SiteMap Page 129

Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 Next


I am performing a thermal transient in Workbench. Convection is defined by using a command object (to define spatially varying convection with SF using tabular loading).
The calculation is correct - but Workbench does not report the mid node results at this boundary, leading to alternating solution: 0 - result - 0 - result as shown in the attached plot.

Postprocessing in classic works well.

Steady state works well in Workbench.

Why does this happen only in WB and only for a thermal transient?


In ANSYS models having SOLID90s using the QUASI solver option, we suppress the mid nodes on any face with applied convections. That is because the QUASI solver uses only a diagonalized convection matrix which does not work well with 8-noded faces (only the consistent formulation does), hence we drop down to 4 nodes.

From the Element Manual for 9.0:

For transient solutions using the THOPT,QUASI option, the program removes the midside nodes from any face with a convection load. A temperature solution is not available for them. Do not use the midside nodes on these faces in constraint equations or with contact. If you use these faces for those situations, remove the midside nodes first. For the time being, the only workaround I can think of is to postprocess in ANSYS.

Note 1: supported WB convection does not have this issue because we apply convection via 152 elements which don't show the problem. If you don't issue the "sf,all,conv,%conve%,22" the results are ok.

Note 2: Feature request 49915 pending as of 10/20/2006

QUESTION:<br><br>I am performing a thermal transient in Workbench. Convection is defined by using a command object (to define spatially varying convection with SF using tabular...

Customer gets the following error when trying to start ANSYS 10.0 on EM64T RED HAT Linux AS 4 update 4 system:

/apps/ansys/ansys_inc/v100/ansys/bin/linem64t/ansys.e100:error while loading shared open shared object file:No such file or directory

Why is he getting this error?

We provide two copies of "" for...

Customer received the following error while trying to install Workbench 10.0 SP2 and WB 11.0 preview.

"error (-5009: 0x8002801d)has occurred while running the setup"

Is there a workaround for this error?

Try the following:<br><br>Rename "C:Program FilesCommon FilesInstallshieldProfessionalRunTime" to RunTime.orig<br><br>Rename "C:Program Files...

Class 3 Error

When Results and Definition files contain different Solution Units, the resulting interpolation is not correct. In some circumstances, the interpolator complains that the meshes do not overlap (when they should), and in others the interpolated field values are scaled by a factor related to the difference in Solution Units.

The problem is that the interpolator cannot read Solution Units from a Definition File (only a Results File), and in this case assumes they are identical in both files (i.e. the raw values are interpolated without appropriate scaling). The resulting behaviour can be mesh and Solution Unit dependent. For example:

- differences in the Length unit may result in apparent partial or non-overlap of meshes (the latter resulting in failure of interpolation).
- differences in any Solution Unit will result in an apparent scaling of the interpolated field values.

For steady state calculations where the interpolation is successful, the solver may still run to convergence. However, for transient calculations the final solution is dependent on the initial conditions and will be incorrect.

Note that when comparisons of two Results files is performed using the Calculate Differences option, the resulting Difference variables

WORKAROUND:<br>No workaround is currently available.<br><br>FIXED IN:<br>Release 11.0.

Under certain circumstances, when injecting particles using the 'Discrete Diameter Distribution' option, an incorrect number of tracks are injected for the specified size bands.

For example, in a particular case with 10 size bands, if a total of 200 tracks are started, then as expected there are 20 tracks per size band. However, if 400 tracks are started, there are 39 tracks in each of the first 9 size bands and 49 tracks in the final size band.

This is due to a rounding error in the particle diameter distribution routines within the CFX Solver.

WORKAROUND:<br>No workaround is currently available.<br><br>FIXED IN:<br>Release 11.0.

Class 3 Error

Some CEL expressions that use the exponentiation operator (^) do not appear to be evaluated correctly. For example, the CEL expression

2^((1 [m] + 1 [m]) / 1 [cm])

is evaluated as 4, whereas its correct value is 2^200.

This is due to a bug in the CEL expression evaluator for expressions that use the exponentiation operator (^) and involve mixtures of different but dimensionally consistent units (for example [m] and [cm]).

This error occurs when all of the following are true:

- the exponent is a constant expression.
- ignoring any units, all terms in the expression are integer constants
- ignoring any units, the expression evaluates to an integer (e.g. in the above case the expression (1 + 1)/1 = 2)
- the expression would evaluate to a different value if units were taken into account

For example, the following expressions will evaluate correctly:

2^(3 [m/m]) --> effectively no units in this case
2^((1 + 2) /3) --> no units involved, exponent resolves to 1
2^((1 + 1) /4) --> no units and exponent resolves to non-integer value
2^((1 [m] + 1 [m]) /2[m]) --> same units used throughout
2^((1 [m] + 1 [m]) /4[cm]) --> exponent does not resolve to integer if units ignored

Note that irrespective of this problem, an exponent expression must be internally dimensionally consistent and must resolve to a dimensionless value. An expression that does not satisfy these requirements will generate an error whe

WORKAROUND:<br>Modify the expression in line with the above recommendations to avoid the error.<br><br>FIXED IN:<br>Release 11.0.

Class 3 Error

For calculations that involve rotating domains, exported results from the Solver Manager show the mesh to be rotated correctly but not the velocity field ('Velocity' and 'Velocity in Stn Frame').

This is due to a bug in the mesh export library and affects export to

WORKAROUND:<br>If the 'Use Initial Mesh for Rotating Domains' option is enabled, the velocity vectors are correctly located on the initial mesh. Otherwise, no workaround is...

Class 3 Error

When a quantitiative calculation using the variable 'Heat Flux' is performed in CFX Post for a simulation including radiation, the result does not match the same calculation performed using the variable 'Wall Heat Flux'.

This is due to a bug in CFX Post. The variable 'Heat Flux' does not include the radiative flux component, whereas its' hybridized counterpart 'Wall Heat Flux' does. Since for quantitiative calculations in CFX Post it is recommended that the variable 'Heat Flux' is used (to avoid hybridization effects), the difference for many simulations (e.g. combustion)

WORKAROUND:<br>No workaround is currently available.<br><br>FIXED IN:<br>ANSYS CFX Release 11.0.

Customer gets segmentation fault when trying to start WB using the ./startpage command.
He goes to the /ansys_inc/v100/aisol directory and types ./startpage and WB then crashes with the following error:

./startpage line 9: 17310 Segmentation fault
/ansys_inc/v100/aisol/workbench -Startpage $*

Why is he getting this error?

We do not support SuSe 10 or 10.1 in WB 10 and WB 11.0.<br>We only support SuSe 9.

How can I get bcd file successfully from TRASYS run in TAS?

After you finish defining the problem properly for TRASYS inside TAS, click File>Export>To TRASYS, make sure that you check the box for 'Run TRASYS' while exporting the...

What license feature do I need to import parasolid files into DM and/or Simulation?

DM uses the parasolid kernel, therefore no license is need to import parasolid into DM.<br>To be able to import parasolid into Simulation directly, you can use either rdpara, or aipips or...

Why could the torque and power calculations be so different when using different method to compute the values?

They should be the same. There are two ways of calculating torque.<br><br>-one is the force acting on all the rotating walls multiplied by its lever arm which is the method used to...

**** Entered By: rlange @ 10/27/2006 02:08 PM ****

Q. I am doing a CFX analysis with adaptive time step control activated. I provided CEL table input to control Pressure vs. time from t=0 to the specified end time of the case. Why does CFX give me this error message: Failure to calculate variable using Expression Language?

A. With adaptive time step control and other cases, the last time step chosen by CFX might take the case beyond the specified end time. If that happens and the user did not specify Extend Max for an interpolation function (table) that uses time as a variable, then the solver produced the above error because it cannot evaluate the pressure at the end time, which would be larger then the user specified time. This behavior is well documented.
Therefore it is a good idea to extend the time range of the CEL interpolation function beyond the specified end time or turn ON Extend Min and Extend Max of the interpolation function.

**** Entered By: rlange @ 10/27/2006 02:08 PM ****<br><br>Q. I am doing a CFX analysis with adaptive time step control activated. I provided CEL table input to control Pressure vs....

Why does the "Animate" program close on some machines in ANSYS 10.0 when opening a .avi file?

We have seen this and it may be an issue with ¿Animate¿ using the DSGStreamN.dll.<br><br>Development is doing a new build of ¿Animate¿ for the 11.0 release...

The solver fails and prints the following message to the output file:

****** FATAL ERROR ******
The orthographic view transformation failed on domain interface "GGI".
Failure may be due to r=0 included in transformed cylindrical coordinates of an interface
with rotational relative motion.

<br><br>Make sure that the axis of rotation is specified correctly.<br><br>As the message indicates, the solver cannot solve on a frame change interface where the radius...

My analysis has 3000 load steps. How can I use the RESCON command
to have the maximum number of restart points?

Per the MAXFILES field documentation of the RESCONTROL command, ¿The<br>total maximum number of .Rnnn files for one run is 999.¿<br>If you want to be able to restart from...

Originally I have 25 properties, and I add two more properties for my application. By accident it turns up to over 6000 properties, most of them are duplicated from the original properties. If I use Delete function, it can only delete one property at one operation. How can I easily delete those duplicated properties?

Finish your problem settings in TAS first, reference those useful properties to related elements. Then click Information>Property to bring up Properties Table window, click...

When a UNC path is used to locate a file or directory in CFX-Pre on Windows (e.g. //mycomputer/CFXFiles/mesh.gtm), it does not work. Typically, CFX Pre will generate an error saying that the specified file cannot be read.

There are some areas of CFX that do not currently support UNC paths on Windows. Examples of this are:

- Mesh Import dialog in CFX Pre.
- Profile Data dialog in CFX Pre.
- Solver temporary working directory

but there may also be several others.

WORKAROUND:<br>No workaround is currently available, other than to specify a mounted drive (e.g. D:/CFXFiles/mesh.gtm).<br><br>FIXED IN:<br>No fix is currently available.

Class 3 Error

If a compressible transient flow is undertaken with only one iteration per time step, then the solution can be incorrect if the Heat Transfer option is not set to Total Energy, or if heat transfer is not included in the simulation.

This is due to the Solver not extrapolating the pressure at the start of the time step in these circumstances. This means that density is not extrapolated, and so the code cannot calculate an accurate value for drho/dt on the fir

WORKAROUND:<br>Either run with at least two iterations per time step, or use the Total Energy option for Heat Transfer.<br><br>FIXED IN:<br>Not fixed yet.

When running the CFX solver on 32 bit Windows, it is possible for the address space to become fragmented. Since the solver requires contiguous address space, this can limit the amount of memory available to the solver. This problem is described in more detail, along with some workarounds in the following solution:

Will this problem still occur on 64 bit Windows (Win64)?

<br><br>In summary, using 64 bit Windows will solve this problem, but unfortunately not using CFX-10.0.<br><br>64 bit Windows has the advantage that much more memory can be...

CFX Mesh creates no 1:1 mesh for a combination of several parts.

Mesh is only 1:1 for the domains within a part. Domains which are not in the same part are meshed independently. Therefore join all solids to a new part in Design Modeller. Choose the parts -...

Q1: For ANSYS 10.0 and 11.0 on 32-bit Windows, the Intel Fortran 8.1 compiler and Microsoft Visual C++ .NET 2003 are required to use User-Programmable Features. However, Intel Fortran 9.1 and Microsoft Visual Studio 2005 are the current versions being sold.

Q2: During linking using Intel Fortran 9.1 and MS VS 2005, I get the following error:
unresolved external symbol ___powr8i4 referenced

A1: It seems that a customer can only purchase MS VS 2005. For Intel Fortran 9.1, the user can, however, download an earlier version from the Intel Download Center:
To use the Intel Download Center:
- login using their Intel Premier credentials
- click on Product Downloads
- click on Intel Visual Fortran Compiler, Standard Edition, for Windows
- set the version pull-down to 8.1
(- set the minor version pull-down to 042)

Note that if the user uses MS VS 2005 and Intel Fortran 8.1, the latter is not integrated in MS VS 2005 (Intel Fortran 8.1 is only integrated under MS .NET 2003). Hence, in this case, the 8.1 compilers would need to be run via the DOS command-line.

A2: This is most likely due to differences in math functions between Intel Fortran 8.1 and 9.1. (The error is related to using the power function, such as X**3.) Per above, it is recommended to use Intel Fortran 8.1, as that is the supported version for using UPFs in ANSYS 10.0 and 11.0 on 32-bit Windows.

Q1: For ANSYS 10.0 and 11.0 on 32-bit Windows, the Intel Fortran 8.1 compiler and Microsoft Visual C++ .NET 2003 are required to use User-Programmable Features. However, Intel Fortran 9.1 and...

What is the difference between SOLID45 and SOLID185?

Regarding the differences between SOLID45 and SOLID185, the main difference is element formulation. As you may know, we encounter 2 main 'locking' type of problems with lower-order...

A program crashed. But still the license is blocked by the server. How can I remove license from license service?

Get the license status (ANSLIC Admin Utility or command line: /home/cfx/ansys_inc/shared_files/licensing/linop64/lmutil lmstat -a -c 1055@server) returns the feature list,<br>e.g. for the...

How can you interpolate results onto a mesh without the CFX 11.0 solver checking the node locations?

If the structure is the same (same topology), you can set the following environment variable:<br>interp_nocrdvx_check = 1<br><br>to perform your interpolation without checking node...