XANSYS Message: 25744 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Subject: Re: STRUC: T-dep CTEs
Author: Steve Groothuis
Date: 2001-07-10 12:25:00Eric,
Normally, I avoid MPAMOD and supply the secant CTE directly to ANSYS. In
using TREF and BF,,TEMP wisely, one can replicate the desired
thermally-induced strains within the model.
The short answer to your loading condition is that in order for ANSYS to
realize the individual temperature values for secant CTE, you must solve the
thermal excursion in steps (e.g., go from T_ref(100C) to T_final (25C) in
steps of 5-10C.) In you can live with the entire model experiencing the
same thermal excursion (i.e., uniform temperature distribution), then I will
propose a simple approach. As an example, the following APDL segment
illustrates a typical temperature cool-down from 100C to 25C.
/solu
tref,100
temp1=75
temp2=50
temp3=25
hitemp=100
lotemp=25
ramp=(hitemp-lotemp)/5 ! substeps for ramp loadsteps (1 substep/5 deg)
!
! First thermal step (100C to 75C)
!
autots,off ! turn off auto time stepping
nsubst,ramp ! set substeps for this load step
bf,all,temp,temp1 ! apply uniform temperature to all nodes
kbc,0 ! linearly ramp bf load temp from previous
load step
time,1 ! set pseudo-time at end of load step
solve ! solve load step
!
! Second thermal step (75C to 50C)
!
autots,off ! turn off auto time stepping
nsubst,ramp ! set substeps for this load stepstep
bf,all,temp,temp2 ! apply uniform temperature to all nodes
kbc,0 ! linearly ramp bf load temp from previous
load step
time,2 ! set pseudo-time at end of load step
solve ! solve load step
!
! Third thermal step (50C to 25C)
!
autots,off ! turn off auto time stepping
nsubst,ramp ! set substeps for this load stepstep
bf,all,temp,temp3 ! apply uniform temperature to all nodes
kbc,0 ! linearly ramp bf load temp from previous
load step
time,3 ! set pseudo-time at end of load step
solve ! solve load step
All of the values above are purely for illustration purposes. You will have
to establish a good ramp rate for your particular temperature-dependent
materials. Go through a simple example of a 2D rectangle in order to prove
that the resultant stresses, strain, and displacements are correct.
Best Regards,
Steve Groothuis
Micron Technology, Inc.
-----Original Message-----
Sent: Tuesday, July 10, 2001 10:57 AM
To: xansys
Cc: egan@n...
Hello All,
When modelling bimaterial plate bending due to temperature-dependent CTE
mismatch, some issues are:
1. If supplied with secant CTE's over individual temperature ranges for one
material, how does one model this? (The MPAMOD command is to convert a
secant
CTE defined relative to one 'definition' temperature to T_ref. How does
this
apply when several secant CTE's are involved?)
2. If you supply ANSYS with several T-dep CTE's to define the CTE over a
given
temperature range, ANSYS outputs a result depending on only the CTE in the
final/destination temperature range. For instance, if you input:
T CTE (E-6/C), (allowing T_ref = 100 C for all CTE's)
25 10
49.9 10
50.0 20
100 20,
ANSYS will display a step function of CTE vs T. Let us assume cooling to 25
C.
The result is the same as if you would input 10 for the CTE in the T-indep.
menu. I believe it is because 10 is the secant CTE over the entire
temperature
range to T_ref = 100 C. Is this correct?
I would be very grateful for comments on this problem. It seems to me that
the
easiest way to model such problems is to find the overall secant CTE over
the
temperature range of interest, for example, T_ref --> T, and input this
secant
CTE into the T-indep menu. If so, for what use is the T-dependent material
property menu??
Best Regards,
Eric Egan