My geometry was created in mm and my density is given in g/cm^3. How do I convert density [g/cm^3] units to match geometry [mm] units? |
If the geometry is modeled in millimeters [mm]. then the density should be in [metric ton / mm^3]. Here are three sets of consistent units:<br>Length (m), Mass (kg), Density (kg/m^3),... |
Hydrostatic pressure for Hyper74 is different from that for Plane183. (Both models use the same neo-Hookean hyperelastic material model, identical mesh, and identical loading.) Other results such as SEQV compare well, however. Why is HPRES different? |
The HPRES values reported for Plane183 are approximately equal in magnitude but opposite in sign to those reported for Hyper74. This difference is due to the use of a different sign convention... |
I want to make use of User Fortran with the CFX-5 Solver on a Windows platform. Can I use the Salford Fortran compiler for this? |
The only supported Fortran compiler on Windows platforms for CFX-5 is the Compaq Visual Fortran compiler. For CFX-5.6 and CFX-5.7 the version of Compaq Visual Fortran required is version 6.5 or... |
My HYPER56 model matches my PLANE182 model for principal stress directions,
but the principal strain directions are different (and seem wrong). Is this a bug? |
Yes, Defect 22500 was filed and corrected in the BETAUP20040609 build of 9.0.<br> The input below demonstrates the error in the HYPER56 principal strain... |
When installing ANSYS 8.1 from the cd, some users have reported the following error:
An error occurred during the move data process -2147024873
Component: CommonFiles-HELP-ENUS
File Group: CommonFiles-HELP-ENUS
File: D:data2.cab
What is causing this error? |
-2147024873 Data error (cyclic redundancy check). The CD Rom is being read incorrectly. Try copying all the files to the local hard drive first, then run the setup off the hard drive. If it is... |
Are Structural Rotations and Rotational Displacements in Radians or Degrees? |
Both are in Radians.<br><br>Structural Rotations:<br>See Commands Reference> D Commands> D command in Notes: "Structural rotations should be input in... |
On a Windows 2000 machine (Service Pack 4) CFX-5.5.1 and CFX-4 run without problems. However with CFX-5.7 the CFX Solver runs but neither CFX-Pre nor CFX-Post can be launched. In both cases a Windows appears with title : PreGUI_ogl.exe and containing the message : Application Error.
The graphics card in the machine is an ATI Fire GL4 Videoaccelerator.
The processor type is IBM RC1000 256-bit / IBM GT1000 |
<br>Successful tests have been carried out with the ATI FireGL X2, ATI FireGL TL128 and ATI Rage Mobility cards. The Fire GL4 card has not been tested yet.<br><br>Things to... |
FE Modeler is bringing in the original Simulation mesh, not the refined mesh
automatically created (due to an inserted Convergence Tool with the Maximum
Refinement Loop greater than zero). Why? |
We are aware of this current limitation in FE Modeler and hope to be able to<br> bring in the refined mesh (i.e., the automatically refined mesh shown with the<br> stress... |
A customer reads .sub files to access reduced matrices, load vectors and DOF definition. When reading .sub files generated by ANSYS 8.1, they get incorrect data. Was the .sub file structure changed? Can you provide details on the new .sub file structure ? |
I don't think there have been any changes to the .sub file that are not documented. Documentation for the .sub file is contained in the FDSUB.INC file that can be found in the... |
Q1) How can I plot a nodal solution of safety factors for different materials having different yield strengths? The element table values are centroidal values, so these are not accurate enough for my situation. The SALLOW and SFACT and related commands do not allow for multiple materials having different yield strengths.
A1) One way to accomplish this is to create a macro, get stresses, manipulate values, then put stresses back into the database. Please note that Full graphics (/GRAPH,FULL) is required, and plotting/listing is only available immediately afterwards - for example, selection commands will reset these values to the original stress values.
The following input file demonstrates the aforementioned technique, and vector commands are taken advantage of for efficiency reasons.
!----------------------------------------
! Put safety factors for different materials
!
! Two cantilever plates subject to pressure loading
! Assume both plates have different yield strength
! Plot factors of safety for both
!
! Assume 1 plate has 10e3, another 20e3
!
! Use *VPUT to modify SEQV values (could use DNSOL)
! Need to use Full graphics to display
!----------------------------------------
finish
/clear
!----------------------------------------
! Graphics settings
! /GRAPH,FULL is important
!----------------------------------------
/graph,full
/vup,1,z
/view,1,1,2,3
/auto
/triad,off
/plopt,minm,off
/prep7
et,1,63
r,1,0.1
mp,ex,1,10e6
mp,nuxy,1,0.3
rectng,,10,,10
rectng,20,30,,10
esize,,5
amesh,all
d,all,uz
lsel,s,loc,y,0
nsll,s,1
d,all,uy
lsel,s,loc,x,0
lsel,a,loc,x,20
nsll,s,1
d,all,ux
lsel,s,loc,y,10
nsll,s,1
sf,all,pres,1000
allsel,all
/solu
solve
finish
/post1
set,last
shell,top
/title,Original SEQV distribution
plnsol,s,eqv
/wait,3
*get,AR20,node,,num,max
*dim,STI_MASK1,array,AR20
*dim,STI_MASK2,array,AR20
*dim,STI_STRSS,array,AR20
!----------------------------------------
! Get masking vectors
!---------------------------------------- |
Q1) How can I plot a nodal solution of safety factors for different materials having different yield strengths? The element table values are centroidal values, so these are not accurate enough... |
Are there any plans for Workbench port to AMD Opteron 64-bit for Windows? |
No, there are no plans for Opteron support at 9.0.
|
Are there plans to support Workbench on the Linux operating system? |
Yes, we plan to port Workbench 9.0 to the Linux platform, which is due out at the end of 2004.<br>No details of operating system version/make are known at this time.
|
Can you provide me with a concise summary of the features of your high frequency electromagnetic solver? |
Below is quick summary of the requested information.<br><br>1) 4 classes of problems<br> a) propagation (of waves along guided structures - waveguides, transmission... |
Q) I run a large LS-DYNA model and get the following error from the solver:
*** Error - Input error found -
*** Error reading 20th control card
A) The "20th control card" is based on structured rather than keyword input. If the following line is added:
*CONTROL_STRUCTURED_TERM
in the keyword file, that will output the structured input file as "dyna.str" then terminate the LS-DYNA solver run.
From the resulting "dyna.str" structured file, you may see the following lines:
$ control card 20 TAURUS DATABASE CONTROL I
$ pltc frcint prtc ndth nsth nstb nsts nstt lc1 lc2 lc3 sst
1.500E-06 0.000E+00 2.143E-07********** 0 0 0 0 0 0 0
Per the LS-DYNA Sturctured Manual, the "ndth" and "nsth" are numbers of nodal and hex time history blocks. The number of requested nodes/elements is too large (**** instead of numbers being displayed). Reducing the number of nodes/elements for the output request, then rerun the solution. |
Q) I run a large LS-DYNA model and get the following error from the solver:<br>*** Error - Input error found -<br>*** Error reading 20th control card<br><br>A) The... |
The RSYS command seems to have no effect on the older element
formulations in a large deformation (NLGEOM,ON) analysis. Why? |
This behavior is documented in Section 5.5.1 "Rotating Results<br> to a Different Coordinate System" in the ANSYS 8.1 Basic Analysis<br> Procedures Guide with the... |
How can I correct an axis-symmetric model that was
incorrectly built based on rotation about the global x-axis? |
ANSYS uses the global y-axis as the axis of rotation.<br>To transfer the areas and mesh issue the commands:<br><br>! if there already is a csys of 11, use a higher... |
Q) In 8.1, under "Utility Menu > MenuCtrls > Font Selection...", users may change the fonts used in the ANSYS GUI. However, where are these settings stored on Windows and UNIX?
A) Regarding font settings, these are saved in the following location:
Windows: Windows registry [HKEY_USERSxxxxxSoftwareAnsys, Inc.ANSYSANSYS 8.1GUIFontsuserFonts]
UNIX: Similar location within ansys8.1.reg file found in ~/.ansys/ansys8.1 |
Q) In 8.1, under "Utility Menu > MenuCtrls > Font Selection...", users may change the fonts used in the ANSYS GUI. However, where are these settings stored on Windows... |
I get the error shown below when trying to plot total strain
for a BEAM189 model. What is causing this?
DISPLAY NODAL SOLUTION, ITEM=EPTO COMP=EQV
*** FATAL *** CP= 599.891 TIME= 14:22:44
Programming error in pagend.F ... Dtemp array overflow
elem=1 idata=2 record_length=8640 array_size=5400.
ALL CURRENT ANSYS DATA WRITTEN TO FILE NAME= ansabort.db
FOR POSSIBLE RESUME FROM THIS POINT |
Defect 24725 was filed and then corrected in the UP20040601 build<br> of ANSYS 9.0. The error was caused by an array overflow in POST1.<br> The statically dimensioned array... |
Why am I having a problem when I use triangular face elements in the mesh for which I want to use the 'AFSURF' macro. The logic works fine when I have a quad mesh (see duct-heat.inp) but when I provide a regular shaped triangular mesh on the surface of the duct, I have a problem. The message refers to changing elements attached to an area, which makes no sense. Why is this happening? |
The AFSURF macro had to try and change the existing MESH200 3 noded triangles to create the SURF152 which a fourth (duplicate of third) node. But the element modify fails. The macro will work... |
**** Entered By: jthompso @ 06/22/2004 02:03 PM ****
Question:
We use ANSYS 8.1 Classic with UPF to get information of the result file. The database access routines as described in
the UPF Guide do not provide enough information to get all the data we need so we are accessing the results file from within the ANSYS program. The headers for the result-file accessing routines (biniqr, binset, binBasePut, binrd and binclo) state that these routines are not intended to be linked with the ANSYS program but only as standalone program. We understand that in the ANSYS program you are using the routines such as bioiqr, bioset, bioBasePut, biord and bioclo for internal processing.
Unfortunately, some of "bio"-routines use less or other arguments than the "bin"-routines. We kindly ask you to provide information about the arguments of the "bio"-routines.
Answer:
As you note in the incident notes, the BIN* routines are not to be linked into the ANSYS code. The reason for this is that the BIN routine use BIOCOM (rather than their own BINCOM) as the common for the binary file data so that a user could be walking on BIO file data with the BIN file calls. The BIN routines are similar to the BIO routines except that ANSYS will provide the scratch memory for the BIO routines whereas the user must supply the scratchmemory for the BIN routines. Other than that they function the same. The documentation in the 8.1 Programmer's manual appears to be adequate to understand the arguments.
Attached is a text file that contains the headers for the BIO routines that are of use to a user programmer.
One other bit of information; the ANSYS rst file is always attached to buffer 6. If you try to write a file on buffer 6, ANSYS may commandeer that buffer for the rst file; overwriting anything that you have there. All other file are attached to the next open buffer so it shouldn't be a problem. I generally use buffer 9 as it is the least likely buffer to be used by A |
**** Entered By: jthompso @ 06/22/2004 02:03 PM ****<br><br>Question:<br><br>We use ANSYS 8.1 Classic with UPF to get information of the result file. The database access... |
When using SHELL181, if the user specifies KEYOPT(8) =2 (storage of layer Data = All layers + Middle), PLNS,S,X gives the correct results at the mid-plane, but PLES,S,X does not. It just plots the same results as TOP. Same is true of SHELL63 with KEYOPT(11)=2. |
This is a code defect dating back to ANSYS 7.1. It has been corrected for the ANSYS 9.0 release.
|
With the simple free meshing of an area as shown by the input below,
ANSYS 8.1 aborts with a division by zero error:
fini
/clear
/plopts,info,1
/prep7
k,1, 0.00, -0.232, 0
k,2,-0.632E-01, -0.232, 0
k,3,-0.892E-01, -0.387, 0
k,4,-0.121 , -0.332, 0
k,5,-0.576E-01, -0.441, 0
k,6, 0.892E-01, -0.387, 0
k,7, 0.576E-01, -0.441, 0
k,8, 0.121 , -0.332, 0
k,9, 0.632E-01, -0.232, 0
l,2,1
l,1,9
l,9,8
l,8,6
l,6,7
l,7,5
l,5,3
l,3,4
l,4,2
al,1,2,3,4,5,6,7,8,9
et,1,shell63
lesize,all,,,2
amesh,all
*** FATAL ***
Divide by Zero in FEM_utNorm.
ALL CURRENT ANSYS DATA WRITTEN TO FILE NAME=
ansabort.db
FOR POSSIBLE RESUME FROM THIS POINT
Is this a known bug? The same input works fine in ANSYS 8.0. |
Yes, Defects 22623, 22725, and 22750 were filed (independently with<br> similar meshing tasks) and subsequently fixed in ANSYS 9.0 Preview 1.<br> The fix has also been included... |
How does Workbench handle shell results? Are they averaged, or unaveraged, at boundaries between:
* material properties
* shell thicknesses
* geometric discontinuities (shell planes intersecting at an angle) |
Workbench does not average shell results across material or shell-thickness boundaries, but it does average across geometric boundaries.
|
QUESTION: We are inputting the datapoints for B-H curve for Emag Analysis for Magnetisation curve (with H ranging from 0 to 150) and Demagnetisation curves (H coming down from 150 to 0). The typical values of B-H are given below for both the curves, wherein there are duplicate points as there are identical H values in both of these curves. So, when the below macro is run, the new values of B in the Demagnetisation curve are overwritten with the B values in the Magnetisation curve. Is there any way to avoid overwriting and to consider both the curves?
! Magnetisation curve
TB,BH,1 ! B-H CURVE FOR MATERIAL 1
TBPT,,0,0
TBPT,,20,.1 ! H AND B RESPECTIVELY
TBPT,,25,.125
TBPT,,50,.250
TBPT,,75,.35
TBPT,,100,.375
TBPT,,150,.425
!Demagnetisation curve
TBPT,,100,.4
TBPT,,75,.375
TBPT,,50,.35
TBPT,,25,.275
TBPT,,0,.15TBPLOT,BH,1
ANSWER:
ANSYS does not account for hysteresis in electromagnetic analyses. There is no way I know of to define two sets of BH data (one for magnetization, the other for demagnetization) for a given material.
You might be able to assign unique material property numbers to each individual iron element and monitor the field in a series of restarts performed in a do loop. When the magnitude of the field begins to decrease, switch to a demagnetization curve. Note that you will need a unique demagnetization curve for each element, because each element in your model will reach a unique level of saturation.
An easier way would be to use manufacturer's data for coreloss. Below is a response for a similar inquiry:
There are 2 components of coreloss that I'm aware of:
1) joule heating (losses due to induced eddy current)
2) hysteresis
We can calculate #1 directly, by including the VOLT degree of freedom
and defining electrical resistivity (RSVX) in eitherPLANE13/53 (2D) or
SOLID97/117 (3D) elements. We cannot directly calculate hysteresis loss
- the power associated with the area enclo |
QUESTION: We are inputting the datapoints for B-H curve for Emag Analysis for Magnetisation curve (with H ranging from 0 to 150) and Demagnetisation curves (H coming down from 150 to 0). The... |
Where in the menus do we find commands for plotting failure criteria? I think they have been moved from 8.0 to 8.1 |
There was a bug in the ANSYS 8.1 GUI that omitted the composite failure plot commands from the contour-plot menus. This bug has been fixed in 9.0.
|