February 23, 2007

The Focus

Issue 55

February 23, 2007

A Publication for ANSYS Users

Issue 55

Spinning up Rotor Dynamics

By Eric Miller

For
decades,
ANSYS software has been
the dominant
tool for people
who make machines that spin. Turbo-machinery, motors,
shafting, all of these industries have relied
upon ANSYS for simulating various aspects of their devices. But through all of
this, one aspect of simulating rotating systems was missing needed capability - Rotor
Dynamics. The last two releases of ANSYS
(10.0 and 11.0) have largely filled in this
important area.
Before we get into the detail of the ANSYS
implementation, you may notice that this
article does not really go into detail on the
history, theory or details of the math. This
is intentional because others do a much
better job at explaining it. We recommend
that you first read a nice paper by Swanson,
Powell and Weissman called A Practical
Review of Rotating Machinery Critical
Speeds and Modes
www.sandv.com/downloads/0505swan.pdf
Next, read section 15.4 in the ANSYS The-

ory manual and Chapter 8 of the ANSYS A very simple test case is shown in the
Advanced Guide. If you want some history, Macro ROTDYN1.MAC. Use it to explore
I recommend A Brief History of Early
Rotor
Dynamics
by
Nelson
findarticles.com/p/articles/mi_qa4075/is_2
00306/ai_n9296359
some of the topics discussed below. Figure
1 shows the model.
Reading these should give you a good foundation. In the following sections we will Supported Elements
talk very briefly about some of the key After doing the homework above you will
things that PADT has learned about these see how you can add RD simulation by
tools that anyone looking to use them calculating the Coriolis Terms [G] and addshould know. But remember that this is a ing them to your damping matrix [C] as is
huge topic and before you do a RD analysis shown by the equation of motion that ANyou need to do more research and some test SYS uses:
cases.
Adding those terms for a simple pipe and
mass element is not so hard and this was
done years ago. The last two releases have
added far more general elements that allow
for much more accurate modeling, which
has been a lot of work for the programmers
in Pittsburgh. A summary is shown in
Table 1.
The CORIOLIS Command and Changing the
Solver

Figure 1: Simple Model

The basic step that you need to do to include

the Coriolis terms in your solution is execute the CORIOLIS, Option,-,-,RefFrame
command. When Option = 1, ANSYS cal(Cont. on Pg. 2.)

Table Operations - A Lost Art

By: Doug Oatis

As many people
have found, working with arrays and
tables is nothing
like riding a bike.
One day you are
writing a complicated macro that scales tables according to
a quadratic function dependent on the table
indices, then one week later you're relearning how to use the *dim (well, maybe not
that bad).

entire table (i.e. your film coefficient values

for a transient thermal analysis). The *m
commands are limited in how you can modify a table. One tidbit that people seem
surprised to hear is that you can use all of
the *vxxx commands to modify tables
(*voper, *vfun, etc.). The only issue you
have to deal with is that the *vxxx commands only operate on a single column.

You can get around this limitation using a

*get and *do loop. Using the *get, you can
pull the number of columns (note, that the
value returned does not include the 0-th
The new year has brought an influx of table column in the total column count). You can
questions. At first glance, it seems like a then use the returned scalar as the limit in
daunting task to go through and scale an your *do loop to operate over an entire
table.
www.padtinc.com

Using this method, you can further improve

it by having your script only operate on
certain columns, rather than the entire table.
If you look closer at the *v commands, you
can also add controls on which rows you
want to operate on. The *v commands
require you to specify the starting point for
the operation, and then they (Cont. on Pg. 5.)

Contents
Rotor Dynamics...............................1
Table Operations ............................1
Tension/Comp Mat Props...............5
Awesome APDL..............................6
Advertising ......................................7
1-800-293-PADT

February 23, 2007

(Rotordynamics, Cont.)

ELEMENT
BEAM4
PIPE16
MASS21
SOLID45
SOLID95
SHELL181
PLANE182
PLANE183
SOLID185
SOLID186
SOLID187
BEAM188
BEAM189
SOLSH190
SHELL281

Station- Rotating
ary Ref- Refererence
ence
Frame
Frame

Table 1: Summary of Rotor Dynamics Elements at 11.0

The Focus
TYPE of STATIC, MODAL, HARMONIC
and TRANS, and HARMONIC and
TRANS include full and modal superposition solutions. The most common type of
RD simulation is a natural frequency simulation (modal) analysis. When the geometry is not spinning a standard modal analysis
works just fine, but when there is spinning
the Coriolis term adds those non-symmetric
terms that introduce forces to the system,
which cause the natural frequencies to split
and shift up and down (see Bearings and
Damping along with Whirling below and
the Practical Review article.) So in order
to design a rotating machine, you need to
know where your natural frequencies are at
a given speed, usually so you can avoid a
nasty excitation at some common speed. As
is shown in example one, you get a feel for
how speed effects frequency by running at
0 RPM and then several speeds up to the
maximum rotational velocity that the system will see.

culates the Coriolis terms for the appropriThe next most common simulation is a harate elements. You then use the RefFrame
monic analysis. Here you sweep through a
argument to specify if you are doing a starange of excitations where the excitation
tionary reference frame (see below).
frequency is also applied as the rotating
The only down side of applying the Coriolis frequency. Again, the Coriolis terms shift
terms in the damping matrix is that damping the frequencies, and damping plays a bigger
matrices are non-symetric, and therefore role, making such a simulation very valurequire a non-symmetric eigenvalue solver able. If your excitation is different from your
for eigenvalue problems. Fortunately, the rotating frequency, you can use the SYNQRDAMP solver has had a lot of work done CHRO command to scale it up or down.
on it and it is very fast and robust at 11.0.
Finally, many people want to know the
So
dont
forget
the
MODOPT,
loads that are exerted on structures, joints
QRDAMP,,,ON to activate the solver and
and bearings when a rotating structure rotell it to do a complex solution.
tates (think jet engine spinning on a wing
Typical RD Simulations
then the plane turns). This can be done as a
You can include the Coriolis terms in AN- static (apply IC commands to specify velocities) or transient dynamic simulation where
the Coriolis effects are
included.

Figure 2: Campbell Diagram for Simple Model

Issue 55
whirling. Sometimes, the deviation can be
significant and you may end up right at an
excitation frequency.
Figure 2 shows the Campbell Diagram produced by ANSYS with the PLCAMP command. There is also a PRCAMP command.
For simple systems these commands work
great and are very quick to generate. If you
have a more complex system, you may want
to take the time to do a set, all to a file,
massage in Excel, and make your own diagram.
Stationary vs. Rotating Frames

Another thing you should have learned

from the above references is the difference
between solving in a rotating vs. a stationary reference frame. If you are having a
tough time with the difference, you may
find the Wikipedia article Rotating reference frame useful from a math perspective.
From a practical standpoint, if the parts are
attached to a fixed structure, then you need
to solve in the stationary frame and if it is
not, use the rotating frame. A key thing to
note when using the rotating reference
frame is that spin softening is automatically
included. This can then cause solution
problems because the stiffness matrix becomes negative when the the rotational velocity is higher than the natural frequency
(see equations 3-77 to 3-79 in the Theory
Manual.) If you run into this problem apply
a negative shift to a modal run or use large
deflection for transient dynamic or static
runs.
Bearings, Damping and Whirling

In the real world, rotating structures are

attached to static structures through some
sort of bearing. Bearings are not infinitely
stiff, and the friction and lubricant in them
introduce damping. So now you can visualize your system rotating on a set of springs,
and often springs that have stiffness that
Campbell Diagrams
varies with speed and direction. The same
The primary post progoes for damping.
cessing tool for people
doing RD work is the You can capture this in ANSYS with a
Campbell
Diagram. spring-damper element like COMBI14,
This workhorse graph which has been around for a while, or with
for turbomachinary de- the newer COMBI214 which was purpose
signers is also very built for modeling bearings in rotor dynampowerful when applied ics. The 214 allows you to specify stiffness
to RD. The big differ- and damping ratio at 0, 45, 90 and 135 if
ence from a traditional you have non-symmetric bearings (just 0 if
Campbell Diagram is it is symmetric) and you can give ANSYS a
that you can see the table for those values. What this means is
modes split because of that you can make the con(Cont. on Pg. 3.)

www.padtinc.com

1-800-293-PADT

February 23, 2007

(Rotordynamics, Cont.)

The Focus
for most rotor assemblies, this is still not
only the most efficient, but the most accurate method. But sometimes a system does
not lend itself to this type of approximation
and ANSYS has provided a unique solution
in the industry to address it.
A good example is the geometry shown in
Figure 4. This is an approximation of a
turbo-molecular pump that PADT worked
on a while ago. The design was never
viable because we could never get through
a critical speed do to size/bearing constraints. But one thing we found was that
spin stiffening and spin softening played an
important role along with bearing damping
and stiffness. There is no way to model that
type of geometry for rotor dynamics without
using
a
full
3D
model.
ROTDYN2.MAC is a simplified sample
problem with stress stiffening turned on.
Figure 5 shows typical results for this geometry.
Most of the examples where you would use
this geometry include rotating structures
where stress stiffening is important or
where the axisymmetric static modes in the
part may interact with whirl modes.
Random RD Thoughts

This article is already too long. But there

are a few more key things users should
know when diving into this great capability:

Figure 3: Whirl Model

tribution dependent on other calculated values like time, speed or temperature. See the
doc on the element for more information.
One of the improvements to the software for
RD was the inclusion of the PLORB and
PRORB commands in POST1 for plotting
the whirling (or orbiting) of beam/mass RD
models. Figure 3 shows an example for
ROTDYN1.mac. In addition, you can use
the ANHARM macro to create an animation
of the whirl for beam/mass and solid RD
models. This has been a very useful tool
because some modes whirl forward and
other backward The ANHARM macro is
a great way to visualize this.

Issue 55

Models in the stationary reference

frame must be axisymmetric. No bladed disks!
When you do Campbell Diagram
plot/list or a orbit plot/list POST11 goes
into a special mode and some post processing commands dont work as expected.
Always do a new SET
command when you are done with

Figure 5: Mixed Cup and Shaft Mode

Campbell and Orbit post processing.

This is especially true for the ANHARM command. It wont work unless
you do the SET.
If you have more than one rotating
group, not a problem. Use the CMOMEGA command to specify the rotating
axis and magnitude. PLORB, ANHARM recognize the different groups
automatically and you can specify a
group to plot in the Campbell diagram
with PLCAMP
If you are going to turn on stress stiffening, you need to tell ANSYS to not save
the stress run in the results file with the
CAMPBELL,ON command. Otherwise
on each run in your loop that steps
through speed, your result file will get
overwritten.
It can sometimes be tricky to constrain
out the rotational degree of freedom on
a stationary reference frame 3D model.
A good way is to create a massless mass
element, constrain its axial rotational
DOF, then use RBE3s to tie that DOF
to the ends of you shafts (See
ROTDYN2.MAC for an example). On
beams, just constrain the rotational
DOF.
If you are building a beam/mass model
and you want to plot results on the 3D
geometry (/ESHAPE,1) then you must
save the element results by setting the
4th argument in MXPAND to yes: MXPAND, ,,,YES

Animations of the whirl solutions along

with the example macros can be downloaded from :

Full 3D Models

For many decades people have been doing

rotor dynamics with in-house and commercial codes that used beams and masses. And

ftp.padtinc.com/public/downloads/roto_dy
n_focus.zip
Figure 4: 3D Representative TMP Model

www.padtinc.com

1-800-293-PADT

February 23, 2007

finish $/clear
/file,rotdyn1
/prep7
! Define geometry params
shftlen = 12
shftdia = 1
dskoff = shftlen*.75
dskdia = 5
dskthk = 1
bstf = 60000 ! Brng Stiffness
nmd = 20 ! Number of modes
nmspn = 5 ! Number of speeds
mxspn = 6000 ! Max Omega
! Build model nodes
n,1
n,5,dskoff-(dskthk/2)
n,6,dskoff+(dskthk/2)
n,10,shftlen
fill,1,5
fill,6,10
! Bearing ground nodes
n,101,
n,201,
n,110,nx(4)
n,210,nx(4)
! Use 188 elements, solid
et,1,188,,,2
sectype,11,beam,csolid
secdata,shftdia/2

The Focus
ROTDYN1.MAC
sectyp,12,beam,csolid
secdata,dskdia/2,18
! Make elements
et,11,14,,2
et,12,14,,3
r,10,bstf
secnum,11
type,1 $mat,1
real,1
e,1,2 $e,2,3
e,3,4 $e,4,5
e,6,7 $e,7,8
e,8,9 $e,9,10
secnum,12 $e,5,6
type,11 $real,10
e,1,101 $e,4,110
type,12 $real,10
e,1,201 $e,4,210
! Fix axial and rotational DOF at ends
d,1,ux $d,1,rotx
d,4,ux $d,4,rotx
! Fix Brng ground nodes
d,101,all $d,201,all
d,110,all $d,210,all
! Plot the model
/view,1,1,1,1 $/vup,1,z

Issue 55
/eshape,1,1 $/pnum,sect,1
/num,1 $eplot
! Dummy mat props
ex,1,10e6 $nuxy,1,.23 $dens,1,.001
finish $/solu
! Setup modal run
antype,modal
coriolis,on,,,on
modopt,qrdamp,nmd,,,on
! Loop on speeds
*do,i,1,nmspn
spn = (i-1)*(mxspn/(nmspn-1))
omega,spn
mxpand,nmd,,,yes
solve
*enddo
finish
/post1
! Plot/List Campbell Diagrams
plcamp,,1,rpm
prcamp,,1,rpm
! Plot orbit
set,5,6
plorb
! Animate whirl
set,5,6
plnsol,u,sum
anharm

ROTDYN2.MAC
finish $/clear
/file,rotdyn2
/prep7

et,200,200,6 !Mesh 2D Area

esize,.125
amesh,all

! Set Geometry Parameters

shftrad = .25
shftlen = 3
webthk = .5
conir1 = 1.75
conir2 = 1.65
conor1 = 2.25
conor2 = 1.75
conoff = -.5
stblen = 1
! Set Speed and Bearing Stiffness
mxspn = 10000*2*3.14159/60
bstf = 20000
nmd = 20 !number of modes
nmspn = 5 ! Number of speeds

et,1,185 ! Setup solid element and

revolve
type,1
esize,,6
vrotat,all,,,,,,1,2,360

! Build model
k,1,0, $k,2,shftlen
k,3,shftlen+webthk

et,10,214,,1,0 ! Make the Brng Element

r,10,bstf,bstf
! find node on shaft
nn1 = node(kx(1),ky(1),kz(1))
nn2 = node(kx(13),ky(13),kz(13))
! Ground node
n,30000,nx(nn1),ny(nn1)+.5,nz(nn1)
n,30001,nx(nn2),ny(nn2)+.5,nz(nn2)
n,40000,nx(nn1),ny(nn1),nz(nn1)
n,40001,nx(nn2),ny(nn2),nz(nn2)

finish
/solu
coriolis,on,,,on ! Turn on RD
campbell,on !tell ANSYS to save rst
*do,i,1,nmspn !loop on speeds
finish
/solu
antype,static !Prestress run
pstres,on
spn = (i-1)*(mxspn/(nmspn-1))
omega,spn
solve
finish
/solu
antype,modal !Modal run
pstres,on
modopt,qrdamp,nmd,,,on
mxpand,nmd
solve
*enddo

type,10 $real,10 $mat,10

e,nn1,30000 $e,nn2,30001
d,30000,all $d,30001,all
! Fake mass elements for fixing ROTX
et,201,21 $r,200
type,201 $real,200
e,40000 $e,40001
d,40000,rotx $d,40001,rotx
! Tie masses to ends of shaft
cmsel,s,nfx1

save
finish
/post1
/view,1,-1,1,1
/vup,1,z
/dist,1,2.937
/focus,1,1.25,.43,0
set,nmspn,3
plnsol,u,sum,1 ! Plot mode shape
anharm
! Animate whirling

k,4,0,shftrad
k,5,shftlen,shftrad
k,6,shftlen+webthk,shftrad
k,7,0+conoff,conir1
k,8,shftlen,conir2
k,9,shflen+webthk,conir2
k,10,0+conoff,conor1
k,12,shftlen+webthk,conor2
k,13,kx(3)+stblen
k,14,kx(3)+stblen,shftrad
a,1,13,14,6,12,10,7,8,5,4
adel,all
lfill,8,9,.25 $lfill,7,8,.25
lfill,3,4,.25 $lfill,4,5,.5
al,all

www.padtinc.com

asel,s,loc,x,0 !Fix ends in axial DOF

nsla,s,1
d,all,ux
cm,nfx1,node
asel,s,loc,x,kx(13)
nsla,s,1
d,all,ux
cm,nfx2,node
nall

nsel,a,,,40000
rbe3,40000,rotx,all
cmsel,s,nfx2
nsel,a,,,40001
rbe3,40001,rotx,all
nall
! Dummy mat props
ex,1,10e6 $nuxy,1,.23
dens,1,.001

1-800-293-PADT

February 23, 2007

The Focus
(Table, cont.)
The fix is to copy DougTable to another
want you to modify the first 3 rows or your temporary table (created using *dim).
table, set your starting point to be the fourth Then, simply set the stored to location
row (i.e. *voper,tablename(4,i),... where i is parameter of the *voper command to be
the column number).
DougTable, as shown below:
This built-in looping of the *v commands
can cause problems if you want to apply
different operations to different rows of
your table. For example, say you wanted to
modify an existing table that is graphically
shown as DougTable.

*dim,d_temp,table,5,5
*mfun,d_temp(0,0),copy,D
ougTable(0,0)
*get,d_column,parm,'doug
table',dim,
*do,i,1,d_column

Issue 55
*voper,DougTable(1,i),d_
temp(1,i),mult,2
*voper,DougTable(3,i),d_
temp(3,i),sub,10
*enddo
A little practice and creativity can help you
to further expand on this macro. Ultimately, you can move your get to the front of the
macro to pull the x (row) and y (columns)
dimensions of your table and use those to
define your temporary table. To really
show off to your fellow analysts (because if
your spouse isn't an engineer, they definitely won't appreciate it just ask my wife),
you can use Jeff Strain's *Varying your
*Vwrite (July 2004) to make your script
more flexible in handling table names and
different scaling operations.

Unequal Tension/Compression Mat. Prop's

By: Rod Scholl

Back in release 9.0, some significant improvements were made to the Cast-Iron
material model. The most remarkable feature of this material model is the ability to
implement different tensile vs. compressive
plasticity curves. However, this capability
can be expanded (with a little effort) to
model a different compressive vs. tensile
modulus by tweaking the plasticity curve.

TB,UNIAXIAL,1,1,3 ,COMPRESSION
TBTEMP,10
TBPT,,0.203E-02,0.300E+05
TBPT,,0.500E-02,0.500E+05
TBPT,,

To extend this power further, one can essentially model materials with unequal tensile
and compressive modulus. Of course the
problem is non-linear, so there is a penalty
here, but it is a fairly cheap resolution to the
case where the moduli are unequal.

higher, the corresponding modulus is applied. This, of course requires a non-linear

solution, which may or may not be an added
expense depending if other non-lienarities
are already present at their relative convergence rate. Test cases on a simple cube
showed very quick convergence with this
model, with 3 iterations required when
forced to a single substep via
(Cont. on Pg. 6.)

The expected usage of the cast-iron input

By making the linear elastic portion of the
looks like the following:
material model negligibly small, one can
accomplish two separate moduli curves.

A Typical input looks like:

TB,CAST,1,,,ISOTROPIC
TBDATA,1,0.04
TB,UNIAXIAL,1,1,5,TENSION
TBTEMP,10
TBPT,,0.550E-03,0.813E+04
TBPT,,0.100E-02,0.131E+05
TBPT,,

www.padtinc.com

Thus up to some small value of strain, such

as 1e-6 in/in the compression and tension
will use the higher modulus. But for strains
5

1-800-293-PADT

February 23, 2007

The Focus

Issue 55

Awesome APDL: Customizing Animation Macros

By Rod Scholl

course this is how the ANSYS gui disSeveral times now Ive posted emails plays it, but I wanted to be able to
looking for some special graphics easily send these mesmerizing videos
script, such as a custom animation, and to the customer.
each time someone has to remind me It was a simple change to make the 2nd
that I can edit the ANSYS scripts them- half of the capture loop backward. The
selves. Not every command is a script, first part of the code to the right shows
but the animation actions are. There is the original APDL chunk that sets the
a collection of them in the installation distortion scale factor (DSC) and does
directory. ANMODE.MAC is a good the replot. In the second half, an *if
example for mode shapes.
statement is added that scales forward
One thing Ive never liked is that it
only shows the 1st half of the modal
cycle in the .avi file. I want it to show
the expansion AND contraction cycle.
That way I can put the created .avi file
loop-play in a media player and stare
ad infintum at the gracefully moving
image. Ahhh, peace at last Of

on the first half and backwards on the

second.

(tension/compression, cont.)

but the interesting part is as follows:

It usually only takes a few minutes of

poking around, making changes, and
seeing what happens to figure out some
minor modifications. So take advantage of these open source goodies
from ANSYS Inc!

NSUBST,1,1,1. This method shouldnt be

a path-dependent (irreversible) solution if
you are simulating linear modulus with this
method.
And finally, one last trick (which took quite
some time to sort out) is that the curve cant
quite be linear it is allowed in KINH for
example, but not in the case of cast-iron.
Thus you must make the curve minimally
decreasing at each point.
The whole script can be downloaded here,

MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,1,,10e6
MPDATA,PRXY,1,,.3
TB,CAST,1,,,ISOTROPIC
TBDATA,1,0.04
TB,UNIAXIAL,1,1,3,TENSION
TBTEMP,10
TBPT,,1e-6,10e6*1e-6
!NOTE the -1 for slightly
! decreasing slope
tbpt,,1e-5,10e6*1e-5-1

Upcoming Training Classes

Month

Start

Feb '07

26-Feb

28-Feb 104 ANSYS Workbench Sim - Intro

Tempe, AZ

Mar '07

8-Mar

9-Mar

Tempe, AZ

12-Mar

13-Mar 203 Dynamics

19-Mar

20-Mar 501 ANSYS/LS-DYNA

Tempe, AZ

28-Mar

30-Mar 902 Multiphysics Sim for MEMS

Tempe, AZ

Apr '07

May '07

End

Title

105 ANSYS WB NL Structural

LV., NV

4-Apr

101 Introduction to ANSYS, Part I

Tempe, AZ

5-Apr

6-Apr

107 ANSYS WB DesignModeler

Tempe, AZ

9-Apr

11-Apr

401 Low Freq. Electromagnetics

Tempe, AZ

12-Apr

13-Apr

604 Introduction to CFX

Tempe, AZ

16-Apr

18-Apr

201 Basic Structural Nonlinearities

Tempe, AZ

18-Apr

19-Apr

204 Adv. Contact and Fasteners

Tempe, AZ

25-Apr

27-Apr

152 ICEM CFD/AI*Environment

Tempe, AZ

2-May

4-May

104 ANSYS WB Simulation - Intro

7-May

8-May

100 Engineering with FEA

LV, NV
Tempe, AZ

Original Script ----*****************

CAPTURE FOR JavaScript
*****************

~tcl,'ansys::report::animImage'
*ELSE ! } {
/REPLOT
*ENDIF ! }
_DSC=_DSC-_DDD
*ENDDO
!--!COM
!COM
!COM

Modified Script ----*****************

CAPTURE FOR JavaScript
*****************

~tcl,'ansys::report::animImage'
*ELSE ! } {
/REPLOT
*ENDIF ! }
*if,_i,lt,AR11/2,then
_DSC=_DSC-_DDD*2
*else
_DSC=_DSC+_DDD*2
*endif
*ENDDO
!NOTE the -2 for slightly
! decreasing slope
tbpt,,1e-1,10e6*1e-1-2
TB,UNIAXIAL,1,1,3,COMPRESSION
TBTEMP,10
TBPT,,1e-6,10e6*1e-6
!NOTE the -1 for slightly
! decreasing slope
tbpt,,1e-5,5e6*1e-5-1
!NOTE the -2 for slightly
! decreasing slope
TBPT,,1e-1,5e6*1e-1-2
nsubst,1,1,1
solve

ANSYS, Inc. Is doing some great regional user group

meetings in North America. These are highly technical
and well attended: www.converge.ansys.com

Location

2-Apr

!--!COM
!COM
!COM

Links

Feel the need for making high quality graphs in batch

mode outside of ANSYS. GNUPlot has been around for
years and keeps getting better:
www.gnuplot.info

- ANSYS 11.0 Products Released

Visit www.ansys.com for details in the near
future. PADT will cover the 11.0 release
heavily at our on-line seminar and in future
issues of The Focus

News

The Focus is a periodic publication of Phoenix Analysis & Design Technologies (PADT).
Its goal is to educate and entertain the worldwide ANSYS user community. More information on this publication can be found at: http://www.padtinc.com/epubs/focus/about

www.padtinc.com

1-800-293-PADT

February 23, 2007

The Focus

Issue 55

The Shameless
Advertising Page

The 2007 PADT CAE Wall Calenders are here!

One year, at a glance. And this year, no froofroo pastel colors!
Do You Read the Focus
and Find it Useful?

Visit:
www.padtinc.com/docs/lit/padt_2007_1.pdf
to download your copy before your wall fills
up!

Does Your Company

Outsource their RP Jobs?
Have you Considered
PADT as your
RP Source?
PADT is a Leader
in High Quality
SLS, SLA, FDM and
Injection Molding

CFD Simulation Services

Let us Quote your

Next RP Job

PADT Knows Flow

Outsource your next CFD Job to the PADTs CFD Experts
www.padtinc.com
1-800-293-PADT

rp@padtinc.com

www.padtinc.com

1-800-293-PADT