XANSYS Message: 37546 [Go back to message list]
[bookmark on del.icio.us]
Average Rating: 5 (1 votes)
Subject: Re: Substructuring - problem in scaling of a load ve ctor
Author: Dan Bohlen
Date: 2002-05-13 11:21:00Do you have temperature dependent material properties. You might need to scale
more than just your loads....
-----Original Message-----
Sent: Thursday, May 09, 2002 2:38 PM
To: xansys@yahoogroups.com
vector
Nick, thanks a lot for pointing out the mistake. Yes it should have
been SFE, all,2,selv,,100. But SFE is for surface loads and I want to
change the temperature which is basically a body force. Will it be
right to use or would you suggest something else?
Can you give me some idea how to modify Steve's load file for
substructuring analysis. One method I think of is to write these
substeps as the individual load steps.
Narendra Singh
Grad Student
SUNY Binghamton
--- In xansys@y..., "Nick" wrote:
> SFE,... ! Apply superelement load vector
> --- ! Load step options
> You have to use SFE with SELV.
> Read the sample input for doing substructuring and SFE manual.
> Steve's input files are not written for Substructuring analysis.
> Hope this gives you a good starting point.
> -n
> -----Original Message-----
> From: narendra_uor [mailto:bj93467@b...]
> Sent: Wednesday, May 08, 2002 5:47 PM
> To: xansys@y...
> Subject: [xansys] Substructuring - problem in scaling of a load
vector
> Hi All,
> I am trying to use the substructuring feature of ANSYS in one the
> electronic package. I am subjecting the package to a thermal cycle
of
> 0C to 100C (10min ramp/5min dwell).
> I used the linear material properties for die, overmold and die
> attach and made them as one superelement in the generation pass. I
> made two load vectors - 1) contains all the symmetry constraints
> 2) 1 degree decrease in temp (373 to 372)
> Now, in the use pass I couldnt figure out how to scale the body
force
> (temperature)on superelement in between the substeps during the
ramp,
> for instance I am using the load file provided by Steve Groothuis
> *do,cycle,1,ncycle
> ! ---- RAMP LOW (LOAD STEPS 1,5,9,...) ----
> autots,off ! turn off auto time stepping
> nsubst,rampstep ! set substeps for this load step
> bf,all,temp,lowtemp ! apply uniform temperature to all nodes
> kbc,0 ! linearly ramp load temperatures
> time,(cycle-1)*tcycle+lowramp ! set time at end of load step
> solve ! solve load step
> ! ---- DWELL LOW (LOAD STEPS 2,6,10,...) ----
> autots,on ! turn on auto time stepping
> nsubst,10,100,1 ! set the size of first time step
> bf,all,temp,lowtemp ! apply uniform temperature to all nodes
> kbc,1 ! use applied load temperature for all substeps
> time,(cycle-1)*tcycle+(lowramp+lowdwel) ! set time at end of load
step
> solve ! solve load step
> I guess, this scaling should be easily done during dwell by simply
> using
> BFE, all,2,selv,,100
> Where,
> a) `all' is the selected superelement
> b)`2' is the second load vector as defined above
> c)100 is the scale factor used for the dwell at temperature 273K
> This is what I have figured out by reading the help on
Substructuring
> topic. So I might have used redundant steps in it like I doubt if
we
> really require two load vectors. So, I would highly appreciate any
> modification/suggestion on this.
> Thanks and regards,
> Narendra Singh
> Grad Student
> SUNY Binghamton