1. Why is the performance of SHELL181 in vibration (modal) analysis poor?
This element appears to be giving consistently stiffer results. Representative percentage errors
are of the order of 5 ' 8 as the mode number approaches 100 in a structure having nearly 19000 DOF.
Other shell elements ' 43, 63 and 93 are performing better.
The foregoing findings are based on studies for a case study problem involving four ANSYS elements and
2 NASTRAN elements. The findings are summarized in the attached XL sheet. |
(From product development) For first order elements, lumped mass approximation usually provides better accuracy.<br>Apparent "poor" performance of 181, in higher modes, that... |
How do I get reaction forces on spot welds in Simulation? |
There is no way to obtain the force in a spot weld through the<br>interface. You can create a named selection for the pair of vertices<br>that define a spot weld and use esln,s,1 in... |
What does the following error message mean?
*** ERROR ***
Error in symeqn called by dstmak. Please send your data to your ANSYS
Technical Support Provider. |
In discussing this error message with Development, they noted that:<br><br> "The symeqn.F routine is simply trying to invert a matrix and it is detecting<br> that the... |
Class 3 Error
SYMPTOM:
The integrated quantity function massFlowAve() when applied to the variable Velocity on either a Sample Plane or a User Surface returns incorrect values.
PROBLEM:
This is due to a memory corruption in CFX-Post, which happens becuase Velocity is used for both mass flow calculations and as the quantity to average. The problem is limited to "external" geometry objects, i.e. surfaces that are not directly mesh related (user surfaces and sam |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>ANSYS CFX Release 11.0.
|
Q. What exactly happens when switching from single to double precision in CFX-11, specifically on a 64bit machine? Also, is the behaviour any different between solver and partitioner? And what about coordinates?
A. When running CFX-11, the solver code (whether partitioning or not) writes out a header to the outfile which includes a line something like this:
| single-int32-64bit-novc6-optimised-supfort-noprof-nospag-lcomp |
The first three entries are of interest and these represent floating point precision, integer precision and then addressing.
If you are running on a 64bit machine with an appropriate OS, then your addressing will be 64bit allowing you to access large amounts of memory and solve larger and more complex models. However, it does not imply anything about the size of floating point variables or integers. With Fortran (which is used for the CFX solver), the default for these is always 4 bytes. If you request double precision, then within the solver this means all floating point is done with 8 bytes precision. It doesnt affect integers or addressing... the latter being a function of the architecture and OS. The drawback with running in double precision is that there is usually a significant CPU overhead from moving around double the quantity of data. However, there is some evidence to suggest that on 64bit operating systems the overhead is much reduced.
With partitioning, the default behaviour is to ignore any double requests because there is very little to be gained within the partitioning methods by having 8byte floating point (the RCB method being a possible exception). However, you can work around this by explicitly specifying the partitioning executable and pointing it to the double precision executable in the installation. Also, you can over-ride the default 4byte integer default for the partitioner by specifying '-part-large' which gives 8byte integers (and logicals) and allows the partitioning of much larger problems up to |
Q. What exactly happens when switching from single to double precision in CFX-11, specifically on a 64bit machine? Also, is the behaviour any different between solver and partitioner? And what... |
SYMPTOM:
When modifications (e.g. name changes, location changes, creation/deletion) are made to Named Selections in DesignModeler, some changes are not reflected in CFX-Mesh after perrforming a Geometry Update action.
PROBLEM:
This happens because there is no internal identifier associated with a Named Selection other then the user specified name. Hence when it is renamed, for example, the association between the Named Selection and the Region in CFX-Mesh is lost.
The following behaviour should also be noted:
If you modify the location of an existing Named Selection in DesignModeler, then when you update the geometry in CFX-Mesh, this modification will be applied to the equivalent Composite 2D Region in CFX-Mesh as long as the new location does not overlap with the location of any existing Composite 2D Regions in CFX-Mesh (including other regions generated from Named Selections).
If you delete a Named Selection in DesignModeler then when you update the geometry in CFX-Mesh the equivalent Composite 2D Region is not deleted. You must delete it manually if you no longer want this region in CFX-Mesh.
If you rename a Named Selection in DesignModeler, then when you update the geometry in CFX-Mesh, the Named Selections will not be re-imported. You must either rename or delete the equivalent Composite 2D Region in CFX-Mesh before updating the geometry. This is because CFX-Mesh effectively sees the renamed Named Selection as a new Named Selection, and then finds that its location overlaps with the existing Composite 2D Region which has the original name of the Named Selection. |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>ANSYS CFX Release 11.0.
|
Why do I receive the following error message when trying to open any agdb files?
Unable to checkout 'Ansys BladeModeler' license:
Product: Ansys blademodeler (feature 'acfx_bldmdlr')
Checkout failed for the above product
Flexlm error message:
No such feature exists |
BladeModeler is not required to run DesignModeler unless the user has selected it as the default 'Modeler' license in Tools - Options - Licensing - License... |
Could you explain how in CFX the Turbulence Eddy Frequency and the Turbulence Eddy Dissipation are
correlated? Is it defined as omega = epsilon /k ? |
From the fact that the turbulence eddy viscosity is defined as<br><br>in k-epsilon model: mu_t = C_mu*rho*k^2/epsilon<br>and<br>in k-omega model: mu_t... |
I am trying to plot wall shear stress as a function of time. To do so, I select the fluid.rfl file for time history postprocessing. Then I add the item 'shear stress at the wall' to the variable list. But irrespective of the node (at the wall) I choose, the min and max values are being shown as 'invalid'. When I do a contour plot for the wall shear stress, it shows me reasonable values, but I cannot plot it as a function of time. What could be the reason for this? |
This is a known issue (bug report 19895) and has not been fixed. At this point it must be considered a code limitation.
|
How can I use CHKMSH to select unmeshed areas or volumes? |
The documentation for the CHKMSH macro does not adequately define<br>how to use it. I did several tests and the GUI creates 3 components in<br>the process of using the CHKMSH... |
The CFX 10.0 solver fails with the following error message in the solver output file:
Error detected by routine PEEKI
CDANAM = /FLOW/ALGORITHM/ZN1 /SYSTEM/VARIABLES/SPHEATP_FL1 /IVAR
CRESLT = NONE |
<br><br>This PEEKI error message occurs because you have not specified a specific heat capacity value for your material(s). Even if your run is isothermal, the specific heat capacity... |
How can I get the Advanced Meshing tab to appear to Workbench 11.0 Preview 5 on Linux? |
You need to run a script called setup.csh appearing in the installation directory aiewb.<br>For example:<br><br>If the installation is... |
I did the following steps in ANSYS Workbench Environment (AWE), but no components are shown in ANSYS Classic Environment.
What else must be done?
1. Imported geometry and created a named selection in Workbench.
2. Chose "Open analysis in ANSYS" from the project pane to transfer
the simulation into Classic ANSYS.
3. Once ANSYS opened, chose the Component Manager from "List -->
Component Manager". The named selections should appear here, but don't. |
In order for ANSYS Workbench Environment to properly create the<br>component definitions for the ANSYS Classic Environment, you need to have the<br>Environment and Solution branches... |
Customer installed Platform LSF 6.2 and ANSYS 10.0 SP2 on a WinXP x64 system.
When he tries to launch products through the product launcher, he receives the error dialogue:
Application Error
Error: invalid commands name "lsf_start"
(complete error message is attached to this solution)
What is causing this error message when trying to set up ANSYS with the Platform LSF integration? |
Unfortunately, LSF doesn't run on Windows x64 at ANSYS 10.0 version.<br>It's been put into version 11.0 for the Intel platform, but even at 11.0, it won't work on the... |
Do you have an APDL example demonstrating how to use [*VFUN,,PRIN] to calculate principal stresses from component stresses? |
C******************************************<br>C*** MAKE UP STRESS... |
QUESTION:
Do you have an example demostrating the procedure to model liquid on top of a piezoelectric device (SAW) using MFS single code to do the FSI?
ANSWER:
Attached is a hypothetical piezo SAW covered in a fluid modeled with FLOTRAN CFD elements with harmonically varying voltage on the two electrodes (180 degrees out of phase). Both the piezo and CFD fields are solved as transients using MFS.
The only thing I can think of adding to this model for completeness is periodic boundary conditions (if the system you are interested in is indeed periodic). I tried to do this but had trouble applying the PERI.MAC macro to the CFD elements ' I suspect the macro does not anticipate duplicate nodes on the FSIN boundary between the CFD and piezo elements.
Verification of the results is up to the user as I really have no idea what kind of fluidic response to expect for such a system. Note that I exaggerated displacement in piezo voltage animation for clarity. Also note that I placed toolbar button calls to macros that this input file creates. The macros are set up for postprocessing either field (the piezo one or the CFD one) in POST1 and POST26 (they're for your convenience to simplify the set up procedure). |
QUESTION:<br><br>Do you have an example demostrating the procedure to model liquid on top of a piezoelectric device (SAW) using MFS single code to do the... |
UESTION:
I'm attempting to simulate the performance of an electric generator using a transient 2D planar circuit coupled electromagnetic model in which I "manually" spin the rotor with either magnetic contact or updated CEINTF on a sliding boundary in the airgap. The model runs fine with linear permeability but fails to converge when I use nonlinear BH data. How do I overcome divergence?
ANSWER:
Try the following user-defined convergence criteri |
<br><br>cnvt,csg,,0.01<br>cnvt,volt,,0.01
|
We will upgrade UG from NX2 version to NX4 version.
Is there any issue with ICEMCFD/UG direct interface(UG--->Tetin) for UG NX4 version? |
ICEM 11.0 currently does not support UG version NX4. There is a patch for Windows platforms in the works.<br>It can be made available for users who wish to try upgrading UG to NX4, but has... |
I recently installed a beta version of Internet Explorer 7.0. After I did this, I encountered problems with various Workbench 10 SP1 modules.
CFX-Mesh sometimes crashes but at least it is usable. Design Modeler and Simulation won't even load. For instance, if I try to load a .dsdb file, the program
gets hung up before launching. If I open up a new file in DM, I can create a sketch and then it freezes. |
There have been various problems encountered with Workbench 10.0 and the beta versions of IE 7.0. If you revert to IE 6.0, the<br>Workbench 10.0 modules should function normally.
|
It is often useful to use a CEL expression as the argument of an integrated quantity function. For example, for monitoring purposes. Unfortunately, it is not possible to use a CEL expression as an argument of an integrated quantity function, as only variables are permitted as arguments to this type of function. A workaround to this problem is described below. Note that this restriction only applies to integrated quantity CEL functions. Other functions can take expressions as arguments. |
<br><br>This restriction can be worked around by using an algebraic additional variable.<br><br>For example, suppose the following expression had been... |
Class 3 Error
SYMPTOM:
If the solution unit for angles is changed from the default of radians to anything else, e.g. degrees, then the Solver continues to interpret all angles in radians, and
consequently gets incorrect answers.
PROBLEM:
This is due to a bug in the CFX Solver in which the code only allows radians as the solution uni |
WORKAROUND:<br>The workaround is to not change the solution unit for angles.<br><br>FIXED IN:<br>Release 11.0.
|
Class 3 Error
SYMPTOM:
Multiphase flows with domain interfaces of type stage give randomly incorrect answers.
PROBLEM:
This is due to a bug in the CFX Solver in setting the stage group mass flows which point to uninitia |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>Release 11.0.
|
Class 3 Error
SYMPTOM:
When particles cross domain interfaces of type stage, they should restart at random angular positions for the same axial and/or radial position on the interface, but in parallel
this does not happen correctly.
PROBLEM:
This is due to a bug in the CFX Solver where the random position is restricted to be in the same partition in which the particle reached t |
WORKAROUND:<br>There is no workaround other than to run in serial mode.<br><br>FIXED IN:<br>Release 11.0.
|
Class 3 Error
SYMPTOM:
The Reitz and Diwakar break-up model gives incorrect droplet diameters when the model is in the stripping regime, that is for high speed sprays.
PROBLEM:
This is due to a bug in the CFX Solver where it uses the dynamic viscosity rather than the kinematic viscosity in the droplet diameter |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>Release 11.0.
|
Will Ansys 10.0 run on SuSE 10 Linux? |
We certified SuSE 10 OS on the EM64T system:<br><br>SuSE LINUX 10.0 (X86-64) OS<br>Kernel 2.6.13-15.8-smp x86_64 GNU/Linux
|