XANSYS Message: 92815 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Rate item:

Subject: Re: [EMAG] [POST] display of magnetic flux lines for 3D
Author: Peter Frei
Date: 2008-04-11 12:37:07

Dave,

CFX Post is also capable of calculating stream lines for any kind of
vector field. It also reads Ansys Classic results files. So I tried
reading the results file of a SOLID117 analysis. Before trying out the
flux lines, I first checked the B field vectors. It appears that some of
the B vectors are read in correctly, but they are mixed with many spurious
vectors pointing in the (-1, -1, -1) direction. I don't know if there's a
way to get rid of those.

Kind regards,

Peter

_________________________________
Dr. Peter U. Frei
Rockwell Automation AG
IEC Power Components, Dept. ER
Buchserstrasse 35
CH-5001 Aarau
Switzerland
pfrei_at_ra.rockwell.com
Phone: +41 62 837 24 47
Fax: +41 62 837 29 17



Peter Frei
Sent by: xansys-bounces_at_xansys.org
11.04.2008 11:55
Please respond to
ANSYS User Discussion List


To
ANSYS User Discussion List
cc

Subject
Re: [Xansys] [EMAG] [POST] display of magnetic flux lines for 3D
analysis






Dave,

Thanks for the hint regarding Flotran. I agree about PLTRAC being not as
general purpose as it could be.

Your post inspired me to try another workaround that does not involve
Flotran. I thought that it might be possible to use the electric field
tracing option of PLTRAC since there's no cross product in the electric
field part of equation (5-258):

> m {a} = {F} = q ( {E} + {v} x {B} ). (5-258)

So it tried *VGETting the B field data into an array and *VPUTting it
subsequently into to electric field EF results data. That seemed to work,
at least PRNSOL,EF showed the correct values. Unfortunately, neither
plotting the EF data itself nor plotting the particle trace worked. When
trying to plot the electric field data, all EF values seem to be zero.
PLTRAC generates an error ("No field data for element #"). The code I used

is attached below.

In case you still have an APDL script somewhere that transfers the B field

data to Flotran, I'd be interesting at having a glance at it. I guess it
would roughly consist of the following steps:
1. *VGET the B field data into temporary array
2. Change all element types to FLUID142
3. Apply the B field as velocity boundary condition (D command, VX, VY, VZ

dofs)
4. Solve
5. Postprocess (PLTRAC)
Unfortunately, I'm not familiar with Flotran so I'd first have to go
through the sample Flotran analyses in the Ansys manuals.

Kind regards,

Peter


!!Retrieve results of emag analysis:
RESUME, [your_db_name_here], db
!!(e. g. results from SOLID117 edge-based analysis)

/POST1
ALLSEL

!!Determine the number of nodes in the model:
*GET, anznode, NODE, 0, NUM, MAX

!!Create arrays for the temporary storage of B data:
*DEL, tempbx, , NOPR
*DEL, tempby, , NOPR
*DEL, tempbz, , NOPR
*DEL, tempbsum, , NOPR
*DIM, tempbx, ARRAY, anznode
*DIM, tempby, ARRAY, anznode
*DIM, tempbz, ARRAY, anznode
*DIM, tempbsum, ARRAY, anznode

!!Fill the temporary storage with the B data:
*VGET, tempbx, NODE, , B, X
*VGET, tempby, NODE, , B, Y
*VGET, tempbz, NODE, , B, Z
*VGET, tempbsum, NODE, , B, SUM

!!Check contents of the temporary storage:
*STATUS, tempbx
*STATUS, tempby
*STATUS, tempbz
*STATUS, tempbsum

!!Add VOLT degree of freedom (not sure if this is needed):
DOF, VOLT

!!Fill the electric field data with the contents of the temp.
storage:
*VPUT, tempbx, NODE, , EF, X
*VPUT, tempby, NODE, , EF, Y
*VPUT, tempbz, NODE, , EF, Z
*VPUT, tempbsum, NODE, , EF, SUM

!!Print the contents of the electric field data:
PRNSOL, EF

!!Check some plots:
/TYPE, ALL, BASIC !!show interior arrows also
PLVECT, B, , , , VECT, ELEM, ON, 0 !!B field vector plot; works ok
PLVECT, EF, , , , VECT, ELEM, ON, 0 !!E field vector plot; does
not work

!!Particle trace test:
TRPDEL, ALL !!delete existing trace points (if any)
TRPOIN, -10E-3, 0, 0, 0, 0, 0, -1.6e-19, 9.1e-31 !!particle =
electron
PLTRAC,ELEC, , ,0,TRAC,1000,1e-008, ,1,1

_________________________________
Dr. Peter U. Frei
Rockwell Automation AG
IEC Power Components, Dept. ER
Buchserstrasse 35
CH-5001 Aarau
Switzerland
pfrei_at_ra.rockwell.com
Phone: +41 62 837 24 47
Fax: +41 62 837 29 17



Dave Lindeman
Sent by: xansys-bounces_at_xansys.org
10.04.2008 19:24
Please respond to
ANSYS User Discussion List


To
ANSYS User Discussion List
cc

Subject
Re: [Xansys] [EMAG] [POST] display of magnetic flux lines for 3D
analysis






No, there is no easy way to plot flux lines (or even pseudo-flux lines)
for a three-dimensional model in ANSYS. The only work-around I've found
is to setup and run a dummy FLOTRAN analysis using the nodal B-field
components to define the nodal velocity components as boundary
conditions. This isn't too painful since you can script it all using
APDL and the FLOTRAN analysis requires only one iteration. Following
solution you can use PLTRAC to create the pseudo-flux lines (i.e., lines
parallel to the B field).

Clearly ANSYS has streamline plotting capabilities. Why they treat some
vector fields as less worthy than others, I don't know. EnSight and
other general-purpose postprocessors don't enforce any such
restricitions -- you can plot streamlines through your displacement
field if you really want to. I'd prefer the code let ME make the choice
and let ME be responsible for interpreting the results rather than
arbitrarily restricting what can be done with a given tool. People have
been asking for support of electromagnetic vector fields in PLTRAC for
at least a decade now. If ANSYS has implemented their streamline
algorithm in anything resembling a general-purpose fashion, then I'd
have to imagine this "enhancement" could be added in ten minutes or
less. But that would be ten minutes less they could spend on Workbench...

Dave

-------------------------
Dave Lindeman
Lead Research Specialist
3M Company
3M Center 235-3F-08
St. Paul, MN 55144
651-733-6383
Posts possibly associated with message #92815AuthorDateScore
92791Re: [EMAG] [POST] display of magnetic flux lines for 3DDave Lindeman2008/04/10 
92811Re: [EMAG] [POST] display of magnetic flux lines for 3DPeter Frei2008/04/11 
92815Re: [EMAG] [POST] display of magnetic flux lines for 3DPeter Frei2008/04/11 
92846Re: [EMAG] [POST] display of magnetic flux lines for 3DDave Lindeman2008/04/14