Class 3 Error
SYMPTOM:
If the discrete transfer radiation model is run in parallel with multigrey or multiband properties, and there are reflective walls with emissivity less than one, then the Solver will silently give an incorrect answer.
PROBLEM:
There is a coding error in communication within the radia |
WORKAROUND:<br>Run in serial mode if possible.<br><br>FIXED IN:<br>Release 10.0 SP1.
|
QUESTION:
I want to expedite solution turnaround on a harmonic response analysis of a periodically arranged array of piezoelectric transducers. The transducer excitation will not be periodic, so symmetry of periodic boundary conditions will not be appropriate (I will have to represent the entire structure or at least several transducers to either side drive transducer). Ultimate objective is to determine cross talk between any given drive transducer and those adjacent to it (information to be used for the purposes of designing a control system).
Would substructuring work? How about doing a mode superposition harmonic response analysis.
ANSWER:
I tried substructuring,but was not successful (VOLT DOFs not allowed in a substructure including mass and damping) so I went with the mode superposition harmonic response procedure. I was leery of using voltage excitation on the drive electrode because when run this way, the drive electrode voltage was reported equal to zero in postprocessing and I also saw statements to the effect that non-zero DOF excitation cannot be used in mode superposition. So I used charge excitation (ANSYS command: F,,AMPS ' misleading but it's negative charge that gets applied with this command).
More details in the readme in the attached zip file.
The attachment to this solution contains an input file, macros, plots and an animation. Input file is completely parametric, creates, solves, and postprocesses model. |
QUESTION:<br><br>I want to expedite solution turnaround on a harmonic response analysis of a periodically arranged array of piezoelectric transducers. The transducer excitation will... |
I am running an unsteady free-surface calculation in CFX 10.0 with Service Pack1. The run works fine on a DUAL Opteron (2 parallel), but crashes on 2-DUAL core procs (4 parallel).
Do you know how I can overcome this problem? |
This problem has been seen before, and it is to do with the Metis partitioning, when the partition boundaries are parallel to the free surface.<br>The workaround would be to partition... |
Class 3 Error
SYMPTOM:
If the User runs a particle transport case within a single rotating frame, it is likely that the tracks observed in CFX-Post will be incorrect. This will probably be apparent. Additionally, the expert parameter 'pt tracks in absolute frame' will have no effect.
PROBLEM:
There is a coding error in the routine that determines how to write the tracks into |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>Release 11.0. Custom executable for Release 10.0 available on request.
|
Class 3 Error
SYMPTOM:
If the coal swelling model is used, then the code will give incorrect answers, because the swelling is coded based upon the masses of all species in the coal particle and not just on the reference material.
PROBLEM:
There are coding errors in the routine that calculates the diameter change, and in the |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>Release 11.0. Custom executable for Release 10.0 available on request.
|
QUESTION:
What is the procedure for using SURF152 to define both convection and radiation acting on one side of a mesh of SHELL131 layered shell elements having throug thickness conduction capability.
ANSWER:
An example input file is listed below. Procedure can be inferred from comments and commands appearing in the listing.
fini
/cle
C*********************************************************
C*** MODEL DEMONSTRATES PROCEDURE TO USE SURF152 TO
C*** CONVECT AND RADIATE HEAT AWAY FROM THE TOP
C*** SURFACE OF SHELL131 LAYERED THERMAL SHELL
C*** HAVING THRU THICKNESS CONDUCTION. A TRANSIENT IS
C*** PERFORMED.
C*********************************************************
/sys,del *.png
/vie,1,3,2,1
/vup,1,z
/esha,1
/pnu,mat,1
/num,1
C*********************************************************
C*** PARAMETERS
C*********************************************************
a=0.100 ! PLATE EDGE LENGTH (m)
t=0.010 ! PLATE THICKNESS (m)
k=1 ! PLATE THERMAL CONDUCTIVITY (W/m-C)
emsvty=1.0 ! EMMISSIVITY
formf=1 ! FORM FACTOR
Hc=5 ! CONVECTION COEFFICIENT (W/m^2-C)
Ht_area=500 ! HEAT FLUX APPLIED TO UNCOOLED SIDE OF PLATE (W/m^2)
T_env=21 ! ENVIRONMENT TEMPERATURE (C)
t_final=500 ! END OF THERMAL XSIENT
z_xtra=-1*t ! Z COORDINATE OF RADIATION XTRA NODE
C*********************************************************
C*** GEOMETRY
C*********************************************************
/prep7
rect,-a/2,a/2,-a/2,a/2
C*********************************************************
C*** ATTRIBUTES
C*********************************************************
et,1,131 ! PLATE (LAYERED THERMAL SHELL)
keyopt,1,3,0 ! UP TO 15 LAYERS, QUADRATIC TEMP THRU t
keyopt,1,4,1 ! ONE LAYER THRU t
sectype,1,shell ! SECTION PROPERTIES
secdata,t,1
mp,kxx,1,k
mp,dens,1,7800
mp,c,1,5
et,2,152 ! CONVECTION SURFACE EFFECT ELEMENTS
keyopt,2,4,1 ! NO MIDSIDE NODES
keyopt,2,8,2 ! EVALUATE Hc @ AVERAGE TEMP
r,2 |
QUESTION:<br>What is the procedure for using SURF152 to define both convection and radiation acting on one side of a mesh of SHELL131 layered shell elements having throug thickness... |
Class 3 Error
SYMPTOM:
If mass sources are included in a calculation using the Reynolds stress turbulence model, then the code will fail. Sometimes the failure will be a quick divergence of the solution, but it may just be an incorrect answer.
PROBLEM:
There is a memory overwrite in the routine that calculates the effect of mass sources in the Reynolds stre |
WORKAROUND:<br>There is no workaround.<br><br>FIXED IN:<br>Release 11.0. Custom executable for Release 10.0 available on request.
|
Does Workbench 10.0 SP1 for the Windows x64 OS support LSF? |
Not at this time. Platform Computing currently does not have LSF drivers for the Windows XP x64 operating system.<br> We are currently awaiting word from Platform Computing as to when... |
Linux AMD Opteron 64bit: User Fortran compiled with PGI compiler fails on machines without PGI compiler:
| Message: |
| |
| Error opening /f1home/heickenb/Fortran/linux-amd64/libPipeValve_i- |
| nlet.so: libpgf90.so: cannot open shared object file: No such fil- |
| e or directory |
| |
Follow the following steps:<br><br>1.) Logon to linux machine with PGI:<br><br> -compile with the command:<br><br> /CFX/CFX-10.0/bin/cfx5mkext... |
How do I resolve the following error when starting Workbench 10.0 SP 1 on my Windows XP 32-bit machine?
Error - "JMService module has encountered an error and needs to be closed" |
Manually register JMService.dll, as follows:<br><br>C:Program FilesANSYS Incv100AISOLCEintel>regsvr32 JMService.dll<br><br>Note: The regsvr32.exe typically resides in C:WindowsSystem32... |
Customer had problems installing Workbench 10 SP1 for Windows x64.
Machine: Intel Xeon 3.60 GHz; RAM: 4GB and O/S: Windows XP 64-bit SP1
The errors are listed below:
First error:
ANSYSWBU.exe - Unable to locate Component
X This application has failed to start because MSVCR71.dll was not found. Re-installing the application may fix this problem
Second error:
Self-Registration Error
1. C:ANSYS_Incv100AISOLCpmmonFileswinx64AGGfxU.dll
The specified module could not be found
2. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsEngineeringDataU.dll
The specified module could not be found
3. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsImageU.dll
The specified module could not be found
4. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsStdCollectionsU.dll
The specified module could not be found
5. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsStdObjectU.dll
The specified module could not be found
6. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsStdUtilU.dll
The specified module could not be found
7. C:ANSYS_Incv100AISOLCpmmonFileswinx64AnsWbu.dll
The specified module could not be found
8. C:ANSYS_Incv100AISOLCpmmonFileswinx64CommentEditorU.dll
The specified module could not be found
<More files...> |
According to the messages above, something prevented our installation from writing<br>the winx64 version of msvcr71.dll, msvcp71.dll and atl71.dll to the system32... |
Why do I receive the "Tcl or TK not Initialized" error if I minimize the window, then restore it when running ANSYS Workbench 10.0? |
This is a defect (42585 ), which has been resolved in version 11.0.<br><br><br>The workaround for previous versions is to resize the window while ANSYS is active.<br>You... |
I have not been able to make CFX run in parallel on 64-bit Windows machine? |
<br><br>On a 64-bit machine, the default CFX installation path is usually:<br>C:Program Files (x86)Ansys IncCFXCFX-%v<br>However, there is a<br> C:Program... |
Even though I'm plotting meridional contour plots, the meridional legend still comes up as a smeared color legend. I can't make the color bands discrete. This makes it hard to match up the color bands on the legend with the bands on the contour plot correctly.
Q: Also, the legend doesn't respond to how many contours I have in my contour plot. I have to make a point to set the number of tics to match the contours. |
Full control of turbo plots can be restored by promoting them out of the turbo mode to the general mode. You can promote your meridional contour plot to general mode by right clicking on it in the... |
Will ANSYS run on SuSE 10.0 on an AMD Athlon 64? |
No. ANSYS is not supported on SuSE 10.0 or the Athlon 64 chip.<br>ANSYS may install and run, but it is an unsupported platform.
|
Class 3 Error
SYMPTOM:
If variable Cp definition is used, then different results of total pressure on Linux and Windows / HPUX are obtained for exactly the same case (same DEF file) .
PROBLEM:
There is an unset variable in the 1D version of the subroutine which generates the h(T) table, hence the platform |
WORKAROUND:<br>There is no workaround<br><br>FIXED IN:<br>Release 11.0.
|
SYMPTOM:
If the MFX solver is used and conservative interpolation is selected for the load transfer between CFX and ANSYS, the results can be incorrect although things appear to run fine.
PROBLEM:
When conservative interpolation is selected, the CFX variable name for the load transfer must be given as "Total Force" or "Wall Heat Flow", as documented in the ANSYS online help for the MFLC command. This applies whether the case is set up through the ANSYS Prep7 user interface (under "Custom" load transfer) or directly through the MFLC command in the ANSYS input file. If the CFX variable name is set up correctly, then cases where the conserved quantity (force or heat flux/flow) is transferred from CFX to ANSYS work fine and give correct answers, but cases where the conserved quantity is transferred from ANSYS to CFX will stop with an error as CFX is not currently capable of receiving "Total Force" and "Wall Heat Flow" quantities.
If the CFX variable name is given as "Total Force Density" or "Wall Heat Flux" for a conservative load transfer, then this is a setup error. However, the case will apparently run without errors. ANSYS will still send "Total Force" and "Wall Heat Flow" quantities, but CFX will read them as if they were "Total Force Density" or "Wall Heat Flux", which means that in effect, the received quantities are incorrect by an area factor. In most cases, this area factor will be sufficiently large that either the case will be impossible to converge or will converge to answers which are wrong by orders of magnitude and clearly incorrectly. However, for some cases the area factor may be close enough to 1 that apparently sensible answers can be obtained which are actually incorrect. |
WORKAROUND:<br>For cases where the conserved quantity (force or heat flux/flow) is transferred from CFX to ANSYS, ensure that the CFX variable is set as "Wall Heat Flow" or... |
How can I adjust the launcher100 and ANSLIC_ADMIN Utility font size on a Unix machine? |
You can put the -fontsize option on the command that you use to start the ANSLIC_ADMIN utility or the launcher100.<br>Try starting with size of 12 (e.g., -fontsize 12), and then go up from... |
Are there any examples of beam-to-beam contact simulating a telescoping cylinder? |
Yes, please see the attached ANSYS 10.0 SP1 input file, "telescoping.inp". It uses<br> tubular BEAM188 elements along with CONTA176 and TARGE170 elements. The<br> ... |
Q. Can we change the font size of ANSYS Help
A. Increasing font sizes in Online Help
The online help system for ANSYS and ANSYS Workbench 10.0 specifies a standard display font at 10 point for body text.
This font size provides optimal viewing of the help.
If you are viewing the help in Unix, you may increase the font size on your display by holding down your Ctrl key and tapping +.
If you are viewing the help in Internet Explorer on Windows, you may increase the font size in the following manner:
1. Select "Internet Options" from the Tools menu. The Internet Options box displays.
2. Click the "Accessibility" button in the dialog box. The Accessibility box displays.
3. Check the "Ignore font sizes specified on Web pages" checkbox.
4. Click OK.
5. To increase/decrease the font sizes, hold down your Ctrl key and slide your mouse wheel up or down.
Note that this will affect all web pages you view in Internet Explorer. |
Q. Can we change the font size of ANSYS Help<br><br>A. Increasing font sizes in Online Help<br>The online help system for ANSYS and ANSYS Workbench 10.0 specifies a standard... |
ICEM CFD 10.0 user is not able to import files from UGNX3, running on HP-UX. |
ICEM CFD developers know about this isse and are currently working on it.<br><br>The workaround would be to export from UG as a Parasolid, then import that into ICEM. The customer... |
Could you provide an explanation for the "Interface Compression" setting in CFX-Pre - what it is, when it should be used and what the consequent values mean. |
<br>Free surface flows use a compressive differencing scheme for the advection of volume fractions to achieve a sharp interface. The Interface Compression level controls which scheme is... |
In a two-phase flow, what's the meaning of "Liquid | Vapor Slip Reynolds Number" that appears in the Solver OUT file? Could a very small value cause problems? |
<br>This is a Reynolds number based on the slip velocity between the two phases rather than the absolute velocity of any one of the phases. It's full definition... |
On a 64-bit system what is the maximum RAM that the Solver can utilize? |
<br>For the 64-bit CFX-10 Solver running on a 64-bit operating system the upper limit in single precision is 8GB RAM per stack. The Solver uses 5 stacks - Real, Integer, Character, Logical... |
Q) I am using the Boussinesq approximation and although the fluid heats up the density does not change.
I have set the thermal expansivity to 0.00011. Do you have any suggestions as to what I should do?
A) When you use Bossinesq, the density change due to thermal expansion is just put in as source terms. i.e. the actual density does not change.
If you want to see the true density change then you can either:
1) Run it as you have and then create a variable in CFX-Post to be the true density and add the density change as an expression.
2) or given the relation that the density is a function of thermal expansion coefficient and delta T, put this relationship in for the density. To do this, use an expression remembering to not toggle the 'density depends on' box. |
Q) I am using the Boussinesq approximation and although the fluid heats up the density does not change.<br>I have set the thermal expansivity to 0.00011. Do you have any suggestions as to... |