Q. How does FLOTRAN handle the elevation head of a fluid?
A. FLOTRAN subtract the "static" or elevation head from the absolute pressure. In the Navier-Stokes equations, this makes the gravity term look like (Rho-Rhoref)*Grav. This is done for natural circulation problems so that the pressure field at the start, before any density changes have occurred, is uniform. This is good for the numerics of the problem.
In order to calculate correct absolute pressures, FLOTRAN includes the static head if gravity is specified. The calculation is rho*grav*height, where height is measured relative to the global coordinate system. Neither that static head nor the absolute pressure is available in the results. This is usually not a significant factor.
Now, let us consider an application. Suppose you have a section of vertical pipe filled with water. First of all, we ask ourselves, is the pipe being held out in the atmosphere (where we would expect the pipe to begin to empty at t=0), or is the pipe itself underwater? This latter case is what FLOTRAN calculates if you specify P=0 at both ends. We expect nothing to happen, and numerical roundoff is the only thing that would drive flow.
For the former case, one can either run it as a pressure driven problem, where the pressure drop is that calculated by the user: rho*grav*height, or one can simulate the gravity with body force terms: rho*grav applied with the BF command in the appropriate direction.
The situation changes when we activate the free surface model where the elevation head IS a part of the FLOTRAN pressure. This is necessary, since the free surface depends on gravity, inertia, and so on. You can look at the pressure field at t=0 in any free surface problem and easily see the elevation head. |
Q. How does FLOTRAN handle the elevation head of a fluid?<br><br>A. FLOTRAN subtract the "static" or elevation head from the absolute pressure. In the Navier-Stokes... |
I'm running Tetra on my HP-UX machine, and I've specifiied 2 processors within ICEM, but it only seems to report one processor in usage. |
"Top" on HP is a little different than on other operating systems. It will only show one "tetra" process. What you need to monitor is the load distribution for both... |
How do you impose a given displacement in a cylindrical coordinate system in ANSYS WorkBench (WB) 8.0? |
Do the following in Simulation : (it is easier than coming back in ANSYS) :<br><br>1. Create a local cylindrical coordinate system at the desired location. The first one which is... |
Customer gets the following error message when launching ANSYS90 in alpha tru64 5.1 |
<br><br>410114:/ansys_inc/v90/ansys/bin/alpha/ansys.e90: /sbin/loader: Error:<br>libpthread.so: symbol "__atdlclose_mutex"... |
How can I list results and coordinates on a surface? |
The surface calculation commands (SUxx) were added at Release 8.0 and provide capability for surfaces that<br>is parallel to the path capability for data along a path (line). A... |
Q. In trying to get a particle flow trace in Flotran, I am having a problem in getting a good flow trace plot. I input the coordinates of points in the flow area and the only plot I get is a trace in the X-Z plane. I do not understand what I am doing wrong.
A. Most likely, the coordinate systems are not properly set. FLOTRAN expects that RSYS, DSYS, and CSYS are all set to 0. This is independent of the coordinate system used to solve the model. |
Q. In trying to get a particle flow trace in Flotran, I am having a problem in getting a good flow trace plot. I input the coordinates of points in the flow area and the only plot I get is a... |
I am having problems using the ANSYS LIcense Manager in CFX-5.7.1 on Linux x86_64. |
This is because only the 32bit license manager installation was shipped with CFX-5.7.1,<br>which installs correctly, but the scripts to launch it realises that<br>they are running on... |
I am getting an error when importing a gridgen mesh file ito CFX-5.7 Pre. I did not have this problem in CFX-5.6.
ERROR
The file named "gtm.gmo" does not exist.
Unable to import the requested mesh:
Type = "Generic".
File = "w:/share1/cfdmpo/support/320645/passage_7x7x20.grd".
Generic options = "-n".
Specific options = "". |
<br>Try the following:<br><br> cfx5gtmconv -ig '-n' -gridgen <sourcefile> -gtm <gtmfile><br><br>[Remove the -ig... |
ICEM CFD 5.1 on Solaris 64 machine fails to start, with the following error message:
var1
var_hexa
var_post
var_cart3d
Determining Licensing Options
ld.so.1: /usr/ansys_inc/v90/icemcfd/5.1-sol64/bin/test_licensing: fatal:
libF77.so.4: open failed: No such file or directory
Killed
HAS_AIENV: Undefined variable
What is causing this error message? |
<br>On the installation CDs only the 32bits libraries are available. The 64 bits libraries are missing.<br>There a service pack due for V5.1, which will resolve the... |
Q. If I run a model containing SHELL181 and using more then one processor NPROC>1, the rst file is very large (>>1Gb).
What is the cause?
A. This is a bug in 8.1. When using more then one processor (NPROC>1) and with models which include shell181 elements, ANSYS was over estimating and then setting internally the size of the rst file. The bug has been corrected for 10.0 and will be incorporated into a 9.0 service pack. |
Q. If I run a model containing SHELL181 and using more then one processor NPROC>1, the rst file is very large (>>1Gb).<br>What is the cause?<br><br>A. ... |
QUESTION:
Do you have an example demonstrating a modeling procedure that can be used to model a current fed voice coil motor, such as those used in computer hard disk drives? This must be a 3D model and I would like to be able to calculate the structural response and coil displacement as a function of time given current excitation in the stranded coil.
ANSWER:
A 3D COUPLED FIELD model of a hypothetical voice
coil motor is attached to this solution. Inspect the
presentation. Place the animations in the same directory that you copy
the presentation to so that they will play in the presentation.
A well documented parameterized input file and post processing macro are
included. The input file creates the model and calculates the transient
magnetostructural response. The post processing macro produces xy plots
of (1) the net MVDI force acting on the magnetic circuit and (2) the net
Lorentz force acting on the coil.
Results quantities calculated:
1) Fields due to permanent magnets in magnetic circuit of current
in coil
2) Forces acting on magnetic circuit (MVDI) and coil (Lorentz)
3) Coil displacement versus time
Assumptions:
1) No eddy currents are induced in magnetic circuit - it is
modeled with SOLID96 MSP elements
2) Coil is current fed (current, rather than voltage controlled).
SOLID62 coupled field elements are used for this.
3) In my example the coil undergoes rigid body motion (due to
nodal coupling on displacement degree of freedom) and so is stress free.
However, by eliminating the nodal coupling, the deformation as well as
the translation of the coil could be calculated.
Note that I tried a number of other modeling strategies, including the
multifield solver. However, to date, this is the only approach I have
been able to successfully implement for this type of simulation. |
QUESTION:<br>Do you have an example demonstrating a modeling procedure that can be used to model a current fed voice coil motor, such as those used in computer hard disk drives? This must be... |
How to calculate the cylindrical velocity components in CFX-Post |
CFX-Post includes additional Turbo post-processing capabilities for turbomachinery. However, also for non-turbo applications the turbo menu provides the capability to create a Cylindrical... |
QUESTION:
I get the error shown below when I use the ANISO material model. My tensile yield in the x, y, and z directions is 102,000 psi and my compressive yield in all 3 directions is 72,000 psi. Satisfaction of the consistency equation only appears to allow for very small differences between tensile and compressive yield strength. For instance, a difference of 15 psi would satisfy it, but that small difference would be useless to me, and to most others for that matter.
*****************************
*** ERROR *** CP= 2.125 TIME= 18:02:13
TheTB,ANISO table for material 1 has anisotropic yield stresses that do not satisfy the consistency equation requirement. Consistency equation equals 1.058823529E-05 which must be less that 1.883239171E-09. Improve the accuracy of the input yield stresses.
*****************************
Are there any other material models that can handle real differences between the tensile and compressive response for materials with a history of cold rolling?
ANSWER:
Unless one is willing to accept the different value of yield stress in one of the directions, the consistent condition will never be met.
If both the elastic and plastic responses of the material are completely isotropic, and the only is difference is yield stress in tension and compression, then use the cast iron option to mimic the behavior by using
TB,CAST,1
tbdata,1,0.5
TB,uniaxil,1,,,TENSION
tbpt,,,,
TB,uniaxil,1,,,COMPRESSION
tbpt,,,,
ADDITIONAL DETAILS:
The option of using tb,cast option was suggested as the properties of this material are very close to isotropic in both elasticity and plasticity. The only special thing here is that they have difference yield stress and tension and compression. It is actually not really an orthotropic material, which is why it was suggested to use the cast option with plastic Poisson's to 0.5. This will ignore the Ra |
QUESTION:<br><br>I get the error shown below when I use the ANISO material model. My tensile yield in the x, y, and z directions is 102,000 psi and my compressive yield in all 3... |
After reading some geometry into DesignModeler 9.0, transferring it to
Simulation, and then returning to the DesignModeler tab, the geometry
can no longer be manipulated. The scaling does not seem to work. Why? |
Defect 30026 was filed and corrected in the UP20050120 build of ANSYS<br> Workbench 10.0 by modifying the dynamic link library "DSGStreamU.dll".<br> This fix... |
I am a university user running Solid Edge v16 and Workbench. When I try to bring SE geometry into Workbench the program tells me there is "no active document". What's wrong? |
Solid Edge has placed certain license restrictions on their academic versions. The API calls necessary to read SE geometry into WB are no longer available.
|
Background:
When using the MPICH distributed parallel feature in CFX 5 on a Windows network, the MPICH daemon must be installed on all the hosts involved in the run. ADDITIONALLY on the MASTER node, details of a user with access to all the hosts must now be set. (DOMAIN username & password). [Refer to the User Manual for more details]
Problem:
Every time the WINDOWS DOMAIN user name and/or password is changed (which most networks require every 30 days etc.), the details should be reset on the MASTER node. On some machines, everything works fine on initial setup. BUT after the password is changed, when trying to reset the MPICH user registration, it causes an error as detailed below:
1) User un-registers info using "cfx5parallel -unregister-mpich-user" from the command line.
2) User changes password, and logs into his master node.
3) User tries to re-register with new password:
-------------------------------------------------------------------------
>cfx5parallel -register-mpich-user
Starting the MPICH User Registration tool. You will need to give a user
account and password which can be used to log into the other machines in a
distributed parallel run. For network users, you can include the domain
name and a backslash preceding the user, eg "BEDROCKFred". If you do not
choose the information to be persistent, you will have to re-do this
process every time the machine is rebooted.
No user information is currently registered.
account: akpaul
password:
confirm:
Do you want this action to be persistent (y/n)? y
SetupCryptoClient:CryptAcquireContext(...) failed, error: -2146893809
Error Text: Object already exists.
Error: Unable to setup the encryption service.
The new password was not successfully encrypted.
------------------------------------------------------------------------
NOTE: A similar error happens if the user tries to re-register without un-registering first.
Solution:
According to the MPI maintenance team at Argonne Nationa |
Background:<br><br>When using the MPICH distributed parallel feature in CFX 5 on a Windows network, the MPICH daemon must be installed on all the hosts involved in the run.... |
Q. How can I control how well a variable is resolved on a chart? I made two lines of different length and they indicated different values of pressure for a certain location on the line.
A. This is controlled during the definition of the line objects themselves by modifying the number of sample points. This is available in the as "Samples" under the Geometry tab when the object is created. By using, for example 50 points on a line 50 cm long, and then 83 points on a line 83 cm long, you ensure representations consistent between the two lines. |
Q. How can I control how well a variable is resolved on a chart? I made two lines of different length and they indicated different values of pressure for a certain location on the... |
Q. How to define a energy source (sink) such that a specified temperature is maintained in the sub-domain?
A. Set an Energy Source term in the required subdomain in the following form:
Source = -C*Cp*(T-Tspec)
where C is a large number (example 10E+8[kg m^-3 s^-1]), Cp is the specific heat capacity of the fluid, T is the local temperature and Tspec is the temperature that needs to be maintained in the sub-domain.
The following CEL demonstrates how to implement this in CFX 5. (NOTE the value of the 'Source Coefficient' in the definition of the sub-domain source)
CEL:
EXPRESSIONS:
Constant = 10E+8 [kg m^-3 s^-1]
Tspec = 300 [K]
heatsink = -Constant *Specific Heat Capacity at Constant Pressure *( T - Tspec)
END
END
SUBDOMAIN: sink
Coord Frame = Coord 0
Location = B41
SOURCES:
EQUATION SOURCE: energy
Option = Source
Source = heatsink
Source Coefficient = -Constant*Specific Heat Capacity at Constant Pressure
END
END
END |
Q. How to define a energy source (sink) such that a specified temperature is maintained in the sub-domain?<br><br>A. Set an Energy Source term in the required subdomain in the... |
I am running a simple laminar pipe flow problem in CFX-5.7. I put in a
developed parabolic velocity profile at the inlet to the pipe. I have using an additional variable
to model a diffusive scalar.
My boundary condition for the additional variable at the inlet is 1 (uniform). My boundary condition for the additional
variable at the outer pipe wall is zero.
I would expect that the flux of additional variable into the wall (or the corresponding normal flux), would decrease
down the length of the pipe as would the value of the additional variable.
I observe the following behavior:
1. For high diffusivities of the additional variable, the behavior is as expected.
The normal flux of the additional variable at the wall decreases continuously with pipe length.
2. For low diffusivities, however, I see a initial region where the normal flux increases with pipe length
before it reaches a maximum value and decreases thereafter. If I look at the conservative value
of the additional variable at the pipe wall, I see that it has a high value initially before decreasing to
zero. The hybrid values of the additional variable at the wall are always zero. The region of non-
zeroconservative additional variable wall values corresponds to the region where the normal flux
increases with pipe length. |
<br><br>The behavior described is caused by an inconsistency between the additional variable boundary<br>conditions at the wall and at the inlet.<br><br>At the inlet,... |
Simulation is sometimes not able to locate result files after the solution is complete. This applies to Asynchronous Solutions (UNIX Server, LSF Cluster, WB Cluster)
This occurs when the user submits the job, then exits Workbench.
If Simulation is able to find the results, there will be green arrows next to the results. You can then retrieve them by RMB on Solution and selecting Get Results.
If Simulation is not able to find the results, then there will be yellow lightning bolts next to the solution results. |
Workaround : Save all model files AFTER you start the remote solution and BEFORE you exit Workbench.<br> In contrast, if you do not save the results AFTER the batch... |
In ANSYS Release 9.0, how can you obtain the total amplitude from a harmonic analysis? |
In a harmonic analysis, the real and imaginary results are stored. The total amplitude is the SRSS of the real and imaginary amplitude. The HRCPLX command can be used to calculate the total... |
Can you confirm if there will be a version of CFX-5.8 for the Opteron chip? If so which version of 64-bit Linux will be supported? |
There will definitely be a version of CFX-5.8 optimised for the Opteron chip. The supported version of 64-bit Linux will be SuSE Linux 9.1. It is also hoped that Enterprise 3.0 will be fully... |
The release notes for CFX-5.7.1 mention that there was an improvement
in the accuracy of the Redlich-Kwong equation of state near the critical point.
The 5.7.1 documentation on the Redlich-Kwong EOS only gives details on the
original formulation. What changes have been made to the RK EOS in
5.7.1? Does this improvement only affect the Redlich Kwong equation of state
or does it involve other property evaluations (entrophy, enthalpy table look up)
near a critical point. |
<br><br>CFX-5.7 contained the standard original implementation of the RK equation<br>that dates back to 1949.<br><br>In 5.71, slightly modified version from Aungier... |
QUESTION:
Do you have an example showing the difference in natural frequency of a piezoelectric transducer when the electrodes are (a) shorted and (b) open?
ANSWER:
Look at VM175 or the attached input file.
Please copy the attached files to a working directory and solve in an interactive ANSYS session. The attached png files will be created. You can see that, as with VM175 in our online documentation, electrical boundary conditions affect natural frequencies. In this case, the lower electrode (z=0) is grounded. Two separate modal analyses are performed:
1) upper electrode grounded (V=0 @ z=dz)
2) upper electrode open circuited (nodes coupled so that the solved-for nonzero voltage is uniform over the electrode surface)
In neither my test model nor VM175 in the online documentation, no prestress steady state analysis was performed prior to the modal analysis. I don't think it's necessary or appropriate to do so, since VOLT and UX/Y/Z are matrix coupled in this element. |
QUESTION:<br>Do you have an example showing the difference in natural frequency of a piezoelectric transducer when the electrodes are (a) shorted and (b)... |
Using the Drop Test Module, (DTM), I made a simple four element beam model and "dropped" it as follows:
gravity = 0 and plenty of initial height so that the model can spin without interference
translational velocity = 0
rotational velocity = -6.28 (rad/sec)
With a solution time of 1 second, I expected the beam's final position to be 6.28 radians (1 revolution) from the starting position. But the solution shows that the beam only gets to 2.47 radians after 1 second. Can you please explain why? |
To define an initial angular velocity, you must simultaneously define an angular velocity AND spatially varying (proportional to radial coordinate) translational velocities. The initial... |