness Variation (Tapered Shells)" of the Structural Analysis Guide of the Mechanical APDL documentation.
Regards Sheldon

[dv0003]

Hello Sheldon thank you very much for this information.
I got through it and would like to share my piece of code for a simple example. There are 4 values of thickness and a grid of x=[30 39] and z=[-16 -11] coordinates.

cmsel,s,Komp_Rechteck_Elem ! select component; needs to be an element component
*dim, _thickvari32, table, 2, 2, 1, x,z ! define table for x and z coordinate
_thickvari32( 1, 0, 1 ) = 30. ! define x positions
_thickvari32( 2, 0, 1 ) = 39.
_thickvari32( 0, 1, 1 ) = -16. ! define z positions
_thickvari32( 0, 2, 1 ) = -11.
_thickvari32( 1, 1, 1 ) = 2.11 ! define values for position (1,1)
_thickvari32( 2, 1, 1 ) = 3.88
_thickvari32( 1, 2, 1 ) = 1.11
_thickvari32( 2, 2, 1 ) = 2.88
sectype,207,shell ! define a new section type (207 is a randomnumber)
secfunction,%_thickvari32%, 0
secoff,mid
emod, Komp_Rechteck_Elem, secnum, 207 ! attach section to elements of the component

This works also for much longer and more complex tables.

Regards David

[samuelP]

Hi David,
Thank you for sharing your script.
Is there some way to confirm that the command has worked? E.g. is it possible to plot the nodal thicknesses with NPLOT or EPLOT, or perhaps write the nodal thicknesses to a txt file?
Thanks, Sam

[dv0003]

Hi Sam I am sorry to tell you that I didn't work it out in such a sophisticated way. I just checked the resulting thickness in the old Mechanical APDL Processor by watching the changes when applying the commands.

Regards David