Users:Structural Optimization/Design Variables/Nodal Thickness

From Carat++ Public Wiki
(Difference between revisions)
Jump to: navigation, search
(Documented Results)
(Theory and Details)
Line 61: Line 61:
  
 
=== Input File ===
 
=== Input File ===
 
bla
 
 
 
== Theory and Details ==
 
  
 
bla
 
bla

Revision as of 07:38, 21 October 2010


Contents

General Description

Short Info

With this design variable, the nodal thickness can be optimized. This is especially usefull in combination with SHELL8 elements since these elements can have a non-constant thickness based on the nodal thicknesses.

The thickness at each node can be initialized in the input file. Alternatively, the initial thickness is computed as the mean thickness of the surrounding elements.

Parameters in the Input File

Compulsory Parameters
Parameter Values, Default(*) Description
TYPE NODE ID, ND-SET ID Linking to previously defined nodes or node sets
SUBTYPE THICKNESS To obtain the desired design variable
Optional Parameters
Parameter Values, Default(*) Description
BOUND OPT-BOUND ID Linking to a bound defined in BOUND block.

Remark: only SCALAR_BOUND is valid.


Example of a Complete Input Block

This design variable can be defined for one node of for a nodal set. It is good practice to add a bound to the thickness design variables to avoid unrealistic values. In general, it is obvious to choose an upper and lower limit. Since the thickness is a scalar design variable, only SCALAR_BOUND is appropriate.

OPT-VAR 1
 TYPE=NODE 100 SUBTYPE=THICKNESS BOUND=OPT-BOUND 1
 TYPE=ND-SET 2 SUBTYPE=THICKNESS BOUND=OPT-BOUND 1

A complete test example

Model description

This example considers the thickness optimization of the Kresge auditorium. In a first step, the shape of the auditorium was optimized for minimal strain energy under dead load (cfr. PhD M. Firl) with constant thickness. In a second step, the thickness of the optimized shape is optimized. Strain energy under dead load is again used as the objective function. The design variables are the thicknesses at every node, constrained between 0.10 and 0.03 m. The shape of the shell is fixed. In the figures, the thickness distribution and the Von Mises stresses in the top plane are shown for 50 optimization steps starting from a constant thickness of 0.05 m. It is clear that the largest part of the shell works very efficient on membrane forces and thus the thickness can be the minimal thickness. Only the support regions are subjected to bending, so the stresses can be reduced significantly by increasing the thickness to the maximal thickness.

Thickness Von Mises top

Input File

bla


References





Whos here now:   Members 0   Guests 0   Bots & Crawlers 1
 
Personal tools
Content for Developers