XANSYS Message: 20639 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Subject: Do I need DNSOL? How to obtain strain components?.
Author: Hector Camacho
Date: 2001-01-30 14:42:00Dear Ansys users,
First of all, I would like to thank Dr. John Zinn for
his advice regarding my email with the subject: How to
use: LDREAD,REAC,,,,,,,RST,? John, it really allowed
me to go forward.
Now, I would highly appreciate if someone could help
me with the following question. In my problem, I am
dealing with three physics environments. One
corresponds to a thermal problem and the other two
correspond to structural problems. First, I solve a
thermal problem. Then, I solve the first structural
problem and the second one.
For both structural problems, I am working with
SOLID45, which is OK, regarding the mathematical form
of the constitutive relations. However, they are quite
different in the sense that, for the first structural
problem (second physics environment) stress is
directly proportional to strain, and, for the second
structural problem, stress is directly proportional to
strain rate (or deformation rate). Therefore, a sum of
'strain components' would be a mistake. WHAT SHOULD
BE ADDED ARE THE STRESS COMPONENTS. Having, the
resulting stress at each node, I would like to
calculate the strain rate (according to ANSYS just
strain) through the constitutive relation of the
viscous sintering problem (second structural problem),
but THERE ARE NODES THAT BELONG TO TWO OR MORE
ELEMENTS WITH DIFFERENT ATTRIBUTES.
Let me talk about some attempts.
1. A fragment of the Ansys code list
/prep7
!*
physics,read,thermal
/solu
solve
!*
physics,read,elast
tref,70,
ldread,temp,,, , ,losa-s-3,rth,
solve
/post1
etable,sx,s,x
avprin,0,0,
!*
etable,sy,s,y
avprin,0,0,
!*
etable,sz,s,z
avprin,0,0,
!*
etable,sxy,s,xy
avprin,0,0,
!*
etable,syz,s,yz
avprin,0,0,
!*
etable,sxz,s,xz
avprin,0,0,
!*
lcdef,1,1,1,
lcdef,stat,all
lcwrite,10, , ,
lcdef,stat,all
/solu
physics,read,struct
solve
finish
/post1
lcdef,stat,all
The point is that, once, a new physics environment is
read, no load cases remain in memory. In addition, I
am afraid that LCOPER,ADD may lead me to erroneous
results because of the reason explained above.
2. DNSOL
/post1
*do,i,1,Nnod,1
v1=sigx(i)
v2=sigy(i)
v3=sigz(i)
v4=sigxy(i)
v5=sigyz(i)
v6=sigxz(i)
dnsol,i,s,x,v1,v2,v3,v4,v5,v6
*enddo
sig__ is the sum of the stresses from the first and
the second structural problems. I expect all other
results to be changed. However, it seems to me that
strain components remain unchanged, since DNSOL
requires that you have placed the data type
(stress/strain) in the element nodal record and a
numerical checking was performed.
Then, as was stated above, I NEED TO KNOW HOW TO
CALCULATE STRAIN COMPONENTS HAVING THE STRESS
COMPONENTS AT EACH NODE for the second structural
problem (third physics environment). ANSYS 5.4 is the
version available at my center.
Let me express my gratitude in advance.
I am a Ph.D. student and I have been learning of and
working in ANSYS for almost a year. I hope that with
some help I will have the model ready shortly. Then,
simulation will start. In resume, I hope to model the
sintering of a ceramic material made of layers of
different materials. The thermal problem is to
calculate the temperature distribution, the first
structural problem is to obtain the stresses presented
in a body made of materials with different dilatation
coefficients. The second structural problem (third
physics environment) is the sintering problem. I think
that a superposition of the stresses of the two
structural problems is a proper way to get the final
result (The Principal of Superposition).
Looking forward to hearing from someone
Sincerely Yours,
MSc. Héctor Camacho Montes
Centro de Investigacion en Materiales Avanzados,
S.C.(CIMAV)
Miguel de Cervantes # 120
Complejo Industrial Chihuahua
Chihuahua, Chih. 31109
México
FAX: 52 14 39-1112
email: hcamacho@y...
hcamacho22@h...
http://www.cimav.edu.mx