- Software
- Zencrack Support
- Bug List
All Issues
Bug List
All Issues
ZENCR555
Applies to: Version 9.3-1.
Fixed in development version: 24 Apr 2024.
Fix included in release version: 9.4-1.
In a remeshing model that uses the crack plane retriangulation option, the remeshing process fails in version 9.3-1 due to an incorrect surface reference resulting from changes to allow multiple crack in a remeshing analysis in version 9.3-1.
Workaround: None.
ZENCR554
Applies to: Version 8.1-1 to 9.3-1.
Fixed in development version: 11 Apr 2024.
Fix included in release version: 9.4-1.
In an Ansys model using large crack-blocks, the surfaces to tie the crack-blocks to the surrounding mesh may be defined using explicit surface elements or ESURF options. If explicit surface elements are used and transition elements are requested between the crack-blocks and the surrounding mesh, the transition elements may be placed on the wrong crack-block faces.
Workaround: Use ESURF surface definitions to tie the crack-blocks to the surrounding mesh.
ZENCR553
Applies to: Version 9.0-1 to 9.3-1.
Fixed in development version: 3 Apr 2024.
Fix included in release version: 9.4-1.
When a Superposition load system is used with method Combination (to combine several time history load systems), the direction of crack growth is reversed by 180 degrees if the crack growth is driven by stress intensity factors. Also, it is not documented clearly in the User Manual section "Combination method for multiple load spectra" that the stress intensity factors used with this load system are always displacement-based values i.e. SIF-integral stress intensity factors are not used even if this parameter is requested to drive crack growth calculations.
Workaround: Use J-integral results to drive the growth calculations if a Superposition load system is used with method Combination.
ZENCR552
Applies to: Version 9.3-1 (and earlier) (Ansys interface).
Fixed in development version: 28 Mar 2024.
Fix included in release version: 9.4-1.
If a model using the Ansys interface contains large crack-blocks with transition elements enabled and there are also pressure or film coefficient surface definitions on the crack-blocks, the update of those surfaces will be incorrect and the Ansys run of the cracked mesh will fail.
Workaround: Do not use the transition element option on any crack fronts in a large crack-block model if there are surface loads on the target crack-block elements.
ZENCR551
Applies to: Version 9.3-1.
Fixed in development version: 14 Feb 2024.
Fix included in release version: 9.4-1.
Any GUI .ini file [COLOURS] entries that set a single colour value are not applied (e.g. MESH_BG_1). Instead, the default colour is used.
Workaround: For some items it is possible to set the colour once the GUI is running via Options > Settings > Colours.
ZENCR550
Applies to: Version 9.1-1 to 9.3-1 (Ansys interface).
Fixed in development version: 12 Jan 2024.
Fix included in release version: 9.4-1.
In the remeshing method, when external surfaces of the model are updated in the region of the end planes of the rings there may be some ring element faces that receive an update from being associated with two underlying faces of the uncracked mesh. Abaqus is able to handle this and there are no issues arising from the update adding the same underlying face to the surface twice. Ansys has problems because the update method creates two separate overlaid faces in the new surface:
- If the surface is used for pressure load, element type 154, there will be a small amount of additional load due to the duplicate face. This does not show in a contour plot of pressure in APDL because each face has the same (correct) pressure.
- If the surface is used for contact, including bonded contact, there may be an apparent
unrelated error from Ansys about element normals (the element numbers are the two
overlaid surface elements). For example:
*** ERROR *** CP = 1061.188 TIME= 12:51:15 The normal of target element 170638 is not consistent with the normal of target element 171365 in real set 10. There may be other elements not consistent in normal direction. Please use the ENORM command to correct it.
If surfaces are defined on the exterior of the model but are not used for any specific purpose, a duplicate surface update does not affect the analysis.
Workaround: None.
ZENCR549
Applies to: Versions up to and including 9.3-1 (Abaqus and Ansys interfaces).
Fixed in development version: 13 Dec 2023.
Fix included in release version: 9.4-1.
The execution of Zencrack for an Abaqus or Ansys run checks for and removes duplicate options from the f.e. command line:
- For Abaqus the "cpu=" and "cpus=" strings are checked.
- For Ansys the "-np" and "-p" strings are checked.
However, if these strings exist multiple times anywhere in the command line, including in filenames, then all but the final one are removed causing the job to fail.
Workaround: Do not include strings -p or -np in Ansys jobnames. Do not include strings cpu= or cpus= in Abaqus jobnames.
ZENCR548
Applies to: Version 9.0-1 to 9.2-2.
Fixed in development version: 11 Aug 2023.
Fix included in release version: 9.3-1.
There is a possible array bounds issue in a remeshing analysis with a high number of external faces on the remeshed tet region. This array bounds issue can occur in versions 9.0-1 to 9.2-2 if the number of external faces on the remeshed (tet) region exceeds 50000. No warning is given and the cracked mesh may contain invalid data if this occurs (e.g. element face updates for pressure, film coefficients and surfaces may be incorrect).
In version 9.3-1 the limit is increased to 200000 faces and an error message is generated if this is exceeded.
Workaround: Reduce the physical size and/or mesh density of the remesh region.
ZENCR547
Applies to: Version 9.2-2 (and earlier) (Abaqus interface).
Fixed in development version: 14 Jul 2023.
Fix included in release version: 9.3-1.
An error message may be produced if a connectivity line in the uncracked mesh has a trailing comma for an element type that is not one of those specifically processed by Zencrack. The error message references the next dataline in the uncracked mesh. For example:
*ELEMENT, TYPE=CONN3D2, ELSET=BUSH_A 23601, 8964, *CONNECTOR SECTION, ELSET=BUSH_A, BEHAVIOR=BUSH_A_BEHAVIOR ***ERROR 3005 Unrecognizable data in the following record reading *ELEMENT data from the uncracked mesh input file: *CONNECTOR SECTION, ELSET=BUSH_A, BEHAVIOR=BUSH_A_BEHAVIOR
Workaround: Remove such trailing commas on the end of affected connectivity datalines.
ZENCR546
Applies to: Version 9.0-1 to 9.2-1 (Abaqus converter).
Fixed in development version: 13 Jul 2023.
Fix included in release version: 9.3-1 (Abaqus converter version 9.3).
There is a possible array bounds issue when converting an Abaqus mesh if an element set used on a *SOLID SECTION keyword has multiple dataline headers to create the set and one (or more) of those headers has no datalines.
This may result in an error when running a Zencrack analysis or importing a model into the GUI.
Workaround: Remove any *ELSET options in the uncracked mesh that have no datalines.
ZENCR545
Applies to: Version 9.2-2 (and earlier).
Fixed in development version: 11 Jul 2023.
Fix included in release version: 9.3-1.
If a set of nodal temperatures contains multiple (different) temperature entries for a particular node, the wrong value is assigned for that node when processing temperature updates. Instead of using the last temperature value found for the node, the first value is used.
Workaround: Remove duplicate node entries withing a block of temperature data.
ZENCR544
Applies to: Version 9.2-2 (and earlier).
Fixed in development version: 11 Jul 2023.
Fix included in release version: 9.3-1.
There is a possible array bounds issue when processing temperature data if a set of nodal temperature data contains more entries than there are nodes in the uncracked mesh (i.e there are duplicate entries or entries for nodes that do not exist in the mesh).
Workaround: Remove any extra temperature definitions.
ZENCR543
Applies to: Versions 9.1-1, 9.2-1 (Linux installer).
Fixed in development version: 3 Apr 2023.
Fix included in release version: 9.2-2.
If the Linux installer runs in command line mode rather than GUI mode, the installer copies files to the chosen directory but does not run the setup program during the installation to configure those files nor does it make entries in system files to enable use of a runzcr911 or runzcr921 shortcut.
Workaround: See section 4.1 of the Installation And Execution Manual for information on the Zencrack setup program and zencrack.ini file. The [SETUP] section of the zencrack.ini file needs to be changed manually - see section 4.1.10 of the Installation And Execution Manual.
ZENCR542
Applies to: Versions 9.01 and 9.02 of utility PROCESS (included as part of Zencrack 9.0-1, 9.1-1, 9.2-1).
Fixed in development version: 19 Apr 2023.
Fix included in release version: 9.2-2 (PROCESS version 9.03).
For models with linear crack front elements the extraction, by process.exe, of the maximum number of crack front nodes on a single crack front may be incorrect for crack-block models with more than one crack front. There is order dependency - if the crack front with the largest number of crack front nodes is defined first the extraction is correct. The consequence of this issue is internal array overwrite of data and unpredictable termination of the Process execution. process.exe is used when extracting data for xy plotting in the GUI so this issue may affect creation of xy plot data. For example, for "along crack front" plot data extraction there is an error message in the GUI "Failed to extract data for list of crack fronts".
Workaround: None.
ZENCR541
Applies to: Versions 9.0-1, 9.1-1 (Ansys interface).
Fixed in development version: 16 Mar 2023.
Fix included in release version: 9.2-1 (Ansys converter version 9.2).
Identification and processing of the TEMP parameter on bfblock and PRES or CONV parameters on sfeblock are case sensitive in the Ansys mesh converter with only upper case versions TEMP, PRES and CONV being identified. Files output by Ansys APDL using CDWRITE contain upper case and are processed correctly. Files output by Ansys Workbench contain lower case are are not processed correctly.
Workaround: Edit the uncracked mesh and ensure the TEMP, PRES or CONV parameters on bfblock or sfeblock are upper case (the keywords themselves may be upper or lower case, only the parameters need to be changed to upper case).
ZENCR540
Applies to: Versions 9.0-1, 9.1-1.
Fixed in development version: 13 Dec 2022.
Fix included in release version: 9.2-1.
Remeshing a region of a mesh that contains collapsed brick elements fails with a traceback if the overall model has multiple disconnected regions.
Workaround: None.
ZENCR539
Applies to: Version 9.1-1 (Linux).
Fixed in development version: 17 Feb 2023.
Fix included in release version: 9.2-1.
There is an erroneous error message when attempting to run a Zencrack job from the GUI on Linux:
Zcr bat file /install_folder/Zencrack/9.1-1/tools/zcrpath911.bat is missing
This file is not part of a Linux installation - it is only present on Windows.
Workaround: Make an empty file zcrpath911.bat in the folder indicated in the message.
ZENCR538
Applies to: Versions 9.0-1, 9.1-1.
Fixed in development version: 8 Feb 2023.
Fix included in release version: 9.2-1.
Remeshing results from a previous step in a growth analysis may be used for the current remeshing step if the current remeshing process fails very early during the remeshing initialisation phase (e.g. due to a remeshing license error).
Workaround: None.
ZENCR537
Applies to: Version 9.1-1 (and earlier).
Fixed in development version: 14 Feb 2023.
Fix included in release version: 9.2-1.
There may be minor changes to j-integral output when running the same model with Abaqus 2022 (or earlier) and 2023.
In crack-block models where both sides of a crack front are modelled there were previously a small number of surface nodes with two *NORMAL definitions in the contour integral region. Abaqus would use only one of the definitions. Abaqus version 2023 treats multiple *NORMAL definitions at a node in the contour integral region differently than previous Abaqus versions if virtual crack extensions are out of the local crack plane. This results in very small numerical differences for some models between Abaqus 2023 and previous versions. These small numerical differences are not enough to change the overall global behaviour of a growth run. In version 9.2-1 all surface nodes in the contour integral region requiring a *NORMAL definition have a single definition giving consistency between Abaqus 2023 and previous versions.
Workaround: None.
ZENCR536
Applies to: Version 9.1-1 (and earlier).
Fixed in development version: 24 Jan 2023.
Fix included in release version: 9.2-1.
A traceback that would occur in the Ansys mesh converter if the uncracked mesh contained lines longer than 256 characters is fixed in version 9.2-1.
Workaround: If possible, shorten any line(s) longer than 256 characters or remove them.
ZENCR535
Applies to: Versions 8.0-1 to 9.1-1.
Fixed in development version: 24 Jan 2023.
Fix included in release version: 9.2-1.
If an input file contains multiple *THRESHOLD keywords only the data on the last *THRESHOLD keyword is retained.
Workaround: None.
ZENCR534
Applies to: Version 9.1-1 (and earlier).
Fixed in development version: 24 Jan 2023.
Fix included in release version: 9.2-1.
The deltaK and Kmax values written to the detailed integration (.bk) output files for fatigue analysis are always based on a conversion from G to K using an alpha conversion factor of 1. If the growth law uses a different factor (via the CONVERSION parameter on *CRACK GROWTH DATA) there is a small inconsistency between the reported deltaK and Kmax and the values used to calculate the da/dn for the result point. This is only a reporting issue in the .bk file and does not affect analysis calculations during the integration process.
Workaround: None.
ZENCR533
Applies to: Versions up to and including 9.0-1 (Ansys interface).
Fixed in development version: 3 Oct 2022.
Fix included in release version: 9.1-1.
Processing of ANSYS SFE load types that are not ones considered for updates by Zencrack but that are defined across two datalines may result in those datalines being separated in the cracked mesh by the insertion of Zencrack updates. For example the first SLEV option in this set of loads may be split:
SFE, 4824, 1,PRES,1,R5.0 1.00000000 0.00000000 0.00000000 0.00000000 SFE, 4824, 1,PRES,2,R5.0 0.00000000 0.00000000 0.00000000 0.00000000 SFE, 4824, 1,SELV,1,R5.0 1.00000000 0.00000000 0.00000000 0.00000000 SFE, 4824, 1,SELV,2,R5.0 0.00000000 0.00000000 0.00000000 0.00000000
Workaround: A workaround is to ensure a comment line splits the non-updateable types from the previous block that has updates i.e.:
SFE, 4824, 1,PRES,1,R5.0 1.00000000 0.00000000 0.00000000 0.00000000 SFE, 4824, 1,PRES,2,R5.0 0.00000000 0.00000000 0.00000000 0.00000000 ! SFE, 4824, 1,SELV,1,R5.0 1.00000000 0.00000000 0.00000000 0.00000000 SFE, 4824, 1,SELV,2,R5.0 0.00000000 0.00000000 0.00000000 0.00000000
ZENCR532
Applies to: Versions up to and including 9.0-1 (Ansys interface).
Fixed in development version: 29 Apr 2022.
Fix included in release version: 9.1-1.
Execution of Ansys from the GUI may fail if Ansys is executed during mesh import or if a Zencrack analysis is initiated from the GUI. This behaviour varies with Ansys versions and may also be affected by existence or otherwise of certain dll files on the machine.
Workaround for import: In the zencrack_gui.ini file set RUNANSYS to NO. This reduces the chance of Ansys executing during mesh import. If the problem persists, remove solid model and ESURF options from the mesh file and try importing the modified file.
Workaround for Zencrack analysis: Execute the analysis from a Zencrack command prompt instead of from inside the GUI.
ZENCR531
Applies to: Versions up to and including 9.0-1.
Fixed in development version: 27 Sep 2022.
Fix included in release version: 9.1-1.
The information message in the .rep file that summarises the open or closed status of nodes on a crack face may be incorrect in some cases (depending upon the crack orientation). The underlying analysis is correct, only the ***INFORMATION message may be incorrect. Specifically, the detailed open or closed status associated with crack front nodes is correct and the calculation of stress intensity factors from displacements is also correct.
***INFORMATION
Displacements indicate closure at 768 nodes out of 768 on this crack face. <- incorrect
The entire crack face is closed. <- incorrect
MODE I, II, III RELATIVE DISPLACEMENTS AND OPEN/CLOSED STATUS AT THE CRACK FRONT.
NOTE - R IS THE DISTANCE FROM THE RADIAL NODE POSITION TO THE CLOSEST CRACK FRONT POSITION.
THIS CLOSEST CRACK FRONT POSITION MAY NOT COINCIDE WITH THE RESPECTIVE NODE POSITION.
Nc CF NODE RADIAL NODE(S) DISTANCE R Vi Vii Viii STATUS
1 321 354 386 1.54862E-02 4.74540E-04 5.32081E-04 -2.65212E-07 OPEN <- correct
2 1443 1476 1508 1.54862E-02 4.72792E-04 5.26828E-04 5.85676E-05 OPEN <- correct
etc.
Workaround: None.
ZENCR530
Applies to: Version 9.0-1.
Fixed in development version: 26 Sep 2022.
Fix included in release version: 9.1-1.
The parameters MAGNITUDE DADN, DIRECTION DADN, MAGNITUDE DADT and DIRECTION DADT were introduced on the *ENERGY RELEASE RATE keyword in version 9.0-1 to allow different fracture mechanics parameters to drive the fatigue and time calculations in a combined fatigue-time analysis. However, they are not processed correctly and the integration in a fatigue-time analysis uses the parameters specified for fatigue to drive both fatigue and time integration.
The analysis results using one parameter for fatigue and time are correct (i.e. the behaviour is effectively the same as version 8.3-1 and earlier releases where one parameter is used to drive fatigue and time calculations).
Workaround: None.
ZENCR529
Applies to: Version 9.0-1.
Fixed in development version: 26 Sep 2022.
Fix included in release version: 9.1-1.
The Gmax values along a crack front have damping applied before being used in the integration scheme. In a fatigue analysis if the parameter MAGNITUDE DADN on *ENERGY RELEASE RATE is given a non-default value (and MAGNITUDE DADT is not defined i.e. left as the default), the damping is applied twice.
This does not have a huge effect on the global bahaviour of the analysis but the detailed numbers would differ from an analysis in which damping is only applied once.
Workaround: define both parameters MAGNITUDE DADN and MAGNITUDE DADT with the same value if a non-default value (e.g. SIF-displacement) is to be used. This workaround of defining the time term, MAGNITUDE DADT, applies even if the analysis is fatigue-only.
ZENCR528
Applies to: Version 9.01 of utility PROCESS (included as part of Zencrack 9.0-1).
Fixed in development version: 6 Sep 2022.
Fix included in release version: 9.1-1.
In a case where a crack front node is closing the Gmax value for the node calculated from j-integrals is correctly assigned a -ve sign in the .rep file to indicate closure. But the extracted values in the .csv file produced by Process.exe are always positive in version 9.0-1 of the Process.exe.
This issue also affects the Gmax (j-integral) values that are generated on xy plots in the GUI as the extraction of the data for the plots uses the Process utility.
Workaround: None.
ZENCR527
Applies to: Versions 8.3-1 and 9.0-1 (earlier versions not checked) (Ansys interface).
Fixed in development version: 14 Jul 2022.
Fix included in release version: 9.1-1.
If surfaces are defined on brick element faces of a crack-block or split element using an element definition, EN, the update of the surface is incorrect after a crack-block transfer has taken place (or in the initial cracked mesh if there are split elements).
Workaround: Use a surface element method other than EN in the uncracked mesh.
ZENCR526
Applies to: Versions up to and including 9.0-1.
Fixed in development version: .
Fix included in release version: 9.1-1.
In some cases the end step number of a search through available results sets may not be reported correctly in an error message. This only occurs if the f.e. analysis did not successfully complete all expected steps (i.e. the Zencrack analysis would stop anyway with this error message and the issue is simply the incorrect reporting of the step number).
***ERROR 6033 The data you want to search for load system 1 cannot be located in the available results sets. Please check your input and the completion status of the last f.e. analysis. The requested end step number (*****) is not available.
Workaround: None.
ZENCR525
Applies to: Versions up to and including 9.0-1.
Fixed in development version: 6 Jul 2022.
Fix included in release version: 9.1-1.
If the number of results sets extracted from the f.e. analysis decreases from one f.e. step to the next the internal handling of the data may cause overwrite. This has also been found to be the cause of an analysis terminating without any traceback or error message.
This issue is only likely to occur in analyses using a full cycle load system in which the f.e. step loading is such that automatic incrementation may produce different (decreasing) number of increments in the f.e. steps as the crack grows.
Workaround: Use controls on the f.e. steps to try to prevent reduction in the number of increments from one f.e. analysis to the next. If fixed time stepping is possible rather than automatic time stepping this would prevent the issue from arising.
ZENCR524
Applies to: Versions up to and including 9.0-1.
Fixed in development version: 29 Jun 2022.
Fix included in release version: 9.1-1.
If there are multiple TSHAP options in the uncracked mesh, the insertion of the TSHAP data into the cracked mesh inserts the last TSHAP type at each position.
Workaround: Consider whether the multiple TSHAP options are required and can be replaced by an alternative option.
ZENCR523
Applies to: Version 9.0-1.
Fixed in development version: 26 May 2022.
Fix included in release version: 9.1-1.
In a remesh model with linear ring elements the initial reports to the .rep file for material information that is set via subroutine user_material_id() contain spurious data for phantom midside positions i.e. the output generator does not recognise that the crack front only has corner nodes.
The analysis is correct and the processing of the corner nodes by the subroutine is also correct - the issue is just invalid messages in the .rep file.
***INFORMATION : MATERIAL ID CHANGE The material id has been changed from 0 (&&&&&&&&&&&&&&&&&&&&) to 1 (Steel) in subroutine user_material_id for the following crack front node: crack front 1, node position 2 (node 0) at x,y,z = 7.367348-312 9.414206E+02 3.774603E+02
Workaround: None.
ZENCR522
Applies to: Versions up to and including 9.0-1.
Fixed in development version: 28 Sep 2022.
Fix included in release version: 9.1-1.
A small number of array re-allocations pass an uninitialised value which may lead to an out-of-bounds error. This may cause the analysis to terminate without any traceback or error message. If there is no such termination the issue does not affect the analysis.
Workaround: None.
ZENCR521
Applies to: Versions up to and including 9.0-1.
Fixed in development version: 7 Feb 2022.
Fix included in release version: 9.0-1 (7 March 2022 update).
An array bounds issue occurs during crack growth integration in a time-only analysis when detailed integration output is requested. This issue does not affect the analysis results or the detailed integration output. If desired it can be prevented in two ways:
Workaround 1: Define a fatigue crack growth law e.g. a Paris law (in a time-only analysis a fatigue law can be defined but is not used, however it prevents this issue arising).
Workaround 2: Do not request detailed integration output.
ZENCR520
Applies to: Versions up to and including 8.3-1.
Fixed in development version: 10 January 2020.
Fix included in release version: 9.0-1.
The datalines of the *ENERGY RELEASE RATE, STEPS option are not treated in the expected way if there is a gap in the step numbers on the datlines. For example steps 2 and 3 can be implied from step 1:
*ENERGY RELEASE RATE, STEPS 1,1 4,9999
This should process steps 1, 2, 3 and 4 with frequency settings of 1, 1, 1 and 9999 but instead processes them with settings 1, 9999, 9999, 9999.
Workaround: Define a frequency for each step. In the above example:
*ENERGY RELEASE RATE, STEPS 1,1 2,1 3,1 4,9999
ZENCR519
Applies to: Versions up to and including 8.3-1.
Fixed in development version: 21 May 2019.
Fix included in release version: 9.0-1.
There are possible unexpected error messages or array bounds issues due to incorrect use of counters related to Ansys types, reals, material ids, tshaps or solution steps. These conditions are unlikely to occur in "normal" input data files - they require high id values or high number of instances to be triggered.
Workaround: Use low id numbers for types, reals and materials ids.
ZENCR518
Applies to: Versions 8.1-1 to 8.3-1 (Abaqus interface).
Fixed in development version: 2 May 2019.
Fix included in release version: 9.0-1.
An uncracked Abaqus mesh that includes a *FILM PROPERTY keyword produces an error:
***ERROR 3014 Unable to interpret "*DLOAD" data
Workaround: Apply film coefficients by their numerical value directly on *FILM or *SFILM keywords instead of by a name via the *FILM PROPERTY keyword.
ZENCR517
Applies to: Version 8.1-1 (only the 32bit Windows version).
Fixed in development version: 12 July 2017 (with an updated 8.1-1 version created on 17 September 2018).
Fix included in release version: 8.2-1.
Creation of the initial cracked mesh may fail with one or more warnings of type:
***WARNING An unexpected NaN value has been encountered during the meshing process.
This issue affected the 8.1-1 32bit Windows executable only. Code changes made on 12 July 2017 removed the issue.
An updated 32bit executable for version 8.1-1 is available on request to customers with the appropriate maintenance status.
Workaround: None.
ZENCR516
Applies to: Versions 7.6 to 8.3-1 (Abaqus interface only).
Fixed in development version: 13 September 2018.
Fix included in release version: 9.0-1.
If the VCE=Normal option is used on *ENERGY RELEASE and the crack front has more than 16 corner nodes along the crack, the analysis fails with a Fortran error "too many records in I/O statement" while generating the first cracked mesh.
Workaround: Use *ENERGY RELEASE RATE, VCE=Direct (which is the default option).
ZENCR515
Applies to: All versions up to and including 8.2-2.
Fixed in development version: 22 May 2018.
Fix included in release version: 8.3-1.
Crack front configurations in which reversed through crack-blocks are adjacent to corner crack-blocks (normal or reversed) may give error 4107 e.g.:
***ERROR 4107 User input has placed crack-block 3 on crack front 1. Zencrack thinks this crack-block should be on crack front 2.
Workaround: None.
ZENCR514
Applies to: Versions 7.9-1 to 8.2-2.
Fixed in development version: 22 May 2018.
Fix included in release version: 8.3-1.
Crack face facets sent to the .vis file for the initial crack face of reversed crack-blocks are for the incorrect region of the crack-block face.
Workaround: None.
ZENCR513
Applies to: Rainflow versions 7.4 to 8.2 (supplied with Zencrack 7.8-1 to 8.2-2).
Fixed in development version: 23 May 2018.
Fix included in release version: Rainflow version 8.3 (supplied with Zencrack 8.3-1).
In the case that there are 2 or 3 reversals the rainflow count may produce a traceback or extract an incorrect cycle.
With 2 reversals:
- Rainflow versions 8.1 and 8.2 can give incorrect results or a traceback.
- Rainflow version 8.0 and earlier correctly stop with an error as they do not support 2 reversals.
With 3 reversals:
- Rainflow versions 7.4 to 8.2 extract an incorrect minimum value for the cycle if Point1<Point3 and Point2 is a peak.
Workaround: None.
ZENCR512
Applies to: All versions up to and including 8.2-2.
Fixed in development version: 31 Jan 2018.
Fix included in release version: 8.3-1.
When the *TIP MODEL option to create multiple nodes at each point on the crack front is used, the additional nodes that are created are not merged/tied at the interfaces between adjacent crack-blocks along the crack front. A displaced plot of the mesh shows apparent gaps opening up along the crack front.
Workaround: None.
ZENCR511
Applies to: All versions up to and including 8.2-2 (Abaqus interface).
Fixed in development version: 31 Jan 2018.
Fix included in release version: 8.3-1.
An error 3032 occurs when using an Abaqus submodel that has a shell-to-solid boundary e.g.:
***ERROR 3032 Error reading first constraint direction as: 0.3
This happens because the second item on the dataline in this case is a real value whereas Zencrack is looking for a degree of freedom.
Workaround: Place the *BOUNDARY keyword and dataline into an included file e.g.:
*Include, Input = boundary.inp ** ** Name: Sub-BC-1 Type: Submodel ** *Boundary, submodel, step=1 ** SET-3, 0.3
ZENCR510
Applies to: Version 8.2-1 (GUI 8.2-a).
Fixed in development version: 31 October 2017.
Fix included in release version: 8.2-2 (GUI 8.2-b).
When tabular data is loaded to the material screen, the fatigue law shows "user subroutine" instead of "tabular". Depending upon what is changed in the materials data screen, an OK may result in tabular data being retained in the internal data structure, or it may incorrectly be converted to "user subroutine".
Workaround: None.
ZENCR509
Applies to: Version 7.8-1 to 8.2-1.
Fixed in development version: 25 September 2017.
Fix included in release version: 8.2-2.
The fix for zencr468 did not cater for some additional conditions in which data written to the .bk file could be incorrect if the integration step produces da=dn=0. As with zencr468 the analysis is correct but the bk file may contain incorrect deltaK and Kmax values in some rows.
Workaround: None.
ZENCR508
Applies to: Version 8.1-1 to 8.2-1.
Fixed in development version: 20 September 2017.
Fix included in release version: 8.2-2.
The output from a crack growth analysis may contain zeros or NaN values for some items if the following conditions are met:
- Multiple materials are defined in the .zcr file.
- At least one material is not assigned to any crack front.
- The un-assigned material has no basic data defined (i.e. no Youngs modulus or Poisson ratio).
- The un-assigned material is subsequently assigned to at least one crack front node in subroutine user_material_id.
This issue is ultimately due to invalid input data (i.e. missing Youngs modulus and Poisson ratio). However, there is no warning or error message to indicate the possible cause of the zero and NaN values after the originally un-assigned material has been assigned via the subroutine.
Workaround: Correct the input data by defining Youngs modulus and Poisson ratio.
ZENCR507
Applies to: Version 7.9-1 to 8.1-1.
Fixed in development version: 15 August 2017.
Fix included in release version: 8.2-1.
The default mapping option for midside nodes when using STANDARD mapping was changed from MIDSIDE=Straight to MIDSIDE=Curved rings in version 7.9-1. But if the explicit '*MAPPING, TYPE=Standard' option is included the old default value MIDSIDE=Straight is used.
Workaround: Include an explicit '*MAPPING, MIDSIDE=Curved rings' option if there is an explicit '*MAPPING, TYPE=Standard' option.
ZENCR506
Applies to: Version 7.8-4 to 8.1-1.
Fixed in development version: 14 August 2017.
Fix included in release version: 8.2-1.
If invalid split set data is defined such that one or more pairs of elements have both of their elements on one side of the split, the analysis may produce an unexpected and unhelpful error while generating the first cracked mesh:
***Error 4089 Unable to find node <> while searching the node list.
where the reported node number is 0 or may be a node number that does not exist in the mesh.
Workaround: Ensure that the split set data is correct.
ZENCR505
Applies to: Versions 7.7 to 8.1-1 (earlier versions not checked).
Fixed in development version: 14 July 2017.
Fix included in release version: 8.2-1.
Input of material data allows for Kc and Kic values to be defined with default values set at 1E20. If only Kc is defined in the input data the calculation of the apparent fracture toughness results in warnings for each crack front node and the analysis stops:
***WARNING The apparent fracture toughness is exceeded at corner node position 1 on crack front 1 during this integration step.
If only Kic is defined in the input data the analysis runs but the input value for Kic is swamped by the defalt of 1E20 for Kc and the apparent fracture toughness is always 1E20.
In the fix introduced in version 8.2-1 , if only one of Kc or Kic is defined, the other takes the identical value.
Workaround: Define the same value for Kic as for Kc.
ZENCR504
Applies to: Versions 7.7 to 8.1-1 (earlier versions not checked).
Fixed in development version: 20 July 2017.
Fix included in release version: 8.2-1.
Temperature dependent threshold data must be defined in order of ascending temperature. No check on the input data was being performed and input of threshold data in order of non-ascending temperatures resulted in incorrect threshold calculations.
Workaround: Define threshold data in order of ascending temperature.
ZENCR503
Applies to: Versions 7.7 to 8.1-1 (earlier versions not checked).
Fixed in development version: 20 July 2017.
Fix included in release version: 8.2-1.
If temperature dependent threshold data is defined using the TYPE=Deltakth, kth or ctth options the threshold value used is that associated with the last temperature value instead of the first if a threshold value is required for a temperature below the first temperature value.
Workaround: Ensure that the lowest temperature used to define threshold data is below the lowest temperature expected as part of the output from the analysis.
ZENCR502
Applies to: Versions 7.7 to 8.1-1 (earlier versions not checked).
Fixed in development version: 14 June 2017.
Fix included in release version: 8.2-1.
A Fortran error (Program Exception - access violation) may occur in routine plane if the crack front becomes too close to the back face of a crack-block.
Workaround: None.
ZENCR501
Applies to: Versions 7.7 to 8.1-1 (earlier versions not checked).
Fixed in development version: 14 June 2017.
Fix included in release version: 8.2-1.
Warnings of the type below may be created when the mesh relaxation region is processed. The node numbers reported on these warnings are not correct and are not useful in identifying the region of the mesh to which the warnings apply. Analysis results are correct, only the warning messages are incorrect.
***WARNING Node 66 has more than 8 connected hex elements with a common node. Check your cracked mesh.
Workaround: None.
ZENCR500
Applies to: All versions up to and including 8.1-1.
Fixed in development version: 6 May 2017.
Fix included in release version: 8.2-1.
Crack growth integration will terminate with a warning message containing conflicting information about re-integrating with a limit of dn=1 if the cycle count for a step exceeds 2147483647. For example:
***WARNING The step will be re-calculated using dn=1, because the tolerance on change in a is met in dn=2.99608425E+10 cycles. Instead of using less than 1 cycle the growth step is re-calculated using dn=1.
The fix for this issue also addresses some format issues for very high cycle counts.
Workaround: Set a limit of the allowable change in dn per step of 2147483646 on the *TOLERANCE keyword (item 9 on dataline 1). This will allow the integration step to complete but may produce small increments of da between f.e. steps.
ZENCR499
Applies to: Versions 7.8-1 to 8.1-1.
Fixed in development version: 28 March 2017.
Fix included in release version: 8.2-1.
If called from a user subroutine for a node that does not exist the utility function utility_coords causes the analysis to stop with an error '***ERROR 4089 - Unable to find node' instead of returning a value of -1 to the user subroutine.
Workaround: None.
ZENCR498
Applies to: Process versions 7.19 to 7.25 (included with Zencrack versions 7.7 to 8.1-1).
Fixed in development version: 5 April 2017.
Fix included in release version: 8.2-1.
Extraction of data from a multi-set f.e. analysis using the option "5) Results set with maximum G from energy" reports zero values for the results set column in the .csv file for all but the last crack front if the model has more than one crack front. The FE time/load column may also contain incorrect data.
Workaround: None.
ZENCR497
Applies to: Version 8.0-1.
Fixed in development version: 27 January 2017.
Fix included in release version: 8.1-1.
A remote transfer run for Abaqus that is set to extract results from the .dat file fails with error 1002, cannot find the .sta file. This happens because the remote transfer run incorrectly tries to transfer the .res file instead of the .sta file.
Workaround: Include this option in the .zcr file: *REMOTE TRANSFER, TYPE=From remote, STA=Yes.
ZENCR496
Applies to: GUI version 8.0-a (released with Zencrack 8.0-1).
Fixed in development version: 13 January 2017.
Fix included in release version: 8.1-1.
An attempt to 'Get data' when extracting results data for XY plots produces an error 'Process execution failed' if the results file is in a folder that contains spaces in the path.
Workaround: Copy the .vis and .rep files for the analysis to a folder that has no spaces in the path. Open the copied results file in the GUI and 'Get data' for the copy.
ZENCR495
Applies to: Process versions 7.21 to 7.24 (included with Zencrack versions 7.8-1 to 8.0-1) (Ansys interface).
Fixed in development version: 12 January 2017.
Fix included in release version: 8.1-1.
Data for the 'FE time/load' column in the .csv file created by the Process utility is not extracted from .rep files containing results from Ansys. Instead the column contains zeros.
Workaround: None.
ZENCR494
Applies to: Versions 7.9-3 and 8.0-1 (Ansys and NX Nastran interfaces).
Fixed in development version: 12 January 2017.
Fix included in release version: 8.1-1.
When multiple datalines are used for the STEPS option on *ENERGY RELEASE RATE, all steps of the f.e. analysis results are processed using the frequency setting defined on the last dataline (i.e. any different frequency settings on earlier datalines are ignored).
Workaround: None.
ZENCR493
Applies to: All versions up to and including 8.0-1.
Fixed in development version: 21 November 2016.
Fix included in release version: 8.1-1.
Some crack-blocks contain a small asymmetry in some node position which may prevent successful creation of a cracked mesh if a mixture of standard and reversed crack-blocks are used on a crack front and the edge ratios are very low or very high. Affected crack-blocks are:
- s02_q38x2
- s02_t19x1
- s03_q46x2
- s03_t23x1
- s_t111x5
- s_t151x5
Workaround: None.
ZENCR492
Applies to: Versions 7.9-3 and 8.0-1.
Fixed in development version: 14 December 2016.
Fix included in release version: 8.1-1.
Any output variables on the dataline for *OUTPUT, TYPE=INTEGRATION VARIABLES that contain leading spaces will generate an error:
***ERROR 2301 - Variable code " xyz" is not valid.
Workaround: Remove the leading spaces in each item on the dataline.
ZENCR491
Applies to: Versions 7.9-1 to 8.0-1.
Fixed in development version: 8 December 2016.
Fix included in release version: 8.1-1.
An apparently correct set of input data may produce an error:
***Error 4077 - Inconsistent edge ratios for merge of crack-blocks X and Y.
This can happen to the second (or higher) crack front in a model if the following conditions are met for the crack front:
- The crack front contains exactly two through crack-blocks.
- The INITIAL= parameter for the crack front is set to ratio or size methods.
- The ratio or size values assigned to the two edges of each crack-block are different.
Workaround: Convert the input for the affected crack fronts to INITIAL=Points.
ZENCR490
Applies to: Versions 7.5 to 8.0-1.
Fixed in development version: 15 August 2016.
Fix included in release version: 8.1-1.
A job that has several hundred crack front transfers may take ever increasing amounts of memory, possibly ending with an error e.g.:
***ERROR 7003 - Memory allocation problem for array ndsplit
Workaround: None.
ZENCR489
Applies to: GUI version 8.0-a (released with Zencrack 8.0-1).
Fixed in development version: 10 March 2016.
Fix included in release version: 8.1-1.
The GUI does not allow spaces in the folder name for .zcr files. Note that spaces should not be used in the job name itself - this is not a bug. The bug relates only to spaces in the folder path.
Workaround: Don't use spaces for locations containing Zencrack files.
ZENCR488
Applies to: Version 7.9-3 and 8.0-1 (NX Nastran interface).
Fixed in development version: 7 June 2016.
Fix included in release version: 8.1-1.
Character strings instead of numbers for SOL types are ignored and may cause the analysis to fail.
Workaround: Use numbers for SOL definitions in the uncracked mesh e.g. SOL 401.
ZENCR487
Applies to: Versions 7.5-8 to 8.0-1.
Fixed in development version: 25 May 2016.
Fix included in release version: 8.1-1.
In an analysis with multiple results sets and only a subset of those results sets used in the load systems, the retention of G data for use in calculating dG/da may be incorrect. Tables in the .rep file may show zero values for sqrt(G) and dG/da with some G error terms being reported as 100%. The consequence is that the step size will constantly reduce from one step to the next which is a clear indication that there is a problem in the analysis.
Workaround: Use the STEPS parameter on *ENERGY RELEASE RATE to ensure that only the specific results sets of interest are processed by Zencrack.
ZENCR486
Applies to: Version 7.9-3 and 8.0-1 (NX Nastran interface).
Fixed in development version: 11 May 2016.
Fix included in release version: 8.1-1.
Load applied to a split element face gives the following type of fatal error in NX Nastran because the change in node number on one side of the split set has not been carried into the load updates:
*** USER FATAL MESSAGE 4578 (GP3C)
PLOAD4 LOAD SET = 5 WITH DIAGONAL CONNECTING GRIDS 1100 AND 1114
IS INCORRECT FOR ELEMENT WITH ID = 1513
Workaround: None.
ZENCR485
Applies to: All versions up to and including 8.0-1 (Abaqus interface).
Fixed in development version: 10 May 2016.
Fix included in release version: 8.1-1.
Abaqus *DLOAD datalines containing element sets with a space in the set name are not processed correctly for load updates. If any crack-blocks exist in the set, load will be applied on them in an unexpected fashion.
Workaround: Remove the space from the affected set names.
ZENCR484
Applies to: All versions up to and including 8.0-1 (Ansys interface).
Fixed in development version: 11 March 2016.
Fix included in release version: 8.1-1.
Transfer of crack face load to split sets occurs twice for Ansys models.
Workaround: None.
ZENCR483
Applies to: All versions up to and including 8.0-1.
Fixed in development version: 11 March 2016.
Fix included in release version: 8.1-1.
Transfer of crack face load to split sets only occurs in the first load step with such update requirements. If a subsequent step has changes to the crack face load those changes are not applied to the split sets. This behaviour applies to the Abaqus and Ansys interfaces. For NX Nastran only the first set of PLOAD options that need split set updates get the updates.
Workaround: None.
ZENCR482
Applies to: Version 7.9-3 and 8.0-1.
Fixed in development version: 25 April 2016.
Fix included in release version: 8.1-1.
If an element set defined on *ELEMENT SET, NAME= via the GUI (or directly in a .zcr file) is passed to the cracked mesh file with EXPORT=Yes and the expansion option EXPAND=Yes is also set, there may be an error that "zcr80.exe has stopped working". This can happen after a transfer of a crack front if the original crack position does not have any elements in the set but the position after the transfer has at least one crack-block element in the set.
Workaround: Set EXPAND=No.
ZENCR481
Applies to: Version 7.9-2 to 8.0-1.
Fixed in development version: 12 April 2016.
Fix included in release version: 8.1-1.
The report in the .dbg file produced by *OUTPUT, CYCLE COUNT OUTPUT=Yes contains sqrt(G) values in the columns headed "G at max" and "G at min". If the analysis is a full cycle analysis, incorrect values (sqrt(sqrt(G)) are then used in the calculations for R, deltaRG and da/dn and incorrect values are reported in the table. These values are local for this report table only - the values used in the integration scheme and the analysis results reported in the .rep file, other output files and other tables in the .dbg file are correct.
Workaround: None.
ZENCR480
Applies to: Version 7.8-1 to 8.0-1.
Fixed in development version: 5 April 2016.
Fix included in release version: 8.1-1.
An analysis may stop without any error message or traceback while creating the first cracked mesh. This can occur if the largest element in the mesh is adjacent to a very high number (thousands) of much smaller elements (such as might be the case if a large crack-block target element is surrounded by many very small tet or tet elements).
Workaround: Set the value tolradius1 on the *CONTROLS keyword to less than 1 (though trial and error may be needed to find a value that is successful as the issue is mesh dependent).
ZENCR479
Applies to: Version 7.7 to 8.0-1.
Fixed in development version: 10 March 2016.
Fix included in release version: 8.1-1.
A Zencrack analysis using a nodelocked license file may fail with an error:
License manager error, failed license checkout from <i>truncated folder path</i>, Code=-102, No such file or directory.
This happens if the path of the folder containing the license file is =>48 characters. The GUI is not affected by this issue.
Workaround: Create a folder with the truncated name shown in the error message and place the license file in that folder as well as in the folder with the full path.
ZENCR478
Applies to: Rainflow version 7.5 and 7.6 (supplied with Zencrack 7.9-2 and 7.9-3 respectively).
Fixed in development version: 1 February 2016.
Fix included in release version: 8.0-1.
Zero values to locate the maximum and minimum position of a cycle may be returned from a rainflow count if there are exactly 3 reversals. If the rainflow analysis was part of a Zencrack integration process the subsequent processing is affected by these zeros.
Workaround: None.
ZENCR477
Applies to: Version 7.9-2 to 7.9-3.
Fixed in development version: 22 December 2015.
Fix included in release version: 8.0-1.
Crack growth integration using the numerical forward predictor scheme is incorrect for load time history combinations. Results are correct using the closed form integration scheme.
Workaround: Use the closed form integration scheme (*INTEGRATION SCHEME, SOLUTION=CLOSED FORM).
ZENCR476
Applies to: All versions up to and including 7.9-3 (Abaqus interface).
Fixed in development version: 10 December 2015.
Fix included in release version: 8.0-1.
If an Abaqus uncracked mesh contains a *CONTOUR INTEGRAL keyword in step 2 or later it is not removed when the cracked mesh is generated. This causes incorrect processing of the contour integral results from the analysis.
Workaround: An uncracked mesh is not expected to contain any *CONTOUR INTEGRAL options so remove them.
ZENCR475
Applies to: Version 7.6 to 7.9-3.
Fixed in development version: 7 December 2015.
Fix included in release version: 8.0-1.
Utility function utility_da_modify does not always set the step size as expected:
- The step size from utility_da_modify is not set if the *TOLERANCE, TYPE=FIXED scheme is used to control step sizes.
- The step size from utility_da_modify is not set during the steps where scaling takes place if the *TOLERANCE, TYPE=ERROR CONTROL scheme is used to control step sizes.
Workaround: None.
ZENCR474
Applies to: Rainflow version 7.6 (supplied with Zencrack 7.9-3).
Fixed in development version: 3 December 2015.
Fix included in release version: 8.0-1.
Erroneous output may occur if the first pass through the reversals list does not identify any cycles. The reported cycle count may include cycles with maximum or minimum values of 888 and -888 due to use of incorrect internal variables.
Workaround: None.
ZENCR473
Applies to: User manuals for version up to and including 7.9-3.
Fixed in development version: 14 November 2015.
Fix included in release version: 8.0-1.
The Input Formats manual lists .res as a valid extension type for Abaqus on *REMOTE TRANSFER but this extension type is not allowed. From 14 November 2015 the .res extension is allowed for both *SAVE and *REMOTE TRANSFER.
ZENCR472
Applies to: Version 7.8-1 to 7.9-3.
Fixed in development version: 14 November 2015.
Fix included in release version: 8.0-1.
The option NEGATIVE R=DELTA K on *CRACK GROWTH DATA, TYPE=PARIS gives Error 2012: Parameter value for "negativer" is invalid.
Workaround: Do not include NEGATIVE R=DELTA K on the keyword line - this is the default value and does not need to be explicitly stated.
ZENCR471
Applies to: Versions 7.9-2 and 7.9-3.
Fixed in development version: 27 August 2015.
Fix included in release version: 8.0-1.
The node set NFACExxxSIDEy containing all crack face nodes on side y of crack xxx is incomplete if the crack face has more than 10000 nodes (i.e. some crack face nodes are missing from the set). For Abaqus and Ansys the set is used for extraction of nodal data from the analysis. This means that some crack face nodes will have zero displacements. If those nodes are used to calculate K values then the Ks will be zero. Note that for Abaqus the .dat and .fil extraction is not affected because a separate node set NCFxxxCTOD is used. The node set NCFxxxCTOD was removed from .odb extraction requests to reduce the number of history requests on 3 July 2014. So K values would be correct in versions 7.9-1 and earlier even though the node set NFACExxxSIDEy is limited to 10000 nodes.
Workaround: None.
ZENCR470
Applies to: Version 7.9-1 to 7.9-3 (Abaqus interface for Windows).
Fixed in development version: 30 July 2015.
Fix included in release version: 8.0-1.
If the Abaqus installation is in a folder containing a space, the extraction of results from the .odb file may fail with a system error in the command prompt window. Error 1002 "Cannot find file: JOBNAMEr-zcr.001" appears in the .rep file.
Workaround: Edit the <Zencrack folder>\tools\zencrack.ini file and place "" around the value of ZCR_ABAQUS e.g.
ZCR_ABAQUS = "c:\program files\simulia\abaqus\6.14-1\code\bin\abq6141.exe"
ZENCR469
Applies to: Version 7.0 to 7.9-3.
Fixed in development version: 12 June 2015.
Fix included in release version: 8.0-1.
If a crack front with two distinct ends has three or more crack-blocks along its length and the first crack-block defined on the keyword data lines is not at one of end locations, a growth analysis may end with error 4029 or 6006 after f.e. step number 1.
Workaround: Ensure that the first crack-block defined for a crack front is at the end of the crack.
ZENCR468
Applies to: Version 7.8-1 to 7.9-3.
Fixed in development version: 12 June 2015.
Fix included in release version: 8.0-1.
In some cases a data line for "end of da/dn increment" output to the .bk2 file may incorrectly include zero value for Kmax, deltaK and da/dn. This can happen if the integration scheme attempts to integrate a block of cycles for which there are no cycles remaining. This is only likely to be an issue near the end of the analysis where the dn increment between f.e. steps is 1 cycle. The analysis is correct - this is only a reporting issue to the .bk2 file.
Workaround: None.
ZENCR467
Applies to: Version 7.8-1 to 7.9-3.
Fixed in development version: -.
Fix included in release version: 8.0-1.
The documentation states that the MINOR CYCLES parameter on *LOAD SYSTEM has default value NO but the program uses a default value of YES.
ZENCR466
Applies to: Version 7.9-3 (Abaqus interface).
Fixed in development version: 4 June 2015.
Fix included in release version: 8.0-1.
The extraction of data from the Abaqus .odb file results in a KeyError traceback failure from the Python script while defining variable incValue if any analysis increment with contour integral data has no field data output.
Workaround: Ensure that all increments for which contour integral data is extracted have at least one field output request by adding appropriate field output request(s) to the uncracked mesh file.
ZENCR465
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 22 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Non-ASCii characters (e.g. Chinese font characters) entered into the GUI cause an error message that files cannot be saved. The GUI may also hang. A partial input file may be created as a result. In the fixed version the non-ASCii characters do not cause any errors but are converted to ASCii characters in the .zcr file and lose their original meaning. It is recommended that such characters are not used, even in the fixed version.
Workaround: Ensure that only ASCii characters are used.
ZENCR464
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 9 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Importing a new uncracked mesh on the Input tab to replace an existing uncracked mesh does not set the "changed data" indicator. If exiting the GUI without saving the data there is no warning about unsaved changed data.
Workaround: Use the File/Save option to save the data.
ZENCR463
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 9 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Importing an uncracked mesh after Results have been loaded fails to show those results in the VS screen.
Workaround: None.
ZENCR462
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 10 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Profile lines and crack front faces may not display fully when viewed from some angles.
Workaround: Try changing the viewpoint.
ZENCR461
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 15 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Picking a single element (e.g. crack-block element) may fail where there are nearby smaller elements.
Workaround: Try changing the viewpoint or removing some elements from the display.
ZENCR460
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 23 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Picking nodes with some elements not displayed may pick nodes from non-displayed elements.
Workaround: None.
ZENCR459
Applies to: GUI version 7.9a to 7.9c (included with Zencrack 7.9-1 to 7.9-3 March 2015 release).
Fixed in development version: 9 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Plotting invalid xy data values (e.g. log of zero) fails with no message.
Workaround: None.
ZENCR458
Applies to: Process version 7.23 and earlier (included with Zencrack 7.9-3 March 2015 release and earlier) only for .rep files created by Zencrack 7.7-2 or later.
Fixed in development version: 6 May 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Raw j-integral data is not extracted correctly for Ansys models when only one side of the crack is modelled. An extra factor of 2 is applied to the extracted values i.e. the reported values in the .csv file are twice as big as they should be. This also affects plotting of these values using the "get xy data" option in the GUI.
Workaround: None.
ZENCR457
Applies to: Process version 7.23 (included with Zencrack 7.9-3 March 2015 release).
Fixed in development version: 6 May 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Raw j-integral data is not extracted correctly for NX Nastran. Values are reported as zero in the .csv file with Ct-integral header instead of J-integral. This also affects plotting of these values using the "get xy data" option in the GUI.
Workaround: None.
ZENCR456
Applies to: GUI version 7.9c (included with Zencrack 7.9-3 March 2015 release).
Fixed in development version: 8 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
An attempt to "get xy data" for all steps of a multi increment f.e. analysis fails if the Zencrack job is still running (e.g. on fe step 2 or 3 if results are requested for step 1).
Workaround: Wait for the analysis to finish and try again.
ZENCR455
Applies to: Version 7.6 to 7.9-3 (March 2015 release).
Fixed in development version: 30 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Utility function utility_crack_front_node does not return a value of -1 if the requested node position is higher than the actual number of nodes on the requested crack front. Instead an apparently random value is returned.
Workaround: None.
ZENCR454
Applies to: Version 7.6 to 7.9-3 (March 2015 release).
Fixed in development version: 10 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
A request to use subroutine USER_DADT_CT instead calls USER_DADT_G and results in a Fortran error.
Workaround: None.
ZENCR453
Applies to: Version 7.8-1 to 7.9-3 (March 2015 release).
Fixed in development version: 8 April 2015.
Fix included in release version: 7.9-3 (May 2015 release).
Ct-integrals are not processed correctly from the Abaqus .odb file. Depending upon the .py script version either zero values are reported (up to and including 7.9-1) or the script crashes with an error (7.9-2 and 7.9-3).
Workaround: Edit the script and set the Ct-integral label to be "Ct" not "C".
ZENCR452
Applies to: Version 7.8-1 to 7.9-2.
Fixed in development version: 24 March 2015.
Fix included in release version: 7.9-3.
An analysis may fail with error 4165 if the following conditions are met:
- The uncracked mesh contains a mixture of hex and other elements (e.g. tets) and the last element defined is not a hex element.
- A full set of temperatures is defined on all nodes in the uncracked mesh and the mesh relaxation option is activated (i.e. temperature interpolation will be required).
Workaround #1: Identify the largest edge length on all hex elements in the mesh and define a tolerance:
*CONTROLS, tolradius2=value
where value=largest edge length/1e20.
Workaround #2: Turn off mesh relaxation (*RELAX, TYPE=NONE).
ZENCR451
Applies to: Various versions.
Fixed in development version: 12 March 2015.
Fix included in release version: 7.9-3.
Several issues were identified with the remote transfer capability:
- Remote transfer variable names were not being picked up correctly from the zencrack.ini file with the result that remote transfer failed. (Versions 7.9-1 and 7.9-2).
- Extraction of results does not work if remote transfer is used and results are to be extracted from the Abaqus .odb file. (Versions 7.8-1 to 7.9-2).
- Binary files were being transferred in ascii mode so were not valid on the local machine. (Versions up to 7.9-2).
Workaround: None.
ZENCR450
Applies to: Process version 7.22 and earlier (included with Zencrack 7.9-2 and earlier).
Fixed in development version:20 February 2015.
Fix included in release version: 7.9-3.
Some data may not be extracted in the following cases:
- Raw contour integral values are not extracted for Abaqus analyses with T-stress output and multiple crack fronts.
- "node error" and "largest error" data are not extracted reliably when the load systems use one or more energy=search options.
Workaround: None.
ZENCR449
Applies to: Rainflow version 7.5 and earlier (included with Zencrack 7.9-2 and earlier).
Fixed in development version: 5 February 2015.
Fix included in release version: 7.9-3.
The output from the rainflow utility program may not report any cycles for some cases.
Workaround: None.
ZENCR448
Applies to: Version 7.8-1 to 7.9-2.
Fixed in development version: 19 February 2015.
Fix included in release version: 7.9-3.
Contour integral results are not extracted correctly from the Abaqus .odb file if ten or more contours are calculated on a crack front. This happens because the internal name of the contour integral variable has 01, 02 etc. if there are ten or more contours but 1, 2 etc. if there are less than ten contours.
Workaround: Extract results from the .dat or .fil files.
ZENCR447
Applies to: Version 7.8-1 to 7.9-2.
Fixed in development version: 18 February 2015.
Fix included in release version: 7.9-3.
T-stress results are not extracted from the Abaqus .odb file for Abaqus 6.14 when using Zencrack 7.8-1 to 7.9-2. This happens because the internal name of the T-stress variable has been changed in Abaqus 6.14.
Workaround: Extract results from the .dat or .fil files or use Abaqus 6.13 or earlier if results must be extracted from the .odb file.
ZENCR446
Applies to: Version 7.8-1 to 7.9-2.
Fixed in development version: 4 February 2015.
Fix included in release version: 7.9-3.
If .bk1 or .bk2 files from a previous analysis are locked when an analysis is re-run (e.g. open in Excel), a Fortran write error may occur.
Workaround: Ensure no .bk1 or .bk2 files are locked and re-run the job.
ZENCR445
Applies to: Version 7.7 to 7.9-2.
Fixed in development version: 15 January 2015.
Fix included in release version: 7.9-3.
Error 4130 reports an incorrect number for the current spline factor value.
Workaround: None.
ZENCR444
Applies to: Version 7.9-1, 7.9-2 (Windows only).
Fixed in development version: 9 January 2015.
Fix included in release version: 7.9-3.
Use of a Zencrack user subroutines fails if the job folder path or subroutine filename contains spaces.
Workaround: Edit zencrack.ini to include double quotes around $ZCR_USER_SUB$ in the definition of the ZCR_FORTRAN variables.
ZENCR443
Applies to: Version 7.9-1, 7.9-2.
Fixed in development version: 9 January 2015.
Fix included in release version: 7.9-3.
If job folder paths are very long the runzcr79 program used to launch Zencrack jobs may fail with a Fortran out of bounds error.
Workaround: Reduce the length of folder names in the path.
ZENCR442
Applies to: Version 7.9-1, 7.9-2.
Fixed in development version: 17 December 2014.
Fix included in release version: 7.9-3.
Element set names on keyword *ELEMENT, NAME= have spaced removed when processing the input data.
Workaround: None.
ZENCR441
Applies to: Version 7.9-a, 7.9-b (Zencrack GUI in versions 7.9-1, 7.9-2).
Fixed in development version: 8 October 2014.
Fix included in release version: 7.9-3.
Meshes do not import to the Zencrack GUI if using Ansys 10.0. The Zencrack analysis is not affected and an uncracked mesh that does not import into the GUI can still be used in a Zencrack analysis.
Workaround 1: Upgrade to Ansys 11 or later.
Workaround 2: A partial workaround is to set RUNANSYS = NO in the c:\zencrack\7.9-x\zencrack_gui.ini file. If there are no solid model commands in the mesh it should then import (although node and element sets will not be available).
ZENCR440
Applies to: Version 7.9-2 onwards.
Fixed in development version: -
Fix included in release version: -
In the table summarising the results of a maximum or minimum energy release rate search the column for "set" is limited to 4 digits. This was sufficient in v7.9-1 and earlier but 9999 may be exceeded in v7.9-2 if the f.e. analysis has more than 10000 steps or the combined spectrum time history in a multiple spectra combination analysis has more than 9999 time points. This is a minor reporting issue - the analysis is correct.
ZENCR439
Applies to: Version 7.9-1 (documentation).
Fixed in development version: 19 September 2014.
Fix included in release version: 7.9-2.
In the Input Formats Manual the EXPORT parameter on keyword ELEMENT SETS is incorrectly documented as EXPAND. The decription for the parameter is correct.
ZENCR438
Applies to: Version 7.1 to 7.9-1.
Fixed in development version: 17 September 2014.
Fix included in release version: 7.9-2.
If the STEPS parameter is used on keyword *ENERGY RELEASE RATE there is no check on whether or not the number of data lines exceeds the upper limit of 16 lines. For v7.8-1 or later the Abaqus interface stops with an error message 2080. For earlier versions with the Abaqus interface and all versions with the Ansys interface, the analysis ends with a Fortran error or continues with some corruption of the data on these data lines.
Workaround: None.
ZENCR437
Applies to: Version 7.9-a (Zencrack GUI in version 7.9-1).
Fixed in development version: 7 July 2014.
Fix included in release version: 7.9-b (Zencrack GUI in version 7.9-2).
If tet elements are excluded from a mesh import process the display of the mesh incorrectly uses all nodes to calculate the display parameters. If the brick elements are a small part of the entire mesh this can lead to a very small initial representation of the mesh in the viewport.
ZENCR436
Applies to: Version 7.9-a (Zencrack GUI in version 7.9-1).
Fixed in development version: 30 May 2014.
Fix included in release version: 7.9-b (Zencrack GUI in version 7.9-2).
Plots saved in the GUI with the File/Capture options may include part of the File drop-down menu.
Workaround: None.
ZENCR435
Applies to: Version 7.8-1 to 7.9-1.
Fixed in development version: 15 July 2014.
Fix included in release version: 7.9-2.
Incorrect scaling of dG/da during the integration process for full cycle load systems. This occurs only if the integration type is forward predictor and minor cycles are requested. The effect of the error on the overall analysis result is likely to be small in most cases.
Workaround: The constant G integration scheme is not affected by this issue.
ZENCR434
Applies to: Version 7.8-2 to 7.9-1.
Fixed in development version: 9 July 2014.
Fix included in release version: 7.9-2.
The fix for bug zencr403 can cause poor performance of the Abaqus analysis and results extraction from the .odb file if the Abaqus analysis has a large number of steps. This update to the zencr403 fix optimises the history output requests and improves performance of extraction of results from the .odb file. Note that this is a performance issue only - the analysis results are correct.
Workaround: None.
ZENCR433
Applies to: Version 7.1 to 7.9-1.
Fixed in development version: 1 July 2014.
Fix included in release version: 7.9-2.
The summary report of load levels in a spectrum file may not report the correct overall highest maximum and lowest minimum load levels correctly. If all maximum load levels are less than zero a value of zero is reported for the highest maximum value. If all minimum load levels are greater than zero a value of zero is reported for the lowest minimum value. Further, the number of load blocks in the spectrum is reported as ****** if there are 1000000 or more load blocks in the spectrum. These reports during reporting of the load system data are for information only and do not affect the analysis.
Workaround: None.
ZENCR432
Applies to: Version 7.9-a (Zencrack GUI in version 7.9-1).
Fixed in development version: 5 June 2014.
Fix included in release version: 7.9-b (Zencrack GUI in version 7.9-2).
Crack-block l05_q1176x14 does not appear in the list of crack-blocks in the GUI (or the Abaqus/CAE plug-in) and so cannot be used when defining a crack front.
Workaround: Select l05_q672x8 instead. When the .zcr file is saved edit the file and change l05_q672x8 to l05_q1176x14. Run the job from the command line.
ZENCR431
Applies to: Version 7.8-1 to 7.9-1.
Fixed in development version: 29 April 2014.
Fix included in release version: 7.9-2.
Use of multiple *OUTPUT, TYPE=INTEGRATION options (e.g. one for .bk1 and one for .bk2 files) with different lists of selected nodes does not behave as might be expected. Instead, the selected nodes listed on the last *OUTPUT, TYPE=INTEGRATION option are used for all .bk1 and .bk2 output files.
Workaround: None.
ZENCR430
Applies to: Version 7.8-1 to 7.9-1 (32bit Windows installations only).
Fixed in development version: 14 April 2014.
Fix included in release version: Version 7.9-1 installer updated with correct file on 14 April 2014.
The version of rainflow.exe supplied with 32bit Windows versions contains a 64bit executable and will not run. It gives a standard Windows error message. For Windows 7 this reads: "This version of rainflow.exe is not compatible with the version of Windows you're running. Check your computer's system information to see whether you need a x86 (32-bit) or x64 (64-bit) version of the program, and then contact the software publisher."
Workaround: None.
ZENCR429
Applies to: Versions up to and including 7.9-1.
Fixed in development version: 13 February 2014.
Fix included in release version: 7.9-2.
"***ERROR 4029 - Arc length out of range" or "***ERROR 4034 - Could not locate point in the local crack plane of crack-block" may occur if crack fronts in a model are located very far from the origin (i.e. if crack front nodes have coordinates with absolute value(s) many orders of magnitude larger than the length of a typical crack front element). Models with large offsets from the origin may generate error 4029 during crack growth. Models with huge offsets from the origin may generate error 4034 when creating the first cracked mesh.
Workaround: Translate the mesh so that crack region is positioned closer to the origin.
ZENCR428
Applies to: Version 7.9-a (Zencrack GUI in version 7.9-1).
Fixed in development version: 7 January 2014.
Fix included in release version: 7.9-b (Zencrack GUI in version 7.9-2).
The Zencrack GUI program will hang if the 'Clear Zcr mesh' option in the File menu is selected while any fe node sets are highlighted in the mesh viewport.
Workaround: None.
ZENCR427
Applies to: Version 7.9-1.
Fixed in development version: 8 January 2014.
Fix included in release version: 7.9-2.
If a keyword data line *ELEMENTS SETS, NAME= is defined in the .zcr file without the EXPORT parameter, the analysis fails with a Fortran error. Input files saved from the Zencrack GUI are not affected by this issue as the EXPORT=YES or EXPORT=NO options are always created by the GUI.
Workaround: Ensure that EXPORT=YES or EXPORT=NO is included on the keyword data line.
ZENCR426
Applies to: Version 7.6 to 7.8-4.
Fixed in development version: 18 November 2013.
The fix is a full fix for DISTRIBUTION=INITIAL and a sensible error message for DISTRIBUTION=UNIFORM.
Fix included in release version: 7.9-1.
A cracked mesh may not be generated correctly if a user defined crack front is used and:
- the SPLINE FIXUP=YES option is used for the crack front, and,
- the crack-blocks are not input in a sequential order along the crack front.
An error and/or invalid mesh (i.e. with inside out elements) is generated. The behaviour may be different if the DISTRIBUTION parameter is changed from INITIAL to UNIFORM.
Workaround : Ensure that the crack-blocks are defined in sequence along the crack front. In some cases it may be acceptable to use SPLINE FIXUP=NO. For version 7.8-1 to 7.8-4 it is possible that the crack-block expansion algorithm will produce a non-sequential order. In this case the full set of crack-block definitions generated in the .dbg file would need to be manually re-ordered along the crack front and then used as a full definition of the crack on the *CRACK FRONT keyword in the .zcr file.
ZENCR425
Applies to: Version 7.1 to 7.8-4.
Fixed in development version: 28 October 2013.
Fix included in release version: 7.9-1.
An incorrect alpha value (for G to K conversion) could be used during threshold calculations if tabular fatigue crack growth data is defined on *CRACK GROWTH DATA and the *THRESHOLD keyword is also used.
Workaround : None.
ZENCR424
Applies to: Version 7.8-1 to 7.8-4.
Fixed in development version: 10 September 2013.
Fix included in release version: 7.9-1.
An incorrect error message is generated if the START STEP, START INC, END STEP or END INC parameters are used on the *LOAD SYSTEM keyword when the load system is TYPE=FULL CYCLE.
The error message produced is:
Parameter <> is only valid for energy=search
but should be:
Parameter <> is only valid for type=minimum or type=maximum
ZENCR423
Applies to: Version 7.1 to 7.8-4.
Fixed in development version: 8 October 2013.
The fix is the inclusion of a meaningful warning message rather than a full fix to the underlying problem.
Fix included in release version: 7.9-1.
If an embedded crack front is defined by a user defined spline that has extra points at the start or end of the spline (e.g. the end of the spline is extended and overlaps the start of the spline), the processing of the spline may generate an error number 4053 in the .rep file.
Workaround : For an embedded crack front spline ensure that the first and last spline points are coincident.
ZENCR422
Applies to: Version 7.8-1 to 7.8-4 (Abaqus interface).
Fixed in development version: 29 July 2013.
Fix included in release version: 7.9-1.
If user subroutine user_contour is used to process contour integral values and the analysis extracted results from the Abaqus .odb file, the program will appear to hang at screen message "Starting crack growth integration" after the first f.e. analysis.
Workaround : If user_contour has to be used, extract results from the .dat or (preferably) the .fil files instead.
ZENCR421
Applies to: Version 7.8-1 to 7.8-4 (with network license).
Fixed in development version: 15 October 2013.
Fix included in release version: 7.9-1.
The ZENCRACK_WAIT parameter defines the number of minutes that the program waits for network license tokens to become available if tokens are in use by another job. The availability of tokens is checked every 10 seconds during this waiting period. The on-screen display converts this information to a number of attempts to acquire license tokens e.g. if ZENCRACK_WAIT is 30 minutes, the report is:
Waiting for license, maximum 180 times: interval 10 seconds...
If the number of attempts exceeds 999 (i.e. ZENCRACK_WAIT>166.5 minutes), an incorrectly formatted message is displayed:
Waiting for license, maximum *** times: interval 10 seconds...
This does not affect the acquisition of tokens and the wait time is as requested - only the formatting of the message is incorrect. For versions 7.5-8 to 7.7-2 the same probelm could occur if the number of attempts exceeded 99999.
Workaround : None.
ZENCR420
Applies to: Version 7.8-2 to 7.8-4 (Ansys interface).
Fixed in development version: 16 October 2013.
The fix is the inclusion of an error message rather than a full fix to the underlying problem.
Fix included in release version: 7.9-1.
If an uncracked Ansys mesh is treated by Zencrack as a "normal" mesh rather than a "solid model" mesh and it contains EBLOCK, EN or E surface element definitions before the target crack-block or split element definitions, Zencrack may crash or stop with error 4090:
***ERROR 4090 Unable to find element <> while searching the element list.
Processing of EBLOCK, EN and E for surface elements was introduced in version 7.8-2 - earlier versions will behave differently. Version 7.8-2 may show the issue for surface element types 154, 170, 173 and 174. Versions 7.8-3 and 7.8-4 may also show the issue for surface element type 152.
Workaround : There are three possibilities:
- Edit the uncracked mesh and cut/paste the surface element definitions to a position after the target crack-blocks and split elements. The numbers of the elements in the mesh can stay the same, only the physical position of their definition in the file needs to be changed.
- Force the use of a solid model run with the command line option "-a solidmodel".
- If the surface elements are far from any elements affected by the crack-blocks, split elements and any relaxation region, try using version 7.8-1.
ZENCR419
Applies to: Version 7.1 to 7.8-4 (documentation).
Fixed in development version: 29 November 2013.
Fix included in release version: 7.9-1.
The Input Formats Manual states that the ACCELERATION parameter on keyword *RETARDATION has default NO when Willenborg retardation is used and the ACCELERATION parameter is omitted. However, the default is YES if the ACCELERATION parameter is omitted. Note that the .rep file correctly reports the actual value of ACCELERATION parameter used in the analysis.
ZENCR418
Applies to: Version 7.1 to 7.8-4.
Fixed in development version: 1 July 2013.
Fix included in release version: 7.9-1.
The value of Kmax is not updated correctly when considering retardation effects for fully compressive cycles during Willenborg retardation. This results in a small underestimate of the acceleration effect during Willenborg retardation with the plastic zone size being larger than it should be (i.e. the acceleration effect of decreasing the plastic zone size does not apply enough of a decrease). This issue only affects analyses in which:
- Willenborg retardation is used with acceleration effects turned on.
- The forward predictor integration scheme is used.
- The load spectrum has fully compressive cycles.
The consequence of this issue is that retardation will last for slightly longer than it should and the final cycle count will be slightly higher than it should be. Tests on a SEN model using the Falstaff spectrum give a typical reduction in final cycle count of 0.1% with the bug fix in place.
Workaround : None.
ZENCR417
Applies to: Version 7.5-8 to 7.8-4 (earlier versions not checked).
Fixed in development version: 1 July 2013.
Fix included in release version: 7.9-1.
When the TIME HISTORY parameter is used to define a load vs time history in a "simplified" time or fatigue and time analysis, there are issues in the handling of negative load levels (which are treated as positive during time integration calculations). The result is that parts of the load history may produce crack growth when the growth should be zero. This issue can occur if:
- The load history in the file referenced by the TIME HISTORY parameter on keyword *LOAD SYSTEM, combined with the value of the SCALE parameter, has any negative load levels.
- A static load system with negative load level is combined with a cyclic load system that has a TIME HISTORY file associated with it.
Note: This issue does not affect load systems of TYPE=FULL CYCLE.
Workaround : None.
ZENCR416
Applies to: Version 7.8-1 to 7.8-2.
Fixed in development version: 28 February 2013.
Fix included in release version: 7.8-3.
The automatic split set expansion method may fail to find all split element pairs if the split contains degenerated 8 or 20 noded brick elements.
Workaround : The pairs or sides method should be used to fully define the split. Bug zencr414 may also occur.
ZENCR415
Applies to: Version 7.5 to 7.8-2 (Documentation).
Fixed in development version: 1 March 2013.
Fix included in release version: Crack-block Library Manual, Issue 78 Revision 1 (released with Zencrack version 7.8-3).
The Crack-block Library User manual should state that crack-block s05_q24x2.sup cannot be used with Ansys. If the crack-block is used with Ansys, Zencrack stops (as it should) with an error. The behaviour with regard to crack-blocks that are not allowed with Ansys is less restricted from Zencrack version 7.8-3.
ZENCR414
Applies to: Version 7.0 to 7.8-2.
Fixed in development version: 22 February 2013.
Fix included in release version: 7.8-3.
If a split set contains degenerated 8 or 20 noded brick elements, the elements will only be split if the collapsed faces are topologically adjacent across the split i.e. if collapsed faces for both elements share the same node on the split interface.
Workaround : The connectivity of the elements must be manually modified so that the collapsed faces line up across the split.
ZENCR413
Applies to: Version 7.1 to 7.8-2.
Fixed in development version: 15 November 2012.
Fix included in release version: 7.8-3.
If the first cycle in the load spectrum is an overload compared to the second cycle, it is not identified as such. This only has a consequence if a retardation model is being used in which case retardation does not occur after the overload.
Workaround : Add a cycle with a small maximum load level and high R value before the overload. The added cycle will affect the cycle count by 1 cycle per spectrum pass with negligible influence on the crack growth. The overload will be correctly initialised.
ZENCR412
Applies to: Version 7.7 to 7.8-2 (Abaqus interface on Windows with nodelocked license).
Fixed in development version: 30 October 2012.
Fix included in release version: Plug-ins version 2.3 (released with Zencrack version 7.8-3).
If a nodelocked license file is being used and the license is in a folder with a space in the path, the plug-ins for Abaqus/CAE will report an unknown error 64 when trying to verify the Zencrack license when /CAE starts up.
For the plug-ins supplied with Zencrack 7.8-1 and 7.8-2, the following message appears in an Abaqus dialog box:
RLM licence given: <path defined by LICENCE in zencrack_plugin.ini> Status (-64L) Unknown error - please check with Zentech: -64 Zencrack plug-in is not available.
For the plug-ins supplied with versions of Zencrack 7.7, the following message appears in an Abaqus dialog box:
Status (64) Unknown error - please check with Zentech: 64 Zencrack plug-in is not available
Workaround : In the zencrack_plugin.ini file include double quotes around the folder definition of the LICENSE variable e.g.:
LICENCE = "C:\Program Files\Zencrack\7.8-2\license"
ZENCR411
Applies to: Version 7.8-1 to 7.8-2.
Fixed in development version: 22 October 2012.
Fix included in release version: 7.8-3.
An analysis on a Windows machine that is set to extract results from the Abaqus .odb file will fail if the Abaqus installation folder contains a space in the path. On trying to access the .odb file the following error is reported to the screen:
The filename, directory name, or volume label syntax is incorrect.
The .rep file contains the following error:
***ERROR 1002 Cannot find file: job-output-name-zcr.001
Workaround : None.
ZENCR410
Applies to: Version 7.1 to 7.8-1.
Fixed in development version: 28 September 2012.
Fix included in release version: 7.8-2.
A crack growth analysis that uses a superposition system (*LOAD SYSTEM, TYPE=SUPERPOSITION) and the constant G integration scheme (*INTEGRATION SCHEME, TYPE=CONSTANT G) does not carry out Kic checks even if Kic values are defined.
Workaround : Use the forward predictor integration scheme.
ZENCR409
Applies to: Version 7.8 Input Formats Manual (Issue 7.8, Revision 0).
Fixed in development version: 2 October 2012.
Fix included in release version: 7.8-2 (Version 7.8 Input Formats Manual, Issue 7.8, Revision 1).
The description of parameters END CYCLE, END PASS, END TIME on the *OUTPUT keyword is incorrect in the Input Formats Manual. The descriptions should state that "the output is only generated if the integrated value is less than the value of this parameter" (instead of greater than). The default values are 1E40 rather than zero. In addition, the END CYCLE option should be described as not valid for time-only analysis (instead of fatigue-only).
ZENCR408
Applies to: Version 7.8-1.
Fixed in development version: 27 September 2012.
Fix included in release version: 7.8-2.
If value for tolerances tolzerotemp, tolzerorf, toltimezero2, tolradius1 or tolradius2 are defined on the *CONTROLS keyword, the values are not passed correctly to internal variables.
Workaround : None.
ZENCR407
Applies to: Version 7.8-1 (earlier versions not checked).
Fixed in development version: 28 September 2012.
Fix included in release version: 7.8-2.
If a value for the tolerance TOLintmet is defined on the *CONTROLS keyword, a crack growth analysis may end with an unexpected message that the analysis stopped because there was a load spectrum pass with no growth and that "All crack front nodes have stopped growing".
Workaround : None.
ZENCR406
Applies to: Rainflow version 7.4 (released with Zencrack 7.8-1).
Fixed in development version: 30 August 2012.
Fix included in release version: 7.8-2 (Rainflow version 7.41).
The Rainflow utility may return numerically zero or huge values if only a single cycle is identified after the processing to extract the reversals.
Workaround : None.
ZENCR405
Applies to: Version 7.8 Ansys Interface Manual (Issue 7.8, Revision 0).
Fixed in development version: 2 October 2012.
Fix included in release version: 7.8-2 (Version 7.8 Ansys Interface Manual, Issue 7.8, Revision 1).
Section 2.11 of the Ansys Interface Manual incorrectly implies target and contact element types 170, 173 and 174 are updated on target crack-blocks for files generated by Workbench. (Note - this type of update is added as an enhancement in version 7.8-2).
ZENCR404
Applies to: Version 7.7 to 7.8-1.
Fixed in development version: 13 September 2012.
Fix included in release version: 7.8-2.
If the TYPE=SIDES method is used to define a split set and an equal number of elements on each side of the split contribute an edge rather than a full face to the split, the program will report an access violation Fortran error at line 174 of routine jacob_2d during the generation of the first cracked mesh.
Workaround : Use the TYPE=PAIRS method as follows:
- Remove the elements contributing an edge from the *SPLIT,TYPE=SIDES definition.
- Add *OUTPUT, SPLIT SETS=YES to the input.
- Run the job and look in the .rep file for a report of the pairs data in a section marked:
***INFORMATION : SUMMARY OF ELEMENT PAIRS EXTRACTED FROM SIDE DEFINITIONS FOR SPLIT SET n - Change *SPLIT,TYPE=SIDES to *SPLIT,TYPE=PAIRS and remove the sides data lines.
- Cut/paste the pairs data to the input file and add extra pairs for the elements contributing an edge (by using the special pair definitions of "0 element number" and "element number 0" as appropriate).
ZENCR403
Applies to: Version 7.8-1 (Abaqus interface).
Fixed in development version: 30 August 2012.
Fix included in release version: 7.8-2.
In the Abaqus interface some expected field and/or history output may be missing from the .odb file if the following conditions are met:
- the Zencrack analysis requests extraction of results from the .odb file
- there is user generated *OUTPUT,FIELD and/or *OUTPUT,HISTORY request in a step of the uncracked mesh file
- other steps are subsequently defined without any *OUTPUT,FIELD or *OUTPUT,HISTORY requests.
or:
- the Zencrack analysis requests extraction of results from the .odb file
- the analysis only has one step and there is no user generated *OUTPUT,FIELD and/or *OUTPUT,HISTORY request in the step
Workaround 1 : Include at least one *OUTPUT,FIELD and/or *OUTPUT,HISTORY request in every step of the uncracked mesh file. For contour plots, the following can be included: *OUTPUT, FIELD, VARIABLE=PRESELECT.
Workaround 2 : Request extraction of results from the Abaqus .fil file instead of the .odb file. In this case there is no need to change the uncracked mesh file.
ZENCR402
Applies to: Version 7.3 to 7.9-3 (Ansys interface).
Fixed in development version: 1 December 2015.
Fix included in release version: 8.0-1.
If an Ansys uncracked mesh is identified as a "solid model" file and it contains more than one /PREP7 command, the "normal" uncracked mesh generated by Zencrack will contain one set of nodes and elements after each /PREP7 command. This results in warnings in the .rep file about duplicate nodes and elements in the uncracked mesh. In most cases these warnings do not affect the analysis and can be ignored. If the uncracked mesh is a thermal stress model in which the element type changes via the ETCHG,TTS command, Zencrack may not process any data from the Ansys run. (Note - the duplicate issue affects versions 7.3 and later, but warnings are only generated in versions 7.6 and later).
Workaround : If possible modify the data in the uncracked mesh to remove the extra /PREP7 section(s) and/or commands.
ZENCR401
Applies to: Version 7.8-1 (Ansys interface).
Fixed in development version: 13 August 2012.
Fix included in release version: 7.8-2.
If an Ansys uncracked mesh which has no 8 noded brick elements is identified as a "solid model" file and it contains more than one /SOLU command, the "normal" uncracked mesh generated by Zencrack will contain multiple sets of element data. This results in warnings in the .rep file about duplicate elements in the uncracked mesh. These warnings do not affect the analysis and can be ignored.
Workaround : None.
ZENCR400
Applies to: Version 7.7 to 7.7-2 (Ansys interface, converter versions up to 1.46).
Fixed in development version: 25 June 2012.
Fix included in release version: 7.8-1 (converter version 1.50).
The Ansys coverter outputs SFE load values using format F16.9. Any load magnitudes which cannot fit into this format will generate **************** entries in the converted file leading to an error later in the processing.
Workaround : The only workaround is to ensure load magnitudes can be written by an F16.9 format.
ZENCR399
Applies to: Version 7.1 to 7.7-2 (Ansys interface).
Fixed in development version: 25 June 2012.
Fix included in release version: 7.8-1.
Zencrack does not read SFE load magnitudes in the format created by a CDWRITE command if the load magnitude fills the entire width of the 16 character field.
Workaround : If the SFE data is in the uncracked mesh, edit the data to include a space as the 16th character of each affected field. If the problem occurs during the intermmediate processing of a solid model run, then try reducing the number of significant digits in the load data in the uncracked mesh file.
ZENCR398
Applies to: Version 7.7 to 7.7-2 (Abaqus interface).
Fixed in development version: 25 May 2012.
Fix included in release version: 7.8-1.
The plug-ins supplied for Abaqus 6.10 and later with Zencrack 7.7 releases are not compatible with Abaqus 6.12 and generate an error when Abaqus attempts to load the plug-ins on start-up of /CAE or /Viewer.
Workaround : Use Abaqus 6.11.
ZENCR397
Applies to: Version 7.3 to 7.7-2 (Ansys interface).
Fixed in development version: 23 May 2012.
Fix included in release version: 7.8-1.
ESURF options get removed from a cracked mesh file immediately after a crack front transfer if one of the crack-block positions before the transfer is:
- the last element defined in the uncracked mesh for a "normal" input file
- the highest numbered element in the uncracked mesh for a"solid model" input file
Workaround : Ensure the uncracked mesh element numbering is such that as the crack front(s) transfer, a crack-block will never replace the last or highest numbered element in the uncracked mesh.
ZENCR396
Applies to: Version 7.7 to 7.7-2 (Ansys interface on Windows only).
Fixed in development version: 22 Mar 2012.
Fix included in release version: 7.8-1.
If a non-default installation folder containing a space in the path name is used, the Ansys converter program that Zencrack uses to check data formats in the uncracked mesh does not run. An error is reported to the screen e.g: for installation within c:\program files:
'c:\program' is not recognized as an internal or external command, operable program or batch file.
The analysis continues using the unconverted uncracked mesh. In some cases this will run and in others it may cause problems of unrecognised data formats.
Workaround : The recommendation is that Ansys users on Windows should install Zencrack in a top level folder without a space in the pathname.
ZENCR395
Applies to: Version 7.1 to 7.7-2 (Windows only).
Fixed in development version: 21 Mar 2012.
Fix included in release version: 7.8-1.
If a non-default installation folder containing a space in the path name is used, the runzcrXX command will produce an error e.g: for installation within c:\program files:
'c:\program' is not recognized as an internal or external command, operable program or batch file.
Workaround : Several lines in the install folder\tools\runzcrXX.bat file require modification (where XX will be the particular program version number). Edit the .bat file, locate the following lines and add double quotes to them, as below. Note that the lines are distributed through the file rather than being consecutive. Ansys users should also refer to bug ZENCR396.
Original:
call %ZENDIR%\tools\runzcr77_network_mode.bat %ZENDIR%\utils\3dmesh.exe %ZENDIR%\utils\process.exe %2 %3 %4 %5 %6 %7 %8 %9 %ZENDIR%\utils\tanh.exe %2 %ZENDIR%\utils\rainflow.exe %2 %raintime% %raindebug%
Modified:
call "%ZENDIR%\tools\runzcr77_network_mode.bat" "%ZENDIR%\utils\3dmesh.exe" "%ZENDIR%\utils\process.exe" %2 %3 %4 %5 %6 %7 %8 %9 "%ZENDIR%\utils\tanh.exe" %2 "%ZENDIR%\utils\rainflow.exe" %2 %raintime% %raindebug%
ZENCR394
Applies to: Version 7.3 to 7.7 (Ansys interface).
Fixed in development version: 10 Jan 2012.
Fix included in release version: 7.7-1.
During temperature processing, output of a BF data line with zero
node number could occur when a single BF option is present and it does
not contain a node number. This type of issue could arise, for example,
if certain constructs are used to apply a temperature to all nodes in
the model e.g.:
nsle
bf,all,temp,20
Workaround : List all nodes explicitly on multiple BF lines.
ZENCR393
Applies to: Version 7.7 plug-ins for Abaqus/CAE.
Fixed in development version: 27 Jun 2011.
Fix included in release version: 7.7-2.
Due to a change in command line options between Abaqus 6.10 and 6.11, the previous unique option "cpu" must be replaced by "cpus" in v6.11. The plug-ins were released before Abaqus 6.11 was available and use the "cpu" option in the "Run Abaqus" screen when Abaqus is executed. This causes the Abaqus job to abort when using v6.11.
Workaround : None. A fix is available. Contact Zentech International and request the fix (stating whether it is required for 32bit Windows, 64bit Windows or 64bit Linux).
ZENCR392
Applies to: Version 7.1 to 7.7.
Fixed in development version: 21 Jun 2011.
Fix included in release version: 7.7-1.
A missing square term in conversion of G to K during Kic checks could result is a value of K that is too small being used in the check against Kic. The crack would then continue growing with K values higher than the specified Kic. This could only occur if the constant G integration scheme was being used.
Workaround : A workaround would be to input modified Kic values:
a) Use the real Kic and E,nu to calculate a value of Gic.
b) Then input the value of Kic/sqrt(Gic) in the .zcr file in place of the the real Kic value.
Note that this has a limitation if there is temperature dependency since
the E and nu at the failure point would not be known for the up-front
calculation of Gic.
ZENCR391
Applies to: Process utility version 7.19.
Fixed in development version: 3 May 2011.
Fix included in release version: 7.8-1 (Process utility version 7.20 and later).
Version 7.19 of the Process utility (released with Zencrack 7.7) is unable to read raw contour integral values from a .rep file created by a Zencrack version earlier than v7.5-13.
Workaround : None.
ZENCR390
Applies to: Version 7.7.
Fixed in development version: 14 Apr 2011.
Fix included in release version: 7.7-1.
In some configurations a split element on one side of the crack face may be connected to two element faces on the other side of the crack face when using the *SPLIT,TYPE=SIDES option. The affected type of configuration should require two entries in the internal pairs list in order to separate the elements correctly, but only one pair was being generated.
Workaround : Use *OUTPUT,SPLIT SETS=YES to get a list of the identified pairs, then manually add any missing pairs to the list and run the job using *SPLIT,TYPE=PAIRS with the revised pairs list.
ZENCR389
Applies to: Documentation Issue 7.7 Revision 0.
Fixed in development version: June 2012.
Fix included in release version: Documentation Issue 7.8 Revision 0.
The following errors have been noted in the documentation:
- The description of the RESTART keyword in the Input Formats Manual should state that Ansys models that are processed using the solid-model option cannot use the RESTART option. A possible workaround for such models is to export the uncracked mesh model from Ansys/Mechanical APDL using the CDWRITE option and then use the exported version of the model as the uncracked mesh. Depending upon the contents of the file, some manual edits may also be required in the exported file.
-
The Examples Manual and supplied input files have incorrect options for
the superposition load system in examples a05d01 and a05d02. The options
used are old formats that should have been updated. Although the
formats are accepted with a warning when the files are run, the syntax
should be:
For a05d01:
*LOAD SYSTEM,TYPE=SUPERPOSITION,METHOD=CYCLIC-STATIC,CYCLIC=2
1
For a05d02:
*LOAD SYSTEM,TYPE=SUPERPOSITION,METHOD=MAX-MIN,MINIMUM=1,MAXIMUM=2 - The tabular data section of the User Manual (section 6.3.3) states that da/dt vs K data has to be given at the same da/dt values for each curve. This is not the case. It also states that (by implication) each curve must have the same number of points. This is also not correct.
ZENCR388
Applies to: Version 7.0 to 7.6 (Marc interface only).
Fixed in development version: 7 Dec 2010.
Fix included in release version: 7.7.
Marc element types with 15 nodes are not processed correctly and the resulting cracked mesh will contain errors.
Workaround : None.
ZENCR387
Applies to: Version 7.0 to 7.6 (Abaqus interface only).
Fixed in development version: 5 Oct 2010.
Fix included in release version: 7.7.
A blank line in the middle of the Abaqus *NODE data for the uncracked mesh is not removed by Zencrack. This causes Abaqus to give an error:
***ERROR: in keyword *NODE, file "filenamer.inp", line XXXX: Node label is
not a positive integer: 0
Workaround : Remove the blank line from the uncracked mesh file.
ZENCR386
Applies to: Version 7.0 to 7.6.
Fixed in development version: 12 Oct 2010.
Fix included in release version: 7.7.
If the forward predictor is used with numerical integration, the analysis may appear to hang during integration pass 2 if the following conditions are met:
- a node falls below threshold during the integration step
- the number of cycles in the step for that node up to threshold is less than the value being used to integrate the step in pass 2
Workaround : Use the constant G integration scheme instead of the forward predictor.
ZENCR385
Applies to: Version 7.0 to 7.6.
Fixed in development version: 4 Oct 2010.
Fix included in release version: 7.7.
An analysis with a superposition load system will abort or appear to hang during the integration phase if both Gmin and Gmax are zero.
Workaround : No workaround as such, but check input data to determine why Gmin and Gmax are both zero.
ZENCR384
Applies to: Version 7.6 documentation.
Fixed in development version: 1 Dec 2010.
Fix included in release version: 7.7.
User Manual:
Section 11.3 Section 11.3 should state "Limits specific to Abaqus are:" in the line of text (rather than Ansys).
On page 87 the contents of figure 7-3 show the example for "multiple block spectrum loading" rather than the intended example of "random spectrum loading". The figure for the random spectrum loading example is not contained in the manual.
On page 89 the sample input is missing the data line containing the
scale factors for the constant amplitude load system. The correct
example is:
*LOAD SYSTEM, TYPE=STATIC, ENERGY=1
*LOAD SYSTEM, TYPE=CONSTANT AMPLITUDE, ENERGY=2
0.4 1.0
*LOAD SYSTEM, TYPE=SUPERPOSITION, METHOD=CYCLIC-STATIC, CYCLIC=2
1
*SPECTRUM
3
Crack-block Manual:
On page 9 the corner crack-block name should be l02_q3968x16 not l02_q3968x8 (top left picture on the page).
User Subroutines Manual:
In section 17 the header line is incorrect and should be "SUBROUTINE user_prefea" instead of "SUBROUTINE user_postfea".
ZENCR383
Applies to: Version 7.5 and 7.6.
Fixed in development version: 6 Apr 2010.
Fix included in release version: 7.7.
If the number of results sets saved from the f.e. analysis is greater than the number of load systems and the input contains a mixture of "search" load systems of type=maximum (or minimum),energy=search and regular load systems of type=constant amplitude (or static or spectrum), some arrays may not be allocated properly resulting in an invalid analysis. The analysis is likely to stop with an unexpected error. Mixing minimum or maximum search load systems with static, constant amplitude or spectrum load systems is not likely to occur in a practical analysis so this issue is unlikely to be encountered.
Workaround : Do not define minimum/maximum search load systems and static/constant amplitude/spectrum load systems in the same input file.
ZENCR382
Applies to: Version 7.1 to 7.6.
Fixed in development version: 9 Mar 2010.
Fix included in release version: 7.7.
An error in a data check prevents use of the constant G integration scheme if there is temperature dependent crack growth data. The program incorrectly stops with error 2168.
Workaround : None.
ZENCR381
Applies to: All Windows versions up to and including v7.6.
Fixed in development version: 15 Dec 2009.
Fix included in release version: v7.6 downloads updated to include this fix on 22 Dec 2009.
The setup program for Windows versions of Zencrack does not allow installation on Windows 7 systems and instead reports an unidentified operating system.
Workaround : No workaround will be supplied for v7.5. Downloads have been updated for v7.6.
ZENCR380
Applies to: setupzcr75_xx (program date 20 Feb 2008 or later) and setupzcr76_xx (32 and 64 bit versions) (Windows only).
Fixed in development version: 15 Dec 2009.
Fix included in release version: v7.6 downloads updated to include this fix on 22 Dec 2009.
The setup program for Windows has two issues with Fortran 10 configuration:
- if option (3) for Fortran 10.0 is used the configuration has the wrong path to the Fortran folder
- input of option (4) for configuration with Intel Fortran version 10.1 is not permitted
Workaround : No workaround will be supplied for v7.5. Downloads have been updated for v7.6.
ZENCR379
Applies to: Process versions up to and including 7.18.
Fixed in development version: 8 Jun 2009.
Fix included in release version: 7.19.
The Process utility program does not extract results correctly if not all f.e. results sets are referenced by load systems.
Workaround : None.
ZENCR378
Applies to: Process versions up to and including 7.18.
Fixed in development version: 3 Jun 2009.
Fix included in release version: 7.19.
The Process utility program does not report correct values of the nodal G errors if they are in the range 1e-5< |error| <1e5.
Workaround : None.
ZENCR377
Applies to: Version 7.1 to 7.6 (Abaqus interface only).
Fixed in development version: 1 Dec 2009.
Fix included in release version: 7.7.
Contour integral results from the Abaqus analysis may not be extracted correctly if all the following conditions are met (some contour values will be zero in the .rep file but are non-zero in the Abaqus .dat file):
- The option to extract results from the .fil file is used
- There is more than one crack front in the model
- Different crack fronts have different numbers of contours due to the use of different crack-block families
Workaround : Define the crack fronts in a specific order from those having lowest to those having highest number of contours.
ZENCR376
Applies to: Version 7.3e to 7.6 (Abaqus interface only).
Fixed in development version: 16 Jul 2009.
Fix included in release version: 7.7.
Error 3103 is intended to trap element sets with a GENERATE option that exceed 2000 data lines. But any element set exceeding 2000 data lines causes the error.
Workaround : Split the element set into blocks of <2000 data lines.
ZENCR375
Applies to: Version 7.1 to 7.6.
Fixed in development version: 20 Mar 2009.
Fix included in release version: 7.7.
Combined time and fatigue integration produce erroneous results with the cycle count getting out of step with the integrated time if the fatigue spectrum contains only one cycle.
Workaround : None.
ZENCR374
Applies to: Versions 7.4 and 7.5 (Abaqus interface only).
Fixed in development version: 12 Feb 2009.
Fix included in release version: 7.6
If a data line in the *BOUNDARY section of the uncracked mesh contains only a node number or set name (without a comma or any other data) then an error is reported but the error message is not formatted correctly. The error should report that Zencrack requires the first degree of freedom field to be defined.
Workaround : Define the first degree of freedom.
ZENCR373
Applies to: Version 7.1 and later.
Fixed in development version: 6 Feb 2009.
Fix included in release version: 7.6
A power factor controls the step size for the first 10 steps of the error control scheme (see data line 2 of the *TOLERANCE keyword). If tolerances are met within all 10 of these steps the power factor should de-activate from step 11. But instead it is also applied in step 11. This results in a slightly bigger step size at step 11 than is expected. The overall analysis is valid but the step size control at step 11 is not as documented.
Workaround : None required.
ZENCR372
Applies to: Version 7.1 and later.
Fixed in development version: 22 Dec 2008.
Fix included in release version: 7.6
In some cases the user defined crack front option may generate a merge error between crack-blocks on opposite sides of the crack face reporting that too many or too few nodes have been merged between a pair of crack-blocks.
Workaround : None.
ZENCR371
Applies to: Version 7.1 and later.
Fixed in development version: 21 Aug 2008.
Fix included in release version: 7.6
If the top level folder name of the Zencrack installation is longer than 62 characters then an error will occur due to the program not being able to correctly access some files within the installation folders. The exact error depends upon the length of the folder name and the Zencrack version.
This does not affect the filename and path length for job files (which may be up to 256 characters).
Workaround : Ensure that the top level folder name is 62 characters or less. If the folder name is changed after running the tools\setupzcr program then the setup program will need to be re-run to update some configuration files.
ZENCR370
Applies to: Version 7.1 to 7.5-12 (Abaqus interface only).
Fixed in development version: 7 Jul 2008.
Fix included in release version: 7.5-13
If an element set name containing the sub-string "TYPE" is defined using the ELSET parameter on a *ELEMENT keyword line the following error may occur:
***ERROR 3051 ELEMENT X IS REPLACED BY CRACK-BLOCK Y BUT ELEMENT X IS NOT AN 8 OR 20 NODED ELEMENT.
Workaround : Change the element set name so that the string does not contain "TYPE" or make the TYPE= parameter the last option on the data line.
ZENCR369
Applies to: Version 7.1 and later.
Fixed in development version: no fix needed - the problem is related to poor midisde nodes in the uncracked mesh. See workarounds below.
If the initial crack is defined by the INITIAL=SIZE option it is possible that an error 4077 will be generated indicating inconsistent initial crack size data even though the input data is correct. This can happen if the replaced elements have midside nodes that are away from the proper midside positions. In such cases there will be warnings in the .rep file like the one shown below in workaround #2.
Workaround #1: Modify the midside node positions so that they are at the midsides.
Workaround #2: Modify the tolerance that determines whether a midside node is considered to be away from the midside position. This tolerance, TOLMIDWARN1, has a default value of 0.9. In the rep file you will see warnings like this one:
***WARNING MIDSIDE NODE 95 BETWEEN CORNER NODES 99 AND 94 ON ELEMENT 10 WHICH IS REPLACED BY CRACK-BLOCK 3 IS AWAY FROM THE MIDSIDE POSITION: DISTANCE CORNER NODE 1 - MIDSIDE = 3.473309E-01 DISTANCE MIDSIDE - CORNER NODE 2 = 2.781758E-01 RATIO (EXPECTED TO BE 1.0) = 0.800896 THIS MAY AFFECT THE INITIAL CRACK SIZE AND ADVERSELY AFFECT ELEMENT DISTRIBUTIONS IN THE CRACK-BLOCK. PLEASE CHECK YOUR INITIAL CRACKED MESH CAREFULLY. YOU ARE ADVISED TO MOVE NODE 95 TO A BETTER MIDSIDE POSITION. A SUGGESTED POSITION IS: 9.943850E+00 9.943850E+00 0.000000E+00
Determine the lowest ratio reported on all such warnings. Then set the
tolerance TOLMIDWARN1 lower than this value. Alternatively just use a
very low value for the tolerance e.g.:
*CONTROLS, TOLMIDWARN1=0.1
ZENCR368
Applies to: Version 7.4 to 7.5-12 (Finas interface only).
Fixed in development version: 9 Jun 2008.
Fix included in release version: 7.5-13
Some boundary condition data lines can be incorrectly formatted in the cracked mesh and cause a Finas error if all the following conditions are met:
- The "from..to" method is used for boundary condition degrees of freedom
- The crack-block(s) transfer to a new location
- There are supports on the crack plane (i.e. symmetry in the crack plane) that need to be removed as the crack-blocks transfer.
Workaround: Re-write the affected data lines in the uncracked mesh with one degree of freedom per data line. For example replace:
BOUNDARY
1 Step 0 Constraint Set
1 1 3
by
BOUNDARY
1 Step 0 Constraint Set
1 1
1 2
1 3
ZENCR367
Applies to: Version 7.5 to 7.5-12.
Fixed in development version: 28 May 2008.
Fix included in release version: 7.5-13
The options on the SAVE keyword to save and re-number the .rst and .rth files from Ansys analyses during crack growth prediction do not save the files.
Workaround: None.
ZENCR366
Applies to: Versions 7.1 to 7.5-9.
Fixed in development version: 23 Apr 2008.
Fix included in release version: 7.5-10
In cases where the crack-block has several curved outer faces the mapping algorithms may fail with an error number 4095 or 4125 depending upon the mapping options used and whether or not temperature interpolation is required to generate the cracked mesh.
Workaround: None.
ZENCR365
Applies to: Version 7.4 and later.
Fixed in development version: 3 Dec 2008.
Fix included in release version: 7.6
The spline fixup option on keyword CRACK FRONT may produce a distorted mesh for step 2 of a growth analysis if the crack-blocks are of significantly unequal lengths along the initial crack front.
Workaround: None.
ZENCR364
Applies to: Versions 7.4 to 7.5-8.
Fixed in development version: 15 Apr 2008.
Fix included in release version: 7.5-9
The spline fixup option on keyword CRACK FRONT may produce distorted external shape to the crack-blocks if the crack-blocks are skewed significantly from the outer surface of a model. This may ultimately result in inside-out elements.
Workaround: None.
ZENCR363
Applies to: Versions 7.5 to 7.5-8.
Fixed in development version: 2 Apr 2008.
Fix included in release version: 7.5-9
The enhanced ring control mapping option may produce poorly distributed elements in the crack-blocks after a reasonable amount of non-planar growth. The elements are biased too far from the crack plane and may ultimately produce an inside-out element error.
Workaround: None.
ZENCR362
Applies to: Versions 7.4 to 7.5-8.
Fixed in development version: 9 Apr 2008.
Fix included in release version: 7.5-9
A model containing 8 noded brick elements and using multiple nodes at each crack front nodal position does not generate a valid cracked mesh after the first crack-block transfer when *BOUNDARY SHIFT, TYPE=TRANSFER is activated.
Workaround: None.
ZENCR361
Applies to: Version 7.5-8 (Windows 32bit version only).
Fixed in development version: 7 Apr 2008.
Fix included in release version: 7.5-9
A model containing 8 noded brick elements and using multiple nodes at each crack front nodal position does not generate a valid cracked mesh.
Workaround: None.
ZENCR360
Applies to: Versions 7.5-2 to 7.5-8.
Fixed in development version: 4 Apr 2008.
Fix included in release version: Process v7.16
The data for columns "error node" and "error largest" is not extracted correctly by version 7.15 of the PROCESS post-processor for .rep files generated by Zencrack 7.5-2 or later. Instead the two relevant columns in the .csv file contain zeros or a Fortran error will occur and no .csv file will be created.
Workaround: None.
ZENCR359
Applies to: Versions 7.0 to 7.5-8 (Windows only).
Fixed in development version: 22 Dec 2008.
Fix included in release version: 7.6
A linker error occurs when using a Zencrack user subroutine if the Zencrack top level folder name contains any spaces.
Workaround: Edit the runzcr7x.bat file in the tools folder and change all occurences of the string:
%ZCROBJ% %ZCRLIB%
to:
"%ZCROBJ%" "%ZCRLIB%"
In addition for v7.5-8 change all occurences of the string:
%ZENDIR%\bin\rlmclient.lib
to:
"%ZENDIR%\bin\rlmclient.lib"
ZENCR358
Applies to: Versions 7.0 to 7.5 (Windows only).
Fixed in development version: 19 Feb 2008.
Fix included in release version: 7.5-8
A linker error occurs when using a Zencrack user subroutine if the Visual Studio environment is VS2005.
Workaround: Contact Zentech.
ZENCR357
Applies to: Versions 7.4 to 7.5 (Finas interface only).
Fixed in development version: 7 Feb 2008.
Fix included in release version: 7.5-7
If more than two sets of temperatues are defined in the model then the second and subsequent sets are not processed correctly for interpolation into the crack-blocks.
Workaround: None.
ZENCR356
Applies to: Versions 7.1 to 7.5.
Fixed in development version: 21 Dec 2007.
Fix included in release version: 7.5-7
If boundary shifting is on in a crack growth analysis the crack-blocks may rotate in the opposite direction to the way the crack is growing. This is likely to result in the analysis failing with distorted elements.
Workaround: This is most likely to occur if absolute coordinate values are small e.g. units are metres. If this is the case change to mm units.
ZENCR355
Applies to: Versions 7.0 to 7.5 (Abaqus interface only).
Fixed in development version: 4 Dec 2007.
Fix included in release version: 7.5-7
If user subroutines zcr-jint or zcr-ctint are used to extract results from the Abaqus .fil file and a step has more than 100 increments the data cannot be processed by Zencrack.
Workaround: Extract results from the .dat file.
ZENCR354
Applies to: Versions 7.0 to 7.5.
Fixed in development version: 6 Nov 2007.
Fix included in release version: 7.5-6
When more than 80 load systems are defined there are some incorrectly initialised arrays. Subsequent lookups during crack growth integration return invalid data resulting in an error e.g. Error 6018.
Workaround: None.
ZENCR353
Applies to: Versions 7.5.
Fixed in development version: 31 Oct 2007.
Fix included in release version: 7.5-6
When the fix for zencr345 was done the value of NUMBERcfnodes was changed to include all crack front nodes, not just corner nodes as stated in the documentation. The code is now changed to be consistent with the v7.5 documentation in which NUMBERcfnodes is the number of corner nodes on the crack front. This affected all user subroutines which have NUMBERcfnodes in the argument list.
Workaround: None.
ZENCR352
Applies to: Versions 7.1 to 7.5.
Fixed in development version: 30 Oct 2007.
Fix included in release version: 7.5-6
The second node in a split pair may not be placed correctly onto the curved outer surface of the element when boundary shift is on and relax and surround fix are off. The problem occurs for the midside node in the element next to the crack-block i.e. the midside node affected by a boundary shift.
Workaround: None.
ZENCR351
Applies to: Versions 7.1 to 7.5.
Fixed in development version: 19 Oct 2007.
Fix included in release version: 7.5-4
In an analysis using Abaqus or Ansys with multiple results sets at different temperatures the internal cross referencing of temperature values used during crack growth integration is incorrect. The temperature is always derived from the final results set rather than the results set being used to drive the integration. If, and only if, there are temperature dependent material properties this will affect the result of the integration. In most cases the effect is considered likely to be small. Only the crack growth integration is affected - the processing of individual results sets and the report of G, K, E and nu values for each result set is correct.
Workaround: None.
ZENCR350
Applies to: Version 7.5.
Fixed in development version: 21 Sep 2007.
Fix included in release version: 7.5-2
The maximum or minimum displacement-based Ki, Kii and Kiii reported in the search table summary in the .rep file are only valid if the analysis is using displacement-based results to drive the crack growth. If energy values are used to drive the growth the K values reported in the search table are their initialisation values of 0 or 1e20 - the analysis itself is correct only the Ks in the search output table are incorrect.
Workaround: None.
ZENCR349
Applies to: Versions 7.1 to 7.5.
Fixed in development version: 10 Sep 2007.
Fix included in release version: 7.5-1
In some circumstances a time dependent or combined time-fatigue crack growth analysis may hang during crack growth integration. If the option *OUTPUT, PROGRESS=YES is used the last message output to the screen is:
Starting crack growth integration
Integrating crack front 1, node 1 : pass 1
Workaround: None.
ZENCR348
Applies to: Version 7.4a to 7.6 (Abaqus interface only).
Fixed in development version: 1 Dec 2009.
Fix included in release version: 7.7.
Zencrack will report error 5020 "end of file found unexpectedly while trying to read temperature data" when a job meets the following criteria:
- there are temperatures in the finite element analysis
- the model has more than one crack front
- results are extracted from the Abaqus .fil file using subroutine zcr-jint or zcr-ctint
Workaround: Extract results from the .dat file instead.
ZENCR347
Applies to: 3dmesh utility program version 7.06 (released with Zencrack 7.5).
Fixed in development version: 26 Feb 2008.
Fix included in release version: 7.5-8
When a Finas output file is requested the program appears to hang after issuing the report "Processing XX crack front positions".
Workaround: At the point where the program hangs it is waiting for input that has no meaning for a Finas output file. Enter a 2 and the program will continue normally.
ZENCR346
Applies to: Version 7.1 to 7.5 (Unix and Linux only).
Fixed in development version: 19 Feb 2008.
Fix included in release version: 7.5-8
If an existing Zencrack executable is used the command line -e option
requires a "./" in front of the executable name. This is not documented
in the manuals or the runzcr75 script file. e.g.
runzcr75 -j jobname -e ./my_saved_exe.exe
ZENCR345
Applies to: Version 7.1 to 7.4a.
Fixed in development version: 11 May 2007.
Fix included in release version: 7.5
The node number and coordinate data passed into crack growth user subroutines is incorrect for models consisting of 8 noded bricks. These are secondary variables passed into the routine for information and are not usually critical. The value of da/dn is handled correctly provided it is not a function of any of these secondary variables.
Workaround: Use 20 noded bricks.
ZENCR344
Applies to: Version 7.0 to 7.4a.
Fixed in development version: 13 Apr 2007.
Fix included in release version: 7.5
The calculation of displacement based SIFs may not be accurate for models that have 8 noded bricks and a curved crack front due to a poor estimate of the distance from a crack face node to the crack front.
Workaround: Use 20 noded bricks.
ZENCR343
Applies to: Version 7.4, 7.4a (Marc interface only).
Fixed in development version: 12 Apr 2007.
Fix included in release version: 7.5
The field width of temperatures output after processing a temperature option is too wide and the temperatures may subsequently be incorrectly read by Marc when analysing the cracked mesh.
Workaround: None.
ZENCR342
Applies to: Version 7.4, 7.4a (Marc interface only).
Fixed in development version: 11 Apr 2007.
Fix included in release version: 7.5
Processing of data from Marc data blocks is not correct for the following cases and could cause incorrect results or the job to fail:
- reading the first data line of DISP CHANGE or TEMP CHANGE
- processing multiple alphnumeric data lines for temperature, fixed displacement and loading updates
- fixed format displacements which fill their field width.
Workaround: None.
ZENCR341
Applies to: Version 7.4, 7.4a (Marc interface only).
Fixed in development version: 3 Apr 2007.
Fix included in release version: 7.5
An array allocation error will occur if Zencrack processes a temperature definition block in which a node set is used rather than a node list.
Workaround: Use a node list rather than a node set.
ZENCR340
Applies to: Version 7.4, 7.4a - Utility program PROCESS.
Fixed in development version: 12 Feb 2007.
Fix included in release version: 7.5
For a time dependent growth analysis or a combined time-fatigue growth analysis the data for the time column (column H) is not read correctly and is always output as zero.
Workaround: None.
ZENCR339
Applies to: Version 7.0 to 7.4a.
Fixed in development version: 11 Jan 2007.
Fix included in release version: 7.5
In a model with multiple crack fronts the determination of whether or not a crack front is allowed to transfer could be incorrect if:
- at least one crack front exists that contains only through crack-blocks, and,
- at least one crack front exists that contains one or more quarter circular crack-blocks.
Workaround: Define all crack fronts that contain only through crack-blocks before any crack fronts that contain quarter circular crack-blocks.
ZENCR338
Applies to: Version 7.4, 7.4a (Finas interface only).
Fixed in development version: 11 Jan 2007.
Fix included in release version: 7.5
Connectivity output is incorrectly written for split elements when the wide data format is used.
Workaround: Use the narrow data format.
ZENCR337
Applies to: Version 7.0 to 7.4a.
Fixed in development version: 3 Jan 2007.
Fix included in release version: 7.5
In some limited cases the numerical integration scheme can get stuck in an infinite loop. The program appears to hang at the command prompt.
Workaround: None.
ZENCR336
Applies to: Version 7.4, 7.4a.
Fixed in development version: 14 Dec 2006.
Fix included in release version: 7.5
If a large number of nodes are listed on a boundary constraint option in the uncracked mesh the update is done correctly but can take an unexpectedly long elapsed time.
Workaround: None.
ZENCR335
Applies to: Version 7.4 (Abaqus and Marc interfaces).
Fixed in development version: 14 Dec 2006.
Fix included in release version: 7.5
If an existing deltau.val output file is set read-only any job run in the same directory stops with a Fortran error when the first attempt is made to use the deltau.val file. The deltau.val file can exist for the Marc interface or the Abaqus interface (only when results are extracted from the Abaqus .fil file).
Workaround: Turn off the read-only attribute for the file.
ZENCR334
Applies to: Version 7.4, 7.4a.
Fixed in development version: 20 Sep 2006.
Fix included in release version: 7.5
The values of the parameters FIRST and STEP are not read correctly for the TYPE=FROM REMOTE option on the REMOTE TRANSFER keyword. Regardless of the input both parameter values are set to 1.
Workaround: None.
ZENCR333
Applies to: Version 7.4.
Fixed in development version: 28 Jul 2006.
Fix included in release version: 7.4a
If there are multiple crack-blocks along a crack front the program may not correctly identify the intersection points of the user defined crack front and the crack-block edges. This can result in an error reported during the mapping process.
Workaround: Reversing the order of points on the user defined crack front may solve the problem in some cases.
ZENCR332
Applies to: Version 7.4 - Utility program PROCESS.
Fixed in development version: 27 Jul 2006.
Fix included in release version: 7.4a
Raw contour values reported by utility program PROCESS are all zero in v7.4 (i.e. the contents of columns BC onwards).
Workaround: Edit the .rep file and change header text on tables listing the raw contour integral values in the rep file from:
CRACK SET NODE CONTOUR 1
to:
CRACK SET NODE CONTOUR 1
ZENCR331
Applies to: Version 7.3 to 7.4.
Fixed in development version: 4 Jul 2006.
Fix included in release version: 7.4a
The error control scheme does control step sizes between f.e. analyses correctly if results sets exist that are not used in load systems. The error can be reported as outside the tolerance value when it isn't. Even if the error is reported as inside the tolerance, the step size is not calculated correctly. For example:
LARGEST ERROR ON G FOR THIS STEP IS 0.078796% [definitely < 5% !!] LARGEST ERROR ON G SO FAR IS -4.793779% AVERAGE ERROR ON G SO FAR IS -0.061326% NUMBER OF ERROR EVALUATIONS 5 ERROR IS OUTSIDE THE TOLERANCE VALUE OF 5.000000% (75.0% OF TOLERANCE USED FOR THIS CHECK) SCALE FACTOR ON da FOR NEXT STEP IS 9.98424E-01
Workaround: For the Abaqus interface a workaround is to use the STEPS option on *ENERGY RELEASE RATE to turn off j-integral requests for all steps except the ones referenced in load systems. Otherwise, use the adaptive or fixed schemes on *TOLERANCE.
ZENCR330
Applies to: Version 7.1 to 7.4.
Fixed in development version: 4 Jul 2006.
Fix included in release version: 7.4a
Possible Fortran error when using a user defined crack front. The error occurs in a mapping routine because the internally calculated crack-block edge ratio values are incorrect. This is due to incorrect calculation of the points where the user defined crack front intersects the crack-block edges.
Workaround: Try reversing the order of the spline points in the input file.
ZENCR329
Applies to: Version 7.0 to 7.4.
Fixed in development version: 4 Jul 2006.
Fix included in release version: 7.4a
If existing output files are set read-only and the job is re-run using the overwrite=yes option, the existing files cannot be deleted and Zencrack stops with a Fortran error when the first attempt is made to use one of the existing files.
Workaround: Turn off the read-only attribute for the files.
ZENCR328
Applies to: Version 7.1 to 7.4.
Fixed in development version: 4 Jul 2006.
Fix included in release version: 7.4a
The value of constraint set by the user in user subroutine user_constraint is not assigned to the internal variable in the calling routine.
Workaround: None.
ZENCR327
Applies to: Version 7.4 (Abaqus interface only).
Fixed in development version: 6 Jun 2006.
Fix included in release version: 7.4a
Surface update on all faces of an element number or element set name on the Abaqus *SURFACE option (i.e. data specified without any face id) causes an array bounds error at line 160 of surfabq if the data line has no comma.
Workaround: Add a comma after the element number or set name.
ZENCR326
Applies to: Version 7.3 to 7.4 - Utility program PROCESS.
Fixed in development version: 27 Jul 2006.
Fix included in release version: 7.4a
Utility program PROCESS searches the rep file for table headings. If any of these headings were included as part of commented data on a '#' line in the zcr file, the process program fails with a fortran error.
Workaround: Locate the "echo of zcr file" section in the rep file and remove any comment lines that include table headings from a rep file. There is no need to re-run the Zencrack analysis.
ZENCR325
Applies to: Version 7.1 to 7.4 (Ansys interface only).
Fixed in development version: 10 May 2006.
Fix included in release version: 7.4a
Zencrack job fails with a Fortran error in greadansys after the first Ansys analysis. This is due to the coordinate fields in the .001 output file from Ansys merging into one another as a result of very large coordinates ( >9999999 or <-999999).
Workaround: Modify the unit system so that the coordinate values are smaller than abs(999999).
ZENCR324
Applies to: Version 7.1 to 7.4 (Marc interface only).
Fixed in development version: 3 May 2006.
Fix included in release version: 7.4a
Energy based results for an embedded crack (i.e. a crack front which forms a closed loop) are incorrect.
Workaround: Use *ENERGY RELEASE RATE, PREDICTION=DISPLACEMENT.
ZENCR323
Applies to: Version 7.3 to 7.4.
Fixed in development version: 20 Apr 2006.
Fix included in release version: 7.4a
Extrapolation of displacement based results to the crack front nodes may be slightly incorrect for some models with only one element along the crack front. Ks are likely to be out by less than 1% if there is a problem.
Workaround: None.
ZENCR322
Applies to: Version 7.4.
Fixed in development version: 10 Feb 2006.
Fix included in release version: 7.4a
The description in the user manual of the variable IFLIPSTATUS in the argument list of user subroutine user_transfercf is inconsistent with what is done in Zencrack 7.4. In version 7.4 the value of this variable must be set to greater than [0.5 x number of crack-blocks along the crack front] if transfer is required and not to 1 as indicated in the manual.
Workaround: Set the return value of IFLIPSTATUS to 0 for "no transfer" or a high number if transfer is required (e.g. 1000).
ZENCR321
Applies to: Version 7.4 (Abaqus interface only).
Fixed in development version: 19 Dec 2005.
Fix included in release version: 7.4a
The request for output of K FACTORS from an Abaqus analysis causes j-integral data to be read incorrectly if all of the following conditions are met:
- There is more than 1 crack front in the model.
- There are multiple sets of j-integral results data (i.e. multiple load steps or multiple increment output within one step in the f.e. analysis).
- Results data is requested from the .fil file using the zcr-jint subroutine.
Workaround:
- Remove the request for K FACTORS ouput, or,
- Extract results from the .dat file.
ZENCR320
Applies to: Version 7.0 to 7.4 (Abaqus interface only).
Fixed in development version: 25 Nov 2005.
Fix included in release version: 7.4a
Please see the note "Using Zencrack with Abaqus 6.5-4 (or later)" on the Abaqus interface support page for information on this item.
ZENCR319
Applies to: Version 7.1 to 7.4.
Fixed in development version: 2 Nov 2005.
Fix included in release version: 7.4a
The program may fail with an error 4034 when a user defined initial crack front is used. This can occur if there is more than 1 crack-block along the crack front and the crack-blocks are distorted to be long and thin in the crack plane (e.g. 10:1).
Workaround: None.
ZENCR318
Applies to: Version 7.1 to 7.4.
Fixed in development version: 11 Nov 2005.
Fix included in release version: 7.4a
If the *CRACK FRONT, INITIAL=POINTS option is used to define the initial crack front the program will stop unnecessarily with error 4077 under two conditions:
- if the user enters values for ratio or size in fields 4 and 5 of data line 2 for the crack-blocks and the values are inconsistent between adjacent crack-blocks
- if there is a mixture of normal and reversed crack-blocks on the crack front and no data is entered in fields 4 and 5.
Workaround: When using INITIAL=POINTS for a crack front definition, ensure that all data items in fields 4 and 5 for the crack-blocks are consistent. Since these values are not used by the analysis, any suitable values can be defined that will pass the data check. The simplest approach that will work in all cases is to use 0.5 for both fields for all crack-blocks.
ZENCR317
Applies to: Version 7.0 to 7.4.
Fixed in development version: 4 Nov 2005.
Fix included in release version: 7.4a
A certain configuration of one normal and one reversed quarter circular crack-block is not identified as being on the same crack front. This occurs if the orientations are such that a 1-4 edge on one crack-block is adjacent to a 1-4 edge on another crack-block. The program stops with errors 4107 and 4108. (The 1-4 edge is the "edge 2" described in the user manual figure "Definition of nodes for crack orientation").
Workaround: None.
ZENCR316
Applies to: Version 7.4.
Fixed in development version: 18 Oct 2005.
Fix included in release version: 7.4a
The option on *MAPPING to activate biasing of nodes within the crack front element rings may cause a run-time error if:
- there is more than one crack-block along a crack front, or,
- both sides of a crack are modelled.
The error occurs only in a growth analysis. The program fails during
generation of the second f.e. mesh (the first mesh is correct). If
boundary transfer is on and there is transfer after the first f.e.
analysis, the error occurs after completion of the first f.e. analysis
for which there was no transfer.
Workaround: Turn off biasing or use Zencrack 7.3.
ZENCR315
Applies to: User manual for version 7.0 to 7.4.
Fixed in development version: 27 Jul 2006.
Fix included in release version: User Manual Issue 7.4 Revision 1 (released with Zencrack 7.4a)
The option *MAPPING, NORMAL=YES / NO is not documented correctly. It is correctly stated that the option applies only to quarter circular crack-blocks when *MAPPING, TYPE=TWO STAGE is used. However, the documentation incorrectly implies that the option will give element edges normal to the crack front in the final mesh. The correct description is as follows:
The option NORMAL=YES activates an additional modification during the intermediate part of the two stage mapping algorithm. In this intermediate part of the mapping process, the unit cube crack-block is modified to contain an ellipse of the correct A and B ratio. If NORMAL=YES is used, the element edges within the controlled element rings are modified to be normal to the crack front. For highly elliptic cracks, or cases where the controlled ring size is large, this may lead to nodes falling outside the crack-block unit cube boundary. The program then reports a warning and reverts to NORMAL=NO for the affected crack-block e.g.:
***WARNING
It is not possible to generate element edges normal to the crack front for crack-block 2.
Mapping for this crack-block will be re-attempted with NORMAL=NO.
The unit cube nodal positions are then mapped into the crack-block "real space". If the crack-plane of the target element is square then the element edges should be close to being normal to the crack front. If, however, the crack plane is distorted or rectangular, the element edges will not be normal to the crack front. Even if the final mesh does not contain element edges that are nowmal to the crack front, it will, in most cases, be better than the mesh obtained by NORMAL=NO.
ZENCR314
Applies to: Version 7.1 to 7.4.
Fixed in development version: 12 Sep 2005.
Fix included in release version: 7.4a
An analysis with quarter-circular crack-block(s) may end with error 4027 if the option *MAPPING, NORMAL=YES is used (note: this is a default option if parameter NORMAL is omitted):
***ERROR 4027
VALUE OF THETA FOR NODE 8332 (INTERNAL NODE 8332) LIES OUTSIDE
THETA RANGE OF CRACK FRONT NODES FOR CRACK-BLOCK 1:
THETA= -8.71667281E-03
THETA RANGE = -1.57079633E-12 TO 1.57079633E+00
Additional information with the error message suggests setting TOLMAPEND on the *CONTROLS keyword. This may allow a mesh to be generated but it may contain some distortion within the crack-blocks. If there are multiple crack-blocks on a crack front, some may be mapped using NORMAL=NO and others using NORMAL=YES. The result may be different looking meshing within the crack-blocks as different mapping options have been used.
Workaround: None.
ZENCR313
Applies to: Version 7.0 to 7.3f.
Fixed in development version: 9 Aug 2005.
Fix included in release version: 7.4
If the number of cycles in an individual load system or the total spectrum exceeds 2147483647, then the program fails with an integer overflow.
ZENCR312
Applies to: Version 7.0 to 7.3f.
Fixed in development version: 9 Jun 2005.
Fix included in release version: 7.4
Kiii values calculated from nodal displacements are twice as big as they should be. This in turn affects the Gequiv term calculated from Ki, Kii and Kiii.
If *ENERGY RELEASE RATE, PREDICTION=DISPLACEMENT is used in a growth analysis and there is significant Kiii, the crack growth rate will be higher than it should be since the Kiii contribution is too large.
If *ENERGY RELEASE RATE, PREDICTION=ENERGY is used in a growth analysis the displacement-based Kiii values are only for report purposes and the growth prediction is unaffected by the bug.
Workaround: None.
ZENCR311
Applies to: Version 7.0 to 7.3f.
Fixed in development version: 13 May 2005.
Fix included in release version: 7.4
Paris and Walker crack growth options do not allow a power of less than 2.0.
Workaround: For Paris use TYPE=DELTA K POINTS. For Walker use TYPE=TABULAR.
ZENCR310
Applies to: Version 7.3f (also probably 7.0 to 7.3e, but these have not been verified).
Fixed in development version: 13 May 2005.
Fix included in release version: 7.4
Some nodes may be reported as having "closed" status when they obviously should be "open". This is only likely to occur for non-planar crack faces that extend 360 degrees of a circular crack front.
Workaround: None.
ZENCR309
Applies to: Version 7.3 User Manual, Issue 7.3 Revision 0, page 99
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0, page 96
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0, page 96
Fix included in release version: Version 7.4 User Manual
Equation 5-15 should contain da/dt on the left hand side, not da/dN. The description following the equation should contain "t - time" in place of "N - numbr of fatigue cycles".
ZENCR308
Applies to: Version 7.1 to 7.3f (Abaqus interface only).
Fixed in development version: 27 Apr 2005.
Fix included in release version: 7.4
Surface update on all faces of an element on the Abaqus *SURFACE option (i.e. element specified without any face id) causes an array bounds error at line 299 (in v7.3f) of surfabq if:
- the element is specified as an element number (with or without a trailing comma), and,
- the element is one that is replaced by a crack-block.
Workaround: Specify the element on six data lines with one face id per data line, rather than just one data line without any face id.
ZENCR307
Applies to: Version 7.3f (also probably 7.0-7.3e, but these have not been verified) (Abaqus interface only).
Fixed in development version: 13 Jul 2005.
Fix included in release version: 7.4
The Abaqus *SYSTEM command is not always handled correctly by Zencrack if points A and B are defined but point C is not defined.
Workaround: To address any potential issue with *SYSTEM, ensure that points A, B and C are fully defined.
ZENCR306
Applies to: Version 7.3f (also probably 7.0 to 7.3e, but these have not been verified) (Abaqus interface only).
Fixed in development version: 26 Apr 2005.
Fix included in release version: 7.4
An array bounds problem can occur at line 635 (in v7.3f) of routine insabq if *CONTACT is used in the zcr file and a split set is defined that has a zero for one of the elements.
Workaround: None.
ZENCR305
Applies to: Version 7.3f (also probably 7.0 to 7.3e, but these have not been verified).
Fixed in development version: 29 Apr 2005.
Fix included in release version: 7.4
An array bounds problem can occur at line 153 (in v7.3f) of routine facetinitialsplit if there are a large number of split element pairs (several hundred). The problem is also dependent on a high number of elements along a crack front in combination with a high number of split pairs.
Workaround: None.
ZENCR304
Applies to: User Manual for versions 7.0 to 7.3.
Fix included in release version: Version 7.4 User Manual
In the description of keyword *OUTPUT, the default for option FE DISPS is stated to be YES. This is incorrect - the default is NO.
ZENCR303
Applies to: Version 7.3f (also probably 7.0 to 7.3e, but these have not been verified) (Marc interface only).
Fixed in development version: 14 Feb 2005.
Fix included in release version: 7.4
Coordinates may be incorrect in a Marc analysis after transfer of crack-blocks to a new location. This is only likely to cause problems when there is large non-planar growth.
Workaround: None.
ZENCR302
Applies to: Version 7.3 to 7.3f (Unix versions and Ansys interface only).
Fixed in development version: 15 May 2006.
Fix included in release version: 7.4a
In the Unix runzcr73 script the symbol FETYPE is incorrectly set to the value "solidmodel" if the Ansys option "-a solidmodel" is used on the command line.
Workaround: Fix this problem by locating these lines in the runzcr73 file:
case $CNVANSYS in
N* | n*) CNVANSYS=normal ;;
S* | s*) FETYPE=solidmodel ;;
and correcting the third of the lines as below:
case $CNVANSYS in
N* | n*) CNVANSYS=normal ;;
S* | s*) CNVANSYS=solidmodel ;;
ZENCR301
Applies to: Version 7.3f (also probably 7.0 to 7.3e, but these have not been verified) (Marc interface only).
Fixed in development version: 26 Jan 2005.
Fix included in release version: 7.4
In the Marc interface elements may be missing from the cracked mesh from fe analysis 2 onwards if crack-blocks with different numberes of elements are used in the analysis.
Workaround: None.
ZENCR300
Applies to: Version 7.3f - HP-UX version only.
Fixed in development version: 14 Jan 2005.
Fix included in release version: 7.4
An error 4089 "can't find a node" may occur when *MAPPING,MIDSIDE=CURVED is used. The same job is okay with the default option of *MAPPING,MIDSIDE=STRAIGHT.
Workaround: None, other than using the default *MAPPING,MIDSIDE=STRAIGHT.
ZENCR298
Applies to: Version 7.4, 7.5 - all interfaces.
Applies to: Version 7.1 to 7.3f - Abaqus and Ansys interfaces only.
Fixed in development version: 23 Apr 2008.
Fix included in release version: 7.5-10
An analysis with temperature interpolation can give an unexpected error 4125 stating that temperature cannot be interpolated for a particular node. This occurs because the node falls fractionally outside the boundary of the uncracked mesh as a result of the mesh modification process.
The fix to bug ZENCR366 should address this issue for the majority of cases.
Workaround for v7.3f, v7.4, v7.5: Use *CONTROLS,tolshape1=x with a value of x of 1.01. Keep increasing x in steps of 0.01 until the problem is resolved. If the problem persists and is for a midside node on the split element part of a crack face, a second option is to user *CONTROLS,itolsplit=1. This may result in some midside nodes on split elements of the crack faces being away from their midside positions, but should allow the analysis to progress.
Workaround for v7.1 to 7.3e: Use *CONTROLS,tolshape=x with a value of x of 1.01. Keep increasing x in steps of 0.01 until the problem is resolved.
ZENCR297
Applies to: Version 7.0 to 7.3e.
Fixed in development version: 2 Dec 2004.
Fix included in release version: 7.3f
Some analyses can end with error 4034 where no error would be expected. This occurs because a point cannot be located/positioned in the crack history. Additional code has been implemented in v7.3f to provide a back-up method of positioning the point if the normal approach fails. This fixes some error 4034 problems.
Workaround: None.
ZENCR296
Applies to: Version 7.3 to 7.3e (Ansys interface only).
Fixed in development version: 30 Nov 2004.
Fix included in release version: 7.3f
Temperatures are not interpolated correctly to internal crack-block nodes if boundary shifting is activated. If the temperature distribution is not too severe or the crack-block does not shift far from its original position, then the temperatures are likely to be acceptable.
Workaround: None.
ZENCR295
Applies to: Version 7.1 to 7.3e - Abaqus and Ansys interfaces only.
Fixed in development version: 30 Nov 2004.
Fix included in release version: 7.3f
Temperature interpolation may produce a temperature of zero on one of the original mesh nodes that is shared by adjacent crack-blocks. This problem only occurs if there are multiple crack-blocks and boundary shifting is activated - even then there are some job dependency issues and the problem will probably not occur. A visual check of the temperature contours in the cracked mesh is recommended (select only the crack-blocks for display).
Workaround: Change the order of the crack-blocks in the input file.
ZENCR294
Applies to: Version 7.3 to 7.3e (earlier versions not checked)
Fixed in development version: 2 Dec 2004.
Fix included in release version: 7.3f
Two issues have been identified in relation to definition of tangents, normals and VCEs for crack front nodes:
- For some crack geometries the VCE definitions may not be as sensible as would be expected. The raw energy release rate results may appear to be discontinuous or negative. The subsequent processing to calculate Gmax will, however, be valid for the majority of affected cases. If there is a problem it will manifest itself as a reversal of the crack growth direction from that which is expected.
- Tangents and normals at the free surfaces may not be as accurate as they should be for cases where both sides of the crack are modelled and the outer surface is curved. This has a small effect on analysis results.
Workaround: None.
ZENCR293
Applies to: Version 7.0 to 7.3e.
Fixed in development version: 25 Nov 2004.
Fix included in release version: 7.3f for PC-Windows, 7.4a for all platforms
The output table in the .rep file that reports displacements for crack face node pairs only allows report of node numbers up to 99999. Higher numbers are shown as ***** e.g.:
CRACK FACE NODES SIDES 1/2 DX side 1 DY side 1 DZ side 1 DX side 2 DY side 2 DZ side 2
1 ***** 6778 7.1474E-03 5.0520E-01 -1.4755E-03 7.2454E-03 5.0430E-01 -2.1202E-03
1 91593 98685 7.1409E-03 5.0520E-01 -2.4826E-03 7.2287E-03 5.0430E-01 -3.0723E-03
1 91597 98688 7.1356E-03 5.0510E-01 -3.4904E-03 7.2101E-03 5.0430E-01 -4.0024E-03
This is a reporting issue only and does not affect the analysis or any results processing.
Workaround: None.
ZENCR292
Applies to: Version 7.3 to 7.3e (earlier versions have not been checked).
Fixed in development version: 18 Nov 2004.
Fix included in release version: 7.3f
There may be an array bounds problem in routine facetinitialsplit while generating the initial cracked mesh if the following conditions are met:
- the model has more that one split set defined
- one of the split sets has more than about 125 split pairs (the exact number is dependent on the model e.g. the number and type of crack-blocks will affect the number of split pairs that can be used before experiencing a problem)
Workaround: None.
ZENCR291
Applies to: Version 7.0 to 7.3e (Abaqus interface only).
Fixed in development version: 5 Nov 2004.
Fix included in release version: 7.3f
The supplied Abaqus user subroutines, zcr-jint and zcr-ctint, which can be used to extract results from the Abaqus fil file (using *ENERGY RELEASE RATE, RESULTS=FIL-USER and *USER), do not work correctly for fil files with one million or more records.
Workaround: A workaround is to remove all requests in the uncracked mesh for output to the fil file. The file should then be less than 1 million records and will be processed correctly.
ZENCR290
Applies to: Version 7.0 to 7.3e (Abaqus interface only).
Fixed in development version: 30 Nov 2004.
Fix included in release version: 7.3f
If the Abaqus *SYSTEM keyword is used in an analysis which uses the Zencrack option *BOUNDARY SHIFT, TYPE=TRANSFER, the nodal coordinates in the cracked mesh after the *SYSTEM option may be incorrect after the crack-blocks have transferred to a new location.
Workaround: Modify the uncracked mesh so that *SYSTEM is not required.
ZENCR289
Applies to: Version 7.0 to 7.3d (Abaqus interface only).
Fixed in development version: 29 Oct 2004.
Fix included in release version: 7.3e
A Fortran out of bounds error occurs in routine modabq if an ELSET option with the GENERATE parameter in the uncracked mesh contains more than 1000 data lines.
Workaround: Repeat the *ELSET line to split up the set definition ensuring that each block of data has less than 1000 lines.
ZENCR288
Applies to: 3dmesh utility version 7.03.
Fixed in development version: 17 Sep 2004.
Fix included in release version: 7.3e (utility version 7.04)
The generated .inp file contains corrupted data instead of ** comment characters if the default of Abaqus is selected by hitting return when prompted:
Do you want an Abaqus, Ansys or Marc file ?
(enter a, n or m - default is a)
Workaround: Enter a when prompted with the above question.
ZENCR287
Applies to: Version 7.0 to 7.3d (Abaqus interface only).
Fixed in development version: 4 Oct 2004.
Fix included in release version: 7.3e
If an element set contains more than 50000 elements but the mesh has less than 50000 elements an array bounds error will occur. Windows versions will stop with a Fortran traceback. SG versions continue and eventually report an error because internal data has been corrupted (error 4054 has been observed).
Workaround: Ensure all element sets contain <=50000 elements.
Note: the limiting value is 40000 in v7.0, 50000 for 7.1 to 7.3d and 75000 from 7.3e
ZENCR286
Applies to: Version 7.1 to 7.3c (Ansys interface only).
Fixed in development version: 27 Aug 2004.
Fix included in release version: 7.3d
If element split sets are defined in the zcr file, the line of data in the uncracked mesh before the EN line for each split element is deleted in the cracked mesh file. The missing line is most likely to be an EMORE line for the preceeding element.
Workaround: Place a comment line before each split element in the uncracked mesh (and before each element that may become a split element if *BOUNDARY SHIFT, TYPE=TRANSFER is used).
ZENCR285
Applies to: Version 7.0 to 7.3c.
Fixed in development version: 7 Sep 2004.
Fix included in release version: 7.3d
The crack growth data option "delta G points" requires input of the stress ratio for which the data applies. It is not stated in the documentation that the equations used are only implemented for R=>0. If a value of R<0 is entered the calculation of da/dN will be unreliable. If R=-1 the program fails with a Fortran error in routine growthnode. The fix prevents use of R<0 with this option.
Workaround: Use an alternative crack growth data option.
ZENCR284
Applies to: Version 7.0 to 7.3c.
Fixed in development version: 7 Sep 2004.
Fix included in release version: 7.3d
If Gmax from the energy approaches in the Abaqus and Marc interfaces is negative for a crack front node, Zencrack assumes that Gmin has been found and makes the G value positive and reverses the growth direction for the node. This can give incorrect crack growth. This can only occur if the analysis has one set of VCEs (i.e. a symmetry model or if *GROWTH CONTROLS,DIRECTION=INITIAL or PREFERRED is used). Note that the Gmax term that is calculated form the f.e. analysis results should always be positive. A negative value suggests a numerical issue or path dependence in the contour results. (This is not the same as Zencrack applying a negative sign to a Gmax value to indicate that the crack is closing).
Workaround: Contact Zentech for details - requires use of zcr-jint subroutine to modify the -ve value before it reaches Zencrack.
ZENCR283
Applies to: Version 7.1 to 7.3c.
Fixed in development version: 17 Aug 2004.
Fix included in release version: 7.3d
A model with split sets may given an unexpected error 4089 "unable to find node xxx" when generating the initial mesh if the uncracked mesh has 8 noded bricks and the crack front contains quarter circular crack-blocks.
Workaround: Redo the uncracked mesh using 20 noded brick elements.
ZENCR282
Applies to: Version 7.0 to 7.3c.
Fixed in development version: 10 Aug 2004.
Fix included in release version: 7.3d
Tab characters in the uncracked mesh may cause the Zencrack mesh reader to report an "unrecognised data" error or the data line containing the tab may not be treated correctly. In the latter case a cracked mesh is created but it may be incorrect.
Workaround: Replace tabs in the uncracked mesh by spaces.
ZENCR281
Applies to: Version 7.3 to 7.3c (Ansys interface only).
Fixed in development version: 4 Aug 2004.
Fix included in release version: 7.3d
Zencrack fails reading in materials data file jobname.mp for jobs that use the "-a solidmodel" command line option. The error occurs because of problems with trailing commas generated in intermediate files by Ansys for materials with temperature dependent properties.
Workaround: None.
ZENCR280
Applies to: Version 7.0 to 7.3c (Abaqus interface only).
Fixed in development version: 6 Aug 2004.
Fix included in release version: 7.3d
Data on a *BOUNDARY line in the following format: node # , dof # , 0
is interpreted by Abaqus as being equivalent to: node # , dof # , dof # , 0
But it is interpreted incorrectly by Zencrack. If fixities on crack-block nodes that are to be updated are applied in this way, the nodes are correctly removed from the *BOUNDARY data but no updates appear in the cracked mesh.
Workaround: Use either of the standard input formats for *BOUNDARY:
node # , dof # , dof # , 0
or
node # , dof # , , 0.
ZENCR279
Applies to: Version 7.0 to 7.3c (Abaqus interface only).
Fixed in development version: 5 Aug 2004.
Fix included in release version: 7.3d
If points B and/or C are left blank on the data lines of the Abaqus *SYSTEM keyword, the coordinate system change may not be calculated correctly by Zencrack.
Workaround: Always define points B and C on the *SYSTEM data lines.
ZENCR278
Applies to: Version 7.0 to 7.3c.
Fixed in development version: 6 Aug 2004.
Fix included in release version: 7.3d
An array bounds error can occur if all the following conditions are met:
- the analysis is a crack growth analysis.
- each f.e. analysis has multiple sets of energy release rate / displacement results.
- the number of results sets varies between f.e. analyses (due to non-linear effects).
- Zencrack is attempting to process all the available results sets (e.g. *ENERGY RELEASE RATE, STEPS is used with a frequency of 1).
Workaround: In the Abaqus interface do not use the *ENERGY RELEASE RATE, STEPS option - the default will then be one set of results per Abaqus load step. The total number of results sets will not change from one f.e. analysis to the next.
ZENCR277
Applies to: Version 7.3 to 7.3c (Ansys interface only).
Fixed in development version: 4 Aug 2004.
Fix included in release version: 7.3d
Zencrack gives an array bounds error at line 1898 of routine modansys if TSHAP,QUA8 is used.
Workaround: None.
ZENCR276
Applies to: Version 7.0 to 7.3c.
Fixed in development version: 3 Aug 2004.
Fix included in release version: 7.3d
A job may unexpectedly give an error 4095 or 4034 while generating the initial cracked mesh. The error that is reported depends upon the mapping options that are selected on *MAPPING. Note that these errors may also occur for valid reasons and not because of the bug.
Workaround: None.
ZENCR275
Applies to: Version 7.3 to 7.3b - Ansys interface on Windows machines only.
Fixed in development version: 13 Jul 2004.
Fix included in release version: 7.3c
The value of ZCRANSYS is not set correctly in the runzcr73.bat file by the setup program setupzcr73 if an Ansys product is specified AND the Ansys location contains spaces. The closing double quote that should be around the Ansys executable is instead placed at the end of the ZCRANSYS string. The comments in the runzcr73.bat file regarding the closing double quote are also incorrect. The program correctly replaces the spaces with the # symbol in the batch file e.g.
set ZCRANSYS="c:\program#files\ansys#inc\v80\ansys\bin\intel\ansys.exe#-p#struct" <- incorrect
set ZCRANSYS="c:\program#files\ansys#inc\v80\ansys\bin\intel\ansys.exe"#-p#struct <- correct
With the incorrect definition for ZCRANSYS the analysis produces a screen error when a Zencrack job attempts to execute Ansys:
'c:\program' is not recognized as an internal or external command, operable program or batch file.
Workaround: Edit the runzcr73.bat file and correct the definition of ZCRANSYS.
ZENCR274
Applies to: Version 7.1 to 7.3b.
Fixed in development version: 29 Jun 2004.
Fix included in release version: 7.3c
If a temperature time history is to be used (by specifying *TEMPERATURE, TYPE=TIME HISTORY), but no temperature time history is defined (on *LOAD SYSTEM) or the temperature time history has less than 2 points, Zencrack will stop due to a fortran file error on unit 91 at line 50 of getinfo_temp.
Workaround: Ensure that the temperature time history file is correctly defined if required i.e. has at least 2 points.
ZENCR273
Applies to: Version 7.1 to 7.5-12 (Abaqus interface only).
Fixed in development version: 7 Jul 2008.
Fix included in release version: 7.5-13
If a trailing comma is included after the last node in the connectivity data for 8 noded bricks, or if the connectivity stretches over more than one data line, the connectivity is not read correctly by Zencrack.
Workaround: Remove trailing commas and place all 8 nodes on the first data line.
ZENCR272
Applies to: Version 7.0 to 7.3b (Abaqus interface only).
Fixed in development version: 24 June 2004.
Fix included in release version: 7.3c
If the *CONTACT keyword is used for an 8 noded mesh, Zencrack generates two keywords in the cracked mesh file that are not required. There is no data line for the unnecessary *SURFACE keyword and this causes Abaqus to stop with an error.
Workaround: A workaround is available by pasting data from the cracked mesh into the uncracked mesh file and removing the incorrect lines. Contact us for more details.
ZENCR271
Applies to: Version 7.1 to 7.3b.
Fixed in development version: 23 June 2004.
Fix included in release version: 7.3c
Some 8 noded models may give the following error: "***ERROR 4089 Unable to find node ************".
Workaround: None.
ZENCR270
Applies to: Version 7.1 to 7.3b.
Fixed in development version: 29 Jun 2004.
Fix included in release version: 7.3c
The error message for an incorrect parameter on keyword *MEMORY states that TIME HISTORY is a valid parameter rather than TIME. The TEMPERATURE parameter is not included in the error message. In addition, some warning messages state that the *MEMORY keyword has a parameter TIME HISTORY. These reports should state that the parameter is TIME.
ZENCR269
Applies to: Version 7.0 to 7.3b.
Fixed in development version: 30 Jun 2004.
Fix included in release version: 7.3c
The user manual states that comment lines starting with a # character are allowed in spectrum, time history and temperature history files and sub-files. But including a comment in one of these files generates an error.
Workaround: Remove comments from affected file(s).
ZENCR268
Applies to: Version 7.1 to 7.3b (Ansys interface only).
Fixed in development version: 25 Jun 2004.
Fix included in release version: 7.3c
A number of issues have been identified in the Ansys interface:
- problems with NSEL command:
- NSEL,NONE is required at the start of generation of NSEL commands for node set components
- generated NSEL commands may have format errors
- some node sets may be created more than once in the cracked mesh
- node set NALL is not written correctly (must be requested using *NODE SETS,ALL)
- problems with ESEL command:
- may cause an out-of-bounds error
- errors for parameters TYPE
- commands that do not require processing cause an error
- problems with the handling of the R keyword
- duplicate surface elements can be written for contact or target surface elements when using the "-a solidmodel" command line option
- element types other than 8 or 20 noded bricks may not be handled correctly
- displacement results are not extracted correctly if there are 3 or more crack fronts in the model or if there are 2 crack fronts in which both sides of the crack are modelled.
ZENCR267
Applies to: Version 7.0 to 7.3b (Marc interface only).
Fixed in development version: 30 Jun 2004.
Fix included in release version: 7.3c
In an analysis with crack-block transfer where only one side of the crack is modelled, the symmetry boundary conditions are not removed after transfer. The "open" crack face will not open properly due to the unwanted constraints. This problem is easily identified in deformed plots of the cracked mesh after transfer.
Workaround: Use a full model rather than a symmetry model.
ZENCR266
Applies to: Version 7.3 User Manual, Issue 7.3 Revision 0, page 265
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0, page 251
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0, page 251
Fix included in release version: Version 7.4 User Manual
In the description of the THRESHOLD keyword, the following statement:
Kth The time based threshold is defined using a single value of deltaKth.
should be:
Kth The time based threshold is defined using a single value of Kth.
The description that follows later in the section "Data Lines" is correct.
ZENCR265
Applies to: Version 7.0 to 7.3a (Marc interface only).
Fixed in development version: 15 Jun 2004.
Fix included in release version: 7.3b
The connectivity in an 8 noded model is always generated as element type 7 even if the uncracked mesh contained a different element type (e.g. 117).
Workaround: None.
ZENCR264
Applies to: Version 7.0 to 7.3a (Marc interface only).
Fixed in development version: 15 Jun 2004.
Fix included in release version: 7.3b
No checks are made on the element type for replaced elements. If an invalid type is specified, there is likely to be an error or the analysis may be incorrect. For example, Zencrack gives an array bounds error at line 260 of routine relaxsurfset1 if all replaced elements are invalid types for replacement. Valid element types for crack-block elements are:
- for 20 noded elements: types 21, 35, 57, 61, 44, 71
- for 8 noded elements: types 7, 117, 43, 123.
Workaround: Only replace the listed element types with crack-blocks.
ZENCR263
Applies to: Version 7.0 to 7.3a (Marc interface only).
Fixed in development version: 15 Jun 2004.
Fix included in release version: 7.3b
Zencrack gives an array bounds error at line 90 of routine parmarc if the Marc ALIAS parameter is used for an element type of greater than 99.
Workaround: Manually modify the element type in the connectivity to remove the need for an alias.
ZENCR262
Applies to: Version 7.1 to 7.3a.
Fixed in development version: 10 Jun 2004.
Fix included in release version: 7.3b
In an analysis with time dependence and superposition method 1 there is a chance of a divide by zero error when calculating the growth direction if the instantaneous total Gmax is exactly zero. If the problem occurs, the program stops with a Fortran error.
Workaround: None.
ZENCR261
Applies to: Version 7.1 to 7.3a.
Fixed in development version: 9 Jun 2004.
Fix included in release version: 7.3b
Zencrack gives a -ve square root error at line 405 of routine integrate_dadt (line 408 in v7.3a) if an analysis that includes time dependent crack growth has a load system of type=superposition in the spectrum.
Workaround: None.
ZENCR260
Applies to: Version 7.1 to 7.3a.
Fixed in development version: 10 Jun 2004.
Fix included in release version: 7.3b
In a combined time and fatigue growth analysis it is possible for a cycle to be integrated at the end of one integration step and again at the start of the next. This is most likely to occur if the growth due to da/dt effects is much lower than that due to da/dn effects.
Workaround: None.
ZENCR259
Applies to: Version 7.1 to 7.3a.
Fixed in development version: 8 Jun 2004.
Fix included in release version: 7.3b
Zencrack gives an array bounds error at line 250 of routine reportkey15 if an analysis that includes time dependent crack growth meets any of the following criteria:
- if a load system is defined without a time history file when one is required; in this case a correct error 2200 is reported before the program fails
- if the time history associated with a spectrum has 2 or less points.
Workaround for item 1: Add a time history file as per the error message 2200.
Workaround for item 2: If there are 0 or 1 data points, the input data is invalid. There must be at least 2 points in a time history. So the data should be corrected. If two points define a full history, a third point should be added (e.g. midway between the existing points).
ZENCR258
Applies to: Version 7.1 to 7.3.
Fixed in development version: 2 Jun 2004.
Fix included in release version: 7.3a
Zencrack may give a Fortran -ve square root error at line 86 of routine kicexceeded if an analysis that includes time dependent crack growth has a non-zero Poisson ratio. The error occurs during integration after the first f.e. analysis. Analyses with zero Poisson ratio are correct. Fatigue-only crack growth is not affected.
Workaround: Set the Poisson ratio to zero on the *MATERIAL keyword. This will give a slightly different solution to a true fix, but the results should be indicative of the correct behaviour.
ZENCR257
Applies to: Version 7.0, 7.1, 7.2 (Abaqus interface only) (with Abaqus 6.4-1 or later).
Fixed in development version: 5 Apr 2004.
Fix included in release version: 7.3
Zencrack may not read crack front nodal temperature values correctly from an analysis carried out with Abaqus 6.4. The same analysis is processed correctly if run with Abaqus 6.3. This is due to a format change in the Abaqus .dat file.
If an Abaqus analysis contains temperatures the crack front nodal temperatures are always reported in the rep file in the table headed "TEMPERATURE VALUES FROM ABAQUS OUTPUT FILE, SET xxx". This table will be incorrect for jobs run with Abaqus 6.4-1 or later.
The incorrect temperature values only affect the processing of the f.e. results if the zcr file contains the option *TEMPERATURE, TYPE=FE and temperature dependent material data has been specified.
Workaround: Use Abaqus 6.3.
ZENCR256
Applies to: Version 7.1, 7.2 (Ansys interface only).
Fixed in development version: 8 Mar 2004.
Fix included in release version: 7.3
If an Ansys jobs fails at f.e. analysis 2 or later during a crack growth run (e.g. due to a "distorted element" error message), Zencrack continues to use the displacements obtained for the last successful f.e. analysis and fails to recognise that Ansys reported an error. The analysis results are incorrect after the first f.e. step that fails to run.
Workaround: This problem is most likely to arise because of element
distortion errors issued by Ansys. These can be turned off by adding the
following command before the connectivity data of the uncracked mesh
file:
SHPP,OFF
ZENCR254
Applies to: Version 7.1, 7.2 (Ansys interface only).
Fixed in development version: 18 Feb 2004.
Fix included in release version: 7.3
Constraint equations generated for crack-blocks that require them are incorrect in the Ansys interface. Results from the analysis will be incorrect. Affected crack-blocks are:
- sq112x4.sup
- sq160x4.sup
- sq38x2.sup
(Reminder: crack-block sq38x2.sup should not be used with the Ansys interface due to a collapsed element configuration inside the crack-block that is not supported by Ansys.)
Workaround: None.
ZENCR253
Applies to: Version 7.1, 7.2 (Ansys interface only).
Fixed in development version: 2 Apr 2004.
Fix included in release version: 7.3
A comment character, !, placed on a data line of the uncracked mesh that Zencrack processes may cause the job to stop with ***ERROR 3006.
Workaround: Remove the comment or place it on the preceding line of the file.
ZENCR252
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0.
Fix included in release version: Version 7.3 User Manual, Issue 7.3 Revision 0.
A high value of ratio for fixed or variable ring size control does not have the expected effect. The ratio obtained is instead 0.4.
This happens because a check is made to prevent the ring size control producing potentially distorted element rings. This uses the control tolerance tolringlimit with a default value of 0.4 and has the effect of limiting the ring ratio values input by the user to this upper bound. To allow any fixed or variable ring size, this tolerance should be set to 1.0 using:
*CONTROLS,TOLRINGLIMIT=1.0
Although this tolerance is listed under the *CONTROLS keyword, it is not described elsewhere in the documentation where it would be more useful.
ZENCR251
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0.
Figure 11-1 of the user manual does not indicate that crack-block sq112ax4 is a large crack-block. It is correctly listed in Figure 11-2.
ZENCR250
Applies to: Version 7.0, 7.1, 7.2.
Fixed in development version: 14 Feb 2004.
Fix included in release version: 7.3
Surface normal pressures (other than on crack faces) are missing on both the original target crack blocks and the updated crack block elements after transfer of crack blocks to new locations using *BOUNDARY SHIFT, TYPE=TRANSFER. When the crack-block is in the original location, all pressure load updates are correct.
Workaround: None.
ZENCR249
Applies to: Version 7.0, 7.1, 7.2.
Fixed in development version: 31 Jan 2004 - for Abaqus interface only, 14 Feb 2004 - for Ansys and Marc interfaces.
Fix included in release version: 7.3
Loads on faces of crack-blocks other than the crack face can cause errors in the load update for split set elements.
Workaround: Try placing the crack face load first in the list of loads in the uncracked mesh.
ZENCR248
Applies to: Version 7.0, 7.1, 7.2.
Fixed in development version: 31 Jan 2004.
Fix included in release version: 7.3
A blank line as input for data line 1 on keyword *CRACK FRONT causes an end of file error at line 218 of routine getkey07.
Blank data lines for other keywords that have required data items are correctly trapped and an error message given. The error message could be more helpful as it indicates "an error in the following data" then echoes a blank line.
Workaround: Correct or remove the blank data line in the zcr input file.
ZENCR247
Applies to: Version 7.1, 7.2 (Abaqus interface only).
Fixed in development version: 31 Jan 2004.
Fix included in release version: 7.3
An analysis with load on multiple faces of the crack-block may cause an out of bounds error at line 315 of routine loadit if transition elements are used.
This is due to incorrect allocation of arrays related to load updates when transition elements are used. This problem can only occur if transition elements are used and is dependent on the selected crack-block that is used and the number of transition elements. The higher the number of transition elements the more likely that the error will occur.
Workaround: None.
ZENCR246
Applies to: Version 7.1, 7.2.
Fixed in development version: 31 Jan 2004.
Fix included in release version: 7.3
Extrapolation of values to nodal positions on the crack front may be incorrect for node position 1 when there is more than 1 element along the crack front. This is most obvious in the extrapolated SIF values calculated using the displacement method. The error is very small - typically fractions of 1%.
In v7.1 and v7.2 the incorrect extrapolation of displacement based SIF values does not affect analysis results because of bug zencr241.
Workaround: None.
ZENCR245
Applies to: Version 7.1, 7.2 (Abaqus interface only).
Fixed in development version: 27 Jan 2004.
Fix included in release version: 7.3
A mesh generation error occurs as a result of an input data error. This leads to an array bounds error at line 361 of routine dldabq. Correcting the input data to remove the mesh generation error solves the problem.
When Zencrack encounters a problem generating the cracked mesh an attempt can be made to write what data has been generated for the cracked mesh even though there is an error. In some cases this can help identify the problem. If two crack-blocks that cannot be merged are placed next to one another an error is generated and an attempt is made to write the cracked mesh data. If there is a *DLOAD option in the uncracked mesh the program fails while writing the mesh with an array bounds error at line 361 of routine dldabq.
Workaround: Correcting the input data to remove the mesh generating error solves the problem. However, if you want to allow the program to write the cracked mesh that it was attempting to generate at the time of the array bounds error, you must remove all *DLOAD options from the uncracked mesh file and retain all remaining (incorrect) input data.
ZENCR244
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The value of the third argument, NFE, that is passed into user subroutine user_transfercf is the number of the next f.e. analysis not the number of the previous f.e. analysis.
Workaround: Take this into account when coding the subroutine.
ZENCR243
Applies to: Version 7.1, 7.2.
Fixed in development version: 27 Jan 2004.
Fix included in release version: 7.3
When a crack front attempts to transfer into the final element(s) before a free surface the program will fail with error 6012 and/or 6013.
Workaround: None.
ZENCR242
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
When there are only 2 corner node positions on a crack front, extrapolation of values to the crack front nodal positions is incorrect. The extrapolated values are an average of the base values used to drive the extrapolation. This may affect the following:
- extrapolation of displacement based stress intensity factors to the crack front nodal positions (in this case, due to bug zencr241 these slightly incorrect values are not used in growth prediction)
- calculations related to G and dG/da during update of crack front positions after crack growth.
This problem arises if a crack front contains a single crack-block of type st16x1, st19x1, st23x1, st26x1, st28x1, st28cx1, st35x1.
Workaround: None.
ZENCR241
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The plane strain stress intensity values derived from displacements are extrapolated to crack front nodal positions and reported at those points. If a crack growth analysis is carried out using displacement based values, the original stress intensity factors are used rather than the values extrapolated to the node positions. The effect of this error is greatest in meshes where the element edges at the crack front are not normal to the crack front.
Workaround: None.
ZENCR240
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
When a displacement based solution is used to drive a crack growth analysis, the values reported in the table of conversion of energy results to stress intensity factors are incorrectly based on displacement derived Gequiv values rather than the Gmax values from the energy solution. This is a reporting issue only and does not affect the analysis results.
Workaround: None.
ZENCR239
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
Part way into the analysis a mixed mode crack growth analysis with 2 crack fronts gives correct growth on one crack front and incorrect growth on the other crack front. At the incorrect crack front the shared crack face nodes on the crack-block / split set interface are incorrectly mapped onto an extension of the original crack faces rather than the calculated growth history.
This problem can only arise if all the following conditions are met:
- at least one split element set is included in the model
- more than one crack front is connected to a single split set
- crack growth is non-planar for these crack fronts
- boundary shifting causes the crack face nodes of the split set elements to fall beyond the edge of the initial crack face definition.
Workaround: None.
ZENCR237
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The value of Gequiv calculated from displacement based stress intensity factors is always reported as positive, even in a mode I problem where the crack is closed. The Ki stress intensity factor has correct -ve sign in such cases. The consequence is that if displacement based values are used to drive the crack growth calculation when Ki is -ve, the values of Gequiv are incorrect and the growth prediction will be incorrect.
Workaround: None.
ZENCR236
Applies to: Version 7.1, 7.2 (Abaqus interface only).
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The TRANSITION parameter on keyword CRACK FRONT causes the analysis to fail if a value of 1 is used. There are two possibilities:
- if there are transition elements attached to only one face of the crack-block then the program gives an array bounds error at line 531 of normal.f
- if there are transition elements attached to more than one face of the crack-block the program reports "***ERROR 4005 - CRACK PLANE IS NOT VALID FOR CRACK-BLOCK n".
Note that in situation 2 it is not physically possible to use a transition element setting of 1 and that situation 1 is not a good idea if sensible results are required from the analysis.
Workaround: Change to a sensible setting for the number of transition elements or turn the feature off (by omitting the option or using a value of 0).
ZENCR235
Applies to: Version 7.0, 7.1, 7.2 (Abaqus interface only).
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The STEPS option on *ENERGY RELEASE RATE for the Abaqus interface does not write out any contour integral data for a step if the frequency value for the step is set to zero. If the first step that the user defines on the STEPS data lines has frequency of zero, then the behaviour up to and including that step is consistent with what the user wants. If the step with frequency of zero is the second or later step that is defined on the STEPS data lines, the behaviour for the step is not what it should be - instead of de-activating contour integral requests with FREQUENCY=0, the value of the FREQUENCY parameter is retained from the previous analysis step.
Workaround: Add the following into the analysis step of the uncracked mesh file that is required to have FREQUENCY=0 for contour integral output:
*CONTOUR INTEGRAL, FREQ=0, CONTOURS=3
data lines e.g. Z$001,1,1,1
The data lines can be obtained by cut/paste from any *CONTOUR INTEGRAL block in a step that has data generated by Zencrack. Alternatively they can be entered manually with one line for each crack front node, starting with "Z$001,1,1,1" and ending with "Z$nnn,1,1,1" where nnn is the number of nodes on the crack front e.g. Z$005 if there are 5 nodes on the crack front.
ZENCR234
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0.
Applies to: Version 7.0 User Manual, Issue 7 Revision 1.
Applies to: Version 7.0 User Manual, Issue 7 Revision 0.
Fix included in release version: Version 7.3 User Manual, Issue 7.3 Revision 0.
An incorrect conversion factor is given in the user manual "Unit Conversions" section:
1 ksi in0.5=0.347487 MPa mm0.5
should be:
1 ksi in0.5=34.748481 MPa mm0.5.
The html unit conversion utility, FMunitconversion.htm, is correct.
ZENCR233
Applies to: Version 7.2 Installation Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 Installation Manual, Issue 7.1 Revision 0.
Applies to: Version 7.0 Windows Installation Manual, Issue 7 Revision 1.
Applies to: Version 7.0 Unix Installation Manual, Issue 7 Revision 0.
Fix included in release version: Version 7.3 Installation Manual, Issue 7.3 Revision 0.
None of the installation manuals is clear about how to set up Zencrack for use with a Zencrack user subroutine.
Additional notes for Windows:
When the setup program is executed after installation it checks for Compaq Visual Fortran being installed in the default location. If it is, the setup configures all files automatically. When the user executes a job with the "-u" command line option, the compile and link process should run smoothly.
If Compaq Visual Fortran is not installed or is installed in a non-default location, the setup program prompts for some input and then continues with the configuration.
Additional notes for Unix:
The runzcr7x script contains a variable called ZCRFORTRAN which is set to "f90". If you use some other command line option to execute your Fortran 90 compiler, you should change this variable.
ZENCR232
Applies to: Version 7.0 onwards.
Fix included in release version: 7.5
In some situations a split pair may be needed in which the crack face extends more on one face than on the other. In such cases where the "short" crack face has no elements other than the crack-block, the cracked mesh is not created correctly and a handful of nodes on opposite crack faces remain tied together.
Workaround: Include a row of dummy elements behind the crack-blocks on the "short" crack face to mate up with the "long" crack face. Then the split pair can be defined. Set the Young's modulus to a very low value for the dummy elements so that it will not affect the results.
ZENCR231
Applies to: Version 7.2 Installation Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 Installation Manual, Issue 7.1 Revision 0.
Applies to: Version 7.0 Installation Manual, Issue 7 Revision 1.
Applies to: Version 7.0 Installation Manual, Issue 7 Revision 0.
Fix included in release version: Version 7.3 Installation Manual, Issue 7.3 Revision 0.
Step 8 of the section "To install the program files" for Windows systems may cause confusion over which file is being referred to depending upon the user's view settings for the folder. The step should be re-worded for v7.2 as follows (with 72 replaced by 71 or 70 for version 7.1 and 7.0 respectively):
8. To make a desktop shortcut for Zencrack:
- right click on the Zencrack shortcut file in the c:\zencrack72\tools directory; this is the file called "Zencrack 72" which has size 1.52kb and not the icon file "Zencrack72.ico"
- drag it onto your desktop
- select "Copy here" or "Create shortcut here"
ZENCR230
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
The program may get trapped in an infinite loop while carrying out numerical integration of crack growth. This is only likely to occur if no proper threshold has been defined and da/dn falls below 10-15. Results are correct up to the point where the program gets stuck.
Workaround: Try defining a meaningful threshold.
ZENCR229
Applies to: Version 7.2 User Manual, Issue 7.2 Revision 0.
Applies to: Version 7.1 User Manual, Issue 7.1 Revision 0.
Applies to: Version 7.0 User Manual, Issue 7 Revision 1.
Applies to: Version 7.0 User Manual, Issue 7 Revision 0.
Fix included in release version: Version 7.3 User Manual, Issue 7.3 Revision 0.
If the same displacement boundary conditions are listed in multiple Abaqus analysis steps only the boundary conditions in the first step are updated. This most likely results in Abaqus failing to run due to having boundary conditions specified on nodes which no longer exist.
If different displacement boundary conditions are listed in multiple Abaqus analysis steps the updates applied by Zencrack may not be correct.
These issues are known limitations but are not explained in the documentation.
ZENCR228
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
Runtime math error may occur at line 150 of imap2d.f if a crack front attempts to transfer into an invalid location when *BOUNDARY SHIFT,TYPE=TRANSFER is used.
Workaround: None.
ZENCR227
Applies to: Version 7.1, 7.2 on Windows platforms.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
If a crack-block name is given with anything other than lower case extension (i.e. .sup) in the .zcr file, then an error will be reported e.g.
***ERROR 1001 SYSTEM ERROR 29 OCCURRED OPENING FILE c:\zencrack72\crack.sup
The file name given in the error message is also incorrect.
Workaround: Use lowercase .sup extension to define the crack-blocks in the .zcr file.
ZENCR226
Applies to: Version 7.1, 7.2.
Fixed in development version: 23 Jan 2004.
Fix included in release version: 7.3
Array bounds error may occur in line 531 of normal.f. The error is caused when *BOUNDARY SHIFT is activated but none of the boundary nodes in the crack plane can be shifted.
Workaround: None.
ZENCR157
Fixed in: version 7.4 (fully fixed)
Fixed in: version 6.0-e (partial fix - see below)
Zencrack does not correctly update boundary conditions on crack-blocks if the boundary conditions are specified using node numbers and the option OP=NEW is used on *BOUNDARY.
Workaround 1: request node sets in the cracked mesh and apply the boundary condition to the appropriate node set(s) (keyword *NODE SETS in v7.0 onwards or set INDSET=1 in earlier versions).
Workaround 2: check that OP=NEW is essential for the correct application of the boundary conditions - if not remove it.
This is partially fixed in version 6.0-e:
- Zencrack now adds all updated boundary conditions into the first *BOUNDARY option that occurs in the mesh. Previously a new *BOUNDARY option was generated for the updated fixities.
Zencrack Support
Bug List
More in this category