Optimization in ansys

[KAM]

Hello!

I am doing an optimization with respect to weight on a sandwich beam in pure bending. The density of the core is unknown.. I have read that DVs are usally geometric parameters. I wonder, if I could use the density of the core, expressed by the youngs modulus as an DV?
I use the shell 91 element for the sandwich, max deflection as SV, and the volume as the objective function.

[Steve Yao]

The youngs modulus can certainly be a DV.
I believe the ansys command to define the modulus would be as follows:

mp, ex, 1, 7200

Where "1" is the element type#, and "7200" is the modulus (make sure your units are consistent!)

Out of curiousity, what are you using to write your optimization code?

[KAM]

http://fsae.com/groupee_common/emoticons/icon_smile.gifhello!
Thanks for your quick reply. I have tryed to do the optimization bought ineraktively, an by usin the log-fil, that I have saved as an txt.file.
Here is my log file:


!
!------------------------------------------------------
! Input:
! Optimization of 3 - point bending sandwich beam
!
! Design variable; thickness of face; 0.7 < t < 2.0
! E-modulus of core; 50 < E < 150
!
! State variable; Maximum deflection; DMAX < 10 mm
!
! Objective; Minimize the volum
!
! Length = 1000 mm
! Width = 150 mm
! t_foam = 40 mm
!
! Force = -2500 N
!
! Quadrilateral Mapped mesh, element lenght 10
!
!
! Analytical Results:
!
! t_aluminium = 1 mm
! E-modulus of core = 95 Mpa (gives density = 80 kg/m3)
!
! Nummeric results:
!
! t_aluminium = 0.7028mm
! E-modulus of core =148.9 Mpa (gives density = 120 kg/m3)
!
!
!------------------------------------------------------FINISH ! Make sure we are at the BEGIN level
! /FILNAME,simba,0
*SET,T_AL,1 !antakelse på face thickness
*SET,TENSILMODU_FOAM,55 !antakelse på youngs modulus for coren
FINISH
/PREP7
ET,1,SHELL91,,1
KEYOPT,1,1,16,
KEYOPT,1,2,1
KEYOPT,1,4,0
KEYOPT,1,5,1
KEYOPT,1,6,0
KEYOPT,1,8,0
KEYOPT,1,9,1
KEYOPT,1,10,0
KEYOPT,1,11,0
*SET,_RC_SET,1,
KEYOPT,1,2,1
R,1
RMODIF,1,1,3,0, , , ,0
RMODIF,1,13,1,0,T_AL,0,0,0, !Design variable, face thickness
RMODIF,1,19,2,0,40,0,0,0,
RMODIF,1,25,1,0,T_AL,0,0,0, !Design variable, face thickness
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,1,,70000
MPDATA,PRXY,1,,0.3
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,DENS,1,,2.7e-6
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,2,,TENSILMODU_FOAM !Design variable, E-module
MPDATA,PRXY,2,,0.32
RECTNG,0,1000,-150/2,150/2,
FLST,5,4,4,ORDE,2
FITEM,5,1
FITEM,5,-4
CM,_Y,LINE
LSEL, , , ,all
CM,_Y1,LINE
CMSEL,S,_Y
LESIZE,_Y1,10, ,0,0,4,-1,-2
CMDELE,_Y
CMDELE,_Y1
MSHAPE,0,2D
MSHKEY,1 !Mapped mesh
CM,_Y,AREA
ASEL, , , , 1
CM,_Y1,AREA
CHKMSH,'AREA'
CMSEL,S,_Y
AMESH,_Y1
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
/UI,MESH,OFF
FINISH
/SOL
ANTYPE,0 ! Static analysis
DL,4, ,UX,0
DL,4, ,UZ,0
DL,2, ,UZ,
F,2596,FZ,-2500 !point load at the middel node of the beam
SOLVE
FINISH
/POST1
AVPRIN,0, ,
ETABLE,EVolume,VOLU,
!*
SSUM !sum of each elemelt
!*
*GET,Volume,SSUM, ,ITEM,EVOLUME
NSORT,U,Z
*GET,DMAX,SORT,,MAX
LGWRITE,'SIMBA4','txt','C:\Users\'
/OPT
OPANL,'SIMBA4','txt','C:\Users\' ! Assign as analysis file
OPVAR,TENSILMODU_FOAM,DV,50,300,0.001 ! YOUNGS MODULUS design variable, min 50 MPA, max 150 MPA
OPVAR,T_AL,DV,0.7,2,0.001 ! tHICKNESS OF FACE design variable, min 0.7 mm, max 2 mm
OPVAR,DMAX,SV,,10,0.001 ! MAX DEFLECTION AS state variable, min VALUE -10 mm
OPVAR,VOLUME,OBJ,,,200 ! Volume as object variable, tolerance 200 mm^2
OPTYPE,SUBP ! Subproblem approximation method
OPFRST,30, ! Max iteration,
OPEXE ! Run

In the results, The E-modulus only converges.. .
(sorry for my english, but I am from Norway)


KAM

[serkan]

thanks your sharing