Professional Documents
Culture Documents
ProCAST User Manual PDF
ProCAST User Manual PDF
Version 2006.0
Page 1
INTRODUCTION
SOFTWARE CAPABILITIES
SOFTWARE ORGANIZATION
USER MANUAL PRESENTATION
7
9
11
WHAT'S NEW
12
VERSION 2006.0
VERSION 2005.0
VERSION 2004.1
VERSION 2004.0
IDENTIFIED BUGS, PROBLEMS AND LIMITATIONS
12
19
24
29
33
GETTING STARTED
36
SOFTWARE LAUNCH
PROBLEM SET-UP
CALCULATION
RESULTS DISPLAY
36
38
46
48
SOFTWARE MANAGER
51
FILE MANAGER
MODULE CALLS
ADVANCED MODULE CALLS
RUN LIST
SOFTWARE CONFIGURATION
CUSTOMIZED INSTALLATION
FLEXLM
51
55
63
66
69
73
74
PRE-PROCESSING
78
INTRODUCTION
GEOMETRY IMPORT
THERMAL
GEOMETRY ASSIGNMENTS
MATERIALS ASSIGNMENT
INTERFACES ASSIGNMENT
BOUNDARY CONDITIONS ASSIGNMENT
PROCESS CONDITIONS ASSIGNMENT
INITIAL CONDITIONS ASSIGNMENT
RUN PARAMETERS ASSIGNMENT
Version 2006.0
78
82
84
85
85
90
95
100
102
104
Page 2
Version 2006.0
106
110
122
131
131
135
141
150
153
156
164
167
182
187
191
205
206
220
221
227
233
234
238
246
248
252
255
257
261
262
266
266
267
268
269
276
282
285
287
290
295
298
299
303
307
308
313
313
314
315
316
317
318
Page 3
319
320
321
322
322
324
SOLVER
TROUBLESHOOTING
324
327
RESULTS VIEWING
331
INTRODUCTION
FIELD SELECTION
DISPLAY TYPES
DISPLAY PARAMETERS
TAPE PLAYER
CURVES
GEOMETRY MANIPULATION
331
337
341
351
369
370
376
RESULTS ANALYSIS
385
CRITERION FUNCTIONS
POROSITY
FATIGUE LIFE INDICATOR
HOT TEARING INDICATOR
CRACKING INDICATOR
FRECKLES INDICATOR
385
394
395
396
397
398
RESULTS EXPORTS
399
INTRODUCTION
GEOMETRY
RADIATION FACES
TEMPERATURE
HEAT FLUX
DISPLACEMENTS
STRESS
399
400
401
402
405
406
407
PARALLEL SOLVER
408
INTRODUCTION
HOW DOES PROCAST PARALLEL WORKS ?
USE OF THE PARALLEL SOLVER
REPEATABILITY
LIMITATIONS
MACHINE CONFIGURATION
408
411
416
421
422
423
Version 2006.0
Page 4
425
428
430
433
INTRODUCTION
ADVANCED POROSITY PRE-PROCESSING
GENERAL INFORMATION
MATERIAL PROPERTIES
GAS AND BUBBLE PROPERTIES
PROCESS INFORMATION
CALCULATION SETTINGS
EXAMPLES PREFIX_PORO.D INPUT FILE
ADVANCED POROSITY SOLVER
ADVANCED POROSITY POST-PROCESSING
433
435
436
436
438
440
444
447
452
454
MICROSTRUCTURES
458
INTRODUCTION
CASE SET-UP AND RESULTS
EXAMPLES
IRON AND STEEL
INTRODUCTION TO IRON AND STEEL
IRON AND STEEL MODELS
CASE STUDIES
458
462
467
474
474
480
484
"2-D"
505
INTRODUCTION
MESHING
CASE SETTING
505
507
514
CAFE-3D
517
INVERSE MODELING
524
INTRODUCTION
MODEL SET-UP
INVERSE RUN
FILE FORMATS
INVERSE APPENDIX
524
527
534
538
542
INPUT-OUTPUT FILES
552
Version 2006.0
Page 5
557
RESTART
CONVERGENCE PROBLEMS
STRESS CALCULATIONS
GAPS IN STRESS MODELS
STRESS VISUALIZATION
557
558
559
560
562
TUTORIALS
565
GENERAL INTRODUCTION
FLOW CHART
MESHCAST
PRECAST
PROCAST
VIEWCAST
RESTART
BOUNDARY CONDITIONS
SYMMETRY
TIME STEPS
MATERIAL PROPERTIES
CASTING MATERIAL
EXOTHERMIC MATERIAL
FILTER MATERIAL
MOLD MATERIAL
THERMODYNAMIC DATABASES
CYCLING MODELLING
VIRTUAL MOLD
PROCESS TEMPLATES
HPDC-CYCLING
HPDC-FLOW
LPDC
GRAVITY-SAND
INVESTMENT
FILTER
STRESS
NON-COINCIDENT MODEL
565
565
567
568
569
569
571
575
581
583
585
586
589
591
591
592
598
604
609
610
636
647
666
687
710
734
751
Version 2006.0
Page 6
INTRODUCTION
SOFTWARE CAPABILITIES
ProCAST is a software using the Finite Elements Method (FEM). It allows the
modeling of Thermal heat transfer (Heat flow), including Radiation with view
factors, Fluid flow, including mold filling, Stresses fully coupled with the thermal
solution (Thermomechanics). Beside that, it includes also microstructure modeling
and porosity modeling.
Special models are included in order to account for thixo casting and lost foam.
Specific features are included to account for processes such as high pressure die
casting, centrifugal, tilt. Finally, customized models for foundry processes, such
filters, sleeves are included.
Thermal calculation
Version 2006.0
Page 7
Thermomechanical calculation
Version 2006.0
Page 8
SOFTWARE ORGANIZATION
The software is organized around a Manager, which calls the different modules :
The following figure is presenting the structure of the software. First, the
geometry, in the form of a CAD model is loaded into MeshCAST, to generate a
FEM mesh. Then, the calculation is configured in PreCAST, the Pre-processor.
PreCAST is linked to Thermodynamic Databases for the automatic
determination of the material properties from thermodynamic databases. Before
the solver ProCAST is launched, a "data conditioner" named DataCAST is run.
Finally, the results can be viewed or exported (for further processing) in the Postprocessor ViewCAST.
The ProCAST solvers are divided in "Physical modules" with the following
capabilities :
Version 2006.0
Page 9
Thermal module
Radiation module
Navier-Stokes equation
Penalization of the flow in the mushy zone and in the solid
Mold filling algorithm, with free surface
Filter model
Newtonian and Non-Newtonian flow
Thixo casting models
Lost Foam model
Tilt pouring
Centrifugal casting
Turbulent models
Stress module
The features linked to specific processes (e.g. cycling, tilt, ...) are embedded in the
corresponding physical modules.
Version 2006.0
Page 10
Version 2006.0
Page 11
WHAT'S NEW
VERSION 2006.0
This section is describing the main news of ProCAST v2006.0.
The links are referring to the corresponding section of the manual for more details.
General / Manager
MeshCAST
Version 2006.0
Page 12
The "Partial layer" option was corrected (it was giving the same results are the
Full layer option).
The "*.bstl" files can now be read on Linux.
PreCAST
Version 2006.0
Page 13
Thermodynamic databases
ViewCAST
Version 2006.0
A new "Pick" function has been introduced, to pick the coordinates, node
number and values (e.g. Temperature, Pressure, ...) of a given point selected
interactively on the model (see the "Geometry Manipulation" section for more
details).
The liquidus and solidus temperature are now shown on the Temperature scale
(see the "Results Viewing/Display Parameters" section for more details).
The previously defined scales (i.e. the scales defined manually during a
previous session) are now automatically stored and loaded (see the "Results
Viewing/Display Parameters" section for more details).
Stored views were introduced (see the "Results Viewing/Geometry
Manipulation" section for more details).
When a model is loaded for the first time (or when no view was stored), the
model is displayed in isometric view, with the gravity pointing downwards.
In case of a Stress calculation, it is now possible to calculate all the Stress
results (e.g. Effective Stress, Principal Stress, SigmaX, ...) all at once before
viewing them (see the "Results Viewing/Field selection" section for more
details).
In the case of stress calculation, the possibility to view the total displacement
was introduced (see the "Results Viewing/Field selection" section for more
details).
The displacements relative to a plane were introduced for the visualization of
deformed geometries (stress calculation) (see the "Results Viewing/Field
selection" section for more details).
When a stress calculation is performed, it is possible to display the
underformed geometry in wireframe together with the deformed contours (see
the "Results Viewing/Display Parameters" section for more details).
The calculation of the solidification time was improved in order to account for
remelting (if any).
The calculation of the Filling time was changed in order to ignore the piping
area.
Page 14
Solver
Version 2006.0
The filling algorithm was improved. A new Run Parameter FREESFOPT was
introduced (see the "Flow Run Parameters" section for more details).
The filling solver was improved in order to better handle Tilt pouring models.
A new Run parameter "TILT=1" was introduced in the case of Tilt models
(see the "Flow Run Parameters" section for more details).
For centrifugal casting, the algorithm was improved by the introduction of the
RELVEL Run Parameter (see the "Centrifugal casting" and the "Flow Run
Parameters" sections for more details).
The convergence, the accuracy and the CPU time of the Stress module was
significantly improved.
Two new models (Norton law and Strain Hardening Creep) were introduced in
the Stress module to account for viscoplasticity and creep (see "Stress models
and Properties" section for more details).
The Plastic stress properties (i.e. the Hardening) can now also be defined as a
set of tabulated tensile tests curves in an ASCII file (see "Digitized Hardening"
section for more details).
In the case of Stress calculations with Vacant or Rigid materials, the memory
management was changed in order to reduce the amount of memory needed
(in the scalar version only).
Stress calculations can now be run with piping (i.e. it is now allowed to set
PIPEFS<1 in the case of stress calculations).
The POROS=1 model was extended to allow the treatment of the Graphite
expansion in the case of SGI (see "SGI Porosity model" section for more
details).
The POROS=1 model was improved for the treatment of the third phase in
HPDC. A new Run Parameter (GATEFS) was introduced (see the "Thermal
Run Parameters"and "Active Feeding" sections for more details).
Page 15
Version 2006.0
A bug in the POROS=1 model was corrected (in the identification of the
independant liquid pockets).
A problem in the POROS=1 model with FREESF=2 filling was corrected.
Microstructure models were improved (see "Microstructures" section for more
details).
The memory requirement was significantly reduced when running a case with
many materials.
Time-dependant Inlet boundary conditions were introduced (see the
"Boundary Conditions Database" section for more details).
New User functions have been introduced (see "User Functions" section for
more details).
A problem of Restart in the case of gravity casting was corrected.
The problem of heat transfer between a shot piston and the casting (there was
no heat transfer) has been solved.
The logic of the flow Switch Off was changed in order to avoid some
unexpected situations. If LVSURF < 1, the value of PIPEFS can not be set to
1.
A problem in the FLOW=9 option was corrected.
It is possible now to run Stress models with materials which have no phase
change.
A bug in the case of Stress calculation in the presence of filters has been
corrected.
It is possible now to run a Microstructure calculations together with the
Turbulent model.
A problem in Cycling with FREESF=1 was corrected.
In the case of moving solids, the translation is also called at step 0.
A problem in the case of DIE COMBO was corrected.
A problem in the calculation of the piping when using Periodic Boundary
conditions was fixed.
A problem in the calculation of the "Freckles predictor" has been corrected, as
well as a problem of Restart with Freckles prediction.
If the "statpro" directory (where the status file is written) is momentarily not
available, the calculation will not crash anymore.
In the case of delayed filling (i.e. if the filling does not start since the
beginning of the analysis), the DTMAXFILL timestep will not be active
before the filling starts.
A bug in the treatment of the TOPEN has been corrected.
A problem in the Units of the user function "externalcompute.c" was
corrected.
A problem of "datacast -u" in the case of stress was fixed (the stresses of
vacant domains were lost).
A problem of convergence with non-coincident meshes was fixed.
A problem of memory allocation was corrected in the case of Radiation
problems.
The Run Parameters Recommendations have been updated in order to account
for the changes in the algorithm (see the "Run Parameters Recommendations"
section for more details).
Page 16
The Stress module has been parallelized. Please note that no specific set-up
should be done for parallel processing of Stress calculations. The set-up
exactly the same as the scalar version.
The "DataCAST" module is now the same between Scalar and DMP (e.g. one
can run DataCAST on a scalar machine and then launch the calculation on N
CPU's with the DMP version or vice-versa).
ProCAST 2006.0 contains the first release of the Windows DMP version (see
the "Parallel solver" chapter for more details).
The Lost Foam and the Non-Newtonian module has been parallelized (same
set-up as the scalar version).
The TRI2QUAD=1 functionnality (for radiation) has been parallelized.
Some problems encountered in the DMP versions were fixed.
CAFE
The quality of the solver was significantly improved in order to have a better
convergence (robustness) and a better accuracy of the results.
The defalult value for the "convergence.tolerance" was modified (from 1e-12
to 1e-10) in order to be compatible with the modified algorithm.
A bug in the storage frequency was corrected.
The pressures calculated by the "Advanced Porosity module" can now be
visualized in ViewCAST (see the "Advanced Porosity Post-processing"
section for more details).
It is possible to view Advanced Porosity results (with the calcososft postprocessor) across plateforms (automatic swap option).
The calcosoft post-processor allows now to view results calculated using Hex
or Wedges mesh.
A bug in the viewing of Advanced Porosity results in the presence of
enclosures has been corrected.
Template files of the poro.d input file were added in the software installation
(see the Template button in the "Copy" of the Manager in the end of the
"Software Manager/File Manager" section for more details)
Manual
(The following sections of the manual were added or changed)
The Stress properties recommendations were modified. One should note that it
is very important to set stress properties according to the "rules" described in
Version 2006.0
Page 17
Version 2006.0
the manual. Othewise, one could have convergence problems (see the "Stress
Models and Properties" section for more details - end of the section).
The Computherm manual (from Computherm LCC) is added in the Software
installation (in the dat/manuals/PDF direcrtory). This manual describes for
each alloying system the phases which are calculated, the limitations as well as
the validations which have been made.
The Run Parameters Recommendations have been updated in order to account
for the changes in the algorithm (see the "Run Parameters Recommendations"
section for more details).
A new section was added in the "Microstructure Module" chapter (see the
"Iron and Steel" section for more details).
Page 18
VERSION 2005.0
This section is describing the main news of ProCAST v2005.0.
The links are referring to the corresponding section of the manual for more details.
General / Manager
MeshCAST
The call of the Volume mesher of MeshCAST in batch was re-introduced (see
the end of the "Software Manager/Module calls" section for more details).
The possibility to read Parasolid files with the .x_t suffix was added.
The automatic "zoom-out" was suppressed when the ADDLINE operation was
performed.
A units problem in MeshCAST-2D was corrected.
PreCAST
Version 2006.0
Page 19
The possibility to activate User functions was introduced in PreCAST (see the
"User Functions" section for more details).
A bug in the units of the "Tilt" axis was corrected.
A bug in the storage of the "K-factor" (non-Newtonian model) was corrected.
The "Extract" was extended in order to allow it in the presence of filters.
The "automatic" setting of FLOW=3 was reviewed.
A "Erase line" was added in the "Chemical Composition" definition of the
material properties.
The TENDFILL Run parameters was introduced in order to automatically stop
a calculation at the end of the filling (see the "General Run Parameters"
section for more details).
New alloys were introduced in the Material database. Some properties of
existing alloys were improved.
The Stress material properties database was complemented.
A bug in the treatment of non-coincident interfaces was corrected (it was
creating a double heat transfer on some nodes in some specific cases).
The COARSEC and COARSEP Run parameters were removed from the
interface as they are no longer used.
When PreCAST is opened, the model is oriented according to the ISO
orientation.
Thermodynamic databases
ViewCAST
Version 2006.0
ViewCAST was adapted in order to allow the viewing of result files coming
from a different platform (e.g. viewing on Windows results computed on
UNIX) (see the "Software Manager/Software configuration" section for more
details).
Automatic creation of AVI animations was introduced (see the
"Resultsviewing/Introduction" section for more details).
The GIF/AVI capture action can also be started from the "File" menu.
Page 20
Export of the "Shrinkage porosity" and of the "Advanced porosity" fields has
been introduced.
An automatic pre-defined setting for "Shrinkage porosity" was introduced (see
the end of the "Results viewing/Display types" section for more details).
Short cuts for the different display modes were introduced (see the "Results
viewing/Display types" section for more details).
The export of "Reversed displacements" was corrected, as well as the X-Y
plots of displacements.
The maximum number of timesteps which could be exported was increased
from 200 to 5000.
In Tilt casting, the particles are now "tilting" with the part. In addition, a new
"Tilt" option in the "Parameters" menu was introduced in order to deactivate
the tilt during the viewing (see the "Display Parameters" section for more
details).
The "Fill time" plot was introduced.
A problem with the calculation of RGL on Linux was fixed.
Solver
Version 2006.0
User functions were introduced for "interface heat transfer coefficients" and
"Heat boundary conditions" (see the "User Functions" chapter for more
details).
A bug with the TCLOSE (DIE COMBO cycling) was corrected.
A problem in the cementite formation of cast iron (microstructure module)
was fixed.
The interpolation of the linear thermal expansion coefficient (stress module)
was changed from linear by intervals to piecewise linear.
A problem with the calculation of stress in the presence of piping (stress
module) was corrected. In such case, one should set PIPEFS = 0 (see the
"Thermal Run Parameters" section, under PIPEFS for more details).
A new possibility was included in order to handle the "Gate feeding" in hpdc
in the presence of a shot piston (see the "Active Feeding" and "Thermal Run
Parameters" sections for more details about the GATENODE Run Parameter).
A bug in the POROS 1 model was corrected (to avoid unexpected behaviors in
some specific cases). Moreover, the display of the macroshrinkage (i.e. holes
inside the casting) has been changed (one will have shrinkage porosity values
of 1, but no internal "Voids" will be shown. "Voids" will be used only for pipe
shrinkage at the surface of the model). Please note that the value of FVOL has
no meaning for the viewing of porosity results.
A problem in the Surface load (stress module) which was appearing in some
cases was corrected.
The influence of the gap width on the interface heat transfer coefficient, with
non-coincident meshes, was introduced.
The viscoplastic model (stress module) was improved, for a better accuracy.
The stress module was extended in order to allow the use of filters.
DIE COMBO was extended for the use with non-coincident meshes.
A problem in the handling of the "RESERVOIR" was corrected in order to
allow to fully empty the reservoir.
Page 21
Parallel processing
The Parallel processing solver was introduced in this version for selected
platforms (see the "Parallel Solver" chapter for more details).
Manual
(The following sections of the manual were added or changed)
A new section was added to list the limits in the chemical compositions for the
Thermodynamic databases (see the "Databases limitations" section for more
details).
Version 2006.0
Page 22
Version 2006.0
Page 23
VERSION 2004.1
This section is describing the main news of ProCAST v2004.1.
The links are referring to the corresponding section of the manual for more details.
General
On some platforms, the graphics performances were very poor (in terms of
display speed). This was corrected. However, it may still happen that on some
graphics cards, the graphical performances are not optimum (in terms of
display refresh). In such cases, the configuration of the X-Server may need to
be adjusted.
The installation structure of the software was complemented in order to allow
the installation in a Read only location. Each user has now the possibility to
set a customized directory for his Preferences (including the MeshCAST user
customized icons) and databases. (Customized installation).
MeshCAST
On some platforms (mainly Windows), MeshCAST was not always able to use
all the available memory (e.g. it was crashing at 1.1 GB, whereas the computer
had 2 GB of RAM). This problem was due to the way "reallocation" of
memory was done in the code and how the operating system was managing the
memory allocation. The memory "reallocation" technique was modified in
order to prevent such problems, allowing to use all (or most) of the available
memory. In such situation, temporary files will be created in the working
directory (see the MeshCAST manual for more details about this option and
the "Module calls" section for the activation of this option).
In the case of problems with Shell meshing with symmetry, a new
functionality was added in order to allow the curing of the problem (see the
MeshCAST manual for more details about this option).
The User defined icons can now be stored in a central location (or in a
customized user defined location) (Customized installation).
Some corrections when reading Parasolid files were made.
ANSYS surface mesh input is now available.
PreCAST
Version 2006.0
Page 24
ViewCAST
Version 2006.0
Page 25
Solver
Version 2006.0
Page 26
cases that the filling time may not be exactly correct (if interface nodes are
selected).
A bug in the micro module for hyper-eutectic cast iron was corrected.
When the microstructure module is used, the material properties definition has
slightly changed (as some data can be automatically obtained from the
thermodynamic databases) - see the "Microstructures/Case set-up and Results"
section for more details.
A print-out bug was corrected in the d.out file.
On some platforms, it was still not possible to generate files larger than 2 GB.
This is now corrected. One should be careful however that the file system
(kernel) should be well configured in order to accept such large files.
However, as it was not possible to generate a 64 bits version of the User
interfaces (and especially ViewCAST) on IBM (due to Tcl/Tk problems), it is
not possible to visualize such large models on IBM at this stage. Work around
: visualize these results on a different platform.
A bug in the calculation of the permeability of a mushy zone at low fraction of
solid was corrected. This has an effect on cold shots or early solidification
during filling (in case of very large interface heat transfer coefficients).
DataCAST execution time was reduced for large models.
Manual
(The following sections of the manual were added or changed)
Version 2006.0
Page 27
Version 2006.0
Page 28
VERSION 2004.0
This section is describing the main news of ProCAST v2004.0.
The links are referring to the corresponding section of the manual for more details.
General
All the user interfaces have been re-developed, using the Tcl/Tk and OpenGL
technology. Thus, NutCracker is not anymore necessary to run the software on
Windows machines
The software is licensed using the FlexLM tool (FlexLM). This gives the
possibility to have floating licenses
Direct access to the on-line help
Manager
MeshCAST
PreCAST
Version 2006.0
PreCAST was totally re-developed in a new user interface, based upon the
same fundamental principles as the previous version (Pre-processing)
Interactive tools to manipulate the geometry (rotate, zoom, center, drag)
(Geometry manipulation)
Different display types (wireframe, hidden, hidden mesh, shade) (Geometry
manipulation)
New symmetry definition, with interactive pick of the coordinates (see the
Radiation section)
Page 29
Enhanced definition of the Virtual mold, with possibility to view the thermal
depth and the mold box (Virtual Mold)
Display of the size of the model (Geometry Assignment)
New Material database design (Material Database)
Coherency about material properties definition (enthalpy, Cp, fs and L)
(Material Properties)
Direct visualization of tables (Material Database)
Direct access to material and boundary condition databases (Materials
assignment, Boundary conditions assignment)
Search and Sort tools for each database (Material Database)
Exothermic material definition (Exothermic)
More explicit definition of the material domain types (casting, mold, foam,
filter, exothermic, ...) (Materials assignment)
Definition whether a material domain is empty or not in the "Materials
assignment" menu
Possibility to remove interfaces which have been previously created
(Interfaces assignment)
More flexible way to define non-coincident interfaces (Interfaces assignment)
Possibility to change interactively the non-coincident interface tolerance
(Interfaces assignment)
In all databases, possibility to Import or Export the tables (Material Database)
Enhanced selection of boundary conditions, automatic propagation, selection
of remainder, selection of interfaces (Boundary conditions assignment)
Copy, Paste and Clip capabilities in the Boundary conditions selection
(Boundary conditions assignment)
A "velocity calculator" was introduced in order to calculate the desired inlet
velocity boundary condition (Boundary Conditions Database)
For radiation problem with an enclosure, the alignment of the surface vectors
is simplified (Radiation)
For all databases, the database management (add, read, copy, delete) can be
done in the same window as the assignments (Material Database)
New design of the Run Parameters layout (Run Parameters)
Automatic settings of up to nine Run Parameters Templates (Pre-defined Run
Parameters)
ViewCAST
Version 2006.0
ViewCAST was totally re-developed in a new user interface, based upon the
same fundamental principles as the previous version (Results viewing, Results
analysis, Results export)
Better rendering of the models and of the free surface, including shading and
transparency
Interactive tools to manipulate the geometry (rotate, zoom, center, drag)
(Geometry manipulation)
Full capability "Tape player" to step forward and backwards in single steps or
animated mode (Tape Player)
Page 30
Version 2006.0
More comprehensive way to select the different view modes of the material
domains (solid, invisible, wireframe, hidden mesh, shaded, transparent)
(Geometry manipulation)
Direct access to the different view modes (snapshot, slice, scan, cut-off)
(Display types)
Interactive slicing is available at any time (Display types)
Enhanced timestep display definition, with the possibility to define time
intervals for the display (in addition to step intervals) (Display parameters)
Full export of stress results (in a Neutral format which can be imported in any
stress software) (Results Export/Stress)
Full export of the displacements (in a case of stress calculation), with the
possibility to "reverse" the displacements in order to perform reverse
engineering of distorted molds (Results Export/Displacements)
Automatic calculation of the solidification time (Field selection)
Enhanced symmetries settings (same panel as in PreCAST) (Display
parameters)
Merge of the former PostCAST capabilities into ViewCAST. This allows to
have only one post-processor for the snapshots and the time evolutions curves
(Results viewing/Curves), as well as for the generation of outputs, such as
Niyama, cooling rates, SDAS, ... (Results analysis/Criterion functions)
The XYPlots settings have been totally reviewed (Results viewing/Curves)
The node selection for XYPlots can now be stored, as well as the scale of the
graph for further use (Results viewing/Curves)
Three different modes for the viewing of the free surface have been introduced
(On, Off, Only) (Display parameters)
The definition of the scale is totally new and interactive. The scale is kept in
memory for each field (Display parameters)
The materials can be "exploded" for a better viewing (Geometry manipulation)
The display of vectors is improved. It is possible to define the length of the
arrows, as well as to have a unique length for all the arrows (Display
parameters). The unit vector length is now shown on the screen during
snapshots and slices
An automatic scan of slices (scanning along the X-, Y- or Z-direction) can be
done at a given step (Display types)
The definition of slices (cuts) is improved. It is now possible to view or not the
geometry behind (or ahead) of a slice (Display types)
The "cut-off" capabilities were improved. It is now possible to specify two
cut-off values (Display types)
The display in cut-off mode in the presence of a free surface is now well
treated (Display types)
The information about the case status (prefix, displayed field), time, timestep
has been reviewed and is now shown on the bottom of the screen (Results
viewing/Introduction)
The free faces information is now generated in DataCAST and thus the
loading of a case is much faster
Full viewing capabilities are now available when symmetries are activated
(cut-off, free surface, ...)
Page 31
Solver
Version 2006.0
The flow solver for filling was extensively reviewed. In particular the flow
along walls was significantly enhanced (WSHEAR=2). The stability of the
solver was thus improved for filling (Flow Run Parameters)
A new Run Parameter (WALLF) was introduced in order to better account for
the surface quality of the mold (Flow Run Parameters)
The EDGE Run parameter was suppressed (due to the WALLF introduction)
The end of the filling can be shortcut with the new Run parameter ENDFILL
(Flow Run Parameters)
A new "Pool" model was introduced (RESERVOIR type) in order to facilitate
the modeling of cases with large horizontal free surfaces, such as tundishes
(Materials assignment)
Improved lost foam model
The inlet boundary condition was improved in order to preserve the right
amount of incoming mass
The Porosity model (POROS=1) was significantly improved in order to give
more accurate predictions (Porosity models)
The density definition in the case of porosity calculations was extended
(Density definition)
The gate feeding (in the case of injection - hpdc - lpdc) was changed in order
to be fully automatic (Active feeding)
Enhanced enthalpy formulation for more accuracy of the thermal solver
Exothermic materials can be modeled (Exothermic)
Large files (over 2 GB) can now be handled
The visco-plastic model for stress was adapted to account for creep (Stress
models and properties)
Fatigue prediction is available (Fatigue life indicator)
Hot tearing calculations are available (Hot tearing indicator)
Cracking indicator is introduced (Cracking indicator)
Page 32
General
On UNIX, the command window which is used to launch the manager is not
available for any further operations (as long as the manager is open).
On some platforms, there may be an error message when exiting the different
modules. This message can be ignored.
Manager
PreCAST
ViewCAST
Version 2006.0
Page 33
When displaying vectors in a slice, the contour of the slice does not appear.
It is not possible to superimpose a contour and vectors (in slices).
Sometimes, the particle tracing is giving wrong results at the last steps of the
filling.
In same cases, the automatic timestep update is freezing ViewCAST on SGI.
Work around : do not activate the automatic timestep update on SGI.
The visualization of result files larger then 2 GB is not possible on IBM (as the
executable of ViewCAST is 32 bits). Due to Tcl/Tk problems on IBM 64 bits,
it was not possible to create an executable of ViewCAST 64 bits. Work around
: use an other platform to visualize such results.
On UNIX, depending on the configuration of the X-Server, it was observed
that the performances of the graphics may be not optimum.
When a case is opened directly from ViewCAST (and not from the Manager),
if this case is located in a different directory (from the directory where
ViewCAST was launched) and if this directory contains a ".", the case can not
be opened. Work around : open the case directly from the Manager.
Particle tracing is not available in Tilt and Reservoir cases (i.e. when there is
no inlet velocity, inlet pressure on inlet BC).
In Tilt casting the velocity vectors will not "rotate" with the part and thus they
will appear in the wrong direction. Work around : deactivate the tilt rotation in
ViewCAST when viewing at velocity vectors.
It is not possible to visualize only a filter. Work around : select also an other
material.
In Lost foam calculations, one see non-zero velocities in the mould at step
zero. Just ignore these velocities which are meaningless.
It is not possible to view the Particle tracing with symmetries (i.e. the traces
are not replicated).
Solver
It is not possible to fill simultaneously two separated cavities. One should fill
only one cavity at a time or set (by hand in the p.dat file) the Run parameter
COLDSHUT = 0 (this will deactivate the cold shut detection algorithm).
In models where there are more than one RESERVOIR domain, they should
all touch each other and they should be all full at the beginning. Otherwise,
problems may be encountered.
PIPEFS must be set to zero if it is to be followed by an Advanced Porosity
calculation.
It is not possible to model temperatures below -273C (below absolute zero)
with the THERMAL=2 model. Work around : use THERMAL = 1.
When a domain defined as RESERVOIR is remaining full during the whole
filling, the Pool model will be activated at the wrong fill fraction (i.e. at
LVSURF minus the volume of the reservoir). Work around : set this domain to
CASTING instead of RESERVOIR, as if this domain remains full, it is useless
to define it as RESERVOIR.
MeshCAST
Version 2006.0
Page 34
On some Windows OS, the Open in MeshCAST is not done in the current
working directory.
Miscellaneous
Version 2006.0
Some error messages may appear during the use of CONVBIN, however, one
can ignore these messages.
Page 35
GETTING STARTED
SOFTWARE LAUNCH
Launch - Windows platform
When the software is installed on Windows, a directory named PAM-SYSTEM is
automatically installed on the Desktop of the computer. Inside this directory, a
"ProCAST 2006.0 Manager" icon is present.
To launch the software, double click on this icon and the ProCAST Manager will
start.
Version 2006.0
Page 36
The "on-line help" can be opened with the "Help" button (1).
The working directory should be set with the browse tools (2) .
The "Case" name (or prefix) should be defined (3).
The different modules can be called with the upper buttons (4).
The software is ready to be used.
Version 2006.0
Page 37
PROBLEM SET-UP
In order to get started, a very simple thermal problem will be set-up, step by step.
To do so, a very simple mesh located in the installation directory will be used.
Please note that the meshing, described in the MeshCAST manual will be
addressed in this section (please refer to the MeshCAST manual for more details).
The mesh can be found in the following directory :
On Windows : %ProCAST20050%\Test\Geometry\Test.mesh
On Unix / Linux : $ProCAST20050/Test/Geometry/Test.mesh
Software launch
This "Test.mesh" files should be copied in the desired working directory
(C:\Temp\Start in the following example). Then, the software manager should be
launched (see previous section) and one should browse to the working directory
and set the Case name (prefix) to "Test".
Pre-processing launch
When the "PreCAST" button is pressed (in the upper black band), the preprocessing module opens and the mesh ("Test.mesh" file) is automatically loaded.
Version 2006.0
Page 38
A window with some information about the loaded model (number of domains, of
nodes and elements, as well as the model size) appears. Any click on the menus
will automatically close this window. The geometry can be rotated interactively
with the mouse. The
icon can be pressed in order to view the geometry in
"Solid" mode.
Materials menu
The first operation to perform is to assign material properties to the two domains.
The first domain should be selected (1) and the desired material properties should
be selected in the database (2). An "H13-steel" is selected in this case. Then the
"Assign" button should be pressed in order to assign the selected material
properties to the given material domain (3). The material property assignment
should be repeated for the second material domain (let's select the "AL-7%Si0.3%Mg-A356" alloy.
Finally, the material type ("Mold" for the first domain and "Casting" for the
second domain) should be specified (4). This is done with a right click in the
"Type" column.
Version 2006.0
Page 39
Interface menu
In the interface menu, one should create the interfaces between the different
material domains and assign the desired interface heat transfer coefficients.
The type of interface should be specified. To do so, one should click once on
"EQUIV" (1), which will change into "COINC" (for coincident interface). The
"Apply" button (2), which is turning orange should be pressed to accept the
coincident interface type. A message will appear on the screen, which should be
acknowledged. The desired interface heat transfer coefficient should be selected in
the database (3) and assigned to the corresponding interface (4).
Version 2006.0
Page 40
Version 2006.0
Page 41
Process menu
The gravity should be defined in the "Process menu". To do so, click on the menu
"Process" and select the "Gravity" sub-menu. This will open the "Gravity" panel.
A gravity in the -Z direction is set and the panel is closed with the "Close" button.
Version 2006.0
Page 42
Version 2006.0
Page 43
Version 2006.0
Page 44
Version 2006.0
Page 45
CALCULATION
The calculation can now be launched in the Manager. To do so, the "ProCAST"
button should be pressed (1), which will open the confirmation window for the
run. The "Execute DataCAST first" should be activated (2) and then one can
confirm the launch of the solver by pressing the "Run" button (3).
On Windows, a Command window will open and the DataCAST and ProCAST
are automatically launched. A PAUSE message is displayed when the calculation
is finished.
Version 2006.0
Page 46
Once the calculation is finished, or anytime during the calculation, one can view
the results in the post-processor ViewCAST.
Version 2006.0
Page 47
RESULTS DISPLAY
To launch the post-processor, the button "ViewCAST" should be activated in the
Manager.
Once the ViewCAST window is opened, one can interactively rotate the mouse
(with a left click in the graphics area (1)). The material domains to be displayed
can be selected with the icon (2). The result to be displayed (i.e. Temperature,
Fraction of Solid, ...) should be selected in the "Contour" menu (3). In order to
display a snapshot, the icon (4) should be active. Then, one can display the
different timesteps with the tape player (5), either as an animation or as a step-bystep display. The scale can be changed by a click on the scale values (6).
Slices through the model can be obtained with the Slice icon (1). When no slices
are defined, it will automatically open the slice definition window (2). Then, one
should select the "Add ->" button and the "XYZ Plane" sub-menu, which will
open the slice selection panel (3). When the desired slice is selected, the Apply
button should be pressed and the slice will be displayed. One can select more than
one slice by repeating the "Add ->" operation. Finally, one can show the different
timesteps with the Tape player (4).
Version 2006.0
Page 48
Time evolutions can also be shown with the "XY Plot" menu. The desired field
(e.g. Temperature) should be selected and the desired locations to be plotted
should be defined (e.g. with the "Interval" sub-menu (a value of 50 nodes interval
was selected in the figure hereafter).
Version 2006.0
Page 49
To Quit ViewCAST, the File/Exit menu should be used. The software is asking
whether you want to store the last view. If it is stored, the current position and
orientation of the model on the screen, as well as the material selection, will be
retrieved at the next launch of the post-processor.
Version 2006.0
Page 50
SOFTWARE MANAGER
FILE MANAGER
When the ProCAST software is started, the Manager is launched, as seen in the
figure below. The Manager allows to browse in the desired working directory, to
create new directories and cases, to copy files from one directory to an other and
to launch the different modules of the software.
The default starting directory can be set in the "Installation settings" (see the
"Software configuration" section).
Firstly, the working directory should be defined (with the "browse"
"create directory"
should be defined.
Version 2006.0
and
Page 51
When several cases are present in the same working directory, the prefix of the
first one (in alphabetical order) is selected. In order to scroll through the other
cases, one can use the F2 key.
The File Manager allows to perform some actions on the files, such as "Refresh"
(to refresh the file list), "create new text file"
"Duplicate file"
, "Rename file"
A set of files can be copied from one directory to the other, using the
button. Then, two windows are opened (see figure
below) and files can be copied from the left window to the right one. If the "Keep
Prefix" check box is activated, the prefix remains unchanged, otherwise, the prefix
is changed to the one defined under "Case" in the lower orange section.
Version 2006.0
Page 52
The extensions "d.dat", "p.dat", "d.out" and "p.out" are well handled by the copy.
However, if it is desired to copy "unf" files, the user is prompted for the "old
prefix" and the "new prefix", as shown below :
Version 2006.0
Page 53
Version 2006.0
Page 54
MODULE CALLS
The different modules of ProCAST can be called from the top bar buttons :
Firstly, the FEM mesh should be generated with MeshCAST. Then the case
should be set-up with the pre-processor (PreCAST). Once all the parameters and
conditions are defined, the calculation files should be prepared with DataCAST.
The calculation can be run at this time by calling the ProCAST solver. Finally,
the results can be visualized with the post-processor ViewCAST. At any time
during and after the modeling, the status of the calculation can be monitored with
Status.
MeshCAST, the meshing module of ProCAST, is launched with the "MeshCAST"
button. Then the desired case should be opened into MeshCAST (there is no
automatic launch as the software does not know whether the .gmrst, the .sm or the
.mesh file should be opened). Please refer to the MeshCAST User Manual for
further details. A Command window is opened in the background in order to allow
the viewing of error messages if necessary. This Command window can be closed
at the end by a simple <RETURN>.
The Manager allows also to launch the "2-D mesh generator" or MeshCAST with
the -M option. To do so, a "right click" on the "MeshCAST" button should be
Version 2006.0
Page 55
done. This opens a sub-menu, where one can select between MeshCAST-2D or
MeshCAST-3D and MeshCAST-3D with the -M option.
This special mode of MeshCAST should be used only in the case of large memory
requirements (see the MeshCAST manuals for more details).
When PreCAST is launched, if a "d.dat" file is present, it will automatically be
loaded. Otherwise, the browser will prompt in order to select a ".mesh" file.
Version 2006.0
Page 56
When a right click in made on the "PreCAST" button, it gives access to a submenu in order to activate the "CAFE pre-processor"
Version 2006.0
Page 57
Version 2006.0
Page 58
Upon "Run", a Command Window will be automatically opened and the solver
will be started. As this window is independent from the Manager, it is possible to
use the Manager (and the other modules, such as the post-processing for instance)
while the calculation is running. When the calculation is finished, a PAUSE will
be prompted and one has to hit any key (e.g. RETURN) to make the window
disappear. If there is a problem during the execution, a warning or an error
message may be printed in this window.
The inverse solver can also be called from the "ProCAST" button (see the
Software configuration section for more details about how to activate the Inverse
option):
Version 2006.0
Page 59
Results can be viewed at any time (even during the run of the calculation), with
the ViewCAST button. If the case name is set, as well as the current working
directory, the case is automatically loaded in the post-processor, as shown below.
A right click on the "ViewCAST" button gives access to the "Advanced Porosity"
and the "CAFE" post-processors (i.e. the calcosoft post-processor) :
Version 2006.0
Page 60
To monitor the calculation, the "Status" button can be used at any time. A
graphical status of the case will be shown with the relevant information
concerning the calculation (i.e. percent filled, solid fraction, timestep, current
time, cpu time, ...) corresponding to the case name indicated in the lower orange
window. The status is refreshed every 15 seconds, unless the "Status" button is
pressed again. To exit the Status window, press the "Return to File Manager"
button. Please note that the Graphical status is working only when activated in the
working directory of the case (as the prefixp.dat file is read).
Finally, all these modules can be called manually from a Command Window. The
"Command Window" button opens automatically a Window in the working
directory.
Version 2006.0
Page 61
Then, the modules can be called in the following way (or via ad-hoc aliases) :
For Windows :
Manager
MeshCAST
PreCAST
DataCAST
ProCAST
ViewCAST
%ProCAST20060%\bin\manager.exe
%ProCAST20060%\bin\meshcast
%ProCAST20060%\bin\precast prefix
%ProCAST20060%\bin\datacast prefix
%ProCAST20060%\bin\procast prefix
%ProCAST20060%\bin\viewcast prefix
$ProCAST20060/bin/manager.exe
$ProCAST20060/bin/meshcast
$ProCAST20060/bin/precast prefix
$ProCAST20060/bin/datacast prefix
$ProCAST20060/bin/procast prefix
$ProCAST20060/bin/viewcast prefix
The Volume mesher of MeshCAST can be called in batch mode with the
following command :
On Windows :
%ProCAST20060%\bin\meshcast -batch prefix.sm Flag
On Unix (tcsh, csh) :
$ProCAST20060/bin/meshcast -batch prefix.sm Flag
"Flag" is an integer value defining :
0 : no layer
1 : partial layer
2 : full layer
Version 2006.0
Page 62
In the "Preferences" tab, one can select which "modules" should be activated
(under "Modules display") in the Manager :
Version 2006.0
Page 63
Version 2006.0
Page 64
Version 2006.0
Page 65
RUN LIST
The "Run List" allows to launch automatically several calculations in a row, as
well as to "chain" different calculations (such as a ProCAST thermal calculation
followed by an Advanced Porosity calculation). This allows also to launch
calculations in batch mode.
When the "Run List" menu is selected the Manager window appears as follows :
If it is create a "Run List" from cases in the same directory, one should check that
the desired case name is specified in the bottom window, then the desired solver
should be selected in the left window (Available Solvers) and the "Add button"
Version 2006.0
Page 66
should be pressed. The corresponding run will appear in the right window (Solvers
to be run). Then, one can add as many runs as desired (see example hereafter
where the Case1 and Case2 are launch sequentially).
If the cases to be run are not all in the same directory, one should first select the
Solver to be run (in the left window) and then press the "Folder" button (above the
"Add" button). This is opening a browser, where one can select the desired
directory and the desired case to be run. Once the "Open" button of the browser is
pressed, the corresponding line in the right window is added. With this method,
one can run sequentially a serie of calculation which are located in different
directories.
It is also possible to remove an entry from the right list, by selecting it and
pressing the "Remove" button.
Once the list is ready, the "Run Computations" button can be pressed. This will
automatically launch the calculations. Otherwise, it is also possible to save the
"Batch file", with the "Create Batch File" button. This file will automatically store
the batch file in the working directory.
Version 2006.0
Page 67
One can also edit the batch file or run it with the corresponding buttons. In these
two cases a browser will open automatically, to select the desired file.
Version 2006.0
Page 68
SOFTWARE CONFIGURATION
The software can be configured in the "Installation Settings" window :
Preferences tab
The Preferences can defined, such as the Default directory which will appear
every time the Manager is launched, the preferred Text editor can be configured.
The access to the "Advanced modules" (Parallel, Advanced Porosity and Inverse)
can be configured in this window (see the "Advanced module calls" section for
more details).
Version 2006.0
Page 69
ViewCAST allows now a full compatibility between UNIX and Windows results.
This means that it is possible to visualize on a Windows machine, results which
have been calculated on a Unix machine, or vice-versa.
To do so, one should "Activate the swap mode", as shown below.
Then, the user can select the desired default mode between the two above choices.
This default mode will be used when the "ViewCAST" button will be pressed.
However, it is still possible to use the other mode (i.e. Swap format), by making a
right click on the "ViewCAST" button. Then, one can chose between the "No
swap format" (which means that ViewCAST and the ProCAST run were
performed on the same plateform) or the "Swap format" (which means that
ViewCAST and the ProCAST run were performed on a different plateform).
Version 2006.0
Page 70
In the "Installation" tab, one can check that the environment variable
ProCAST20060 is well configured. It should point towards the v2006.0 directory,
above the "bin", "dat" and "lib" directories.
The "Command window Launch Option" check box (in the bottom of this
Installation Settings window" allows to activate the Command windows when
PreCAST and ViewCAST are launched. This mode should be activated in case of
problems in order to see the possible error messages which may occur. In normal
situation, this mode should be deactivated.
The "DOS Path Converter" (on Windows only) can be useful to convert "long
directory names" into short "DOS" ones.
In ViewCAST, the time is displayed at the bottom of the screen. By default, the
time is shown in seconds. However, in the case of large casting modeling, it may
be useful to see the time in hours:min:sec, whereas for high pressure die casting,
milliseconds are preferred. The default settings can be set in the file "manager.tcl",
which is located in the "bin" directory of the installation.
At about 30 lines from the top of the file, there is the following lines :
### Viewcast ###
global viewcastOption
set viewcastOption "-T1"
Version 2006.0
Page 71
Version 2006.0
(display in seconds)
(display in hours:min:sec)
(display in milliseconds)
Page 72
CUSTOMIZED INSTALLATION
In some cases, the software is installed in a centralized location of the company.
In previous versions, it was necessary the set the installation directory of the
software with Read/Write access rights, in order to allow the access to the
databases and to the status files.
From version 2004.1, ProCAST allows to install the software in a Read only
location and to customize the location of the files which have to be modified by
the user (databases, status files, preferences).
To do so, two optional Environment variables are available :
ProCAST20060_DB for the location of the databases.
ProCAST20060_USER for the status files and preferences.
If a user, or a group of users, would like to have the databases in a different
location then the software installation, one should copy the "db" directory in the
desired location (e.g. /my_db_location/db). Then, the ProCAST20060_DB
environment variable should be set (manually) to the /my_db_location directory.
In the same way, if a user would like to set it own preferences and status files, one
should copy the "pref" and "statpro" directories in the desired location (e.g.
/my_pref_location/pref and /my_pref_location/statpro). Then, the
ProCAST20060_USER environment variable should be set (manually) to the
/my_pref_location directory.
Please note that the "Preferences" directory is containing preferences of the
Manager, as well as the User customized icons of MeshCAST.
Version 2006.0
Page 73
FLEXLM
The software is locked using the FlexLM tool.
For "Node locked" licenses, the following procedure can be used (for floating
licenses, please refer to the general FlexLM documentation).
The following is valid for Windows only. For Unix, please refer to the FlexLM
documentation.
In order to activate your license, a license file should be requested to ESI. The first
step is to identify the machine on which the software has to run. To do so, run the
"FlexLM License Manager", by a double click on the corresponding icon in the
"PAM-SYSTEM" directory which is on your desktop.
This will open the following panel. Go to the second tab "System Settings" (1) and
send this panel to your local distributor. You can also transform this information
in a text file, using the "Save HOSTID Info to a File" button (2).
Version 2006.0
Page 74
Then, the licence file which is received from ESI should be called "pam_lmd.lic"
(usually, it is called "*****.cry" and you should rename it to "pam_lmd.lic" previously it was called "license.dat"). Then, the "pam_lmd.lic" file should be
placed in the C:\flexlm directory.
When one of the following modules (PreCAST, ProCAST or ViewCAST) is
called for the first time, the following screen will be prompted :
Version 2006.0
Page 75
The location of the "pam_lmd.lic" file should be specified. The "Browse" button
could be used to find the "C:\flexlm\pam_lmd.lic" file :
Version 2006.0
Page 76
When the "Finish" button is pressed, the software is unlocked and all the modules
can be used.
Version 2006.0
Page 77
PRE-PROCESSING
INTRODUCTION
To start the Pre-processor, the "PreCAST" button should be used. If a mesh file
(case.mesh) or a "d.dat" file (cased.dat) is present, the case will be automatically
opened.
If there is no case present in the working directory, then the browse window will
open so that the user can select the desired input files (see the Geometry import
section).
When the case is loaded, a window appears with some information about the
model, such as the number of materials, the number of nodes and elements, as
well as the model size. Then, the pre-processor is ready to set-up a case.
Version 2006.0
Page 78
The top bar menu is divided in 9 menus which allow to perform all the operations
to set-up a case :
File
Geometry
Materials
Interface
Boundary Conditions
Process
Initial Conditions
Run Parameters
Inverse
Help
First the model should be opened or saved in the File menu. It allows also to quit
the Pre-processor.
Then, in the Geometry Menu, symmetries can be defined, as well as the virtual
mold characteristics. Moreover, some features of the FEM mesh can be checked,
such as negative Jacobians, or the volume of each domain.
Version 2006.0
Page 79
In the Material menu, the characteristics of each domain (or each material) can be
defined. In addition to the material properties, one can specify the type of the
domain (casting, mold, filter, foam, ...), as well as if it will be empty or not at the
beginning of the calculation (for mold filling).
The Interface menu has no sub-menus. It opens the window which allows to
define the interactions between the different materials, such as heat transfers.
The Boundary condition menu allows to define all the interactions between the
different materials and the outside world (i.e. on the outside surfaces of the
model), such as external cooling, velocities at the surface of the model for flow
calculations, displacements or constraints for stress calculations, etc...
The Process menu gives access to the definition of the gravity, as well as the
definition of the motion of the different domains or enclosures.
Version 2006.0
Page 80
The initial temperatures of each materials are defined in the Initial Conditions
menu.
The Run Parameters menu, as well as the Help menu have no sub-menus. All
the calculation parameters are defined in the Run Parameters window. The on-line
Help can be access from the Help menu.
Below the menus, icons allows to perform a number of operations linked to the
display of the model on the screen. These icons are described in the Geometry
manipulation section.
The next sections are presenting the set-up of a case, according to the following
flow chart.
Version 2006.0
Page 81
GEOMETRY IMPORT
When PreCAST is started from the Manager, it is automatically reading a mesh or
a d.dat file (if they are present in the working directory, with the selected prefix).
The priority is first a "d.dat" and then a ".mesh" file.
If there is no mesh file or d.dat file present in the working directory with the
corresponding prefix, the pre-processor opens with the browser window.
Then the user has the choice of the input format, through the following filter :
The Pre-processor is able to read PreCAST input files (*d.dat), also called
"Restart" files, or meshes coming from MeshCAST (*.mesh), from PATRAN
(*.out) or I-DEAS (*.unv).
Version 2006.0
Page 82
Version 2006.0
Page 83
THERMAL
Thermal model
The Thermal module allows to perform a heat flow calculation, by solving the
Fourier heat conduction equation, including the latent heat release during
solidification. The typical results which can be obtained are the following :
Temperature distribution
Fraction of solid evolution
Heat flux and thermal gradients
Solidification time
Hot spots
Porosity prediction
Flow chart
This section describes the set-up of a thermal case. It is also the opportunity to
introduce the general work flow of the pre-processor, as well as some aspects
which may be used by different modules (e.g. symmetry).
Each step described in the above flow chart is described in the following sections.
Version 2006.0
Page 84
Geometry assignments
Once the geometry is loaded, the following operations can be performed on the
geometry :
Symmetries can be defined at this stage (see the "Thermal/Radiation" section for
more details). The definition of the Virtual mold is also done at this level. (see the
"Virtual Mold" section for more details).
In the "Check Geom" menu, the following features are accessible :
Materials assignment
Once the model is loaded (see the "Geometry import" section), the first operation
is to define the different materials with their properties and attributes. This is
performed in the Material/Assign menu.
On the right of the window, two frames are shown. The top one contains the
material list (or domain list), whereas the bottom one corresponds to the material
database.
Version 2006.0
Page 85
When one clicks on the different materials in the material list, the corresponding
domains are highlighted (in the picture below, the hidden mode with mesh
was selected - see the "Geometry manipulation" section for the other display
modes).
Version 2006.0
Page 86
In the lower frame ("Material database" list), the list of all available material
properties in the material database is displayed. To manage the database entries,
please refer the "Databases" sections. The {T} or {F}which are indicated before
the material name are telling wheter material properties are present in this material
for Thermal only calculations (T) or for Thermal and Fluid flow calculations {F}.
If a {*} appears, it means that the material properties definition is uncomplete and
that this material entry can not be used for a calculation at this stage.
In the top frame ("Domain list"), all the domains (or materials) present in the mesh
are listed. When a mesh is loaded, it appears as follows :
Version 2006.0
Page 87
Then, one should assign a Material to each domain, to define the type of each
domain and to specify whether the domain is empty or not at the beginning of the
calculation.
To assign Materials, (1) one should select the desired domain in the upper list, (2)
select the desired material in the material database list, and (3) click on the Assign
button. This should be repeated for each domain.
Then, the "Type" of each material should be defined. To do so, make a right click
on the "CASTING" word and the available list of possible selection will appear :
Version 2006.0
Page 88
Mold : the mold material should be set to "Mold". This will be used for cycling
calculation (in die casting) in order to allow the calculation of the heating of a die
during cycling (i.e. the temperature of the mold domains will not be reset to the
initial temperature at the beginning of each cycle).
Casting : the casting material should be set to "Casting". This setting is necessary
in particular for all the domains where fluid flow will occur. For a cycling
calculation, the casting domains initial temperatures will be reset at the beginning
of each cycle.
Filter : filter domains should be set with the "Filter" type (see the "Filters" section
for more details).
Foam : for lost foam calculations, the domains where the foam is present at the
beginning of the calculation should be set to "Foam". Of course, during the filling,
the casting material will replace the foam, as it burns.
Insulation : this type has no specific effect on the solver. It will correspond to a
"Mold" type of material. At this moment, this is for information purposes.
Exothermic : this will activate the Exothermic properties of the sleeve (if they are
defined in the corresponding material properties). If the material properties are
containing the exothermic information, but the "Exothermic" type is not activated,
the exothermic model will not be activated (see the "Exothermic" section for more
details).
Core : a core type material should be defined in the case of cycling, where cores
are placed into the mold at each cycle. This means that unlike mold materials, the
initial temperature of the cores will be reset at the beginning of each cycle.
Reservoir : a Reservoir type material is a domain where the free surface will
always be perpendicular to the gravity. This allows to simplify the free surface
computation and it is especially useful in the case of tundish modeling. The
"RESERVOIR" domains should have an "EQUIV" interface with the other
CASTING materials.
Version 2006.0
Page 89
Please note that if there is more than one RESERVOIR domain, they should all
touch each other and they should be all full at the beginning. Otherwise, problems
may be encountered.
If one want to empty the reservoir, no special BC should be specified. If one
would like to keep the reservoir full, a pressure BC should be set at the surface of
the reservoir.
Finally the user has to specify which domains are empty at the beginning of the
calculation (for mold filling calculations). One should make a left click on the
"No" to turn it to "Yes" (which means that Yes the domain is empty). On
additional click returns to No. Of course, more than one domain may be empty (if
the casting is made out of several mesh domains).
When one makes a right click on the material name in the upper list, the lower list
is pointing on this particular material. This is very useful if one wants to see the
material properties of this material.
Interfaces assignment
Once the Materials are defined, one should define the Interfaces, with the
Interface menu.
As for the Material window, two frames appear on the right of the window. The
top one contains the list of all the possible interfaces, whereas the lower one
shows the Interface database. To manage the database entries, please refer the
"Databases" sections.
Version 2006.0
Page 90
Firstly, one should define the type of interfaces in the upper right window :
On the left, the "Material Pair" are shown. "1 and 3" means that there is an
interface between material 1 and material 3. By default, the Type of the interface
is set to "COINC". By clicking on the "COINC" text, one can toggle between
"COINC", "NCOINC" and "EQUIV".
When a mesh is generated with MeshCAST (or with most common mesh
generator), the elements which are on either side of an interface (i.e. adjacent
elements which belongs to two different domains) are sharing the same nodes.
This is called a coincident mesh.
Version 2006.0
Page 91
EQUIV option
When two domains are part of the same entity (i.e. they both belong to the casting
with the same material properties, but they were meshed separately for technical
reasons), one will set an "equivalenced" interface between them (EQUIV). It
means that there will be a continuum between the two domains, with a continuous
temperature profile across the interface, as well as continuous velocity field. In
such a case, the nodes at the interface (shown in orange in the figure below) are
shared by the elements on both sides. This EQUIV option can also be used if one
has different materials in the two domains, but the materials are welded together
(i.e. with a total bounding between the two materials).
COINC option
At an interface between two different materials, such as the casting and the mold,
there is usually a temperature drop. In this case, the nodes at the interface should
be doubled (for a coincident interface), in order to distinct temperature on each
side of the interface. As during the mesh generation, there is one node at the
interface, it is necessary at this stage to duplicate all the interface nodes (as shown
in green in the figure below). This duplication operation is performed when
"COINC" is selected (for "coincident nodes"). The interface, which is shown in
yellow in the figure below has in fact a zero thickness.
NCOINC option
It is also possible to generate a non-coincident mesh (i.e. where the elements on
both sides of the interface are not matching, which means that they are not sharing
the same nodes), by adding different meshes together (see the "Advanced
Version 2006.0
Page 92
features" section for more details on non-coincident meshes). In this case, one has
to specify that the interface is non-coincident, with the "NCOINC" option.
When one toggles between the different options, the interface appears in red and
green. It is thus possible to well identify whether it corresponds well to the desired
interface (see figure below, which was obtained in hidden mesh mode
An other way to view the desired interfaces is to click on the "Material Pair" and
the material on both sides will be highlighted in red and green respectively - the
first material in the list is in red and the second one in green (see figure below,
which was obtained in hidden mesh mode
Version 2006.0
Page 93
Once the desired selections (between COINC, EQUIV and NCOINC) are done for
each possible interface, the STORE button (which is highlighted in orange) should
be pressed.
Then the pre-processor will automatically create the double nodes and a message
will appear to confirm that the number of nodes of the model has increased.
Version 2006.0
Page 94
As it is now possible in version 2006.0 to go from EQUIV to COINC and viceversa, the user should be careful that as nodes were duplicated or removed, some
boundary condition assignments may be corrupted, as well as the extracted initial
conditions. Thus, in this case, they should be re-assigned.
Once the types of interfaces are defined, one has to apply the corresponding heat
transfer coefficients (for COINC and NCOINC only, as nothing as to be specified
for EQUIV). To assign Interface heat transfer coefficients, (1) one should select
the desired Material Pair in the upper list, (2) select the desired interface heat
transfer coefficient in the interface database list, and (3) click on the Assign
button. This should be repeated for each coincident or non-coincident interface.
Version 2006.0
Page 95
After the definition of the interfaces, the Boundary conditions should be specified.
This is done in the "Boundary Conditions" menu.
Assign Surface
The principles of Boundary conditions definition are described in the following
figure.
Version 2006.0
Page 96
Firstly, the desired boundary conditions should be "Added" (1) in the upper list
(2). With the "Add ->" button, one has the following choices :
Version 2006.0
Page 97
For some boundary conditions, such as "Heat" or "Symmetry", faces of the Finite
Element Mesh are selected. For other boundary conditions, such as
"Temperature", "Velocity", ... the boundary conditions are applied on nodes. The
choice between faces or nodes is automatic.
Select all.
Deselect all.
Select and propagate. All the faces or nodes which have an angle with the
neighbors smaller than the specified propagation angle will be selected.
Deselect and propagate. All the faces or nodes which have an angle with the
neighbors smaller than the specified propagation angle will be deselected.
Select remainder. The remaining faces or nodes which have not yet been
selected are selected.
Select interface. For "Heat" boundary conditions, the selections are applied
on external faces only. However, for cycling, one would like sometimes to apply a
Heat BC on faces which lie at interfaces. This allows to select those interfaces
automatically. In this case, the following panel is opened. It is proposing the list
Version 2006.0
Page 98
off all active materials. One could select one or more material and all the faces of
the selected materials which are lying on an interface will be selected.
Copy of selection. All the nodes or faces which are applied on the geometry
for the active boundary condition will be copied in the memory.
Assign Volume
In the "Assign Volume" menu, the following entities can be assigned :
Version 2006.0
Page 99
Volumetric Heat corresponds to a "heat source" in the entire volume (see the
"Boundary Conditions Database" section for more details).
Surface Heat allows to specify how the free surface inside a given material
domain is cooling down. It is possible to assign a "Heat" boundary condition to
the free surface inside the domain. With this, it is possible to assign a convective
heat transfer coefficient and an external temperature which will drive the cooling
condition of the free surface itself. For instance, this could be important to take
into account the cooling of the free surface of a swimming pool. Please note that it
is not possible to activate the radiation model with view factors on a free surface.
Momentum Source corresponds to a force term on the liquid (momentum force) in
the entire volume (see the "Boundary Conditions Database" section for more
details).
Mass Source corresponds to a different way of applying an inlet of metal in the
case of filling (see the "Boundary Conditions Database" section for more details).
Filter Heat allows to define the interface heat transfer coefficient between the
liquid metal and the filter material. Thus, the cooling of the hot liquid, when
passing through an initially cold filter can be modeled using a "Filter heat". The
interface heat transfer coefficient (which is the same as the one defined in the
"Interfaces assignment" section) is applied to the entire filter domain.
Assign Enclosure
In the case of radiation problems with View Factors, the ambiance or the furnace
can be modeled with an enclosure. See the "Pre-Processing/Radiation" secton for
more details. This menu is not used for Thermal problems without "View Factor
radiation".
Gravity
Version 2006.0
Page 100
For Thermal only problems, it is important to define the gravity direction for the
calculation o of the porosity (using the POROS=1 model).
The gravity is defined in the following panel. For standard problem, a constant
gravity (1) is defined. The three components of the gravity are defined in (2). If
one clicks on the X, or the Y, or the Z letter, automatically, the gravity vector is
set in the X-, Y- or Z- direction. With two clicks, the negative direction will be
set.
The "Rotate" is used only for Fluid flow problem (see the "Fluid Flow & Filling"
section for more details).
Assign Volume
This option is used to specify a translation, a rotation or a revolution to the
material domains. As this is mainly used in the case of Radiation problems, please
look for more details in the "Pre-processing/Radiation" section.
Assign Enclosure
This option is used only for radiation problems (see the "Preprocessing/Radiation" section for more details.
Version 2006.0
Page 101
The initial temperature can be defined either as a Constant value throughout the
material domain, or as an Extracted temperature field coming from a previous
calculation.
In the case of "Constant", the list of all the Material domains is displayed and one
can enter the initial temperature (in the lower white field) for each one.
In order to apply an extracted temperature field (e.g. temperature at the end of the
cycle N 10, as initial condition for a filling calculation), one should perform a
first calculation on the same mesh. Then, one could specify the case, as well as the
step to be considered for the initial conditions. In the case of an "Extract", the
following screen appears.
Version 2006.0
Page 102
All available material domains are listed in (1). The desired materials are selected
in this list and the case name from which the thermal field must be extracted (2)
should be selected with the Browse button (3). The directory (4) and the prefix
should be selected (5). Then, the timestep (6) should be defined in the field (7).
When the prefix (and the directory) of the desired case is selected, if one clicks on
it in the "prefix" column (2), the extracted temperatures are shows (see picture
below).
Version 2006.0
Page 103
Version 2006.0
Page 104
Version 2006.0
Page 105
Filling behavior
Free surface evolution
Natural and forced convection currents
Dynamic pressure of the liquid
Entrapped gas
Filter behavior
Flow chart
This section describes the additional set-up necessary for fluid flow and mold
filling. For the set-up of a thermal case, please refer to the "Thermal" section of
the Pre-processor.
Materials menu
Version 2006.0
Page 106
Concerning the material assignments, the only requirement is that the domains in
which fluid flow calculation will be performed are defined as "Casting" domains
(or "Filter" or "Foam" - see the "Advanced features" section for more details).
Then, one should define whether the corresponding domain is Empty or not (in the
case of mold filling). One can change from No to Yes and vice-versa by clicking
on the text directly.
Moreover, the material properties of the fluid material should have the flow
properties (e.g. viscosity). This can be checked in the Material database list. The
Material name should be preceded by an {F} (for Fluid properties).
Interface menu
Concerning the interfaces, nothing special should be done concerning the fluid
flow. One can however notice that if the casting (i.e. the flow domain) is noncoincident with the mold domain(s), one will need to set a zero velocity boundary
condition all around the casting domain (in order to prevent "leaks"). See the
"Advanced features" section for more details on non-coincident meshes.
Version 2006.0
Page 107
Process menu
For fluid flow problems, the gravity has to be defined (see the "Thermal/Process
Conditions Assignment" section for more details).
Version 2006.0
Page 108
The fluid flow module should be activated with FLOW = 1 or 3, as well as the
free surface model (FREESF) for mold filling and the gas model (GAS). The
storage frequency of the fluid flow results (VFREQ) should be specified. Then,
information about the reference pressure (PREF) and pressure driven inlet
(PINLET), about the final fill fraction (LVSURF) and the filling parameter
(WSHEAR, WALLF) shall be defined.
Please refer to the "Flow Run Parameters" section for the full description of all the
Run Parameters.
Version 2006.0
Page 109
RADIATION
Radiation model
The radiation module allows to perform complex radiative problems (e.g.
investment casting), with the calculation of the shadowing effects (view factor
calculations). The typical results which can be obtained are the following :
Version 2006.0
Page 110
The figure below presents a casting within a surface enclosure shown in grey (for
symmetry reasons, only half of the geometry is modeled)
Flow chart
This section describes the additional set-up necessary for radiation calculations.
For the set-up of a thermal case, please refer to the "Thermal" section of the Preprocessor.
Version 2006.0
Page 111
Geometry/Symmetry menu
Firstly the symmetries (if any) should be defined. ProCAST is able to deal with
symmetry implying one mirror, two orthogonal mirrors, a single rotation of n
sectors and a combination of them.
In order to illustrate the different possibility of symmetry, consider the simple but
explicit example of two concentric cylinders (the inside cylinder is the casting and
the outside cylinder is a solid enclosure). The full geometry is shown in the figure
hereafter.
Version 2006.0
Page 112
Version 2006.0
Page 113
Version 2006.0
Page 114
Two examples of simplified geometry by a rotation (left: 8 sectors, right: 4 sectors) and
one mirror.
A rotation of n sectors (R) associated with two orthogonal mirrors (M1 and
M2)
Version 2006.0
Page 115
Version 2006.0
Page 116
Nothing specific should be defined at the level of the Materials and Interface
menus. The complex radiation method should be set in the "Boundary conditions"
and "Process" menus.
The enclosure can be divided in different sets (1) (in order to apply different
temperatures and/or emissivities). One can Add new sets (2). Then, using the
selections tools (3) (see the "Pre-processing/Thermal" section for the full
description of the selection tools), the desired surfaces of the enclosure can be
"painted" (4) and stored (5). Finally, the desired temperature and emissivity can be
selected in the database (6) and assigned (7). Please note that one should assign to
each set both a Temperature entry and an Emissivity entry.
In order to compute well the View Factors of the enclosure, one should make sure
that the surfaces of the enclosure (i.e the triangles or the quadrangles) are well
oriented (i.e. the surfaces are pointing inwards). To do so, the following icons are
available next to the selection tools :
Version 2006.0
Page 117
If all the arrows are pointing outwards, the "Reverse" icon allows to reverse the
orientation of all arrows.
If only a few arrows are pointing in the wrong direction, one should use the
"Align" icon in order to have all the arrows pointing in the same direction. Then,
one may need to use the "Reverse" icon to point the arrows inwards.
Process menu
The process menu allows to define the motion (if any) of the enclosure (Assign
Enclosure) with respect to the casting or of material domains (Assign Volume).
Version 2006.0
Page 118
In the case of "Assign Volume", the same type of panel appears, but instead of the
enclosure sets, the different material domains are listed. Again in this case, it is
possible to specify a Translation, a Rotation or a Revolution. Any combination of
the three is possible, however, please note that the user should check that there is
no conflict between these motions and prevent any inter-penetration of the
different materials and/or enclosures.
On the bottom right, the motion database (Process database) appear (see the
"Process Database" section for more details).
Version 2006.0
Page 119
Version 2006.0
Page 120
Version 2006.0
Page 121
STRESS
Stress models
The stress module allows to perform a thermo-mechanical calculation. The typical
results which can be obtained are the following :
Stress distribution
Deformations (elastic and plastic)
Displacements
Gap formation
Elastic springback
Die fatigue
Hot tears
Cracks
In order to address these different aspects, five different stress models are
available.
Linear Elastic
Elasto-plastic
Elasto-viscoplastic
Rigid
Vacant
Flow chart
This section describes the additional set-up necessary for stress calculations. For
the set-up of a thermal case, please refer to the "Thermal" section of the Preprocessor.
Version 2006.0
Page 122
Materials menu
The stress properties should be assigned in the Materials/Stress menu.
This is opening the following window. Firstly, each domain should be selected
(1). Then, the desired properties should be chosen in the database list (2) and then
the "Assign" button (3) should be pressed in order to link the material with the
corresponding stress properties.
The stress properties, as well as the different stress models are described in the
"Stress database" section.
Version 2006.0
Page 123
Interface menu
Nothing specific should be defined at the level of the interfaces.
Version 2006.0
Page 124
One should be careful not to over constrain the model with too many
displacements boundary conditions. If this is the case, this may induce artificial
stresses locally, as shown in the figure hereafter.
Version 2006.0
Page 125
Version 2006.0
Page 126
One should be aware that the application of only a "Surface load" on the top die
(as shown below) is not enough.
This can be explained with the following sketch.. In the left case (blue), only two
loads are applied to the part. This part is not constraint and it can move in any
direction. In order to make sure that the part will not move, one has to apply in
addition the appropriate zero displacements (right case in green). In the center
Version 2006.0
Page 127
case (orange), the displacement constraints are not sufficient and the part could
rotate around the bottom right point.
Finally, "Point load" can be applied locally to model the effect of a local force.
Version 2006.0
Page 128
The stress module should be activated with STRESS = 1. The storage frequency
of the stress results (SFREQ), as well as the calculation frequency of the stresses
(SCALC) should also be specified.
Please refer to the "Stress Run Parameters" section for the full description of all
the Run Parameters.
In addition, when a Stress calculation is performed, one should set PIPEFS = 0.
This allows to prevent an unexpected effect of the piping on the stress calculation.
Version 2006.0
Page 129
opened (right picture), the corresponding mold resistance is removed and the
stresses are relaxed. As the elastic strain is released, the shape of the part is
changing (on the right picture, both the part and the mold are represented,
although the mold is not anymore present).
To model the Elastic Springback effect, the user just needs to define a function for
the interface heat transfer coefficient (at all the interfaces which will not be
anymore in contact when the mold opens). One should define a zero interface heat
transfer coefficient when the mold opens and automatically, the mechanical effect
of the mold will be removed in the model.
However, if one want to continue to simulate the cooling of the casting, while in
the air, the user should not forget to set a Heat boundary condition on the casting
surface (and eventually on the inside mold surface) with a non-zero value from the
opening time (the technique is the same as the one used in cycling calculations).
Version 2006.0
Page 130
DATABASES
Material Database
All the material properties are stored in Databases. This section describes the
database containing the material properties for Thermal and flow calculations. The
Database containing stress properties is described in the "Databases / Stress
Database" section. Moreover, the "Thermodynamic" databases are described in the
"Databases / Thermodynamic databases" section.
In the Material Assignment menu (see "Material assignments"), the content of the
material database is shown (see below). Above the material list, the database
management buttons are present :
Read : the database entry can be read and modified (if the
user has the appropriate rights)
Add : a new entry can be created
Copy : an existing entry can be copied in order to create a
new entry
Del : an exisiting entry can be deleted
Sort : the list of materials is sorted by alphabetical order
Search : a search on the material name can be done
When an existing material is opened (with the READ button), a window appears
with the following content :
Version 2006.0
Page 131
Firstly, the material name, the user, as well as the creation date of the material is
shown. Then, the material properties are organized in different tabs, with a
hierarchical structure :
Composition
Thermal
Conductivity
Density
Specific Heat
Enthalpy
Fraction Solid
Latent heat
Liquidus-Solidus
Exothermic
Fluid
Viscosity
Surface Tension
Permeability
Filter
Comments
The yellow tabs indicate that values are defined for the corresponding properties.
The white tab shows the active one. Once the "Thermal" or "Fluid" tab is selected,
a second level of tabs becomes active (and so on). The figure below shows the
definition of the thermal conductivity :
Version 2006.0
Page 132
The above figure shows the standard panel for material properties definition.
Firstly, the user has to define whether the material property is defined by a
Version 2006.0
Page 133
Version 2006.0
Page 134
Beside these properties, some of them are defined only by constants, such as the
latent heat, the liquidus and solidus temperature or the filter properties. In this
case, only the corresponding constant(s) should be entered (see figure below).
The following section "Material Properties" describes the different properties and
when it should be defined.
Material Properties
Good material properties are the best base for a good simulation. Properties could
be found in several locations, such as litterature, material suppliers, universities,
web, ...
www.matweb.com is material properties website which contains many useful
data.
www.matdata.net is a search engine for material properties.
An other way to obtain material properties is through the Thermodynamic
Databases which are embedded into ProCAST (see the "Thermodynamic
Databases" section for more details).
Thermal problems
Version 2006.0
Page 135
No phase change
For thermal problems (with or without solidification), the minimum data which
are required are the following (typically for mold materials) :
Thermal conductivity
Specific heat
Density
When solidification is present (i.e. for casting materials), one should define in
addition the following properties :
Fraction of solid
Latent heat
Liquidus and Solidus temperatures
where cp(T) is the specific heat as a function of temperature, L is the latent heat
and fs is the fraction of solid.
As there are two ways of defining the phase change, the software is automatically
detecting if there is a conflict in order to have either :
Version 2006.0
Page 136
specific heat
Latent heat
or
enthalpy
The user will need to resolve the conflict, by selecting which data are to be kept
(i.e. either enthalpy or specific heat/latent heat) for each material which has this
duplicate definition.
Density
The density is used in thermal calculation (it multiplies the specific heat, the
enthalpy and the latent heat), as well as in fluid flow calculations and in porosity
calculations. Please refer to the "Porosity models" section for more details about
the density definition.
Version 2006.0
Page 137
For fluid flow problems, it is mandatory to define the viscosity. Then, optionnal
definitions are available, such as "Surface Tension", "Permeability" and "Filter".
Viscosity
with :
strain rate
zero strain rate viscosity
infinite strain rate viscosity
n
a
phase shift
Power law coefficient
Yasuda coefficient
The above parameters can be defined in the database (as constants or as function
of temperature) as follows :
This option is not described as it has not been validated at this stage. One can say
that in conventional casting processes, the surface tension effects are certainly
negligible in comparison to the simplifications made in the free surface
algorithms.
Version 2006.0
Page 138
Permeability
The void fraction (Fv) corresponds to the amount of "porosity" or void inside the
filter. This value is dimensionless [-]. The definition of this value is mandatory in
all cases.
The Surface area (Sa) corresponds to the amount of "interface" between the filter
material and the air (when the filter is empty) per unit volume (see example
below). This value is used for the calculation of the thermal exchange between
the filter and the liquid metal going through, as well as for the automatic
permeability calculation. The units are the reversed of a distance (e.g. [1/m]). The
definition of this value is mandatory in all cases.
Version 2006.0
Page 139
The Permeability of the filter (i.e. its resistance to the flow) can be calculated in
three different ways.
a) Automatic permeability calculation
From the Void Fraction (Fv) and the Surface Area (Sa) definitions, the
permeability can be automatically computed (based upon Karman-Cozeny),
according to the following relationship :
This mode is activated if the "Pressure Drop" and "Permeability" tabs are not
defined.
b) Pressure drop calculation
If the "Pressure Drop" tab is defined, then, the permeability is calculated, using
the following values (coming from simple experiments) :
Version 2006.0
Page 140
Thermodynamic Databases
Version 2006.0
Page 141
Material properties, such as the enthalpy curve and the solidification path (i.e. the
fraction of solid curve versus temperature), density, viscosity and thermal
conductivity can be computed automatically from thermodynamic databases.
ProCAST has an automatic link with thermodynamic databases to calculate these
properties. It is thus possible to compute the enthalpy curve, the fraction of solid
curve, the density, the viscosity and the thermal conductivity, based upon the
chemical composition, for the following systems and the following alloying
elements (the elements shown in blue were added in version 2006.0) :
The other alloying elements which are not present in this list are not available in
the database and will have no effect on the computed material properties. More
details about composition limitations are given in the next section.
To activate the thermodynamic database, one should go in the "Composition" tab
of the material properties window.
Version 2006.0
Page 142
Then the Base alloy (i.e. Al, Fe, Ni, Ti or Mg) should be entered, as well as each
alloying element with its concentration (in weight percent). Once the chemical
composition is entered, the "Apply->" button is pressed and the "Scheil" or
"Lever" option is selected to start the computation of the fraction of solid and of
the enthalpy.
Version 2006.0
Page 143
For most alloys, it is recommended to use the Scheil model, except for low alloy
steels where the diffusion in the solid is very fast.
Version 2006.0
Page 144
In the example above, when the Scheil button is selected, the following curves
appear :
One could see the fraction of solid curve, as well as the fractions of the different
phases as a function of temperature.
In the same time, automatically, the fraction of solid curve, the liquidus and
solidus temperatures, the enthalpy curve, the density, the viscosity and the thermal
conductivity are stored in the database, as shown hereafter (of course the value are
finally stored only when the "Store" button is pressed, before exiting the
database).
Thermal conductivity
Version 2006.0
Page 145
Density
Enthalpy
Version 2006.0
Page 146
Fraction of solid
Version 2006.0
Page 147
Liquidus-Solidus
Viscosity
Version 2006.0
Page 148
For some chemical composition, it may happen that the software which extracts
the data from the Thermodynamic database is not able to find the right set of
stable phases at low temperature.
Usually, this does not affect the determination of properties of interest (which are
more near the solidification range). Thus, it is possible to use the calculated values
Version 2006.0
Page 149
as such. If such situation occurs, one should check the calculated data in order to
be sure that it covers at least the temperature range of interest.
In very few cases, it is possible that the density calculation does not give relevant
results, as shown in the figure hereafter. In this case, these result should not be
used (or the wrong values should be erased).
As a general rule, if the results are not realistic, it is advised to suppress (i.e.
ignore) the elements which are present in very low concentrations. This is
especially true for traces of Sulfur (S) and Phosphorus (P) in steels, which are
sometimes "corrupting" the results.
Version 2006.0
Page 150
The user has the choice of either not calculate the Stress properties, to create a
new entry in the Stress database or to substitute/Add the data to an existing entry.
If a new entry is created the user has to specify its name (without spaces).
If the user would like to substitute or add the calculated data to an existing entry
of the stress database, one should select the desired entry with the Browse button.
Once these choices are made, the computation can be started (of both the thermal
and stress properties) with the "Compute" button.
Please note that the other Stress properties (i.e. Yield stress, hardening,
viscoplastic,...) can not be calculated at this stage. Thus these properties will
remain empty.
The following figures are showing examples of computed Stress data from an
A356 alloy.
Version 2006.0
Page 151
Young's modulus
Version 2006.0
Page 152
Poisson's ratio
Databases limitations
The Computherm databases can be used for the following elements and in the
following ranges. More information can be obtained on the
www.computherm.com web site.
The recommended composition ranges mentioned hereafter are not strict limits.
These are ranges which were extensively tested.
The Computherm manual (from Computherm LCC) is added in the Software
installation (in the dat/manuals/PDF directory). This manual describes for each
alloying system the phases which are calculated, the limitations as well as the
validations which have been made.
Al database
Version 2006.0
Page 153
Developed for Al-rich alloys such as commercial casting and wrought alloys.
Tested with more than 40 commercial Al alloys.
20 Components :
Major alloy elements: Al, Cu, Fe, Mg, Mn, Si, Zn
Minor alloy elements: Ag, B, C, Cr, Ge, Hf, Ni, Sc, Sn,
Sr, Ti, V, Zr
Major phases: Liquid, Fcc_A1(Al), Diamond_A4(Si),
Al5Cu2Mg8Si6, Al8FeMg3Si6, Eps, Sigma-(Al,Cu,Zn)2Mg, T(Al,Cu,Zn)49Mg32, Al20Cu2Mn3, Al23CuFe4, Al7Cu2Fe, SAl2CuMg, TAO(t), a-AlFeSi, b-AlFeSi, AL15_FeMn3Si2(aAlMnSi), AlMnSi-Beta, AlCu_Theta(q), Al13Fe4,
AlMg_Beta, Al11Mn4, Al12Mn, Al4Mn, AL6_FeMn, Al3Ni1,
AlSr4, Mg2Si,Al3Zr, Al3Sc_x
Recommended composition range (in wt%):
Al 80 ~ 100
Cu 0 ~ 5.5
Fe 0 ~ 1.0
Mg 0 ~ 7.6
Mn 0 ~ 1.2
Si 0 ~ 17.5
Zn 0 ~ 8.1
other
0 ~ 0.5
Fe database
Developed for Fe-rich alloys.
18 Components:
Al, C, Co, Cr, Cu, Fe, Mg, Mn, Mo, N, Nb, Ni, P, S, Si,
Ti, V, W.
59 Phases: Liquid, BCC_A2 (ferrite), HCP_A3,
FCC_A1(austenite), TCP phases, Carbides, and so on.
Recommended Composition Limits (wt%):
Fe > 50
Ni < 31
Cr < 27
Co, Mo < 10
V, W < 7
C, Cu, Mn, Nb, Si, Ti < 4
Al, Mg, N < 0.5
P, S < 0.05
Version 2006.0
Page 154
It was observed that alloy elements which are present is very small quantities
(such as P and S) may cause problems in the phase determination. As these
elements do not affect significantly the material properties (although it may have
important effects in other fields), it is recommended to remove these elements for
the computation.
Mg database
Developed for commercial Mg-rich alloys
16 components:
Mg, Ag, Al, Ca, Ce, Cu, Gd, Li, Mn, Nd, Sc, Si, Sr, Y,
Zn, Zr
Contains more than 200 phases.
Recommended composition range (in wt%):
Not yet available
Ni database
Developed for commercial Ni-rich alloys.
17 Components: Al, B, C, Co, Cr, Fe, Hf, Mo, N, Nb, Ni,
Re, Si, Ta, Ti, W, and Zr.
63 Phases: Liquid, Fcc_A1(g), L12_Fcc(g), TCP phases,
Carbides, and so on.
Recommended Composition Limits (wt%):
Ni > 50
Al, Co, Cr, Fe < 22
Mo, Re, W < 12
Hf, Nb, Ta, Ti < 5
B, C, N, Si, Zr < 0.5
Ti database
Developed for commercial Ti-rich alloys such as alpha, alpha+beta, and beta
alloys.
18 Components: Al, B, C, Cr, Cu, Fe, H, Mo, N, Nb, Ni,
O, Si, Sn, Ta, Ti, V and Zr
108 Phases: Liquid, BCC_A2(b), HCP_A3(a),
DO19_Ti3Al(a2), Laves, and so on.
Recommended Composition Limits (wt%):
Version 2006.0
Page 155
Ti > 75
Al, V < 11
Mo, Nb, Ta, Zr < 8
Cr, Sn < 5
Cu, Fe, Ni < 3
B, C, H, N, O, Si < 0.5
When the Scheil model is used, with the above composition, 10 phases are found
(in addition to the liquid phase) :
Version 2006.0
Page 156
Version 2006.0
Page 157
AlSi9Cu3Mg0.3
Version 2006.0
Page 158
AlSi9Cu3Mg0.3Fe1.3
(the effect of Fe is mainly visible at the liquidus)
AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2
(Mn, Ni and Zn are mainly affecting the second half of the curve)
Version 2006.0
Page 159
AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15
(Cr is rising the liquidus temperature from 612 to 640C)
Version 2006.0
Page 160
AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15Ti0.15
(When Ti is added, the Al3Ti phase appears, with a very high liquidus temperature
above 760C)
The following figure are showing the solid fraction curves for all the alloys
together.
Version 2006.0
Page 161
In order to see the effect near the liquidus, the same figure is shown, with a
different vertical scale. Only the first 5% are shown.
Version 2006.0
Page 162
One can see very well on this latter figure the effect of Ti. Ti is added in order to
create this Al3Ti phase which is stable at very high temperature, which is acting as
inocculant. The amount of this phase is very small (around 0.5%).
The above example is showing that one should be careful with the use of
Thermodynamic databases. In this case, for instance, it would be advisable to
ignore the Ti for the thermodynamic computation, in order to avoid this
"artificially" high liquidus temperature.
This explains also why there are differences observed between literature values
(measurements) and computed values for liquidus and solidus temperature. This is
due to the fact that such values are measured usually by Thermoanalysis and that
small amount of solid (like the few percents due to Fe, Cr and Ti near the liquidus
temperature) can not be detected. One should note that the usual literature value of
the liquidus for this alloy is 588C which corresponds to a computed value of
about 5% of fraction of solid. The measurement of the solidus temperature is even
more difficult and thus, it is normal to observe differences.
Version 2006.0
Page 163
Interface Database
The "Interface" database contains two types of entries :
Interface
Die Combo
Interface
For the definition of the Interface heat transfer coefficient, the following panel is
opened.
The structure of the above panel is described hereafter (see the numbers in the
panel below). Firstly, the type of the database entry is shown on the top of the
panel (1). Then, the entry can be labeled with a Keyword (2) which will help to
recognize it's content for later use. If a constant heat transfer coefficient is to be
defined, it's value should be entered in the corresponding field (3). Then the units
can be defined (4). The available choice of units is proposed when one clicks on
the unit box. If it is desired to define a "Temperature-" and/or a "Time-"dependant
heat transfer coefficient, the corresponding table can be defined in the
"Temperature" or "Time" buttons (5) (for more details about the definition of
Version 2006.0
Page 164
If the interface heat transfer coefficient is non-constant, the value defined in (3)
and the table(s) defined in (5) are all multiplied to get the final value (at a given
temperature and a given time). It is allowed to define only a table without a
constant (in this case, it is like if the constant is equal to 1). If a User function is
specified, neither a constant value, nor a table can be defined.
Die Combo
The "Die Combo" definition is used only in the case of die casting, to handle
automatically the sequence of closed and opened dies (see the "Cycling" section
for more details).
The "Die Combo" database entry corresponds to the following panel.
Version 2006.0
Page 165
It allows to define :
a) the interface heat transfer coefficient (when the die is closed) in the "Constant"
field and/or as a Temperature-dependant table ("Temperature" button).
b) when the die opens, the interface heat transfer coefficient between the die and
the air is defined by a heat transfer coefficient (Air Coeff) and by an ambient
temperature (Air Temp).
c) if there is a Spray cooling stage, it can be defined by the corresponding heat
transfer coefficient (Spray Coeff) and the Spray temperature (Spray Temp).
The check-box "Attached until Ejection" will tell whether the switch between the
interface heat transfer and the air cooling will be done when the dies opens or
upon ejection of the part. These times are given in the "Run parameters", in the
"Cycles" tab, as shown below.
Version 2006.0
Page 166
Temperature
"Temperature" allows to specify (or impose) a temperature at a given point. It is
especially used to specify inlet temperatures in filling calculations.
Version 2006.0
Page 167
As most of the panels of the Boundary Conditions database have the same
structure, the principles will be described for this one only.
Firstly, the type of the database entry is shown on the top of the panel (1). Then,
the entry can be labeled with a Keyword (2) which will help to recognize it's
content for later use. If a constant value of temperature is to be defined, it's value
should be entered in the corresponding field (3). Then the units can be defined (4).
The available choice of units is proposed when one clicks on the unit box. If it is
desired to define a "Time-"dependant Temperature, the corresponding table can be
defined in the"Time" button (5). Finally, the data are saved with the "Store" button
(6).
The following figure shows the choice of available units (when one clicks on the
unit box).
Version 2006.0
Page 168
Heat
The "Heat" boundary condition allows to define the heat transfer between the
outside faces of a given domain and the outside world (air, water, ...). The
following panel is available for Heat database entry definition.
Version 2006.0
Page 169
The following equation shows the three possible contributions to the "Heat"
boundary condition (all the contributions are added, although they may not be all
active in the same time or in the same cases).
The first term corresponds to a specified flux. It can be used if a given heat flux
was measured for instance. The second term corresponds to Convective cooling.
This is the most common definition of the cooling of an external face. It is defined
by a heat transfer coefficient with the ambiance and by an external temperature (of
the ambiance). The third term is used at high temperature, when radiation
becomes important. In this case, the transfer is proportional to the StefanBolzmann constant and the emissivity and the difference of the fourth power of
the temperatures (surface temperature and ambiance temperature).
The above equation contains the four values which are shown in the panel below
(1), such as the "Film Coeff", the "Emissivity", the "Ambient Temp" and the
"Heat Flux". These values can be also "Time" and/or "Temperature" dependant
(2). In this case, the value of the constant (if any) is multiplied by the
corresponding table(s). It is also possible to define the different parameters by the
corresponding "User Functions" (3).
Version 2006.0
Page 170
Velocity
The definition of the velocity is made in the following panel. In this case, the three
components of the velocity vector should be defined. If a Time- or a Pressuredependant velocity is set, each component of the vector will be multiplied by the
corresponding table(s).
If a gate should be active until a given filling fraction is reached, the "Fill Limit"
slider can be used for that purpose. This Fill limit has not the same effect as the
Run parameter LVSURF (see the "Run parameters" section for more details).
LVSURF allows to stop the filling, but more important it allows to switch off the
Fluid flow solver when the filling is finished. On the other hand, the Fill limit
defined in this boundary condition should be used if multiple gates are present. It
allows to switch off automatically one gate when a given fill fraction is reached an
the other gate will continue until LVSURF is reached.
Although it is possible to define a Pressure-dependant velocity, one should be very
careful in the use of such capability, as the pressure solution is very sensitive.
Version 2006.0
Page 171
When an inlet velocity is set to fill a casting, there are situations where the user
would like to know which value of velocity should be set in order to achieve a
given filling time. To do so, a "Velocity calculator" is available.
Version 2006.0
Page 172
To use the velocity calculator, the following operations should be done (see the
numbers in the upper figure). Firstly, one should "Add" a new velocity boundary
condition (1) and assign the corresponding nodes on the geometry. Then, open the
"Velocity" boundary condition entry (2) and click on "Velocity calculator" (3).
The "Calculator" window will appear (4) and one should select all the domains
which correspond to the casting (5). In the above example, the casting is made of
one domain only. Then, the expected filling time should be entered in (6) and the
velocity is automatically calculated by pressing the button in (7). The velocity
magnitude appears then in red (8). The value should be reported in (9), in the
appropriate component of the vector. Finally, the values should be stored (10).
One should be aware that the velocity calculator will give accurate results if all the
nodes on the inlet surface are selected. If only a few nodes are selected, the inlet
surface will "overflow" to the neighboring elements and the filling time will be
shorter than the one specified. In order to prevent that, it is advised to design a
small additional volume, corresponding to the jet of liquid entering the casting.
Pressure
The pressure is defined in the following panel, which has the same principles as
the "Temperature" panel described above.
Version 2006.0
Page 173
Inlet
Inlet boundary conditions are used to specify in the same time an "equivalent"
inlet velocity together with an inlet temperature. The inlet Flow rate can be time
dependant.
Turbulence
In the case of a Turbulent model, one can specify the "Turbulent" state at the entry
of the liquid in the calculation domain.
Version 2006.0
Page 174
Vent
If the "Gas" model is activated, the air can escape through vents. Vents are applied
on nodes of the casting domains. Each vent is represented by an "equivalent tube"
by which the air can escape. Thus, a vent is described by the diameter and the
length of the "equivalent tube", as well as a roughness and the exit pressure.
Inject
Version 2006.0
Page 175
In the case of low pressure die casting (lpdc), it is possible to model the liquid
bath of metal and to apply a pressure on it (in order to drive the metal in the mold)
via injected air. The inject BC allows to specify the amount of injected air. Please
note that this method may be delicate to converge.
Displacement
In the case of stress calculations, the displacement of the model should be
constraint. The Displacement BC allows to define these constraints. Please note
that if a field is not filled, the displacement is not constraint in this direction (e.g.
it is possible to specify only a "0" in the x-field which will mean that the node can
not move in the X-direction, but it is free to move in the Y and Z-directions.
Version 2006.0
Page 176
Point Load
For stress calculations, it is possible to apply a load (or a force) at a given
location. The three components of the load should be defined.
Version 2006.0
Page 177
Surface Load
In the case of stress calculations, one can define a load applied on a given surface.
In this case, it corresponds to a pressure. The direction of the load is specified by
the three components of the load vector.
Version 2006.0
Page 178
Volumetric Heat
In some thermal problem, it may be necessary to define the generation of heat
inside a given material (e.g. to model the effect of an electrical resistance). In this
case, the thermal power which is generated per unit volume should be defined.
This heat source will be applied on the entire specified material.
Version 2006.0
Page 179
Momentum Source
In the case of flow problems, one would like to model a source of momentum in a
given material. This could be the case to take into account electromagnetic forces
or the effect of a propeller. In this case, the momentum force vector should be
defined. Please note that this force will be applied in the whole specified material.
Mass Source
In the case of filling calculations, one would like to define the amount of metal
which enters into the material with a Mass source (instead of a velocity or an inlet
BC). In this case, one can specify that a given amount of metal (Flow rate) is
"appearing" at the given location (X, Y and Z coordinates) at the specified
temperature. The metal will "appear" with a zero velocity.
Version 2006.0
Page 180
Emissivity
For radiation calculations, the emissivity of the mold, the casting top surface, the
furnace or the enclosure should be specified.
The emissivity should be defined between zero and one.
Version 2006.0
Page 181
Process Database
It is possible to define Translations, rotations and revolutions of domains. This is
used especially for radiation calculations with view factors (where the furnace is
moving with respect to the part, or vice-versa.
The following panels allow to define these data.
Translation
The translation of a given material domain or of an enclosure is defined with the
following panel. The translation can be defined in three different ways :
x(t) - Translation vector or position (i.e. position vs time)
v(t) - Translation velocity as a function of time (i.e. velocity vs time)
v(x) - Translation velocity as a function of position (i.e. velocity vs position)
x(t) - Position vs Time
Version 2006.0
Page 182
The velocity of the domain vs time can be defined with the panel hereafter. The
direction of the velocity vector is defined by the U, V and W components. This
velocity vector can be either constant or it can be multiplied by a Time function.
When v(t) is used, the software is transforming it in x(t) automatically in order to
have a translation position which is independant from the timestep.
Version 2006.0
Page 183
The velocity of the domain vs Position can be defined with the panel hereafter.
The direction of the velocity vector is defined by the U, V and W components (it
should be a "unit vector"). This velocity vector can be either constant or it can be
multiplied by a "Distance" function. The "Distance" is the relative distance with
respect to the initial location of the domain which moves.
When v(x) is used, it is mandatory to have a non-zero velocity for the initial
position (i.e. the zero position). Please note that it is also mandatory to have
always non-zero velocities defined (i.e. the velocity should never be zero).
Version 2006.0
Page 184
Rotation
The following panel allows to define the rotation of a material domain or an
enclosure. The axis of rotation should be defined by two points.
Version 2006.0
Page 185
The rotation is defined in the "Time" button, which opens the following panel, for
the definition of the angle as a function of time.
Revolution
The "Rotation" definition is used when less than one turn is done during the entire
process. When one has several turns, the "Revolution" definition is used. In this
case, the axis of revolution should be defined by two points.
Version 2006.0
Page 186
Stress Database
To access the Stress database, one should to in the Material/Stress menu
Version 2006.0
Page 187
Then, the list of available materials (stress properties) is shown in a window in the
lower right corner (as for the standard material database. The database can be
managed in the same way with the "Read", "Add", "Copy" and "Del" capabilities.
When a new stress database entry is created, the following window appears.
Version 2006.0
Page 188
The first thing to do is to select the desired model among the five following
options :
Vacant
Rigid
Linear-Elastic
Elasto-Plastic
Elasto-ViscoPlastic
"Vacant" is used to specify that the domain will not participate to the stress
calculation. Thus, no stress and no strain will be calculated and the domain will
not participate to the contact algorithm (i.e. the domain will not create any
resistance to the neighboring domains). No properties should be defined for a
Vacant domain.
No stress calculation will be done in a "Rigid" domain, however, the domain will
participate to the contact algorithm (i.e. the neighboring domains will not be
allowed to penetrate in the Rigid ones). No properties should be defined for a
Rigid domain.
For the three other models, the following data are required :
Version 2006.0
Page 189
The three models, as well as the meaning of the corresponding properties, are
described in the "Stress Models and Properties" section.
Warning
For stress calculations, the thermal material properties must include the phase
change (i.e. fraction of solid curve). Otherwise, no stresses will be calculated in
Version 2006.0
Page 190
this material (as the fraction of solid will not be defined, it will be considered as
"zero" and no stress will be calculated).
Before defining the stress properties, one should be aware of the following. For
stress calculations, the thermal material properties must include the phase change
(i.e. fraction of solid curve). Otherwise, no stresses will be calculated in this
material (as the fraction of solid will not be defined, it will be considered as
"zero" and no stress will be calculated).
The three stress models available in ProCAST can be summarized in the following
figure.
Version 2006.0
Page 191
Beside the Young's modulus, one should define the Poisson's Ratio and the
Thermal Expansion coefficients. The value of the Poisson's ratio is usually around
0.3 for metals.
Version 2006.0
Page 192
The figure above shows how to transform a measured "Strain curve" (shown in
blue) in a Secant thermal expansion coefficient. To do so, one needs to define a
Reference Temperature (Tref), shown in the above figure as To, which corresponds
to a zero strain. The "Secant" thermal expansion coefficient at temperature T2
corresponds to the slope S2 of the line between a zero strain (To) and the strain at
the temperature T2.
In the case of a constant Thermal Expansion coefficient, the strain curve is a
straight line and the coefficient corresponds to the slope of this line. One does
NOT need to define a reference temperature in the case of a Constant coefficient.
Version 2006.0
Page 193
The Reference Temperature is defined in the lower part of the "Secant" panel (see
above).
Elasto-Plastic model
For the Elasto-Plastic model, the properties described above for the Elastic model
(i.e. the Young's modulus, the Poisson's ratio and the Thermal Expansion
coefficients) should also be defined.
In addition, one should define the Yield stress and the Hardening coefficient.
Version 2006.0
Page 194
The Yield stress corresponds to the stress at which plastic deformation starts. It
can be temperature dependant.
The hardening coefficient corresponds to the slope of the stress-strain curve in the
plastic range. Four different models of hardening are available in ProCAST :
Version 2006.0
Page 195
Version 2006.0
Page 196
The Hardening can also be defined freely in an ASCII file where one can enter
any "digitized" curve. More information about this possibility is in the "Digitized
Hardening" section.
In order to take into account for the "Kinematic" non-isotropic hardening behavior
(Bauschinger effect), the Amstrong-Frederick model is available :
x is called the back stress. It corresponds to the "movement" of the center of the
Yield surface.
Isotropic and Kinematic models can be used either individually or together.
Elasto-ViscoPlastic model
For the Elasto-ViscoPlastic model, the properties described above for both the
Elastic model (i.e. the Young's modulus, the Poisson's ratio and the Thermal
Version 2006.0
Page 197
Expansion coefficients) and the Elasto-Plastic model (i.e. the Young's modulus
and the Hardening coefficient) should also be defined.
This model allows to describe the secondary (steady-state) creep with threshold.
Version 2006.0
Page 198
Norton model
Version 2006.0
Page 199
As there is no threshold value, the data for the Yield Stress and the Hardening will
be ignored.
This model allows to describe both the primary (strain hardening) creep and
secondary (steady-state) creep regimes with a possible threshold.
Version 2006.0
Page 200
For the three models, one have (for the plastic model) :
or
Summary
Version 2006.0
Page 201
At high temperature, one can have a viscoplastic behavior. In this case, setting a
hardening parameter (H in case of linear hardening) to zero will "cancel" the
plastic model, whereas setting the "Viscous parameter", will activate the viscoplastic model. On the other hand, at low temperature, if the hardening parameter
will be set to a "high" value, the plastic model will be active, whereas one can
"deactivate" the visco-plastic model by setting a "Viscous parameter" to 0 (it is
allowed to set it to zero, because the "0" is considered as a "deactivation flag" and
not as a "zero" value). Please note that it is advised to set the transition of the
"Viscous parameter" between 0 (at low temperature) to a given value (at
intermediate temperature), within a very small range of 0.001C.
One should however notice that in most cases, the use of an elastic-plastic model
gives very realistic results, without the burden of finding appropriate data for the
visco-plastic behavior.
The principles of the determination of plastic and visco-plastic properties from
experimental measurements are described in the "Viscoplastic properties
determination" section.
Version 2006.0
Page 202
recommendations will allow the best compromise between the physics and the
convergence of the stress solver.
Please note that the definition of these parameters in the mushy zone will
influence the hot tearing indicator prediction.
The value of the Young's modulus in the mushy zone (below a fraction of solid of
20%) should be set as a constant value corresponding to the value of fs = 20%.
Usually, this corresponds to values between 50 and 500 MPa.
Please note that the Youngs's modulus can be automatically calculated in
PreCAST as function of the chemical composition, using the thermodynamic
databases.
Version 2006.0
Page 203
The value of the Yield stress in the mushy zone (below a fraction of solid of 50%)
should be set as a constant value corresponding to 5-10 MPa. One should not set
values below 5 MPa.
If no data are available, the Hardening (in case of linear hardening) can be set to
about 1/20 of the Young's modulus. In all cases, the Hardening should be set to 0
MPa for fraction of solid smaller than 50%.
Remark :
If a tensile test is made on a sample which will be fully mushy at a given time, a
non-zero value of the Hardening should be set for fraction of solid below 50%.
However, this case is never occurring in usual casting processes.
Concerning the Poisson's ratio, if a value of 0.5 is set, it is automatically changed
in the software to 0.48.
Together with these data, the corresponding Run parameters are recommended
(see the "Stress Run Parameters" section for more details) :
CRITFS = 0.5
CONVS = 0.01
PENALTY = 0.01
AVEPEN = 0.1 mm (for large casting, this value can be increased)
SCALC = 5
Version 2006.0
Page 204
Digitized Hardening
Instead of defining the Hardening either as Linear or with a Powerlaw, it is
possible to enter in an ASCII file digitized hardening curves (i.e. the plastic part of
tensile test curves at different temperatures).
To activate this mode, one should select the "Table" tab in the "Hardening" tab :
Then, the number of tables specified in the ASCII file should be set.
The corresponding ASCII file should be called : *ss1.dat (e.g. : prefixss1.dat) and
it should have the following structure :
STRESS_UNIT 4
CURVE 1
POINTS 2
TEMPERATURE 1 293.
0.
8.1159e+01
0.02 1.7684e+02
0.03 2.2671e+02
0.04 2.8462e+02
CURVE 2
POINTS 2
TEMPERATURE 1 823.
0.
9.838
0.01 3.89105e+01
0.05 4.73871e+01
Version 2006.0
Page 205
The stress units are defined first (the unit code corresponds to the one of the
ProCAST d.dat file - e.g. 4 corresponds to MPa).
Then, the number of the curve should be specified, followed by the number of
points in the curve.
The temperature at which the curve is defined (with the unit code before it - e.g. 1
corresponds to degrees Kelvin, 2 to Centigrade and 3 to Fahrenheit), followed by
the curve itself (strain - stress).
Please note that only the plastic part of the curve (i.e. the hardening) is defined by
the tables. This means that each curve should start by a zero strain (this is
mandatory) and that the corresponding stress value is the Yield stress at this
temperature. Thus, the Yield stress defined in the "Yield Stress" tab in PreCAST
will not be used in the case of Tables (the values will be ignored).
Each curve can have a different number of points. Above the last point, the Stress
is extrapolated as a constant (i.e. perfect plasticity).
Version 2006.0
Page 206
Version 2006.0
Page 207
Low Temperature
At low temperature (usually T1 < Tm/3), the tensile test curve is independent from
the strain rate. This means that different tests made at different RAM speeds are
giving the same curve (see the green curve in the figure below).
From this curve, one can get the Yield stress (o) and the Hardening coefficient
(H).
The value of the viscous parameter () should be set to zero in order to "disable"
the viscoplasticity. The Power (n) can be set to any value as it will not be used.
Version 2006.0
Page 208
High Temperature
At high temperature (usually T3 > Tm/2), the tensile test curve is dependent from
the strain rate and the stress level is constant (no hardening) after the transition
stage (see the red curves in the figure below). As there is no hardening, the value
of H should be set to zero. Therefore the threshold (y) is equal to the Yield stress
o.
Version 2006.0
Page 209
In order to obtain the viscoplastic parameters ( and n), one should first determine
the value of the Yield stress (o), which can be equal to zero. To do so, one
should perform a tensile test at the lowest possible strain rate (possibly zero !) and
see the value of the stress on the plateau. If this value is very low, one can
consider that the Yield stress (o) is zero. In order to determine this value, one can
also make loading-unloading tests, with increasing loads until plastic deformation
is obtained.
Then, for each measured tensile test curve, the value of (- Y)/* (which is the
stress level of the plateau minus the Flow stress, Y, which is equal to o in this
case, as H = 0) and d/dt should be plotted in a log-log graph.
The value of and n can be deduced from this graph (see below).
Version 2006.0
Page 210
Intermediate Temperature
At intermediate temperature (usually Tm/3 < T3 < Tm/2), the tensile test curve is
dependent from the strain rate and the stress level is not constant (hardening) after
the transition stage (see the blue curves in the figure below).
The hardening (H) corresponds to the slope of the curves (after the transition
stage). In this model, the hardening is the same for all strain rates.
In order to obtain the viscoplastic parameters ( and n), one should first determine
the value of the Yield stress (o), which can be equal to zero. To do so, one
should perform a tensile test at the lowest possible strain rate (possibly zero !) and
see the value of the stress zero plastic strain. If this value is very low, one can
consider that the Yield stress (o) is zero. In order to determine this value, one can
also make loading-unloading tests, with increasing loads until plastic deformation
is obtained.
Then, for each measured tensile test curve, the value of (- Y)/* (which is the
extrapolated stress level at zero strain minus the Flow stress, Y. In this case, Y
is equal to o, as pl = 0 - because the extrapolation is performed a zero strain)
and d/dt should be plotted in a log-log graph. The value of and n can be
deduced from this graph (see previous figure).
Version 2006.0
Page 211
Version 2006.0
Page 212
Low Temperature
At low temperature (usually T1 < Tm/3), the following curves are measured at
different loads.
Version 2006.0
Page 213
From these curves, one can get the Yield stress (o) and the Hardening coefficient
(H). To do so, the values (shown in the graph below) of and (at the different
loads) should be plotted in a - graph. A linear or Power Law Hardening can be
fitted on those points.
Version 2006.0
Page 214
Beside the above, the value of the viscous parameter () should be set to zero in
order to "disable" the viscoplasticity. The Power (n) can be set to any value as it
will not be used.
High Temperature
At high temperature (usually T3 > Tm/2), the slope of each curve represents the
strain rate, d/dt (see the red curves in the figure below).
Version 2006.0
Page 215
In order to obtain the viscoplastic parameters ( and n), one should first determine
the value of the Yield stress (o), which can be equal to zero. To do so, one
should perform a creep test at decreasing loads. The Yield stress is reached when
there is no more plastic strain. If this value is very low, one can consider that the
Yield stress (o) is zero.
Then, for each measured creep curve (shown above), the value of strain rate
(d/dt) should be plotted as a function of (- Y)/* (which is the stress level of
the plateau minus the Flow stress, Y, which is equal to o in this case, as H = 0)
in a log-log graph (see below).
The value of and n can be deduced from this graph (see below).
Version 2006.0
Page 216
Beside the above values, as there is no hardening, the value of H should be set to
zero. Therefore the threshold (y) is equal to the Yield stress o.
Intermediate Temperature
At intermediate temperature (usually Tm/3 < T3 < Tm/2), the creep test curves are
dependent from both the load (stress level) and time (see the blue curves in the
figure below).
Version 2006.0
Page 217
From these curves, one can get the Yield stress (o) and the Hardening coefficient
(H) as described above in the "Low Temperature" section. To do so, the values
(shown in the graph below) of and (at the different loads) should be plotted in
a - graph. A linear or Power Law Hardening can be fitted on those points.
In order to obtain the viscoplastic parameters ( and n), one should first determine
the value of the Yield stress (o), which can be equal to zero. To do so, one
should perform a creep test at decreasing loads. The Yield stress is reached when
there is no more plastic strain. If this value is very low, one can consider that the
Yield stress (o) is zero.
Then, for each measured creep curve (see blue curves in the zoom below), the
values of the INITIAL strain rate (d/dt) should be plotted as a function of (Y)/* (which is the load, i, minus the Flow stress, Y. In this case, Y is equal
to o, as pl = 0 - because the extrapolation is performed at time=0, where the
strain is zero) in a log-log graph (as described in the "High Temperature" section).
Version 2006.0
Page 218
Version 2006.0
Page 219
RUN PARAMETERS
This section contains the comprehensive description of all Run parameters. The
ones shown in blue are either new or have been modified or further described,
with respect to previous versions.
The different categories of Run Parameters are organized in Tabs. The first Tab
layer (1) corresponds to the different modules of ProCAST. For each modules, the
Run Parameters are divided in a "Standard" tab (2), with the most often used
parameters and one or two "Advanced" tabs (2) with less often used Run
Parameters. The values are then defined in (3).
The values of the Run Parameters are stored in a file named "prefixp.dat" (often
called the "p.dat" file). The Run Parameters can thus be changed either in the
PreCAST interface, or directly "by hand" in the "p.dat" file (see example below).
Version 2006.0
Page 220
The default values are not anymore specified in the manual, as they are set in the
"default_p.dat" file (see the "Pre-defined Run Parameters" section for more
details).
Version 2006.0
Page 221
The Run Parameters of the "General/Advanced" tab are usually not changed from
the default values, except in very specific cases. It corresponds mainly to the
parameters of the different general solver algorithms.
Version 2006.0
Page 222
General
AVEPROP
specifies the method to be used in calculating the properties for each element.
ProCAST will calculate the properties at each Gauss point or you may specify that
the properties be calculated only at the element center and that this value will be
used as an average for the whole element. This averaging reduces, somewhat, the
finite element integration time required. This averaging does not apply to the
specific heat or enthalpy calculations.
Choose from:
0 to calculate at each point, or
1 to use the average
CGSQ
specifies the Conjugate Gradient Squared solver flag. The values specified in this
parameter may be added together. This allows you to "build" a customized solver
approach for your simulation.
Version 2006.0
Page 223
Choose from:
0 = Use the default iterative solver ( TDMA ),
1 = Use the CGSQ solver on the U momentum equation,
2 = Use the CGSQ solver on the V momentum equation,
4 = Use the CGSQ solver on the W momentum equation,
16 = Use the CGSQ solver on the energy equation,
64 = Use the CGSQ solver on the turbulence intensity equation,
128 = Use the CGSQ solver on the turbulence dissipation equation, or
512 = Use the CGSQ solver on the density equation for compressible flow
CONVTOL
specifies the convergence tolerance which will be used in conjunction with the
default non-symmetric iterative solver.
Enter a floating (real) value.
DIAG
specifies the diagonal preconditioning flag for the symmetric solver.
Choose from:
0 = use partial Cholesky preconditioning for everything,
8 = use diagonal preconditioning for pressure,
16 = use diagonal preconditioning for energy, and
16384 = use diagonal preconditioning for radiosity
DT
specifies the initial time step size. Setting DT to zero when INILEV > 0 will cause
ProCAST to use the DT at step INILEV.
Enter a floating (real) value. The default is 1.0000e-03.
Select the units of time from: {sec | min}
DTMAX
specifies the maximum time step size.
Enter a floating (real) value.
Select the units of time from: {sec | min}
DTMAXFILL
specifies the maximum time step size which will be used during the filling stage
only. Once the filling is finished, the DTMAX value will be used. If
DTMAXFILL is not defined (i.e. it is set to zero), the value of DTMAX will be
used for the whole calculation.
Enter a floating (real) value.
Select the units of time from: {sec | min}
INILEV
specifies the initial time level. When an analysis is first started, INILEV should be
equal to zero. When you are resuming an analysis, INILEV should be set to the
time step from which you would like to continue. Note: You must have results for
that time step.
Version 2006.0
Page 224
LUFAC
specifies the preconditioning parameter for the CGSQ solver. This parameter may
speed-up calculations of large models
Choose from:
0 to use diagonal preconditioning, or
1 to use partial LU factorization preconditioning
NEWTONR
turns on the NEWTON Raphson technique for the energy equation.
Choose from:
0 to turn off the Newton Raphson technique,
1 to turn the Newton Raphson technique on, or
2 to turn on the Newton Raphson technique and use bsplines
The default is 0.
Option 2 results in using b-splines instead of linear line segments in the
representation of the thermal properties. It is suggested that all thermal input data
be smoothed before attempting to use b-splines.
Enter an integer value.
NPRFR
specifies the printout frequency. This controls the time step interval at which
results are output to the prefixp.out file.
Enter an integer value.
NRSTAR
specifies the number of allowable restarts before the entire run is abandoned. A
restart occurs when the maximum number of corrections is reached. If too many
restarts are taking place, it could indicate problems with the model setup.
Enter an integer value.
NSTEP
specifies the number of time steps to take in the current run and is used in
conjunction with TFINAL. ProCAST will terminate the run when it reaches this
limit or the TFINAL value, whichever occurs first.
Enter an integer value.
PRNLEV
specifies the level of nodal results to be printed out. The values specified in this
parameter may be added together. This allows you to collect combinations of
nodal information in a single run.
Choose from:
0 = no printout,
1 = nodal velocities,
8 = nodal pressures,
16 = nodal temperatures,
Version 2006.0
Page 225
SDEBUG
specifies the level of solution debugging messages to be captured. These messages
are written to the p.out file.
Choose from:
0 to capture no solution debugging messages, or 1 to obtain information
concerning, solver performance, time step control, and the free surface model
TENDFILL
specifies the delay after the end of the filling at which to terminate a ProCAST
analysis. If this parameter is zero, this parameter will not be active. If one wants to
stop the calculation right after the end of the filling, one should set a small value,
different from 0.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TFINAL
specifies the simulated time at which to terminate a ProCAST analysis. If this
parameter is zero, the run will be stopped by the TSTOP or NSTEP parameter. If
NSTEP and TSTOP and TFINAL parameters are set, the simulation will be
terminated based upon which parameter is reached first.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TSTOP
specifies the temperature at which to terminate a ProCAST analysis (i.e. when all
the temperatures at all nodes are below the TSTOP temperature). If this parameter
is zero, the run will be stopped by the TFINAL or NSTEP parameter. If NSTEP
and TSTOP and TFINAL parameters are set, the simulation will be terminated
based upon which parameter is reached first.
Enter a floating (real) value.
Select the units of Temperature from: {C | K | F}
TMODR
specifies the time step modification factor for restarts. If MAXCOR correction
steps are taken without convergence, the time step is multiplied by TMODR.
Therefore, this number should be less than 1.
Enter a floating (real) value.
TMODS
Version 2006.0
Page 226
specifies the time step modification factor for normal stepping. If the number of
correction steps is less than or equal to NCORL, the subsequent time step is
multiplied by TMODS.
If the number of correction steps is greater than or equal to NCORU, the
subsequent time step is divided by TMODS.
Enter a floating (real) value.
Units
PUNITS
specifies the pressure units to be used in the outputs.
Choose from: {N/m**2 | Pa | KPa | MPa | bar | dyne/cm**2 | atm | psia | Ksi |
lb/ft**2}
QUNITS
specifies the heat flux units to be used in the outputs.
Choose from: { W/m**2 | cal/cm**2/sec | cal/mm**2/sec | Btu/ft**2/sec |
Btu/in**2/sec | cal/cm**2/min | cal/mm**2/min | Btu/ft**2/min | Btu/in**2/min}
TCUNITS
specifies the thermocouple units to be used in the outputs and is only used for
inverse modeling.
Choose from: {C | F | R | K}
TUNITS
specifies the temperature units to be used in the outputs.
Choose from: {C | F | R | K}
VUNITS
specifies the velocity units to be used in the outputs.
Choose from: {m/sec | cm/sec | mm/sec | ft/sec | in /sec | m/min | cm/min | mm/min
| ft/min | in/min}
The choice of the units does not have any effect on the calculated results.
Version 2006.0
Page 227
Version 2006.0
Page 228
Thermal
CINIT
Reserved for future use.
CLUMP
specifies the capacitance matrix lumping factor.
Enter:
0 to use consistent matrix, or 1 to use diagonal matrix
CONVT
specifies the convergence criterion for temperature. A value of around one degree
is generally appropriate. Values larger than the mushy (liquidus--solidus) zone
range are not recommended.
Enter a floating (real) value.
Select the units of temperature from: {C | F | R | K}
Version 2006.0
Page 229
CRELAX
specifies the heat capacity relaxation parameter.
Enter a floating (real) value.
FEEDLEN
"Feeding length" (distance), at which macroporosity can occur, beyond the
MACROFS isosurface.
Enter a floating (real) value.
GATEFEED
allows to specify whether liquid can be fed at the ingate or not. In the case of
injection (i.e. hpdc or lpdc), the shrinkage at the gate is compensated by the liquid
pushed by the piston (for hpdc). On the other hand, in gravity casting, there is no
feeding at the top of the risers and piping will occur. GATEFEED=1 tells the
software to activate the feeding at the ingate, thus leading to no piping at this
location. GATEFEED=1 will automatically activate the "Active feeding" where
inlet velocities or pressures are set (see the "Active Feeding" section for more
details).
Enter an integer value.
GATEFS
GATEFS allows to control the critical fraction of solid above which there is no
more feeding during the third stage pressure in HPDC (see the "Active Feeding"
section for more details). This Run Parameter has to be added or modified
manually in the p.dat file (it does not appear in PreCAST). A higher value of
GATEFS will involve less porosity.
Enter a value between zero and one. The default value is 0.95.
GATENODE
GATENODE is an alternative to GATEFEED. When a gate feeding should be
applied, but there is no external surface of the model where an inlet surface or a
pressure can be applied. This is the case for instance when a shot piston is
modeled. In such case, one can define a location in the volume where this gate
feeding is applied (typically in the middle of the biscuit). In order to define this
location, the corresponding node number should be specified. In such case, no
pressure or inlet velocity BC needs to be specified. To find the desired node
number, it is advised to use the "pick" selection in ViewCAST (Run DataCAST
first) (see the "Active Feeding" section for more details).
Enter an integer value.
LINSRC
specifies the source term linearization parameter for micromodels. This parameter
may be used in conjunction with micromodels that control the evolution of the
solid fraction and thus the release of latent heat. The default value of zero
indicates that the heat generation will only appear in the right hand side source
Version 2006.0
Page 230
term. A value of one will give some contribution to the diagonal terms of the left
hand side matrix.
This improves numerical stability, but does require that the LHS be factored,
which would normally happen anyway.
Enter:
0 = no linearization, or
1 = for linearization of the source term
MACROFS
Parameter for the macroporosity calculation. It sets the limiting fraction of solid
between the macroporosity and the microporosity formation. The value should be
set between 0 and 1.
Enter a floating (real) value.
MOBILE
Enter a floating (real) value.
MOLDRIG
The rigidity of the mold has an influence on the amount of porosity in the case of
expanding alloys. If the mold is totally rigid, the casting can not expand and thus
the alloy expansion will be "available" for the "refill" of the existing porosity. On
the other hand, if the mold is very soft (or weak), the casting will expand and thus
there will be no "refill" of the porosity (of course, the reality is more complex as
the solid shell is thick enough it will act as a "rigid" mold, even if the sand mold is
weak.
In order to take the mold rigidity into account, the Run Parameter MOLDRIG is
introduced (see the "SGI Porosity model" section for more details). MOLDRIG
should be defined by a value between 0 and 1. All the net expansion is multiplied
by MOLDRIG. Thus, with MOLDRIG=1, corresponding to a rigid mole, the
expansion will be fully accounted. On the other hand, no expansion will be taken
into account if MOLDRIG=0. The expansion will be compensated by the mold
movement because the mold is too weak to hold the expansion in this case. For
real situations, the value of MOLDRIG should be set somewhere between 0 and 1
depending upon the casting processes. MOLDRIG should be added "manually" in
the p.dat file by the user. The default value is 1.
PIPEFS
Parameter for the piping calculation (during a porosity calculation). It sets the
limiting fraction of solid for piping formation. The value should be set between 0
and 1.
In order to disable the porosity calculation, one should set POROS = 0. However,
this may still lead to some piping calculation. In order to disable also the piping,
one should set in addition PIPEFS = 0.
For stress calculations, PIPEFS must be set to 0. This is needed in order to prevent
the unexepected influence of the pipe on the stress calculation. One should note
that if in the physical sitution, piping must occur, this will be taken into account in
the calculation as "macropores" when PIPEFS = 0.
Version 2006.0
Page 231
POROS
specifies the porosity calculations to be performed (see the porosity models
description).
Choose from:
0 for no porosity calculation,
1 - most advanced porosity model
4 - evolution of the POROS=8 model, which allows to handle multiple free
surfaces
8 - old porosity model from version 3.2.0
In order to disable the porosity calculation, one should set POROS = 0. However,
this may still lead to some piping calculation. In order to disable also the piping,
one should set in addition PIPEFS = 0.
QFREQ
specifies the time step interval for writing heat flux data to the unformatted results
file. This parameter can be used to reduce the size of the prefixq.unf file. Heat flux
results may not be of interest to everyone, so it may be desirable to minimize the
size of this file.
Enter an integer value.
TFREQ
specifies the time step interval for writing temperature data to the unformatted
results file. This parameter can be used to reduce the size of the prefixt.unf file,
which can become quite large for problems with many nodes and time steps. Note
that it is only possible to restart a run from one of the time steps that was written
out.
Enter an integer value.
THERMAL
specifies the thermal analysis to be performed.
Choose from:
0 for no thermal analysis. Solve flow equations alone,
1 to perform thermal analysis, using temperature as the primary variable, or
2 to perform thermal analysis, using enthalpy as the primary variable
TRELAX
specifies the temperature relaxation parameter. This is used for computing the
initial guess for the temperature field in the predictor step. TRELAX should be
greater than or equal to zero and less than or equal to one.
Enter a floating (real) value.
USERHO
specifies the how the density in the mushy zone is calculated.
Choose from:
Version 2006.0
Page 232
NCYCLE
specifies the number of casting cycles to be simulated in a continuous mode. This
parameter is used along with TCYCLE. Both NCYCLE and TCYCLE must be
set. This parameter is typically used in die casting and permanent mold problems.
Enter an integer value.
Version 2006.0
Page 233
TCYCLE
specifies the time of casting cycle to be simulated in a continuous mode (duration
of a cycle). This parameter is used along with NCYCLE. Both NCYCLE and
TCYCLE must be set.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TOPEN
specifies the time at which the mold opens, during one casting cycle. This is used
in conjunction with the "Die Combo" interface definition.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TEJECT
specifies the time at which the part is ejected from the mold, during one casting
cycle. This is used in conjunction with the "Die Combo" interface definition.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TBSPRAY
specifies the time of the beginning of the spray sequence, during one casting
cycle. This is used in conjunction with the "Die Combo" interface definition.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TESPRAY
specifies the time of the end of the spray sequence, during one casting cycle. This
is used in conjunction with the "Die Combo" interface definition.
Enter a floating (real) value.
Select the units of time from: {sec | min}
TCLOSE
specifies the time at which the mold closes (before the start of the next filling),
during one casting cycle. This is used in conjunction with the "Die Combo"
interface definition.
Enter a floating (real) value.
Select the units of time from: {sec | min}
Version 2006.0
Page 234
Version 2006.0
Page 235
ANGTOL
specifies the angle tolerance to be used with VFLIM. Radiation faces which are
grouped using VFLIM tolerance are further differentiated by their solid angle.
Enter a floating (real) value.
ENCLID
specifies an enclosure identification number. This parameter is used in
combination with VFDISP for updating view factors by a displacement interval.
ENCLID indicates which enclosure set is to be tracked, in case all the enclosure
elements are not moving at the same rate.
Enter an integer value.
EPTOL
specifies the emissive power tolerance to be used with VFLIM. Radiation faces
which are grouped using VFLIM tolerance are further differentiated by their solid
angle.
Enter a floating (real) value.
Version 2006.0
Page 236
RDEBUG
specifies the user debug parameter for printing detailed view factor information.
Various combinations of these files may be obtained by adding together these
numbers. For example, RDEBUG = 7 gives all three files. Note that these files can
be quite large, especially the prefix.vf.
Enter an integer value based upon the following:
1 for face to face view factors after symmetrization, in the prefix.vf file,
2 for face to group view factors after symmetrization, in the prepfix.view file
(necessary to see FACE TO GROUP in ViewCAST), or
4 for row sum errors before symmetrization, in the prefix.serr file (necessary to
see ROW SUM ERRORS in ViewCAST
RFREQ
specifies the radiation update frequency. This provides a mechanism for
recomputing the radiosities at some time step interval other than one. This is
particularly useful if you are performing a filling transient along with the view
factor radiation model. In this case, the time step size may be small due to the
filling whereas the mold temperature may not be changing very rapidly. You can
save some computational time by recomputing the radiosities at every tenth step,
for example.
Enter an integer value.
TRI2QUAD
specifies the option to group or not triangles into quadrangles for the radiation
view factor calculation. When TRI2QUAD is set to 1, adjacent triangles are
grouped in order to obtain quadrangles (providing the angle between the triangles
is not too large). This has the effect of reducing the number of radiative face (by
about 50%) which is cutting down the CPU time significantly (by about 75%).
VFDISP
specifies the displacement interval for updating view factors in the radiation
model if there are moving relative surfaces. This is used in conjunction with
ENCLID and will be used in preference to VFTIME if both are specified.
Enter a floating (real) value.
Choose the units of length from: {m | cm | mm | ft | in}. The default is m.
VFLIM
specifies the view factor limit. This parameter is used to agglomerate faces in the
view factor calculations. This reduces the size of the radiosity matrix and speeds
up the radiation calculations.
VFLIM
can be set to a fraction between zero and one. If one face occupies less than this
fraction of the total view space, as seen from another face, the first face is
combined with some others. A value of 0.01 is a good starting point.
Enter a floating (real) value.
Version 2006.0
Page 237
VFTIME
specifies the time interval for updating view factors in the radiation model if there
are moving relative surfaces.
Enter a floating (real) value.
Choose the units of time from: {sec | min}. The default is sec.
Version 2006.0
Page 238
Version 2006.0
Page 239
ADVECTW
specifies the weighting of advection velocities and controls the degree of nonlinearity of the momentum equations. ADVECTW can take on values between
zero and one. Velocities at the last time step are used as the advecting velocities if
a value of zero is used. Velocities at the current time step are used as the advecting
velocities if a value of one is used. Numerical experience has shown that the
accuracy of natural circulation flows can be enhanced by using a factor of 0.5. For
most filling analyses, a value of zero works fine and requires much less
computational time.
Enter a floating (real) value.
COMPRES
specifies whether this is an incompressible flow problem or a compressible flow
problem.
Choose from:
0 to specify an incompressible flow problem, or
1 to specify a compressible flow problem
Version 2006.0
Page 240
CONVV
specifies the convergence criterion for velocity. The value given here is a fraction
of the maximum velocity calculated at each step. Generally, .05 or 5% is
appropriate.
Enter a floating (real) value.
COUPLED
specifies whether the energy and fluid solutions should be coupled or decoupled
within a time step. When the analysis is decoupled, the momentum and pressure
equations are solved repeatedly until convergence. Subsequently, the energy
equation is solved until convergence, assuming the flow field is fixed. With a
coupled analysis, the energy equation is solved in the same loop with momentum
and pressure. Both the momentum and temperature convergence criteria have to
be met to terminate the loop. This method is more accurate, but usually takes more
computational time.
Choose from:
0 to decouple energy and fluid solutions within a time step, or
1 to fully couple energy and fluid solutions within a timestep
COURANT
specifies the courant limit on time step size. This parameter is only used for filling
problems. If COURANT is set to 1.0, the time step will be adjusted so that the
fluid will advance no more than one element length. This is a fairly severe limit on
time step size, but will give the most accurate results for filling transients.
Acceptable results can usually be obtained with values around 100.
Enter a floating (real) value.
DETACHTOP
allows to have a better treatment of the detachment of liquid which is in contact
with horizontal top walls (like "roofs"). A value of 1 allows a better detachment of
such liquid regions, however it takes some more CPU time. This Run parameter
should be added/changed manually in the p.dat file.
EDGE
This Run parameter is not anymore available.
ENDFILL
Sometimes, the user may not be interested by the filling of the last percents (e.g.
the end of the filling of a riser or an overflow). If ENDFILL = 0.98, once 98% will
be reached, the remaining 2% will be filled in one timestep.
FFREQ
specifies the flow update frequency. This provides a mechanism for re-computing
the velocities at some time step interval other than one. This might come into play
if you were solving a conjugate heat transfer problem where the velocity field is
Version 2006.0
Page 241
changing on a longer time scale than the temperatures. This option is not
appropriate for free surface problems.
Enter an integer value.
FLOW
controls the use of fluid equations.
Choose from:
0 do not solve fluid equations,
1 to solve fluid equations,
3 to solve fluid equations during filling, but switch over to thermal only analysis
when the LVSURF fill limit is reached and NCYCLE = 1,
9 to solve fluid equations during filling, but switch over to thermal only analysis
when the LVSURF fill limit is reached and NCYCLE > 1
The default is 0 if there are no "Casting" materials. If "Casting" materials exist,
the default is 1.
FLOWDEL
specifies the delay time between the end of fill and a switch to a thermal only, in
the case of FLOW = 3 simulation. This option is used in conjunction with velocity
boundary conditions with active fill limits. The time delay corresponds to the time
for the fluid to completely settle down in the casting before the thermal only phase
begins.
Enter a floating (real) value.
Select the units of time from: {sec | min} The default is sec.
FREESF
specifies the free surface model number to be used.
Choose from:
0 = no free surface model activated
1 = use the momentum dominated movement of free surface, rapid filling model,
2 = use the gravity dominated movement of free surface, slow filling model.
FREESFOPT
The filling algorithm (in the case of FREESF = 1) was significantly improved in
version 2006.0. From now on, three filling algorithms are available :
FREESFOPT = 0 : it corresponds to the filling algorithm of version 2005.0. It is
less precise than the two other algorithms, however it is the most robust algorithm.
FREESFOPT = 1 and 2 corresponds to two improved filling algorithms. The
difference between these two algorithm is the "numerical balance" between the
mass conservation contribution to the free surface and the momentum
contribution. With FREESFOPT = 1, the mass conservation contribution is more
important than the momentum contribution. This algorithm is more robust, but
may lead to slightly less precise results. With FREESFOPT = 2, the momentum
contribution is predominant over the mass conservation contribution. This model
is supposed to be the most precise, however it is more sensitive to the quality of
the mesh. In order to have good results with FREESFOPT = 2, the mesh should be
well adapted to the local free surface (e.g. one should have enough nodes in the
Version 2006.0
Page 242
stream of the liquid in the case of a thin jet of liquid). This means that the
FREESFOPT = 2 algorithm is less robust as the 1 (or the 0).
As a conclusion, it is strongly advised to use FREESFOPT = 1 in most cases. If
more precision is required, the mesh should be refined and well tuned so that
FREESFOPT = 2 could be used. The FREESFOPT = 0 option is provided for
backwards compatibility purposes, however, it is not advised to be used.
The default value is 1.
GAS
specifies whether or not to consider the trapped gas effects. If the option to
consider trapped gas effects is chosen, trapped gas effects will be considered even
when the model contains no vents, gas injection, or gas diffusion through the
mold. When features normally found in a gas problem (vents, injection, or
gas diffusion through the mold ) are present in a model, GAS will be set
automatically.
Choose from:
0 to not consider trapped gas effects, or
1 to consider trapped gas effects
HEAD_ON
specifies the approach to be used when calculating gravitational term in the
momentum equation for flow problems without free surfaces.
Choose from:
0 = calculate as rho - rho_ref, or
1 = calculate as rho * g
HIVISC
specifies different solution methods for viscosity in the flow problem.
Choose from:
0 = normal flow problem,
1= high viscous flow problem. To be used when the Reynolds number is less the
one. This method only
works for viscosity less than 104 poise. In this case, the advection terms are
neglected, symmetric solvers are employed on the momentum equations, and large
degrees of pressure relaxation are utilized, or
2 = very high viscous flow problem. To be used when the Reynolds number is less
the one. This method is always preferred. In this case, the advection terms are
neglected and momentum effect on implicitly included within a Poisson pressure
equation. This option usually allows for much larger time steps than HIVISC = 1
LVSURF
provides a way to switch from the filling transient to a mode where advection is
due to buoyancy and shrinkage. LVSURF turns all inlets off. It is assumed
thereafter that the free surface is perpendicular to the gravity vector. This allows
the time step to increase significantly. The number represents the fraction of the
total casting volume which is to be filled before changing modes.
Enter a floating (real) value.
Version 2006.0
Page 243
MLUMP
specifies the mass matrix lumping factor.
Choose from:
0.0 to use a consistent matrix, or
1.0 to use a diagonal matrix
NNEWTON
specifies whether the flow is newtonian or non-newtonian.
Choose from:
0 to indicate Newtonian flow, or
1 to indicate non-newtonian flow, where viscosity is a function of shear rate
PENETRATE
Flag to activate the algorithm of inter-penetrating meshes (with a value of 1). This
is used especially for the modeling of a shot piston in hpdc.
Enter an integer value.
PFREQ
specifies the "Particle tracing" launch frequency in the solver. Particles are
launched at each node of the inlet (defined by a velocity BC, an inlet pressure BC
or an Inlet BC), every PFREQ steps. A value of 50 is recommended (see "Display
parameters" for the description of Particle tracing)
Enter an integer value.
PINLET
specifies a pressure driven inflow. Setting PINLET to 1 indicates that all the
pressure boundary conditions are also inflow boundary conditions. Use of this
option allows one to avoid using thin filled regions at the inlets of pressure driven
problems. It allows for filling of metal without having an initial layer of fluid.
Enter an integer value of 0 (off) or 1 (on).
PLIMIT
specifies the pressure cutoff limit. You can use this parameter to turn off an inlet
velocity when the back pressure exceeds the given value. This is useful
particularly in cases where cold shuts are occurring. Otherwise, the program will
keep trying to force more mass into the fluid region, even though there is no place
for it to go, and the pressure will continue to rise.
Enter a floating (real) value.
Choose the pressure units from: {N/m**2 | Pa | KPa | MPa | bar | dyne/cm**2 |
atm | psia | Ksi | lb/ft**2}
PREF
specifies the pressure which is to be subtracted from any boundary condition
pressure in order to convert an absolute pressure into a gauge pressure. This
parameter comes into play when: (1) there is trapped gas, (2) a pressure boundary
condition drives the flow, (3) there are vents, and/or (4) there is gas injected. For
Version 2006.0
Page 244
PRELAX
specifies the pressure relaxation factor. PRELAX, to have an effect, should be
greater than zero and less than one. If it is left to the default value of one,
ProCAST will automatically compute an appropriate relaxation factor.
Enter a floating (real) value.
RELVEL
For centrifugal casting, the fluid flow should be solved in a Relative velocity
reference frame (i.e. in a rotating velocity reference frame). Thus, for centrifugal
casting cases, one should set RELVEL to 1. This Run parameter should be added
manually in the p.dat file (it does not appear in PreCAST).
0 = standard case - no centrifugal (default value)
1 = activation of the relative velocity reference frame - to be used for centrifugal
casting.
Enter an integer value.
TILT
For Tilt pouring problems, one should activate the "Tilt" mode, by setting TILT to
1. This improves the appearance of undesired sticking in the pouring cup.
This Run parameter should be added manually in the p.dat file (it does not appear
in PreCAST).
0 = standard filling mode (default value)
1 = activation of the Tilt mode
Enter an integer value.
TPROF
This parameter indicates that a thermal boundary layer profile is used at the wall
for the energy equation with advection. This has been found to reduce false
diffusion errors.
Choose from:
0 = do not use boundary layer profile, or
1 = use boundary layer profile
Enter an integer value.
TSOFF
This parameter specifies the time at which to switch off the flow solution. For
example, TSOFF 1 42, indicates that the flow solution will be turned off 42
seconds into the simulation. If a cyclic analysis is being performed, then the flow
solution will be turned off 42 seconds into each cycle.
Choose from:
0 = turns this option off, or
Version 2006.0
Page 245
VFREQ
specifies the time step interval for writing velocity and pressure results to the
unformatted files. This parameter can be used to reduce the size of these results
files, which can become quite large for problems with many nodes and time steps.
Note that it is only possible to restart a run from one of the time steps that was
written. Only the steps that are written can be viewed with post-processing.
Enter an integer value.
VPROF
This parameter indicates that a flow boundary layer profile is used at the wall for
the momentum equation with advection. This has been found to reduce false
diffusion errors, although it is not very useful with WSHEAR=2.
Choose from:
0 = do not use boundary layer profile, or
1 = use boundary layer profile
Enter an integer value.
WALLF
specifies the wall slip behavior. The default value is 0.9. A value of 0.98
correspond to more slip along the wall, whereas a value of 0.8 will act as if the
mold surface is rougher (more friction). It is advised to use a value of 0.8 for sand
gravity casting, a value of 0.9 for gravity die casting and a value of 0.98 for high
pressure die casting (HPDC). For LPDC, this value is not used, as WSHEAR=0
must be used (in this case the WALLF algorithm is not activated).
WSHEAR
specifies whether or not the wall shear formulation will be used. The wall shear
formulation will convert no-slip boundary conditions into wall traction conditions.
Choose from:
0 to indicate that wall shear formulation will not be used, or
2 to indicate wall shear formulation will be used.
The use of the Wall shear formulation allows to have non-zero velocities at the
mold wall, which is more representative of the reality (slip of the liquid along
walls).
For HPDC and Gravity casting, a value of WSHEAR = 2 should be used. For
LPDC, a value of WSHEAR = 0 must be used.
The WSHEAR = 1 option is not anymore supported.
Version 2006.0
Page 246
CMU
specifies the proportionality constant used in the turbulent viscosity equation.
Enter a floating (real) value.
CONE
specifies the proportionality constant used in the production of turbulent energy
dissipation.
Enter a floating (real) value.
CTWO
specifies the proportionality constant used in the destruction of turbulent energy
dissipation.
Enter a floating (real) value.
KAPPA
specifies the Von Karman's constant, usually taken as 0.4
Enter a floating (real) value.
Version 2006.0
Page 247
SIGMAE
specifies the diffusivity modifier used in the turbulent energy dissipation transport
equation.
Enter a floating (real) value.
SIGMAK
specifies the diffusivity modifier used in the turbulent kinetic energy transport
equation.
Enter a floating (real) value.
TBRELAX
specifies the turbulence relaxation parameter.
Enter a floating (real) value.
TURB
specifies whether the turbulent flow model is turned on or off. A model started
with TURB = 1 can be restarted at a later time with TURB = 0. This allows
laminar conditions to be considered during mushy or natural circulation flows.
Once TURB has been set to zero, the turbulence model can not be restarted at a
later time. Setting TURB to one for a flow problem which has no turbulence
boundary conditions assigned is OK; the software will automatically define them.
Enter:
0 to turn the turbulent flow model off, or
1 to turn the turbulent flow model on
Version 2006.0
Page 248
Version 2006.0
Page 249
AVEPEN
AVEPEN corresponds to the "Average Penetration". This corresponds to the
maximum average penetration which is allowed during the calculation.
During the calculation, the PENALTY is automatically changed in order that the
penetration is not larger than AVEPEN. The goal is to have the lowest possible
PENALTY number to speed-up convergence, and AVEPEN allows to set the
upper limit.
The default value of AVEPEN is 0.1 mm. For large casting, it is advised to
increase the value of AVEPEN for faster convergence.
CRACK
The cracking indicator model is activated with CRACK = 1 or 3. By default, the
value of CRACK is set to zero. This Run parameter is not available in the Run
Parameters menu and it should be added manually in the p.dat file.
CRACK = 1 activates the cracking model, without feedback on the stress
properties
Version 2006.0
Page 250
CRITFS
This corresponds to the critical fraction of solid where the stress calculation starts.
By default, it is set to 0.5 (i.e. 50% fraction of solid). This critical value is also
used for the computation of the hot tearing.
CONVS
specifies the convergence criterion for the stress calculation.
Enter a floating (real) value.
GAPMOD
specifies the treatment of the interface heat transfer coefficient. With a value of 1,
the interface heat transfer coefficient is automatically modified to account for the
air gap formation (an additional thermal resistance is computed as a function of
the gap width, taking into account conduction and radiation through the air or
vacuum). With a value of 0, the interface heat transfer coefficient which is defined
in PreCAST will not be modified during the calculation.
LOADSCL
When a load is applied in non-linear problem, it has to be applied incrementaly
(within a timestep). LOADSCL is the number of increments for this loading. The
higher the value of LOADSCL, the more accurate is the result (and the higher is
the CPU time). This has to be used mainly for structure analysis type of problems
(e.g. tensile test). For usual casting, as the loading is progressive (due to gradual
temperature changes), it is not necessary to apply such increments. The default
value of LOADSCL is 1. This value has to be added manually in the p.dat file.
PENALTY
PENALTY controls the level of "penetration" allowed by the "contact algorithm".
As the displacements are computed numerically at interfaces, there is always some
penetration between two bodies (as long as they are touching).
High values of PENALTY means that the allowed penetration is very small and
this leads to a more difficult convergence of the algorithm. Small values of the
PENALTY allows to have more penetration (which means a more easy
convergence).
The default value for Penalty is 1, but for thin sections, it could be advised to set it
to 0.01.
During the calculation, the PENALTY is automatically changed in order to
optimize the CPU time and in order to be within the AVEPEN limit.
When a calculation is restarted, it is advised to set the PENALTY to zero. In this
case, the last PENALTY which was used in the calculation will be automatically
set. If a non-zero value is defined, it will be used for the restart step. One should
note that if the PENALTY was decreased during the calculation, to leave the
default value may lead to a very long convergence at the restart step. Thus, it is
Version 2006.0
Page 251
strongly advised to set it to zero for a restart. The PENALTY at each stress
timestep is indicated in the p.out file.
SCALC
specifies the time step interval the stress calculation is performed. It is thus
possible to perform the stress calculation only every 10 thermal steps. SCALC and
SFREQ are independent values and one will not affect the other.
Enter an integer value.
SFREQ
specifies the time step interval for writing stress results to the unformatted files.
This parameter can be used to reduce the size of these files, which can become
quite large for problems with many nodes and time steps. Note that it is only
possible to restart a run from one of the time steps that was written. SCALC and
SFREQ are independent values and one will not affect the other. One should
however be careful that SFREQ is a multiple of SCALC.
Enter an integer value.
STRESS
specifies whether the stress calculation is turned on or off.
Enter:
0 to turn the stress calculation off, or
1 to turn the stress calculation on
VACUUM
When GAPMOD = 1, the interface heat transfer coefficient depends upon the gap
width. With VACUUM = 0, air conduction is taken into account (i.e. a heat
resistance corresponding to heat conduction through air is taken into account).
With VACUUM = 1, no heat conduction is taken into account. In both cases,
radiative transfer through the interface is taken into account (when there is a gap).
This Run parameter should be added manually in the p.dat file.
Version 2006.0
Page 252
EQNMAX
First nucleation parameter of the dendritic primary phase. Maximum density of
nuclei of the Gaussian distribution.
EQSTD
Second nucleation parameter of the dendritic primary phase. Standard deviation of
the Gaussian distribution.
EQUNDER
Third nucleation parameter of the dendritic primary phase. Average undercooling
of the Gaussian distribution.
EUNUCL
First nucleation parameter of the eutectic phase. Nucleation factor.
EUPOWER
Second nucleation parameter of the eutectic phase. Nucleation exponent.
Version 2006.0
Page 253
EUGROW
Eutectic growth kinetics constant.
MICRO
Activation of the microstructure module. The module is activated with a value of
1.
See the "Microstructures" chapter for more details about the models.
It is possible to select the default values for a given system with the "Select
Default Values" button.
Version 2006.0
Page 254
Version 2006.0
Page 255
When one clicks on the "Select Pre-defined Set" button, a list of the available sets
are displayed. Then, the user has to select the desired one. In the case below, six
sets are available, among ten possible sets.
Version 2006.0
Page 256
The values contained in the Pre-defined sets correspond to the Recommended Run
Parameters for each process. It is thus strongly advised to use these pre-defined
sets or to customize them according to your specific processes (see below).
The configuration of the Pre-defined sets can be easily performed by the user. One
should just save the a "d.dat" file, with the desired Run Parameters in the
"dat/pref" directory of the installation, under the name "predefined_x_p.dat",
where x is a number between 1 and 10 (see below).
The label which appears in the Run Parameter window above should be included
in the predefined file as follows (TITLE line) :
Version 2006.0
Page 257
However, for an everyday use, only a few Run parameters have to be set or
modified. This section is presenting the most "popular" Run Parameters that
should be set, with proposed values, for each main family of processes.
These parameters recommendations may be slightly different from previous
versions, as the solver algorithms have been modified.
These recommended Run parameters correspond to the one which are pre-defined
in the "Pre-defined Run Parameters" window. It is thus advised to activate the
"pre-defined" set corresponding to the process and then to set the appropriate
stopping criteria.
(it is advisable to set a stopping criterion in order to limit the CPU time and
avoid unnecessary storage of results)
TFINAL
TSTOP
Porosity
Gravity casting
Timestep handling
DT = 1e-3
DTMAXFILL = 1e-1
DTMAX = 0.5 - 5 (depending upon the size of the model and thus the
solidification time)
Porosity
PIPEFS = 0.3
GATEFEED = 0
Filling
WSHEAR = 2
FREESFOPT = 1
WALLF = 0.8
LVSURF = 0.98
Version 2006.0
Page 258
DT = 1e-6 to 1e-4 (it depends upon the initial velocity of the first stage)
DTMAXFILL = 1e-2
DTMAX = 0.2 - 1 (depending upon the size of the model and thus the
solidification time)
Porosity
PIPEFS = 0.0
GATEFEED = 1
Filling
WSHEAR = 2
FREESFOPT = 1
WALLF = 0.99
LVSURF = 1.0
PINLET = 1 for a pressure filling or PINLET = 0 for a velocity/inlet filling
DT = 1e-3
DTMAXFILL = 1e-2 (it is important to limit the timestep during the filling
of an LPDC part. A value of 1e-2 is recommended for filling time of about
5-20 s.).
DTMAX = 0.2 - 1 (depending upon the size of the model and thus the
solidification time)
Porosity
PIPEFS = 0.0
GATEFEED = 1
Filling
Tilt casting
Timestep handling
DT = 1e-3
DTMAXFILL = 1e-1
DTMAX = 0.5 - 5 (depending upon the size of the model and thus the
solidification time)
Porosity
PIPEFS = 0.3
GATEFEED = 0
Filling
WSHEAR = 2
FREESFOPT = 1
WALLF = 0.8
TILT = 1
DETACHTOP = 1
Version 2006.0
Page 259
Centrifugal casting
Timestep handling
DT = 1e-3
DTMAXFILL = 1e-1
DTMAX = 0.5 - 5 (depending upon the size of the model and thus the
solidification time)
Porosity
PIPEFS = 0.3
GATEFEED = 0
Filling
WSHEAR = 2
FREESFOPT = 1
WALLF = 0.8
RELVEL = 1
Version 2006.0
Page 260
POROSITY MODELS
The different porosity models available in ProCAST are summarized hereafter.
Then, a full description of each model is provided.
POROS = 1
The POROS = 1 model corresponds to the latest porosity model of ProCAST. It
accounts for coupled micro and macroporosity, as well as pipe shrinkage. It can be
applied to both gravity casting and injection (either hpdc or lpdc).
This model can be used with or without flow calculations.
POROS = 4
The POROS = 4 model corresponds to the same model as POROS=8 (see below),
with the additional treatment of multiple free surfaces for piping.
This model can be used with or without flow calculations.
POROS = 8
The POROS = 8 model corresponds to the porosity model which was available in
version 3.2.0 of ProCAST (using the POROS=1 Run Parameter of v3.2.0).
Although this model is less sophisticated than the current POROS = 1 model, it
was kept in this version for the users who have calibrated the model to their
casting and who obtained good results in the past.
The only difference between the current POROS=8 model and the one available in
version 3.2.0 is the method used to compute liquid pockets. Thus, this may lead to
slight differences between the versions.
This model can be used with or without flow calculations.
One should notice that even if POROS = 0 is set, piping will be still calculated
and displayed. In order to disable the piping calculation (in the case of a
THERMAL only calculation), one should set FREESF to zero.
Version 2006.0
Page 261
POROS=1
Parameters of the model
POROS
MACROFS
PIPEFS
USERHO
FEEDLEN
1
(porosity model)
0.7 (critical fs for porosity formation)
0.3 (critical fs for piping formation)
1
(flag to set the density model)
1 5.0e-3 (critical feeding length)
It is necessary to define the gravity vector in order to have meaningful results with
POROS=1.
Model
When a casting solidifies, pockets of liquid are created, surrounded by a mushy
zone and then a solid shell. Automatically, the casting is divided into "regions"
within which the fraction of solid is lower than one or that are bounded by walls
(or symmetry planes). As solidification proceeds and depending upon the
complexity of the geometry, the number of "regions" may increase with time. A
region can thus be split in more regions. A region can disappear when all nodes
have completely solidified.
Version 2006.0
Page 262
b)
Find the highest point of the region that is on a free surface and has a
fraction of solid higher than PIPEFS. In this case, macroshrinkage occurs at that
point. This will also correspond to piping (same result as a) above). However,
instead of showing an empty volume, it will have a "Shrinkage porosity" value of
1.
(For display purposes, in the macroshrinkage, the shrinkage porosity value is set
to 1 and FVOL is set to a value above 0.5, so that it will not exhibit piping, i.e.
"empty nodes", but shrinkage).
Version 2006.0
Page 263
c)
No nodes of the "region" are on a free surface with a fraction solid lower
than PIPEFS. In this case, no more piping can occur and macroshrinkage in the
bulk of the casting will appear. The macroporosity will appear at the highest most
liquid point of the region (e.g. if there is a pocket of liquid which is surrounded by
a mushy zone, itself surrounded by a solid shell, the macroporosity will start at the
highest point of the liquid pocket).
(For display purposes, in the macroporosity region, the shrinkage porosity value is
set to any value and FVOL is set to a value above 0.5, so that it will not exhibit
piping, i.e. "empty nodes", but shrinkage).
Version 2006.0
Page 264
c)
on the other hand, if FEEDLEN=0, the amount of microporosity is the same
throughout the part (except where there is macroporosity). In this case, the amount
of microporosity is everywhere equal to the density change between
Fs=MACROFS and Fs=1.
Version 2006.0
Page 265
d)
if FEEDLEN is set to a very large value (larger than the casting side), the
region between MACROFS+FEEDLEN is not existing anymore. Thus, no
microporosity is created until the whole pocket is above MACROFS (i.e. no
microporosity created according to b1) above). Then, microporosity can start to
occur, according to b2) above.
The amount of porosity is displayed in ViewCAST under the "Shrinkage porosity"
Contour. The unit is volume fraction [-]. In general, values which are higher than
0.01 can be considered as macroporosity, whereas regions with a value lower than
0.01 correspond to dispersed microporosity.
Symmetry planes are taken into account in the liquid pocket calculation (i.e.
symmetry planes "close" the pockets, even if there is liquid at the symmetry wall.
See the "Density definition" section for more details about how to define the
density.
See the "Active feeding" section for more details about feeding from the piston or
the ingate in the case of hpdc and lpdc.
POROS=4
Parameters of the model
POROS
MACROFS
4
0.7
(porosity model)
(critical fs for porosity formation)
Model
This model is based upon the same algorithm as the POROS=8 model (see below).
In addition, multiple piping can be considered. This means that piping occurs at
the highest free surface of each region. Thus, one can have piping at the top of
risers which are at different levels (this was not the case in version 3.2.0 and thus
does not occur with POROS=8)
See the "Density definition" section for more details about how to define the
density.
POROS=8
Parameters of the model
POROS
MACROFS
Version 2006.0
8
0.7
(porosity model)
(critical fs for porosity formation)
Page 266
Model
When the casting solidifies, pockets of liquid are created, surrounded by a mushy
zone and then a solid shell. As soon as a pocket of liquid is surrounded by a zone
which has a solid fraction higher than MACROFS, the density of each node inside
the pocket (from fs=1 to fs=MACROFS) is recorded (as a "critical density").
Then, the amount of porosity of each of these nodes is equal to the density
variation between this "critical density" and the density of the solid.
Piping is occurring at the highest free surface of the model. Thus, if there are two
separate regions with risers at different heights, only the highest one will exhibit
piping, even if the lower one is in an isolated region. POROS=4 corrects this
situation (see above).
Symmetry planes are taken into account in the liquid pocket calculation (i.e.
symmetry planes are "closing" the pockets, even if there is liquid at the symmetry
wall.
The amount of porosity is displayed in ViewCAST under the "Shrinkage porosity"
Contour. The unit is volume fraction [-].
See the "Density definition" section for more details about how to define the
density.
Density definition
For most alloys, the density at the liquidus is lower than the density at the solidus,
thus leading to porosity. By default, the porosity module of ProCAST is using the
density curve in the mushy zone which is defined in the material database (see the
green curve in the Graph A below). However, if this density change in the mushy
zone is not well known, it is possible to automatically calculate the density in the
mushy zone, as an average of the liquid and solid densities, weighted by the
fraction of solid. The density of the liquid and of the solid is calculated as a
function of temperature, by extrapolating the density slopes at the liquidus and
solidus respectively.
Version 2006.0
Page 267
The user has the possibility to activate this automatic density computation with the
Run parameter USERHO. If USERHO is set to 0 (instead of 1 which is the default
value), the density curve which is defined in the material properties, between the
solidus and liquidus, will be ignored and the automatic density computation will
be activated. (Please note that in version 4.x.x, the default value was 0).
This is valid for all the Porosity models (POROS = 1, 4, 8).
Active Feeding
In the case of injection (e.g. high or low pressure die casting), for a while, the
shrinkage is compensated by the piston in the case of hpdc and by the liquid bath
for lpdc, thus leading to no piping.
ProCAST can account for such "active feeding", by setting the Run parameter
GATEFEED=1. In this case, no piping will occur, but liquid will feed the ingate,
as long as the fraction of solid is lower than GATEFS (i.e. there is feeding in all
regions which are within the same GAFEFS isosurface as the gate). The gate is
defined as being the region where in inlet velocity is applied, or where a pressure
is applied (or where GATENODE is applied).
One should note that the porosity level will NOT depend upon the value of the
pressure. The Active feeding is there only to compensate for the shrinkage of the
region in contact with the piston, as long as the fraction of solid is lower then
Version 2006.0
Page 268
GATEFS (i.e. within the same GATEFS isosurface as the gate). This does not
correspond fully to a real third stage pressure, however, the value of GATEFS
could be adjusted with pressure if desired).
For HPDC and LPDC, it is advised to set PIPEFS = 0.0 in order to prevent piping
at the top of the casting, if the gate is closing too early.
The active feeding is valid only with the POROS = 1 model.
In the case of a thermal only calculation (for HPDC or LPDC), one should set a
Pressure boundary condition at the ingate (in order to activate the active feeding)
and one should disable the FLOW run parameter (as normally a Pressure BC
would automatically switch ON the flow solver). Of course, GATEFEED should
also be set to 1 in this case.
In order to activate GATEFEED, one should apply a pressure of inlet velocity BC
on external faces. There are cases where there are no external faces on which to
apply these boundary conditions. This happens in the case of a filling with a shot
piston. To account with such situation, one can apply instead a "Gate feeding"
condition on an inside location of the casting (defined by its node number). In
general, it is advised to select a node inside the final biscuit, which will remain
liquid during most of the casting process. To define this node number, the
GATENODE Run Parameter should be specified, followed by the node number.
In order to find the node number corresponding to the desired location, it is
advised to use the "pick" capability of ViewCAST. This can be done before the
run of the case, just after DataCAST.
Warning :
In the case of a thermal only calculation, a Pressure BC should be set at the ingate
in order to switch ON the active feeding. In such a case, the Pre-processor
automatically is activating the flow solver (FLOW=3). As a consequence, one
should set FLOW=0 manually in the d.dat file in order to deactivate the flow
solver (but the porosity calculation will be performed with active feeding).
Version 2006.0
Page 269
In order to account for such expansion, the POROS=1 model was adapted in the
way describe hereafter.
In the "pockets" of liquid and mushy zone which have a solid fraction lower than
MACROFS (+FEEDLEN), the integral of the density change at each timestep is
performed. In the case of expanding material, some location will expand and some
will shrink. If the total density change corresponds to a net shrinkage, the
"regular" model applies (i.e. this macroshrinkage will occur as the highest most
liquid point). If it corresponds to a net expansion, two scenario may occur :
a)
Find the highest point of the region that is on a free surface and has a
fraction of solid lower than PIPEFS. In this case, the free surface of the casting
(usually in the riser) will go up by the amount corresponding to the net expansion.
b)
Find the highest point of the region that is on a free surface and has a
fraction of solid higher than PIPEFS. The amount of expansion is applied
proportionally to all of the nodes in the region that have pre-existing porosity
(thus, the macroporosity which appeared earlier, when there was a net shrinkage,
will be "refilled"). For example, if the amount of expansion is enough to "refill"
50% of the total porosity in the region, then the porosity of each node is reduced
by 50%.
c)
No nodes of the "region" are on a free surface with a fraction solid lower
than PIPEFS. The expansion is distributed to all the nodes in the region as
described in b).
Version 2006.0
Page 270
c)
On the other hand, if FEEDLEN=0, the amount of microporosity is the same
throughout the part (except where there is macroporosity). In this case, the amount
of microporosity is everywhere equal to the density change between
Fs=MACROFS and Fs=1. Again if the density change between the fs=MACROFS
and fs=1 is positive (expansion), there is no microporosity formation. The already
formed micro porosity can be refilled partially or fully during expansion
depending on the degree of the density change.
d)
If FEEDLEN is set to a very large value (larger than the casting size), the
region between MACROFS+FEEDLEN is not existing anymore. Thus, no
microporosity is created or refilled until the whole pocket is above MACROFS
(i.e. no microporosity created or refilled according to b1) above). Then,
microporosity can start to occur or disappear, according to b2) above.
Mold rigidity
The rigidity of the mold has an influence on the amount of porosity in the case of
expanding alloys. If the mold is totally rigid, the casting can not expand and thus
the alloy expansion will be "available" for the "refill" of the existing porosity. On
the other hand, if the mold is very soft (or weak), the casting will expand and thus
Version 2006.0
Page 271
there will be no "refill" of the porosity (of course, the reality is more complex as
the solid shell is thick enough it will act as a "rigid" mold, even if the sand mold is
weak.
In order to take the mold rigidity into account, the Run Parameter MOLDRIG is
introduced. MOLDRIG should be defined by a value between 0 and 1. All the net
expansion is multiplied by MOLDRIG. Thus, with MOLDRIG=1, corresponding
to a rigid mole, the expansion will be fully accounted. On the other hand, no
expansion will be taken into account if MOLDRIG=0. The expansion will be
compensated by the mold movement because the mold is too weak to hold the
expansion in this case. For real situations, the value of MOLDRIG should be set
somewhere between 0 and 1 depending upon the casting processes. MOLDRIG
should be added "manually" in the p.dat file by the user. The default value is 1.
Density curve
For expanding materials, the density defined in PreCAST should be not anymore
monotonic. The density can increase (locally) with increasing temperature.
One should note that such density curve will not be obtained when the material
properties are computed with the Computherm database. This is because the
expansion is depending upon the microstructure which itself depends upon the
cooling rate (see next section). Thus, such curve should be defined manually by
the user, based upon experiments.
Version 2006.0
Page 272
Cooling curves calculated with the Microstructure module of ProCAST (top) with the
corresponding computed density curves (bottom) (SGI with 3.5% C, 2.39% Si, 0.1% Mn,
0.023% Cr, 0.045% Mn, 0.03% Ni, 0.01% Cu, 0.0055% V, 0.0056% Ti, 0.117% Co)
Version 2006.0
Page 273
The above example is showing the advantage of using a coupled Microstructure Porosity calculation. If the Microstructure module is not available, one should use
the density curve which corresponds to the "closest" average cooling rate.
Example
The figures hereafter are showing an example of an expanding alloy. One can see
very well that the level of the metal in the left riser (which has an insulating
sleeve) is first going down (overall shrinkage) and then it is going back up (overall
expansion).
Version 2006.0
Page 274
The following figure is showing an enlargement on the riser where the liquid level
is going down and then up.
Version 2006.0
Page 275
VIRTUAL MOLD
ProCAST offers the capability of modeling a mold without meshing it, with the
Virtual Mold option. This is especially useful in the case of large sand casting. It
can also be used in permanent mold casting, if one is mainly interested in the
filling behavior.
When a Virtual mold is used, one should define the dimension of the mold (which
is an orthogonal box, aligned with X, Y and Z, the material properties of the mold
and the interface heat transfer coefficient between the different part of the casting
and the mold.
The Virtual mold model considers that the thermal diffusion is occurring in the
mold, according to "half diffusion distances". Thus, the model calculates for each
face of the casting what is the half diffusion distance. This distance is either the
one between the face and the limit of the mold or half the distance between the
face and an other face in front. At the limit of the mold, an adiabatic boundary
condition (i.e. no flux) is considered. Thus, it is advised to defined a "box" which
is large enough, in order not to "saturate" the virtual mold.
A good start for the Virtual Mold definition is to look at the minimum and
maximum size of the model in the "Geometry/Check Geom/Min-Max" menu.
Then, one should open the "Geometry/Virtual Mold" menu, to get the following :
Version 2006.0
Page 276
The size of the Virtual Mold box can be defined in the following fields. One has
the possibility to define automatically the virtual mold box size, using the "Default
Size" button. One should first select on the right, whether the virtual mold box
should be 1, 2, 3, 4 or 5 times larger than the part, in the X, Y and Z directions
respectively. For small components, it is advised to use a virtual mold 5 times the
component size. For large castings, a virtual mold with a size of 2 o r 3 times the
component size will be enough. In case there is any doubt, it is advisable to take a
large virtual mold rather than the reverse, in order to avoid saturation.
Version 2006.0
Page 277
Please note that one should NOT use the real size of the mold for the Virtual mold
computation. As the model considers that there is no cooling of the sides of the
virtual mold, the box should be large enough in order to avoid satuaration of the
mold (and thus too slow cooling).
The size of the Mold box can be visualized with the "Show Mold" button (in the
case of the figure hereafter, the virtual mold of a size 1 time larger than the
component was selected, for viewing purposes).
Version 2006.0
Page 278
When the Mold size is good, one can "compute" the Virtual Mold, using the
"Compute Mold" button. Please note that the computation may take sometime,
depending upon the size (i.e. the number of surface elements) of the model. The
computation time is independent upon the size of the mold box itself (in the case
of the figure hereafter, the virtual mold of a size 1 time larger than the component
was selected, for viewing purposes).
Once the Virtual Mold is computed, the "Thermal depth" can be visualized with
the "Show Depth" button. The color scale can be changed with the "Set Scale"
button, which opens the following panel :
Version 2006.0
Page 279
Finally, if needed, it is possible to erase the Virtual mold, using the "Remove
Mold" button.
Before setting a Virtual mold, a few precautions should be taken :
When a virtual mold is used in conjunction with a symmetry, it is mandatory to set
first the symmetry and then to generate the virtual mold. Otherwise, the symmetry
of the virtual mold will not be applied.
If the mesh is made out of several material domains, it is mandatory to define first
the interfaces between the different domains and then to generate the Virtual
mold. If interfaces are changes after the Virtual mold creation, it will be
destroyed.
Once the Virtual Mold is generated, it appears in the Material list in the Material
properties assignment. One should assign the desired material properties to this
Virtual mold domain.
In the same way, the interfaces between the Virtual Mold and the different
material domains are automatically generated. These are labeled "Virtual". One
should assign an interface heat transfer coefficient to each "Virtual" interface.
Version 2006.0
Page 280
Version 2006.0
Page 281
FILTERS
If Filters are present in the process, they can be modeled.
Firstly, the filter should be meshed as a separate material domain, as shown
hereafter.
Version 2006.0
Page 282
Version 2006.0
Page 283
Finally, the interface heat transfer coefficient between the liquid metal (casting
material) and the filter material can be defined in the "Boundary
Conditions/Assign Volume/Filter Heat" menu. The list of the Filter materials is
shown, as well as the interface heat transfer coefficient database, ready for an
assignment.
Version 2006.0
Page 284
EXOTHERMIC
Exothermic sleeved can be modeled in ProCAST, with the appropriate heat
generation.
Firstly the exothermic sleeve should be meshed as a separate material domain.
Then, the corresponding material properties should be assigned and the
"EXOTHERMIC" type should be defined (if the EXOTHERMIC type is not
defined, the exothermic energy, as defined hereafter will not be released) :
Version 2006.0
Page 285
The "standard" properties (thermal conductivity, density and specific heat) should
be defined as usual (1) for the sleeve material. In addition, the exothermic
properties should be defined in the "Exothermic" tab (2). The Exothermic energy
(3) corresponds to the amount of energy which is generated during the burning of
the sleeve. The Ignition Temperature (3) corresponds to the temperature at which
the exothermic reaction is initiated.
The "burning kinetics" is defined in the table (4-5), as a fraction of burning, which
is a function of time. Once the exothermic reaction is started (i.e. when the
temperature is going above the ignition temperature), the exothermic energy will
be released according to the burnt fraction.
Version 2006.0
Page 286
CYCLING
ProCAST allows to model the cycling sequence in the case of die casting (i.e. to
account for the heating of the die mold during the first cycles of casting). The
figure below shows the principles of cycling, as well as the different sequences.
In order to model cycling, one should be able to take into account the fact that
when the mold is closed, there is an interface heat transfer between the casting and
the mold and when the mold is opened, one should consider a cooling of the die
and of the casting with the ambiance. One should thus switch between an interface
heat transfer coefficient and two "Heat" boundary conditions.
Version 2006.0
Page 287
The above diagrams are showing in green the interface heat transfer coefficients
as a function of time for both the interface between the casting and the fixed die
(solid green line) and between the casting and the mobile die (dashed green line).
One could see that the interface heat transfer coefficient goes to zero when the
casting and the mold are not anymore in contact. On the reverse, as soon as the
mold opens, a "Heat" boundary condition should be set (the red curve on the
above figure corresponds to the heat transfer coefficient and the dashed blue curve
corresponds to the ambient (or external) temperature. One could see on the above
figure that during the spray sequence, the heat transfer coefficient increases and
the spray temperature is also taken into account.
When the die is closed, before the next filling, the interface heat transfer
coefficient between the dies is again activated and the interfaces between the
casting and the die(s) are set to zero.
One should note that the Pre-defined Run parameters for HPDC-cycling are not
designed in order to peform a Porosity calculation during the cycling calculation.
If this is desired, one should set to :
POROS
Version 2006.0
Page 288
PIPEFS
0.
GATEFEED 1
Then, a pressure BC should be set at the ingate in order to trigger Gate feeding.
Version 2006.0
Page 289
LOST FOAM
ProCAST provides the capability to model "lost foam" casting or "evaporative
pattern" casting.
The cavity is filled with foam and is surrounded by a sand mold. During the
filling, the hot liquid metal is heating up the foam which is burning, leaving the
space for the liquid metal.
In the Lost Foam process, the filling is controlled by the rate of burning of the
foam, as well as the escape of the gas through the ingate or through the permeable
sand mold.
The set-up of a Lost Foam case should be done as follows.
Firstly, the geometry should contain at least three components : a part of the
downsprue which is empty, the cavity filled with the foam and a sand mold (see
figure below).
The following figure shows how the materials should be set. The downsprue
should be assigned with the casting material and it should be set as EMPTY=YES.
The cavity, which is filled with foam should be assigned with a "Foam" material
(EMPTY=NO) and the mold should be set with a sand material (EMPTY=NO).
Version 2006.0
Page 290
There are no special requirements for the material properties definition of the
casting material.
For the Foam material, the thermal conductivity, the density, the specific heat and
the latent heat (of burning) of the foam should be specified. Moreover, under
Tsolidus and Tliquidus, one should specify the temperature at which the foam
starts and ends to burn respectively. One should note that the burning kinetics (and
thus the filling rate) is influenced by the density, the specific heat, the latent heat
and the burning temperature. The larger these values, the slower the burning
kinetics will be. This is due to the fact that the foam should be heated up by the
liquid metal and that if these quantities are larger the amount of heat required to
burn the foam will be larger and thus it will take more time.
For the sand, in addition to the usual thermal properties, it is necessary to define
it's permeability (in the "Fluid/Permeability" tab). Typical values range from 1e-6
to 1e-7 cm**2.
The interface between the casting material and the foam material should be set as
"EQUIV" (i.e. no interface), whereas the other interfaces should be set as
"COINC".
In addition the usual "Heat" boundary condition (to cool down the outside of the
mold) and the inlet Temperature, one should set two Pressure BC. One for the top
of the down sprue (on the whole surface) and the other for the outside of the mold.
It is recommended to set a value of 1 atm. on the outside of the mold and a value
slightly larger (e.g. 1.05 atm) on the top of the downsprue.
Version 2006.0
Page 291
Version 2006.0
Page 292
Version 2006.0
Page 293
FOAMHTC 0.02
FOAMHTCMAX 0.25
BURNZONE 1.0
GASFRAC 0.1
DIAG 262144
These parameters are regulating the rate of transfer of heat between the liquid
metal and the foam. When the liquid metal front is at a distance of "BURNZONE"
from the foam, the "interface heat transfer" between them is equal to the value of
FOAMHTC/BURNZONE. This heat transfer is increasing as the liquid get closer
from the foam. When the liquid is "touching" the foam, a maximum heat transfer
coefficient (equal to FOAMHTCMAX) is set. FOAMHTC and FOAMHTCMAX
are defined in the p.dat file in CGS units and BURNZONE is in "cm".
BURNZONE corresponds to the "usual" distance between the foam and the liquid
metal. One should be careful that the mesh size must be finer than the value of
BURNZONE. One should note that the burning kinetics will change when the
value of BURNZONE is changed (as the interface heat transfer coefficient is
equal to FOAMHTC divided by BURNZONE).
As the rate of heat transfer between the foam and the liquid metal is not well
known, it may be needed to calibrate the filling time with the values of
FOAMHTC and FOAMHTCMAX. To shorten the filling time, one should
increase these values.
The value of GASFRAC corresponds to the fraction of the foam which is
transformed into gas (during the burning). The rest is mainly transformed in liquid
traces. A value of 0.1 (10%) is recommended.
In order to allow for smooth Restarts, the Run Parameter DIAG should be set to
262144. This setting is not necessary if no Restart is performed.
Version 2006.0
Page 294
THIXO CASTING
ProCAST has a dedicated model for thixo casting (or semi-solid casting).
When a semi-solid material is injected in a mold cavity, it's viscosity is depending
upon the shear rate (as well as the shear rate which was encountered by the metal
previously during the injection). When the shear rate is high, the solidifying
dendrites are broken and the fluidity is increasing (i.e. the viscosity is decreasing).
In order to account for such a behavior, a specific "Power cut-off" model was
designed.
In order to set-up a Thixo casting case, one should define the appropriate material
properties of the casting material and define an additional simple input file
(prefixg0.dat).
In the Material properties definition, all the thermal properties should be defined
as usual. In the Fluid tab, the Viscosity should be defined with the "Power-Cutoff"
tab.
The values of the "Zero viscosity", the "K Factor" and the "Power", which can be
Temperature dependant, should be calibrated with experiments. Please note that
the value of n should be negative (in order to have a decreasing viscosity with an
increasing shear rate).
Version 2006.0
Page 295
The principle of the Thixo casting model is to divide the mesh in different regions.
For each region, it is possible to define the critical "cut-off" value of the shear rate
. This definition is done in a small input file named prefixg0.dat.
The format of this ASCII file is the following (one line per domain of the mesh) :
Domain_number
critical_shear_rate
The following example shows a g0.dat file for a case with 4 domains. One should
note that domain 2 is the mould and a dummy value (of 1) should be set. The units
of this critical cut-off shear rate is [s].
The principle for the definition of the critical cut-off shear rates is described
hereafter. It is based upon the fact that dendrites are broken in locations where the
shear rate is high and then, even if the shear rate is decreasing afterwards (in an
opened cavity), the viscosity will not increase anymore (as the dendrites have been
broken). This means that the viscosity in a cavity which is following a gate will
remain at a rather low level, corresponding to the shear rate of that gate. One can
illustrate this principle with the following figure.
In domain 3, the viscosity of the semi-solid material is about the one which
corresponds to the shear rate that was encountered in the gate (i.e. domain 2).
Thus, one should set a "cut-off" value in domain 3 which corresponds to the mean
shear rate of domain 2. In the same way, the "cut-off" value of domain 5 should
Version 2006.0
Page 296
Finally, in order to activate the Power cutoff model, the two following Run
parameters should be set :
HIVISC = 2
NNEWTON = 2
Version 2006.0
Page 297
CENTRIFUGAL CASTING
Starting from version 2006.0, centrifugal casting cases should be set-up in the
following way.
Only the CASTING domains should be set with a revolution velocity (i.e. not the
mold materials). This must be defined in the "Process/Assign Volume" menu, with
a "Revolution".
The Run Paremeter RELVEL should be set to 1. This Run parameter is activating
the resolution of the fluid equations in a "Relative velocity reference frame" to
handle the rotation.
The rest of the set-up should be done as for a standard gravity casting (i.e.
WSHEAR = 2, WALLF = 0.8, FREESFOPT = 1 (or 2)).
Version 2006.0
Page 298
If the different domains are meshed separately, they can be loaded in PreCAST
using the "File/Multiple Meshes" menu.
Version 2006.0
Page 299
Once the meshes are loaded, the set-up of the case in PreCAST should be done as
for conventional meshes, except for the definition of the interfaces (see figure
below). Firstly, all the non-coincident interfaces, where at least one element is
coincident on each side, are automatically listed in the interface list. One should
change the "Type" from EQUIV to NOCOIN (1). Then, for the non-coincident
interfaces which are not listed, one should "create" manually these interfaces, with
the Add button (2). This will open the"Add Interface Pair" window, where the
material numbers of each side can be defined (3). The "Master-Slave" concept is
used in the calculation of the heat transfer across the non-coincident interface. It is
advised to set the casting as the Master and the mold as the Slave (although it does
have only a minor importance on the accuracy of the results). Then, one should
check the "Non-coincident Interface" button (4). Once all the interfaces are
Version 2006.0
Page 300
defined, one can "Apply" the whole selection (5), which will create the
appropriate interfaces.
The "In-plane Tolerance" corresponds to the maximum distance between the two
surfaces in order to have a contact (distance normal to the plane of the interface).
The "Perimeter Tolerance" corresponds to the maximum distance around an
element of the master surface where a node of the slave surface can be found.
Version 2006.0
Page 301
By default, the In-plane and Perimeter tolerances are taken as a fraction of the
smallest edge of the whole mesh. Thus, if the mesh has a quite heterogeneous
mesh size, these tolerances may be too small (and thus, there will be "no contact"
and thus no heat transfer at these non-coincident interfaces). If such a case occurs,
one can change (i.e. increase) these tolerances. One should however be careful not
to use too large tolerances so that nodes beyond the opposite surface will be taken
into account. A good tolerance value should be about half of the mesh size of the
corresponding surface.
Version 2006.0
Page 302
GEOMETRY MANIPULATION
The geometry can be manipulated and can be displayed with different modes,
using the following icons.
Manual rotation of the model (it opens the Rotate window). Otherwise,
the model can be rotated interactively with the mouse at any time (as long as the
Center or Drag icon is not activated). If the model is rotated interactively with the
mouse, while the "Shift" key is pressed in the same time, the model is rotating
along the axis perpendicular to the screen (the mouse should be rotated
horizontally).
Auto-scale (automatic scaling of the model so that it fits into the graphics
window)
Center of the model (the location of the model which is selected will
move to the center of the screen)
Version 2006.0
Page 303
Version 2006.0
Page 304
Version 2006.0
Page 305
Version 2006.0
Page 306
MESH OPTIMIZATION
The speed of a FEM calculation depends upon the structure of the mathematical
matrices which have to be solved. This depends upon the way the FEM mesh is
numbered (i.e. which nodes belong to which element and vice-versa).
It is possible to optimize the FEM mesh numbering in order to minimize the CPU
time. This was done previously in MeshCAST. However, as the addition of
interfaces is creating new nodes, it was necessary to know already in MeshCAST
which interfaces had to be created.
In order to simplify this operation, it is now possible to optimize the mesh in
PreCAST, just before SAVING a case. This will take automatically into account
the created interfaces and will thus guarantee that the mesh is optimum.
As the optimization operation may be rather long for large meshes, it is not done
by default at every save of the model. In order to activate the optimization, one
should set it in the "File" menu :
This optimization will be done before exiting PreCAST (it is not performed with a
"Save" or "Save as" operation, but in this case a warning is displayed upon Exit).
One should note that if an optimized model is loaded again in PreCAST, it is not
necessary to optimize it again, as long as the interface settings are not changed.
In the case of initial conditions EXTRACTion, it is not possible to optimize the
mesh when the extracted temperatures were coming from a non-optimized mesh.
Version 2006.0
Page 307
USER FUNCTIONS
ProCAST allows the use of User Functions. User Functions can be used to define
boundary conditions in a very flexible way, such as combined time, temperature
and/or space dependant heat transfer coefficients.
Three categories of Functions are available in ProCAST :
Time
Local temperature
Local fraction of solid
Local coordinates
Material number
Boundary condition ID
This allows to define conditions as a function of any of the above arguments (e.g.
a time- and space-dependant interface heat transfer coefficient).
In addition, "external functions" allow to get the value of the main fields (e.g.
temperature, fraction of solid, velocities, ...) at any location of the model. This
Version 2006.0
Page 308
In addition, it is possible to get the node number by giving the coordinates, using
the following function :
where xin, yin and zin are the coordinates specified by the user, xout, yout and
zout the coordinates of the closest nodes which corresponds to the returned node
number. domain# is the domain number in which the search node should be found.
Version 2006.0
Page 309
On Windows, the compiler and linker of the Microsoft Visual C++ 6.0 package
has been used to create the executables. An "reduced" version of this compiler and
linker, called "Microsoft Visual C++ 2005 Express Edition" is available for free
on the web (information valid in April 2006). It can be downloaded from the
following link : http://msdn.microsoft.com/vstudio/express/visualc/download/
If one would like to use ProCAST with "Microsoft Visual C++ 2005 Express
Edition", one should "manually" set the following (if the software is installed in
the "C:\Program Files\Microsoft Visual Studio 8\" directory (default location)) :
Add to the PATH environment variable, the location of the "bin" directory of
the Visual tool kit (e.g. "C:\Program Files\Microsoft Visual Studio 8\VC\bin")
Add to (or create) the LIB environment variable, the location of the "lib"
directory of the Visual tool kit (e.g. "C:\Program Files\Microsoft Visual Studio
8\VC\lib")
Add to (or create) the INCLUDE environment variable, the location of the
"include" directory of the Visual tool kit (e.g. "C:\Program Files\Microsoft
Visual Studio 8\VC\include")
The version of April 2006 is containing a bug which requires to copy the
"mspdb80.dll" from the "C:\Program Files\Microsoft Visual Studio
8\Common7\IDE" directory to the "C:\Program Files\Microsoft Visual Studio
8\VC\bin" directory. Otherwise the compiler and linker will not work.
Linux
C compiler version 6
Version 2006.0
Page 310
SUN
Units
ProCAST allows to define the inputs with almost any kind of units. In order to
define which units should be taken into account into the software, two possibilities
are provided to the user.
1. Default units can be specified in an installation file (either the main central
installation or in the local user preference file (see the "customized
installation" section for more details). These units will be used in all the cases
run by the user. This ASCII file is called "UserFunctions_units.dat" and is
located in the "dat/pref" directory.
2. Specific units can be used for a given case. These specific units should be
specified in a text file in the local execution directory. This ASCII file is
called "prefix_units.dat".
In both cases, the same units should be used in all the User Functions of the same
case.
The files "UserFunctions_units.dat" or "prefix_units.dat" have exactly the same
format as follows :
time 1
length 1
Version 2006.0
Page 311
temperature 2
velocity 1
heatflux 1
heattransfercoefficient 1
massflowrate 1
Each type of variable available in user functions are mentioned in the above list.
Then, a unit code is following (as an integer value). The above values correspond
to the default units (set at the installation) used by the user routines of the
ProCAST solver (SI Units and degree Celsius).
One can change these units, with a text editor, using the following nomenclature
(corresponding to the standard unit codes of ProCAST, used in the d.dat file) :
temperature
1 = Kelvin
2 = Celsius
3 = Fahrenheit
length
1
2
3
4
5
=
=
=
=
=
m
cm
mm
ft
in
1
2
3
4
5
6
7
8
8
=
=
=
=
=
=
=
=
=
m/s
cm/s
mm/s
ft/s
in/s
m/min
cm/min
ft/min
in/min
velocity
time
1 = sec
2 = min
heatflux
1
2
3
4
5
6
7
8
=
=
=
=
=
=
=
=
W/m2
cal/cm2/sec
cal/mm2/sec
Btu/ft2/sec
Btu/in2/sec
cal/cm2/min
Btu/ft2/min
Btu/in2/min
heattransfercoefficient
Version 2006.0
Page 312
1
2
3
4
5
6
7
8
=
=
=
=
=
=
=
=
massflowrate
1 =
2 =
3 =
4 =
5 =
6 =
W/m2/K
cal/cm2/C/sec
cal/mm2/C/sec
Btu/ft2/F/sec
Btu/in2/F/sec
cal/cm2/C/min
Btu/ft2/F/min
Btu/in2/F/min
Kg/sec
g/sec
lb/sec
Kg/min
g/min
lb/min
Version 2006.0
Page 313
#include <stdio.h>
#include <stdlib.h>
#define real double
#ifdef WIN32
#define EXPORT _declspec(dllexport)
EXPORT real func_convehtransfer(char*, int, real, real, real, real, real,
real, int);
#else
real func_convehtransfer(char*, int, real, real, real, real, real, real,
int);
#endif
extern
extern
extern
extern
extern
extern
/*
*
*/
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real*,real*,real* );
real func_convehtransfer(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* boundary condition ID number */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Version 2006.0
Page 314
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real*,real*,real* );
real func_texternal(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* boundary condition ID number */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Emissivity Function
This function is called : emissivity.c
Version 2006.0
Page 315
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real*,real*,real* );
real func_emissivity(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* boundary condition ID number */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Version 2006.0
Page 316
extern
extern
extern
extern
extern
extern
/*
*
*/
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real*,real*,real* );
real func_heatflux(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* boundary condition ID number */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Version 2006.0
real
real
real
real
usertemp1(int);
userfs1(int);
uservx1(int);
uservy1(int);
Page 317
real func_interhtransfer(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* boundary ID number */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
return 0;
Version 2006.0
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real* ,real* , real* );
Page 318
real func_masssourceflowrate(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numMat)
/* number of the domain */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real* ,real* , real* );
real func_xmasssource(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real time,
/* current time */
int
numMat)
/* number of the domain */
{
/* ------------- Do not change anything above this line ------------- *
Version 2006.0
Page 319
------------- */
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real* ,real* , real* );
real func_xtranslation(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real time,
/* current time */
int
numMat)
/* number of the domain */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Version 2006.0
Page 320
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real* ,real* , real* );
real func_vximposed(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* number of boundary condition */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
Version 2006.0
Page 321
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real* ,real* , real* );
real func_vxsolidtransport(
char prefix[],
/* case name */
int
dimension,
/* 2 = 2D ; 3 = 3D */
real temp,
/* current temperature */
real fs,
/* current fraction of solid */
real time,
/* current time */
real x_coor,
/* local coordinates: x */
real y_coor,
/* local coordinates: y */
real z_coor,
/* local coordinates: z */
int
numBC)
/* number of boundary condition */
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
External Function
This function is called : externalcompute.c
Version 2006.0
Page 322
This function is called at the start of the calculation, at the start of each timestep,
at the end of each timestep and at the end of the calculation.
#include <stdio.h>
#include <stdlib.h>
#define real double
#ifdef WIN32
#define EXPORT _declspec(dllexport)
EXPORT void func_externalcompute(char*,int,int,real);
#else
void func_externalcompute(char*,int,int,real);
#endif
extern
extern
extern
extern
extern
extern
/*
*
*
*
*
*
*
*
*/
real usertemp1(int);
real userfs1(int);
real uservx1(int);
real uservy1(int);
real uservz1(int);
int nodNum (real,real,real,int,real*,real*,real* );
void func_externalcompute(
char prefix[],
/* case name */
int
loop,
/* loop value : 0/1/2/3 */
int
timestep,
/* current timestep
*/
real time)
/* current time
*/
{
/* ------------- Do not change anything above this line ------------- *
* ------------- Program your function below this line ------------- */
}
Version 2006.0
Page 323
Run DataCAST
Run ProCAST
DataCAST
DataCAST converts the input data stored in ASCII in the prefixd.dat file into
binary data ready for the solver. DataCAST also cleans any previously existing
result files and prepare the new ones. This means that when DataCAST is run, all
results will be erased.
Some error checking is also performed in DataCAST.
DataCAST is called with the DataCAST button of the Software manager. When
DataCAST is finished, one should press Return to close the corresponding
command window (see the Module calls section of the Software Manager chapter
for more details).
See the "DataCAST -u" option at the end of this section.
ProCAST
ProCAST is running the solver. No message is displayed by ProCAST, except if
the solver is stopping for an unexpected reason.
ProCAST is called with the ProCAST button of the Software manager. When
ProCAST is finished, one should press Return to close the corresponding
command window (see the Module calls section of the Software Manager chapter
for more details).
Status
Version 2006.0
Page 324
During the calculation, the status of the model can be monitored. To do so, use the
"Status" button of the Manager (see the Module calls section of the Software
Manager chapter for more details). The prefix of the case to be monitored should
be present in the "Case" field and when the Status button is pressed, a window
with the calculation status is displayed :
The status information is stored in the software installation directory (in the
dat/stat directory), under the name prefix.stat. This means that the status of any
calculation is available at any time, by any user who is using the same software
installation. As a consequence, if two calculations with the same prefix are run in
the same time, there will be a conflict and the status will be once the one of
calculation A and once of calculation B. Such a conflict however has no influence
at all on the calculations themselves.
Batch calculations
It is possible to run several calculations sequentially in batch mode. To do so, a
simple script should be written and launch in a Command window, as shown in
the example hereafter (for Windows) :
cd calculations_A
cd case_1
%ProCAST20060%\bin\procast case_1
cd ..
cd case_2
%ProCAST20060%\bin\procast case_2
cd ..
cd ..
cd calculations_B
cd case_3
%ProCAST20060%\bin\procast case_3
cd ..
cd ..
Version 2006.0
Page 325
DataCAST -u option
It is possible to run a calculation until a given step (e.g. step 500) with given
settings and then to continue the calculation with different settings (e.g. if one
would like to test different heat transfer coefficients from step 500).
To do so, run the calculation as usual until step 500. Then, in order to prevent any
error, copy the results in a safe place (as if a wrong operation is performed with
DataCAST, all the results files will be erased).
Load the d.dat file in PreCAST and make the desired changes (e.g. change a heat
transfer coefficient). Change the Run parameter INILEV to 500 and set DT = 0 (so
that the last stored timestep will be used as the first one of the restart) and exit
PreCAST. Then, in order to update the calculation binary files, DataCAST has to
be run. However, in order to prevent DataCAST to erase the existing result files, it
is necessary to activate the -u option.
To use the "-u" option of DataCAST, the corresponding check box should be
activated (see figure above). One could also launch it from a Command window
with the following syntax :
datacast
prefix
-u
Then, the solver ProCAST can be launched and the calculation will be continued
from the step 500, with the new settings. Please note that it is not necessary to run
DataCAST if only the Run Parameters are changed (i.e. only the content of the
p.dat file is changed).
Version 2006.0
Page 326
TROUBLESHOOTING
During the execution of the solver, information are stored in the prefixp.out file. In
case of problem, it may be useful to check the convergence of the solver in this
file.
At each timestep, a number of information are printed in the p.out file. The most
important ones (hereafter in the case of a thermal only calculation) are the step
number (1), the timestep (2), the fraction which has already solidified (3) and the
iterations (4).
If a calculation seems not to run well, it can be useful to look at the evolution of
the timesteps, during the course of the calculation. If the timestep is significantly
reduced, it means that the convergence of the calculation is poor. In addition, the
number of iterations per timesteps (i.e. the number of line shown in (4) in the
figure below, "T" indicating the resolution of the Thermal model) is a good
indication of the quality of the convergence. The least number of iterations, the
better the convergence.
In the case of a mold filling calculation, the prefixp.out file is looking like the
following figure. One can check that the "free surface model" is well activated (1),
that the model is filling with the filling percentage (2). The iterations (3) for the
Thermal model ("T"), for the pressure solver ("P") and for the three components
of the velocity ("U", "V" and "W") are shown. When more than 5-10 iterations are
necessary, one can anticipate a convergence problem. Finally, the memory used
by the calculation (for the current timestep) is shown (4).
Version 2006.0
Page 327
One of the Run Parameter is COURANT. This parameter is limiting the timestep
in order to prevent the free surface to advance more than N elements during one
timestep. When the mesh has a large number of elements and when the elements
are small, this limitation may be quite drastic for small COURANT values. When
this limit is reached, the following message appears in the prefixp.out file :
Version 2006.0
Page 328
The Run Parameter LVSURF allows to switch off the filling calculation when the
casting fill fraction has reached the LVSURF value. A message is printed in the
prefixp.out file to acknowledge this switch.
The fluid flow model will still be solved until the next stored timestep and then it
will be switched off (see figure below). The calculation will then continue with a
thermal only calculation (and stress if activated).
Version 2006.0
Page 329
When the calculation is not converging, the following message is printed in the
prefixp.out file and the calculation of the given timestep is restarted with a
reduced timestep. This can happen a few times during the calculation and one
should not worry about it (typically, if the inlet velocity is suddenly increased, the
timestep may be too large and it should be reduced through this mechanism).
However, if the calculation does not converge several consecutive times, it may
indicate a problem in the problem set-up or in the input data (non relevant material
properties for instance).
Version 2006.0
Page 330
RESULTS VIEWING
INTRODUCTION
Once the calculation is performed, the results can be viewed with the ProCAST
post-processor, called ViewCAST. ViewCAST can be called from the Manager as
follows :
Depending upon the type of calculation which was performed, the following
entities (or variables) can be viewed (raw results) :
Version 2006.0
Scalar fields :
Temperature
Fraction of solid
Page 331
Vector fields :
Velocities
Heat fluxes
This chapter is describing how these results can be viewed. In the Result analysis
chapter, the display of more entities, such as porosity, solidification time,
secondary arm spacings, etc... will be described. The export of the calculated
results, to be used in other softwares are described in the Results Exports chapter.
Once ViewCAST is started, if the prefix of the case is well defined, the case is
automatically loaded in the post-processor and the initial temperature field is
displayed :
The background of the screen can be switched from black to white in the
"Parameters/Reverse Video" menu :
Version 2006.0
Page 332
Firstly, the case should be opened in the File menu (if it is not opened
automatically when the ViewCAST is launched from the Manager). Then, the
field (scalar) to be displayed (e.g. Temperature) is selected in the Contour menu.
In the case of vectors, the desired field is selected in the Vector menu - see the
Field selection section for more details.
The times (or steps) to be displayed are selected in the Steps menu, whereas all
the parameters to configure the display (e.g. symmetry, cut-off values, type of
display, ...) are defined in the Parameters menu - see the Display parameters
section for more details. Then, the Picture menu allows to select the type of
Version 2006.0
Page 333
display (i.e. snapshot, slice, scan slices or cut-off view) - see the Display types
section for more details.
Once these selections are made, the display is activated with the Tape player,
either as individual pictures or frame, or as animations (see the Tape Player
section for more details).
The X-Y Plots menu allows to draw curves (time evolutions) of the different
fields (see the Curves section).
The Actions menu allows to calculate some parameters from the raw results (such
as SDAS, isochrons, Niyama, etc...) - see the Results analysis chapter for more
details.
Please note that the Actions and X-Y Plots menus are replacing the former
module PostCAST (version 400 and before), as well as the File/Export menu.
The Geometry manipulation, as well as the different viewing modes is described
in the Geometry manipulation section.
To File/Exit menu allows to quit the post-processing session. Upon Exit, the user
is asked whether he would like to keep the information about the current
orientation of the model as well as the definition of the material display (i.e. which
materials are active and in which mode) for the next launch of the post-processor.
GIF/AVI capture
All the pictures which are displayed in ViewCAST can be captured as GIF/AVI
files or GIF/AVI animations. To do so, the
opens the following window
Version 2006.0
Page 334
The file name can be specified and one can browse in the desired directory (by
default it is stored in the working directory and the file name corresponds to the
case prefix) - the suffix (.gif or .mpg) is optional. One has to select between a Still
picture or an Animated picture and to select between the GIF or the AVI format. It
is also possible to activate the Reverse Video (with the "Reverse Video" check
box) in order to have a white background.
Then the picture is stored when the "Apply" button is pressed. For an animation, it
will start to store the animation when the Forward button of the Tape player is
activated. The "Cancel" button allows to exit without any action.
If the filename is already existing, the following prompt window is shown :
One has the choice between "Append" (i.e. to append or add the picture to the
previous one(s)), to "Overwrite" (i.e. to replace the existing file by this one) or
"Cancel" to abort the capture operation.
The AVI files are compressed using a "Motion jpeg" compressor. It has been
observed that with old versions of Powerpoint, such "Motion jpeg" files were not
recognized, unless they are renamed as prefix.mpg (instead of prefix.avi).
Version 2006.0
Page 335
When an animation is stored in the GIF format, it can be interrupted at any time
(and the GIF animation will be stored until the interruption). However, it is not
possible to interrupt an animation when an AVI file is created. If the animation is
interrupted before the end, the AVI file will not be created. Thus, one has to
specify the starting and ending steps before the creation of an AVI animation.
Version 2006.0
Page 336
FIELD SELECTION
The entities to be displayed have to be selected in either the Contour or the
Vector menu.
Contour corresponds to scalar variables, such as Temperature, Fraction of solid,
Heat flux intensity, Pressure, Velocity magnitude, Principal stress, Xdisplacement, etc...
Vector corresponds to vector variables, such as Velocities or Heat fluxes.
It is possible to display either one Contour or one Vector field at a time, or to
superimpose one Contour field, with one Vector field.
Contour
The Contour menu is divided according to the different models which can be
activated in the ProCAST solvers.
The different "Contour" fields which are available are shown hereafter
One should note that the "Solidification Time" corresponds to elapsed time
between the liquidus and the solidus temperatures.
The Shrinkage Porosity is only calculated in the Run Parameter POROS is set to a
non-zero value.
Version 2006.0
Page 337
As the stresses (i.e. the stress tensor) are calculated at Gauss points in the solver, it
is necessary to compute in the post-processing the different stress entities that can
be viewed (e.g. Effective stress, Principal Stress, Sigma X, ...). As these
calculations can be rather long, it is possible to compute all the desired one in one
shot, with the "Prepare Stress Results" option. In this case, the following panel
appears :
Version 2006.0
Page 338
The stress result files which are present and up-to-date are shown in Green. The
ones which are not present are shown in Red. If there are files which are present
but not up-to-date (i.e. these files are older than the latest computed stress results
by the solver), they appear in Orange. In this panel, it is possible to select all the
desired stress results to be computed and to press "Apply" to start the
computation. The results which are in green will not be recomputed (unless the
"Overwrite Existing Files" is checked).
Remarks on the Stress results
The "Average Normal Stress" corresponds to the average of the normal stresses
along the X, Y and Z planes. This does not correspond to a stress normal to an
external surface (thus it is possible that the Average Normal Stress is not zero
on a casting surface).
The "Displacement -> Plane" option corresponds to the difference (between the
initial timestep and the selected timestep) of distance between a User specified
plane and a given point (the distance is taken normal to the plane).
When Stress results are to be viewed from a CD-ROM (on Windows only), the
calculated stress results (e.g. Principal Stress 1) can not be stored on the CD (as
it is Read-only). Thus, when the directory where the case is located is in Readonly mode, automatically, the stress results are created and stored in a local file
Version 2006.0
Page 339
Vector
Vector results, such as Heat Flux and Fluid velocity can be selected in the Vector
menu.
It is possible to view both vectors and contours, if desired. In this case, the vectors
will be displayed in white (respectively black).
If vectors only are shown, the color of the vector can be set in the Parameters
menu (see the Display Parameters section for more details) to the vector
magnitude, to the temperature or the pressure. In order to shown the vectors only,
the Contour/None option should be selected.
Version 2006.0
Page 340
DISPLAY TYPES
The type of display can be selected either with the following icons
Version 2006.0
Page 341
Examples of Snapshots (during filling and after), of slice and CutOff (X-Ray) view
Each Display type, except Snapshot, has ad-hoc settings (available in the
Parameters menu), as described hereafter (Slice Data, Scan Data and Scan Data).
Version 2006.0
Page 342
Slice definition
For Slice Data definition (see figure below) :
1. Select "Slice Data" in the Parameters menu
2. Select "Add" and then XYZ Plane (for orthogonal slices). This will open the
slice selection window (3)
3. Select the X, Y or Z plane and then define the location of the plane with the
slider
4. The plane is shown on the model
Version 2006.0
Page 343
5. Press "Apply", the corresponding selection will appear in the list (6) and the
slice will be displayed.
It is possible to use the Tape Player at this stage to see this slice over the
timesteps, either as single pictures or as animations.
As shown in the figure below, several slices can be selected. Moreover, the slice
characteristics can be stored in a file (file named "prefix.clip" in the local
directory) with the "Store" button and can be retrieved in a later session with the
"Read" button. The "Delete" button allows to delete the selected slice. The Show
"Yes"/"No" toggle allows to activate or deactivate slices, without deleting them.
Version 2006.0
Page 344
In the following case, two slices are activated, as shown in the frame above.
When one slice only is activated, one can draw this slice with or without the
Background (using the "Display Background" button.
Version 2006.0
Page 345
With "Display Background : None", only the slice is shown (upper right). With
"Display Background -", the slice is shown together with the background (upper
left). With the "Display Background +", the foreground is shown, up to the slice
(lower left). When the model is rotated (for instance around a vertical axis in the
above example), the background becomes the foreground and vice-versa. This
option is not available with more than one slice.
With the "Add->" button, it is also possible to select "Any Plane" :
In this case, the following window opens. One should define the slice plane by the
coordinates of three points. To select interactively the points on the mesh, the "Get
Co-ord" button can be used (to select the point which is highlighted in red).
Once the three points are selected, it is possible to move this plane along its
normal, with a given Offset.
Version 2006.0
Page 346
It is also possible to rotate this plane with the "Rotate" button. In this case, the
following window allows to define the rotation axis, as well as the angle of
rotation.
Finally, the "Copy Plane" is used to copy a plane which was already defined in
order to modify its characteristics. This is especially useful when one would like
to define several parallel planes which are not in the X, Y or Z directions. One
will define one plane with the standard Anyplane definition procedure and then,
one will copy this plane and offset it to the appropriate value. This can be repeated
for the other parallel planes.
Scan definition
The definition of the Scan parameters is done in the Parameters/Scan Data
menu, which opens the following window :
Version 2006.0
Page 347
The Scan direction is selected, as well as the number of planes (or slices) which
will be shown. The largest dimension of the model in the selected direction is
divided by the number of planes.
Then, the display and the scan through the planes is activated with the Tape
Player.
Beside this slice mode, it is possible at any time to slice interactively the model,
using the "Scan" slider (see figure below).
Cut-off definition
Version 2006.0
Page 348
The Cut-off (or X-Ray) parameters are defined in the Parameters/CutOff Data
menu, which opens the following window :
The Cut-off parameters are defined by two values (e.g. two temperatures). Then
the user has to select among five different display possibilities :
Above Red : only the zones which are above the upper value (in red) are
shown
Below Blue : only the zones which are below the lower value (in blue) are
shown
Inbetween Bounds : only the zones which are in between the upper and the
lower values are shown
Outside Bounds : only the zones which are below the lower value and above
the upper value are shown
Isosurfaces : the isosurfaces corresponding to the two values are shown
Version 2006.0
Page 349
Version 2006.0
Page 350
DISPLAY PARAMETERS
Firstly, the timesteps to be viewed should be defined in the "Steps" menu, which
opens the following window :
The user can choose between Steps and Time increments. In blue, the available
minimum and maximum values are shown. The time values are always indicated
in seconds. If the selected Step increment is smaller than the available stored one,
it will take the first available step for the display.
If ViewCAST is launched during a calculation, the "Update" button allows to
"refresh" the available steps and to view what has been calculated between the
launch of ViewCAST and the current instant.
One can also set "ViewCAST" such that the update is made in a "continuous"
mode during the calculation (this is useful when ViewCAST is launched while the
ProCAST calculation is still running). To do so, the "Setup" button (of the above
screen) should be pressed and the Update interval can be specified (please note
that if it is intended to leave the ViewCAST session opened for a few hours, it is
advisable to set an Update interval not too small, as this will load the CPU) :
Version 2006.0
Page 351
Reverse Video
By default, the background of the screen is black. It can be reversed to white with
the Reverse Video option.
Free Surface
To visualize a mold filling four options are available to view the free surface (i.e.
the surface between the liquid metal and the air) :
Version 2006.0
Page 352
The option "Free Surface (Foreground)" is used to view the free surface even
when it is hidden by the liquid or the rest of the casting. One can see in the upper
figure hereafter (Back face) the free surface in the ingate. If we rotate the casting
(lower figures), the free surface is behind. If the Free Surface is ON, we do not
see it. If the Free Surface is set as "Foreground", one can see it although it is
behind. Of course, one can also see ONLY the free surface.
Version 2006.0
Page 353
The "Foreground" option can be quite useful to view entrapped gas pockets
(which are fully surrounded by liquid).
Enclosure
In the case of a radiation problem with an enclosure (this is valid only for surface
enclosures as solid ones are considered as materials) the display mode of the
enclosure can be selected among six choices : Invisible, Wireframe, Hidden,
Solid, Translucent or Shade :
Version 2006.0
Page 354
Feature Angle
When the geometry is displayed in Wireframe, one can tune the "density" of lines.
In the following example, a Feature Angle of 10 and 60 degrees has been used.
Version 2006.0
Page 355
Vector settings
When vectors are selected, it is possible to color them according to the vector
magnitude or according to Temperature or Pressure. By default, the color is
White. When the "Parameters/Vector Settings" menu is called, the following
panels opens.
Version 2006.0
Page 356
One can choose between an automatic vector length or a manual vector length. It
is also possible to define a Unique vector length (i.e. all the vectors will be
displayed with the same length, regardless of the vector magnitude (except if it is
zero).
When vectors are shown together with a Contour Plot (i.e. a scalar variable such
as Temperature), it is not possible to color the vectors.
The vectors can be plotted with or without the arrow at the tip of the vector.
Displacement Mag.
In the case of Stress calculations, the display of the deformations can be
artificially increased in order to better view them. The real displacements are
multiplied by the "Displacement Magnitude" value.
Version 2006.0
Page 357
Titles
Titles (text), arrows and cirlces can be added to the graphics, using the
Parameters/Titles menu, which opens the following window :
Then, one can "Add ->" either a String (text), an Arrow or a Circle. When a String
is selected, the text has to be typed in the corresponding field and then the text
should be placed with the mouse. When an Arrow is selected, it should be located
with the mouse. The first click should be at the starting point of the line and the
second click should correspond to the end of the arrow. For circle, the two
opposite corners of the square enclosing the desired circle should be defined. To
delete an item, click on the desired one (in the left grey zone), which will highlight
it in red and then press the Delete button.
All the information of this panel can be Stored in an ASCII file and retrieved (with
the Read button) for later use.
Version 2006.0
Page 358
Symmetry
For cases which have been calculated with symmetry, it is possible to
"reconstruct" the full geometry, using the symmetry menu :
The figure below illustrate a mirror symmetry. For more details about the
definition of the symmetries, please refer to the Radiation section of the Preprocessing chapter.
Version 2006.0
Page 359
Display Foam
In the case of a lost foam calculation, one can display or not the foam with this
toggle menu.
Display Pipe
In case of porosity, if piping or macroshrinkage occurs, the corresponding
volumes will be shown as empty. However, in the visualization of the "Shrinkage
porosity", it may be useful to view together the porosity and the pipe shrinkage.
To do so, the "Display Pipe" OFF mode will allow to see the empty regions as
"Shrinkage porosity" with a value of 1 (i.e. 100% void). This mode is only
applicable for the Contour "Shrinkage porosity" and will not affect the viewing of
the other fields, such as the Temperature, Fraction of solid, ...
Particle tracing
The particle tracing allows to follow streamlines during the filling. The particle
traces are calculated during the filling calculation (as it is needed to have access to
all the timesteps, which is not the case during the post-processing). To do so,
Version 2006.0
Page 360
"virtual particles" are launched at each inlet node (defined by a velocity BC, an
inlet pressure BC or an Inlet BC) at given intervals specified by PFREQ (see
"Flow Run Parameters"). This means that it is not possible to compute particle
traces in the case of tilt casting or casting with a filled reservoir of material (as
there are no inlet boundary conditions where particles could be launched).
Moreover, if the inlet is applied to a few nodes only, the number of traces is not
very significant. This will be improved in future versions with more sophisticated
means to launch particles.
To view the Particle traces, activate "Particle Tracking" in the "Parameters" menu.
The traces can be viewed either alone or together with any contour or vectors. One
should note that the traces are always displayed in the foreground of a contour plot
(no hidden view of the traces).
In order to allow the viewing of successive batches of particles, the "tail" of the
traces are erased, as the traces are progressing, leaving the place to new traces.
Version 2006.0
Page 361
Depending upon the geometry and the mesh size, it is possible that one or two
particles are going out of the model. Please note that this does not harm the other
results.
Tilt
In case of a tilt pouring calculation, it is possible to disable the viewing of the
model rotation. By default, the display of the rotation is active.
Display Undeformed
Version 2006.0
Page 362
Version 2006.0
Page 363
Scale
The scale can be modified by a direct click on the scale itself (1). Then, the Scale
settings panel opens (2).
Version 2006.0
Page 364
Version 2006.0
Page 365
For the temperature scale, if there is a material where the solidification range is
defined (i.e. a casting material), there is also the possibility to select automatically
the Solidus-Liquidus interval (Tsol-Tliq).
In the same line, when the Temperature is displayed, the Solidus and Liquidus
temperatures are indicated on the color scale :
Version 2006.0
Page 366
Version 2006.0
Page 367
One should note that the scale and the color map is stored in memory for each
field (e.g. temperature, fraction of solid, velocity magnitude, ...). Thus, if the scale
is changed for the display of temperature and then fraction of solid is shown, when
the temperature field is displayed again, the modified scale is kept.
Version 2006.0
Page 368
TAPE PLAYER
To activate the display, the Tape Player must be used. This allows to display
either individual pictures or animations.
As for a Tape recorder, the Tape Player has the following functions.
Displays the first define time or time-step (Goto first)
Plays the animation backwards (Rewind)
Displays the previous step (Step back)
Pause (stops the animation)
Display the next step (Step forward)
Plays the animation forwards (forward)
Display the last define time or time-step (Goto end)
The definition of the first and last times or time-steps, as well as the time or timestep increment between two frames is made in the Steps menu (see the Display
parameters section for more details).
Version 2006.0
Page 369
CURVES
Curves can be viewed in ViewCAST, using the X-Y Plots menu. This allows to
visualize time evolutions of the calculated fields, such as Temperature, pressure,
velocity, etc...
Please note that these capabilities are replacing the former PostCAST module.
Interval
When Interval is selected, the user has to specify the Nodal interval. This means
that curves will be selected every N nodes (1, 501, 1001, 1501, ...). Only the nodes
of the materials which are active (as Solid) will be selected.
In addition, the user can select the Units which will be used for the display of the
curves. When the Apply button is pressed, the curves are read in the files and
displayed.
Version 2006.0
Page 370
Nodes
When "Nodes" is selected, the following node selection window appears :
Version 2006.0
Page 371
button.
In addition, the user can select the Units which will be used for the display of the
curves. When the Apply button is pressed, the curves are read in the files and
displayed.
The node numbers (as well as the defined scale) can be stored (Export) and
retrieved (Import) in a file. To do so, the "File" button should be pressed and the
following window is opening :
In the case of an Export, the file name should be specified and it will be stored in
the working directory. For Import, one can "Browse" the desired file.
The format of the Export file is the following :
-1 0 3.56
10
30
Version 2006.0
10. 1550. 0
Page 372
100
234
The first line contains the scale. As the scale is optional, the line should start by a
"-1" and should be followed by xmin, xmax, ymin, ymax. Then, there is a "0" or a
"1" to specify whether the Y-scale is defined automatically of not. A value of "0"
means that the specified ymin and ymax are taken into account, whereas a value of
"1" means that the Y-scale is set automatically.
Then, the list of the nodes should be specified.
Finally, a file named "prefix.tt" is created when the curves are displayed, with the
x and y coordinates of the displayed curves. These data can be then either viewed
together with other results (see the External section below) or exported to
spreadsheets (e.g. Excel).
External
When curves are selected with the Nodes mode (see above), the displayed curves
are stored in the same time in a text file named "prefix.tt". The fist line of the file
(see below) contains the number of time-steps (132), the number of stored curved
(3) and the corresponding node numbers (567, 9587 and 23129). Then, each line
contains the time and the field values for each curve.
Any file which as the above format can be loaded and displayed in ViewCAST
with this External option :
Version 2006.0
Page 373
The desired files should be selected with the Browse. Please note that if the format
should be as described above, the name of the file is free.
Once this selection is accepted with the Apply button, the selected curves will be
displayed in the same time as the ones selected with the Nodes option.
Version 2006.0
Page 374
Version 2006.0
Page 375
GEOMETRY MANIPULATION
The geometry can be manipulated with the following icons :
Interactive zoom (zoom up when the mouse is move towards the bottom of
the screen)
Rotation
When the Rotate icons is activated, the following Rotate window is opened :
Version 2006.0
Page 376
With the X, Y and Z sliders, the rotation can be prepared. Exact angles can be
manually entered in the fields on the right. The rotation will be applied only when
the "Rotate" button is pressed. If the "Rotate" button is pressed again, a new
rotation (with the same values) is applied again. The "Reset" button resets all the
sliders to zero. By default, rotation are with positive angles. If one clicks on the
"+" sign, it becomes negative (for negative rotation).
, in the Y-direction
and n the
Z-direction
can be activated, as well as a standard isometric view
.
When the left mouse key is used, the model is oriented with the X, respectively Y
and Z, axis perpendicular to the screen, pointing outwards, whereas the opposite
orientation (pointing inwards) can be obtained with a right mouse key click.
Finally, the X, Y and Z keys can be used to rotate the model by 10 around the x,
y and z axis respectively. When the <ctrl> key is pressed in the same time, the
rotation direction is reversed. When the <shift> key is pressed in the same time,
the rotation is by 30 instead of 10.
Interactive Rotation
At any time, the model can be rotated interactively with the mouse. The left
mouse key should be pressed and the model is rotated when the mouse is moved.
The center of rotation corresponds to the center of the screen (green arrow).
If the "Shift" key is pressed during the rotation, the model will rotate around an
axis which is perpendicular to the screen and located in the center of the screen
(green cross). The rotation is activated upon the horizontal motion of the mouse
(while the left mouse key is pressed).
Zoom
Version 2006.0
Page 377
Stored Views
It is possible to store 6 different views in the "Parameters/Store View" menu. The
views are containing the orientation of the model as well as the material selection.
When a small arrow is following the View-X, it means that a view is stored
already. To store a view in a View-X which is available, just click on the
Version 2006.0
Page 378
Material selection
The Material selection icon
opens the Material selection window, which
allows to define which material are display as well as the type of display.
The list of the all the available materials is displayed. On the right, the colored
columns allow to specify the display attribute of each material, according to the
following nomenclature :
Version 2006.0
Page 379
TR : Transparent material
For each material, one should click in the desired column and a "X" will be
displayed. To define all the materials with the same selection, one could click on
the label (e.g. click on "IN" will make all the materials invisible).
Finally, when the apply button is selected, the display will be updated with the
corresponding selections.
Explode materials
In the Materials selection window, there is the "Explode Materials" button, which
opens the following window.
For each material, it is possible to define an offset for its display. If some
materials have no interface between them (EQUIV), they will be considered in the
same group and they can not be exploded separately.
Version 2006.0
Page 380
When the Apply button is pressed, the corresponding offsets are applied, leading
to a picture as follows :
Version 2006.0
Page 381
One can enter a node or an element number. By pressing the "Node" button, all
the elements which are connected to the specified node are displayed :
If the "Element" button is pressed, the nodes which belong to this element are
listed :
Pick option
Version 2006.0
Page 382
The icon
allows to pick a location on the geometry and display the node
number, the node coordinates as well as the field value (e.g. Temperature, velocity
magnitude, ...).
One should first click on the icon (which becomes outlined in red) and then click
on any point on the surface of the model. The information are displayed in a
yellow frame.
In the case of Stress models (i.e. deformed models), the displayed coordinates will
depend upon the "Displacement Magnitude" which is set. In order to have the real
deformed coordinate, one should set a Displacement Magnitude to 1 (with a value
of 0, it will correspond to the initial location of the mesh at the beginning of the
calculation).
The Pick option is also available in PreCAST :
Version 2006.0
Page 383
ViewCAST exit
Upon Exit of ViewCAST, a message is asking to the user whether he would like
to keep the Last View. If Yes is answered, a file named "prefix.lv" will be stored
in the working directory. This file is containing the model orientation, zoom and
position, as well as the material selections.
Version 2006.0
Page 384
RESULTS ANALYSIS
CRITERION FUNCTIONS
ViewCAST allows to process the results which are calculated in the solver in
order to built criterion functions or metallurgical results (Remark : this
corresponds to the functionalities which were present in PostCAST in previous
releases).
The Actions menu give access to these functionalities.
R G L calculation
RGL is a generic menu which allows to calculate the solidification rate "R", the
cooling rate "L" and the gradients "G", as well as any combination of those.
The solidification rate corresponds to the velocity of a given isotherm (e.g. the
liquidus isotherm). Thus, the user has to specify at which temperature he would
like to calculate the solidification rate (see "R,G Temp" in the figure below). One
can notice that it is allowed to calculate the velocity of any isotherm for other
purposes than a solidification rate.
The cooling rate is calculated as a linear interpolation between two temperatures
(e.g. between the liquidus temperature and 10C above). Thus, the user has to
specify these two temperatures (see "L Upper Temp" and "L Lower Temp" in the
figure below).
The gradient (magnitude) is also calculated at a given temperature, which should
be specified by the user (and which corresponds to the same temperature as the
one used for the solidification rate) - see "R,G Temp" in the figure below.
When the RGL menu is activated, the following window opens :
Version 2006.0
Page 385
Firstly, the user has to choose between two methods to calculate the solidification
rate. In Method 1, when each node reaches the specified temperature, a point is
located along the temperature gradient some distance away and the time that it
takes for the isotherm to reach that point is determined. R is then calculated as that
distance divided by the difference in time.
In Method 2, R is calculated as the cooling rate divided by the temperature
gradient. Method 1 takes longer to compute, but it does not depend on the cooling
rate. The results obtained by Method 2 are affected by the temperature levels used
to calculate L.
Four options are available for the calculation of the thermal gradient. One can
either compute the "Total" gradient (i.e. all components of the gradient) or in
either the X, Y or Z directions. Please note that the gradient is calculated at each
node, when the "R,G Temp" is reached. This means that the gradient corresponds
to a different time at each node.
Version 2006.0
Page 386
Then, the user has the possibility to combine the R, G and L variables together to
obtain a criterion function (called Mapping factor). To do so, the coefficients a, b,
c and d should be specified
= 1.0
= 0.0
= 1.0
= -0.5
Finally, the units (length and temperature) used in the RGL calculation, as well as
the output format (*m.unf, Patran or I-DEAS) should be defined.
When the *m.unf output format is used, the Mapping Factors (M), the cooling
rates (L), the temperature gradient (G) and the Isotherm velocity (R) can be
selected for visualization in the Contour Menu.
Version 2006.0
Page 387
Feeding length
The Feeding length provides the capability to calculate the distance between the
solidus and some user defined critical temperature which represents some fraction
solid beyond which feeding is impaired. This distance is then compared with a
"critical feeding length," which is a simple linear function of the hydrostatic
pressure. If the feeding distance exceeds the critical length, then porosity would be
likely. The Feeding length was designed for the cases of directional solidification
(such as DS or SX casting). Thus, it is not guaranteed that the Feeding length will
work well in other cases.
The critical Feeding Length is calculated as :
Version 2006.0
Page 388
Please note that as the hydrostatic pressure is used in the Feeding length
calculation, the gravity vector should be defined in PreCAST (but it is not
necessary to run fluid flow). If a thermal only calculation was run and the gravity
was not defined in PreCAST before the run of the calculation, it is still possible to
load the case again in PreCAST, to set the gravity vector, to save the case and to
exit. Then, "DataCAST -u" should be used (see the "Solver" chapter for more
details on DataCAST -u, as well as the special care which should be taken in the
use of this capability in order not to scratch the results).
Then, the units used in the calculation, as well as the output format have to be
selected. Once the Apply button is pressed, the Feeding length is calculated. It can
then be viewed in the Contour Menu.
Isochrons
An Isochron is corresponding to the elapsed time from the beginning of the
calculation until a specific temperature is reached. The Isochrons can be
calculated in the Actions / Isochron menu.
Version 2006.0
Page 389
In order to select the desired temperatures, two possibilities are available : "Semi
Auto" and "Specify Temps".
In Semi Auto, the user can select 15 temperatures (to calculate 15 isochrons plots)
by specifying the starting temperature, as well as a "delta" increment between the
temperatures (then the units and the output format should also be specified).
With the "Specify Temps" mode, the user can select any temperature (up to 15
isochrons). Again, the output format and the units should be selected.
Version 2006.0
Page 390
Once the Apply button is pressed, the Isochrons are calculated. They can then be
viewed in the Contour Menu.
Alpha case
Alpha case corresponds to the thickness of alpha case for the surface nodes
adjacent to the ceramic shell in Titanium alloy investment castings.
The necessary parameters, as well as the output format should be specified in the
following window.
Once the Apply button is pressed, the Alpha case is calculated. It can then be
viewed in the Contour Menu.
Version 2006.0
Page 391
The secondary dendrite arm spacing (also called SDAS or 2 ) can be calculated
according to the following equations, where tf is the local solidification time and
M a constant which depends upon the alloy properties.
Version 2006.0
Page 392
Tstart and Tend are the temperatures which are used to compute the local
solidification time. Usually Tstart corresponds to the liquidus temperature,
whereas Tend should be the temperature just above (e.g. 1 degree above) the first
eutectic transformation. The Exponent corresponds to the power 1/3 in the above
equation (it is thus recommended to set the Exponent to 0.33333) and M is the
coarsening constant according to the above equation (which is alloy dependant).
The units for M should be [microns3/sec].
For an Al-7%Si-0.3%Mg (A356), the coarsening constant M is equal to 680
[microns3/sec].
For an Al-2%Cu, the coarsening constant M is equal to 1400 [microns3/sec].
For an Fe-0.09%C (Low carbon steel), the coarsening constant M is equal to
29250 [microns3/sec].
For an Fe-0.6%C (Cast iron), the coarsening constant M is equal to 6050
[microns3/sec].
For an Fe-10%Ni (Stainless steel), the coarsening constant M is equal to 20600
3
[microns /sec].
Once the Apply button is pressed, the SDAS is calculated. It can then be viewed in
the Contour Menu.
Version 2006.0
Page 393
POROSITY
To analyse the porosity in a casting, several options are available :
Porosity
When POROS > 0 is used, one can visualize the Contour called "Shrinkage
porosity". Values corresponding to a level below 0.01 shall be considered as
microporosity (to be viewed best with slices) and values above 0.01 are
considered as macroporosity (to be viewed best with cut-off mode).
See the "Porosity models" section for more details.
Niyama
See the Criterion Functions section for more details.
Version 2006.0
Page 394
Version 2006.0
Page 395
Version 2006.0
Page 396
CRACKING INDICATOR
The "Cracking indicator" model of ProCAST corresponds to cracking occurring
after completion of solidification. The model is based upon the "modified Gurson"
model. It corresponds to a plastic strain driven model.
The cracking indicator model couples the stress calculation with the porosity
calculation. It is valid with any porosity model. The cracking indicator is
corresponding to a void fraction indicator (i.e. it calculates the amount of voids
created by cracks, which have nucleated and grown under the influence of
stresses).
Plastic strain (and plastic strain only) will allow cracks to nucleate and grow (thus
to create "crack void fraction"). The presence of porosity will increase the amount
of crack nucleation and growth. All the plastic strain is taken into account
(including the one forming in the mushy zone).
There is no inputs needed for that model. One should note that the constants which
are in the modified Gurson model, are not temperature dependant (as there is no
indication in the literature about how these values should change with
temperature). However, the temperature dependency is taken into account in one
way by the stress material properties and thus the plastic strain (e.g. there will be
less plastic strain at low temperature).
This model is an "Indicator" which can be used for any material. The value of the
scale of the indicator will strongly depend upon the material and the stress data.
Thus, one should not compare two different materials, but one should use it to
compare the same material in different situation (i.e. different cooling conditions,
different stress state, different porosity level).
This model corresponds to a "damage" model, which couples for the first time
stress calculation with defects in the casting (i.e porosity). This model is very new
and should be used with care as there is very little experience in this field at this
stage. However, it could be very interesting to see the effect of design decisions
on the level of damage (e.g. a reduced cooling will change the amount of porosity
and/or change the amount of stress and thus, it will change the amount of cracks
which may appear).
The cracking model is activated with the Run Parameter CRACK = 1 or 3. The
default value is 0. See the "Stress Run Parameters" for more details.
Version 2006.0
Page 397
FRECKLES INDICATOR
The "Freckles indicator" of ProCAST is based upon a local Rayleigh Number
calculation. The algorithm is described in "Development of a Freckle Predictor via
Rayleigh-Number method for Single-Crystal Nickel-Based Superalloy Castings",
by C. Beckermann, J.P. Gu and W. Boettinger, Metallurgical and Materials
Transactions A, Vol. 31A.
To activate this model, the casting material properties should have been created in
PreCAST with the Thermodynamic databases (see the "Thermodynamic
Databases" section for more details), using the Ni database. When the properties
are calculated, a file named "prefixls.dat" is created. This file (which contains the
liquid solute concentrations) should be present in the working directory in order to
be able to calculate the Freckles indicator.
Moreover, the Run Parameter FRECK with a value of 1 should be added manually
in the p.dat file.
Version 2006.0
Page 398
RESULTS EXPORTS
INTRODUCTION
Different kind of calculated results can be exported in ViewCAST. The
File/Export menu should be used for this purpose.
The results can be exported either in Patran or I-DEAS format, except the Stress
results which can be exported in an ASCII format.
Version 2006.0
Page 399
GEOMETRY
When the File/Export/Geometry menu is selected, the following window opens
Firstly, the format of the output file should be selected (1). Then, as Patran or IDEAS files do not contain the unit information, it is necessary to specify in which
units it is desired to export the geometry information (2).
In the case of stress calculations, as the geometry is deforming, one can choose to
export the deformed geometry at a given timestep (3). It is also possible to export
the geometry with the "reversed displacements" (4). This latter option is useful if
one would like to run a stress model with a "negative" distortion in order to see
whether the final shape will correspond to one of the original drawing. In this
case, one should export the model in the d.dat format.
Version 2006.0
Page 400
RADIATION FACES
For a radiation model, the radiation faces can be exported in the Patran or I-DEAS
formats. One should also specify the units which will be used in the exported file.
Version 2006.0
Page 401
TEMPERATURE
For the export of Temperatures, three options are available for the selection of the
time-steps to be exported, as shown in the menu below.
Interval
With the Interval option, one can export temperatures according to the steps
which are defined in the Steps menu (only the Steps selection will be used and not
the Time selection) :
Then the format of the exported file (Patran or I-DEAS) and the units should be
selected.
Version 2006.0
Page 402
Specify Steps
Any step to be exported can be selected with the "Specify Steps" option. The
selected time-steps should be entered in the following list :
The above window allows also to select the format of the export (Patran or IDEAS), as well as the units. The data are exported when the Apply button is
pressed.
Select Steps
The "Select Steps" option allows to choose the desired steps in the list of the
available timesteps :
Version 2006.0
Page 403
The selected steps are highlighted in red. The above window allows also to select
the format of the export (Patran or I-DEAS), as well as the units. The data are
exported when the Apply button is pressed. The clear button allows to clear all the
selections.
Once the data are exported, a log file with the exported steps is created. This file is
named prefixt.log :
Version 2006.0
Page 404
HEAT FLUX
The heat flux can be exported in exactly the same way as the temperature - see the
"Results Exports / Temperature" section for more details.
Version 2006.0
Page 405
DISPLACEMENTS
The displacements can be exported in exactly the same way as the temperature see the "Results Exports / Temperature" section for more details.
In order to Export the geometry with "reversed" displacements, please refer to the
"Results Exports / Geometry" section for more details.
Version 2006.0
Page 406
STRESS
The stress results can be exported in exactly the same way as the temperature - see
the "Results Exports / Temperature" section for more details.
The only difference is that the stress results are exported in a "Neutral" ASCII file
format in order to be imported in any stress software.
where x, y and z are the node coordinates and s1 to s6 are the six components of
the stress tensor at the node.
As ProCAST is calculating the stress data at the Gauss points, these values are
interpolated at the nodes.
Version 2006.0
Page 407
PARALLEL SOLVER
INTRODUCTION
The goal of Parallel processing is to accelerate the computing time of a
calculation, by distributing it on several processors.
Different techniques are available and are described hereafter.
SMP : Symmetric MultiProcessing
In the SMP technique (often called Shared Memory Processing), the calculation is
split in different CPU's which are sharing the same memory (see figure below).
This technology is usually limited to a maximum of 32 CPU's.
Version 2006.0
Page 408
When each computer (node) has more than one CPU, sharing the same memory,
one can combine the SMP and DMP principles to make a SMP/DMP combined
architecture. Such a configuration (see picture below) requires a very fast network
between the CPU's.
The Parallel version of ProCAST is based upon the DMP technology, using MPI.
Depending upon the machine/cluster which is used, one can use either a DMP or a
combined SMP/DMP architecture. One should note that it is also possible to use a
single computer with several CPU's sharing the same memory.
Version 2006.0
Page 409
As the MPI technology is based upon the "Message Passing Interface", it means
that the CPU's have to communicate between them. This emphasizes the
importance of the network linking the different nodes. One can use either a
network with a 1 Gigabit/s Ethernet switch to link the nodes. It is possible to have
faster communications with for instance a Myrinet network (from Myricom). This
latter solution is more expensive, but the gain in performances can be quite
significant.
As communications are critical for the performances, it is very important to use
the appropriate switch as well as the appropriate cables (e.g. for a 1 Gigabit/s
Ethernet switch, one should use Category 6 cables - or above).
Version 2006.0
Page 410
Then, each sub-domain is allocated to one processor, as shown in the figure below
(each color represents a sub-domain). The communication between the subdomains is done automatically, through the fast network.
Version 2006.0
Page 411
The following pictures illustrate the partitioning of different cases. In the lower
geometry, the partitioning for 2 and 4 CPU's is shown on the left and right,
respectively.
Version 2006.0
Page 412
If such a partitioning is very efficient for thermal, radiation and stress calculations
(as well as for flow calculations with an already filled cavity), it is not the case for
filling calculations. As illustrated in the figure below (where the filling starts from
the biscuit in the bottom of the picture), the first stage of the filling will occur only
in the red sub-domain (CPU 1), whereas the final stage of the filling will take
place in the purple sub-domain (CPU 2). This means that the CPU 2 will not do
anything during a while at the beginning of the calculation, whereas CPU 1 will
"sleep" near the end of the calculation. This will lead to a very poor balancing
between the processors and thus a very bad performance.
Version 2006.0
Page 413
However, as one does not know how the cavity will be filled (this is why a
calculation is performed !), a "dynamic partitioning" algorithm has been
developed. At a given step, the flow front is at a given location (in blue in the
figure below). The software is creating a layer ahead of the the flow front (the
width of this layer corresponds to a number of elements, defined by the Run
Parameter NFFWID). Then, both the liquid region and this layer are partitioned in
sub-domains. Once the liquid is reaching the end of this layer (i.e. when the
distance between the flow front and the end of the layer is less than NFFSAF
elements), a new partitioning is performed. Thus, the frequency of the partitioning
Version 2006.0
Page 414
is defined by NFFWID. The smaller the value, the larger the numbers of
partitioning. As the partitioning is taking some CPU time, one should specify a not
too small value for NFFWID).
One should note that in the case of a filling problem, in the presence of a mold,
two distinct partitioning are made. One for the thermal and radiation problem
(which is done only once at the beginning of the calculation) and one for the
casting domain only, which is dynamically re-partitioned.
In order to perform this automatic domain decomposition, to manage the
communication between the processors and to solve the linear system in a
distributed fashion, several libraries are used :
MIPCH or LAM/MPI
A Portable Implementation of MPI (Message Passing Interface). Depending upon
the platforms, MPI can be called with different names.
Parmetis
A Parallel graphic partitioner
PetSc
The Portable, Extensible Toolkit for Scientific computation
Version 2006.0
Page 415
Version 2006.0
Page 416
Version 2006.0
Page 417
One can see in the above execution windows that a "Parallel" section is appearing.
For the ProCAST solver, one still have in both cases the possibility to "Execute
DataCAST first".
On Linux, the execution in batch mode can also be used (in this case, the Batch
log file can be defined - if nothing is specified, a file named DMP_runlog.txt will
automatically be created (not available on Windows).
Then, for the Parallel solver, some parameters and options should be defined,
before the Run can be launched.
Firstly, the number of processors or the processors selection should be done. If the
radio button "Number of Processors" is selected, the user must specify on how
many processors the calculation should be distributed. Then, the machine will
automatically select the processors.
If the "Processors Selection" is chosen (not available on Windows), the user can
define specifically on which processors the calculation should be run. The list of
processors should be specified in between comas (","). A dash ("-") allows to
define a list (i.e. "2-4" means "2,3,4"). Please note that the processors numbering
is starting at 0.
Then, the Run option should be specified :
On Linux and Unix,"Options(1)" correspond to the MPI options, whereas
"Options(2)" corresponds to the ProCAST Parallel options.
On Windows, no options needs to be specified, except if the data are on an other
disc than the C:/ drive (see the synthax hereafter).
For the MPI options, it is advised to use the default options which are proposed (ssi rpi usysv for Linux). These options are improving the MPI communications
when LAM-MPI is used.
Version 2006.0
Page 418
At this stage, the ProCAST Parallel options are only for debugging purposes and
thus none should be specified by the user.
If one would like to change these options, it can be done directly in the
corresponding fields.
The Parallel solver (on Linux only) can be launched from the "Run List" of the
Manager (see the "Software Manager/Run list" section for more details). It is also
possible to launch the Parallel solver "manually", from a Command Window.
In such case, the syntax is the following :
For Linux :
mpirun -options(1) CPU_selection procastparallel prefix -options(2)
For Windows :
mpiexec -options(1) CPU_selection procastDMP.exe prefix
Version 2006.0
Page 419
Then, the calculation can be launched from the Manager (or from a DOS
command window) from the M: drive :
Please note that the Mapping options (i.e. -map m:\\Disk1\Data) can be configured
in the Installation settings of the Manager.
CPU_selection : selection of processors
-n N
selection of N processors
procastDMP.exe : ProCAST Parallel executable
prefix : case name
Examples :
mpiexec procastDMP.exe prefix
or
mpiexec -map m:\\Disk1\Data procastDMP.exe prefix
Version 2006.0
Page 420
REPEATABILITY
The goal of a parallel solver is to obtain the same results as the scalar version and
regardless of the number of processors. This is called "Repeatability".
One should note that Parallel processing (using an implicit solver like in
ProCAST), involves the resolution at once of a linear system on distributed
processors, with the appropriate communications between the processors. As a
first consequence, the algorithms which are used in the parallel version (especially
for the pre-conditioner) can not be exactly the same as the ones used in the scalar
version. This may lead to small differences in convergence (and thus in timesteps)
and in round-off. Moreover, as iterative solvers are used, the solution is never
exact and it may slightly depend upon the way it is solved (which can be slightly
different on 1, 2, 4 or 8 processors).
Thus, one should expect to have very similar results between the scalar and the
parallel version and between runs made with different numbers of processors,
however, most of the time, one will not have exactly the same result. This is
especially true in filling calculations where very small differences may have an
effect on the filling pattern (as different meshes or different COURANT values
may have too).
Version 2006.0
Page 421
LIMITATIONS
In ProCAST 2006.0, not all the modules and features of ProCAST have been
parallelized yet. This section is describing the limitations of the current version, as
well as the precautions in the use of the parallel solver. As the use of a nonparallelized feature may degrade completely the performances of the solver, these
features are made non accessible and the solver will stop in such case. The tests
are mainly done through the values of the corresponding Run Parameters.
However please note that when a Parallel calculation is terminated, many MPI
print-out may appear. This means that the ProCAST error message may be
"embedded" into MPI messages.
In ProCAST 2006.0, only the THERMAL, RADIATION, FLOW and STRESS
modules are parallelized. The parallelization of the other modules is not planned
at this stage.
The following features are not parallelized (the list may not be exhaustive) and
thus should not be used with the parallel solver :
THERMAL module :
THERMAL = 2
POROS = 4 and 8 (only POROS = 1 is available)
User functions
Freckles indicator
FLOW module :
Thixo casting with the Cut-off model. However the non-newtonian model is
parallelized
Interpenetrating mesh (e.g. shot piston) (PENETRATE = 1)
RESERVOIR option
Turbulent model
Version 2006.0
Page 422
MACHINE CONFIGURATION
In order to have an efficient parallel processing, it is important to set-up the
machines in the right way. This section explains the basics principles of the
machines installation and configuration in order to run the ProCAST Parallel
solver. Please note that the following is not necessary for the Windows DMP
version.
Cluster of processors
During the Parallel calculation, the software is communicating with the different
nodes (computers of the cluster) through "Secured shells" (ssh). Thus, the
machines should be configured in order to allow "ssh" communication between
the nodes. If a "paraphrase" (password) is set to activate the ssh, one should
perform the two following commands when a new Command Window is used :
ssh-agent $SHELL
ssh-add
then, the "paraphrase" should be given. If this operation is not done, the
paraphrase will be asked for each node when the ProCAST Parallel solver will be
launched. If ssh is configured without paraphrase, then, this is not necessary.
If the Parallel solver is launched from the Manager, it is necessary to perform the
above operation in the Command window before the Manager is launched (in this
same command window).
Version 2006.0
Page 423
The disk on which the software and libraries are installed, as well as the disk(s) on
which the data will be located (it can be different disks) should be mounted in
such a way that they can be accessed in the same way (i.e. with the same path)
from any machine (node).
Version 2006.0
Page 424
Linux
Windows
Linux
LAMHOME environment variable
LD_LIBRARY_PATH
LAMBOOT
On Linux machines, the "LAM/MPI" service should be started once when the
machine is booted. To do so, the command "lamboot" should be run (only on node
1). This lamboot will be valid for all further DMP runs. Sometimes, it may be
needed to reset the lamboot. To do so, one should first stop the service with the
"lamhalt" command. Then "lamboot" can be launched again. One should be
Version 2006.0
Page 425
careful that a "lamhalt" or a "lamboot" command will stop all currently running
parallel jobs (using LAM).
Hostname specification
The list of machines (nodes) on which the parallel processing will be performed
should be specified in a file named "lam-bhost.def". This file is located in the lam7.1.1/etc directory.
One should put in the "hostnames" on which the parallel executable should run. If
the machines called "hostname1" and "hostname2" contain each 2 processors, the
lam-bhost.def file should contain the following lines :
hostname1 cpu=2
hostname2 cpu=2
You can also search for this patch in the "Windowsupdate" web site of Microsoft :
Version 2006.0
Page 426
Version 2006.0
Page 427
HARDWARE AND OS
The pre-requisites and requirements for the hardware and for the operating
systems (OS) are the following (at this stage only Linux and Windows are
available) :
Windows
Windows XP and Windows 2000 Professional only.
Please note that the installation of MPICH2 requires the prior installation of the
patch "Microsoft .Net Framework version 1.1".
kernel :
glibc :
>= 2.4.20
>= 2.3.2
In order to get the level of the kernel or of glibc, the following commands should
be run on the machine :
rpm -q kernel
rpm -q glibc
Version 2006.0
kernel :
glibc :
>= 2.4.21
>= 2.2.1
Page 428
In order to get the level of the kernel or of glibc, the following commands should
be run on the machine :
rpm -q kernel
rpm -q glibc
Version 2006.0
Page 429
Version 2006.0
Page 430
Version 2006.0
Page 431
their employees, makes any warranty express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents that its use
would not infringe privately owned rights.
Version 2006.0
Page 432
ADVANCED POROSITY
CALCULATIONS
Porosity calculations are performed as a "post-processing" of the thermal results.
Thus, one should first configure and run a thermal calculation as usual and then to
configure and run the porosity model.
INTRODUCTION
The porosity module of ProCAST is based upon the model developed by Pequet,
Gremaud and Rappaz. For further details, please refer to :
Ch. Pequet, M. Gremaud, M. Rappaz, Modeling of Microporosity, Macroporosity
and Pipe Shrinkage Formation during the Solidification of Alloys using a MushyZone Refinement Method, Met. Mater. Trans. 33A (2992) 2095-2106.
The model can be summarized as follow (abstract of the above mentioned paper) :
A microporosity model, based on the solution of Darcy's equation and
microsegregation of gas, has been developed for arbitrary three-dimensional (3D)
geometry and coupled with macroporosity and pipe shrinkage predictions. In
order to accurately calculate the pressure drop within the mushy zone, a dynamic
refinement technique has been implemented: a fine and regular Finite Volume
(FV) grid is superimposed onto the Finite Element (FE) mesh used for the heat
flow computations. For each time step, the cells, which fall in the mushy zone, are
activated and the governing equations of microporosity formation are solved only
within this domain, with appropriate boundary conditions. For that purpose, it is
necessary to identify automatically the various liquid regions that may appear
during solidification: open regions of liquid are connected to a free surface where
a pressure is imposed; partially-closed liquid regions are connected to an open
region via the mushy zone and closed regions are totally surrounded by the solid
and/or mold. For partially-closed liquid pockets, it is shown that an integral
boundary condition applies before macroporosity appears. Finally, pipe
shrinkage (i.e., shrinkage appearing at a free surface) is obtained by integration
of the calculated interdendritic fluid flow over the open region boundaries, thus
ensuring that the total shrinkage (microporosity plus macroporosity and pipe
shrinkage) respects the overall mass balance.
Version 2006.0
Page 433
Version 2006.0
Page 434
The calculation condition file (prefix_poro.d (ASCII format)) contains all the
necessary information to perform a calculation. It is divided in blocks of data
containing different type of information :
General information
porosity.param
Material properties
viscosity.param
coarsening.param
Gas and bubble properties
nucleation.growth.param
gas.param
Process information
pressure.param
feeding.param
gravity.param
Calculation settings
grid.param
calculation.param
Each block starts by the "block.name" and ends by "end". Inside a block,
commands are inserted. Each command is always followed by one argument
(always one argument). The order of the blocks or of the commands within a
block is free.
Version 2006.0
Page 435
Please note that only SI units must be used for the porosity calculation definition
(however the thermal calculation can be set-up as usual with any units).
The frames shown hereafter in light blue correspond to examples of settings.
The following sections are describing these different inputs.
General Information
porosity.param
Firstly, the type of model should be selected. As in ProCAST only the "standard"
model is available, the above block should always be specified as such.
Material Properties
When the "Advanced Porosity Model" is used in conjunction with a ProCAST
calculation, all the necessary thermo-physical properties (specific heat, enthalpy,
latent heat, fraction of solid and density) are automatically extracted from the
ProCAST input data. Thus, nothing has to be specified and no special care has to
be taken (as in previous versions).
Of course, in order to allow porosity calculations, one has to have a density which
is defined as a function of temperature (in this case a message saying that you
have used a constant "specific.mass.mixture" will be printed). This version of the
"Advanced Porosity Model" does not allow to have "expanding" materials (i.e.
density which increases with increasing temperatures).
In addition to the ProCAST thermo-physical data, the following data must be
specified :
viscosity.param
Version 2006.0
Page 436
For the calculation of the D'Arcy equation (i.e. flow through the mushy zone), the
viscosity is an important quantity. The viscosity can be taken either as a constant,
or as a function of temperature. In this latter case, an Arrhenius law is considered :
: activation.energy
: reference.viscosity
The following frames are showing the two possibilities available for the user :
a) Temperature dependant viscosity
b) Constant viscosity
coarsening.param
Version 2006.0
Page 437
Beside the fraction of solid (gs), the permeability is a function of the secondary
dendrite arm spacing (2), which can be calculated according to the following
equation :
where tL is the time at which the liquidus temperature was reached and M(t) is the
coarsening constant.
The coarsening constant can be calculated as follow, for a binary alloy :
The following frames are showing the two possibilities available for the user :
a) The coarsening constant (M(t)) is specified
The nucleation and growth parameters of pores (i.e. bubbles) are defined in this
above block.
The cavitation pressure corresponds to the pressure of the liquid within a closed
pocket of liquid at which macro-shrinkage will start to appear. If no cavitation
pressure is defined, the Power law for the gas solubility (see gas.param) is also
used for the calculation of the macro-shrinkage. The cavitation pressure is usually
equal to 0.8 to 0.9 of the atmospheric pressure (i.e. 0.8-0.9 bar or 80'000 to 90'000
Pa).
Version 2006.0
Page 438
When a pore will nucleate, it will appear with the radius defined under
"nucleation.radius". As the pore is usually nucleating on a substrate, its radius will
remain constant until the bubble is not wetting anymore the substrate. Then, the
pore radius will increase. Please note that this initial pore radius value should not
be defined with smaller values than 10 microns. Otherwise, the pressure inside the
bubble will be so high that it will never nucleate and the calculation will not run
well.
The nucleation of the pores is defined by a "pore.density".
Finally, the gas.metal.surface.tension will determine the pressure inside a pore,
due to it's curvature.
gas.param
Version 2006.0
Page 439
One should be careful that the initial gas concentration should not be larger than
the solubility limit of the corresponding gas (in such case, bubbles would be
created in the liquid metal before the pouring of the metal and the corresponding
amount of dissolved gas would decrease). In order to prevent such case, a warning
will be printed at the beginning of the calculation (see figure below) if the
following situation arises (then it is advised to reduce the initial gas concentration
or to change the gas parameters).
Process Information
pressure.param
The type of process, as well as the pressure definition are specified in this block
(casting.type).
Version 2006.0
Page 440
Firstly, one should specify the process which is modeled, among three choices :
gravity, injection or continuous.
gravity :
For all type of gravity casting (sand, permanent mold, investment, ...), the
"gravity" type should be used. In this case, the software is automatically setting
the pressure to the highest surface of the casting. This means that the user does not
need to specify anything, except the value of the pressure (p.imposed). The
highest surface is determined with the help of the gravity vector.
With this mode, the metal at the top of the casting will be free to pipe, according
to the value of mobility.limit.
injection :
The type "injection" has to be used for hpdc, lpdc and squeeze casting, as well as
all processes where the liquid metal is injected. In this case, the injection point has
to be specified with the "x,y,z.pressure.coordinate" (see below) - it is advised to
select the last point to solidify as the feeding will be stopped when this point will
solidify.
As the metal is injected, no piping will occur, as there will be a continuous feeding
of metal (however, depending upon the geometry and the thermal conditions,
macroshrinkage may occur). In this case, the mobility.limit does not need to be
specified and it will be automatically set to 1.
continuous :
For all continuous casting processes, the "continuous" type should be used. As for
"injection", one will have a continuous feeding of metal and thus, the
mobility.limit does not need to be defined. In this case, it will automatically be set
to 1. Moreover, the feeding point (usually the last point to solidify) should be
specified with the "x,y,z.pressure.coordinate" (see below). In addition, in the case
of a vertical continuous casting, one should specify the liquid level (with p.level).
In this case, the value of the applied pressure will be automatically adapted to take
into account the appropriate metallostatic pressure.
In the case of "injection" or "continuous", the location where the pressure is
applied should be defined (whereas it is automatically set to the highest surface is
the case of "gravity"). This is done with the keywords "x.pressure.coordinate",
"y.pressure.coordinate" and "z.pressure.coordinate". It is strongly advised to
specify a location at the injection point which will solidify last as the feeding will
be stopped when the fraction of solid at this location will reach a value of 1.
The applied pressure is specified with "p.imposed". This pressure is either applied
on the highest surface (in the case of "gravity"), or at the
"x,y,z.pressure.coordinate" location in the case of "injection" or "continuous". The
value of p.imposed can be defined either by a constant or by a function of time
(with a value block F(TIME)).
Version 2006.0
Page 441
In order to define the pressure as a time table, one should replace the pressure
value by a pointer (i.e. a name starting by a "*" - see below).
Then, an additional value block should be specified with the time function. The
name of the value block should be the same as the one of the pointer (without the
"*"). Then, the following syntax should be used :
If the level of the liquid is not corresponding to the level of the coordinate where
the pressure is applied, this can be accounted for with the keyword "p.level". This
has to be used in continuous casting (as the "effective" level of the liquid is
changing during the process) or if a part of the casting is not modeled. When both
"p.level" and "x,y,z.pressure.coordinate" are defined, the pressure applied at the
p.level altitude (in absolute axis coordinated. in the direction of the gravity)
corresponds to the applied pressure (defined by "p.imposed"). Below p.level, the
metallostatic contribution is added to p.imposed. This can be described with the
following piece of code (considering that the gravity is in the -Z direction) :
if (z > p.level) then
pcell = p.imposed
else
pcell = p.imposed + rho * g * (p.level - z)
endif
where z is the current altitude of the cell for which
the pressure (pcell) is calculated.
Version 2006.0
Page 442
The mobility.limit corresponds to the critical solid fraction from which the liquid
at the top free surface can not move anymore (for piping). If the mobility.limit is
very low (0.01), it means that as soon as the solid fraction is equal to 1%, the
liquid can not move down anymore to feed the liquid pockets below. Thus one
will have almost no piping, but macroshrinkage in the center of the casting. On the
other hand, if a larger value of mobility.limit is used (0.2), the liquid can move
down up to a fraction of solid of 20%. This will induce a significant piping and
thus reduce the amount of macroshrinkage which may occur in the center of the
casting. The value of mobility.limit is dependent upon the type of microstructure.
In the case of columnar structure, the dendrite are linked to the solid shell and thus
can not move too much, leading to a small value of the mobility.limit. On the
other hand, in the case of equiaxed grains, they can move down more freely and
thus, the mobility.limit is higher.
If one would like to prevent piping (such as in high pressure die casting, squeeze
casting or continuous casting), the mobility.limit is automatically set to 1.0 In this
case, the liquid will continuously feed the casting.
The value of mobility.limit can be defined either by a constant or by a function of
time (with a value block F(TIME)). This can be used in order to model the end
stage of a DC casting process. In this case, the mobility limit will be changed from
1.0 to 0.2 when no more liquid metal will be poured in the ingot.
The "permeability.solid.fraction.cutoff" is used to cut-off the permeability at high
solid fractions. As it can be seen in the following equation :
when the solid fraction (gs) is becoming zero, K is equal to zero. As in the D'Arcy
equation, one is using 1/K, it is going to infinity. In order to prevent that, the
permeability is taken as a constant for fractions of solid above the value of
"permeability.solid.fraction.cutoff". Values around 0.95 to 0.98 are recommended.
If values higher than 0.99 are used, automatically a value of 0.99 is set. If nothing
is specified, a value of 0.98 is used.
At very high solid fractions, the interdendritic flow becomes very small and the
pressure drop calculation becomes very difficult (on a mathematical point of
view). In order to improve the convergence of the model, the feeding calculation
is stopped at a fraction of solid above "feeding.solid.fraction.cutoff". Values
around 0.95 to 0.98 are recommended. If values higher than 0.98 are used,
Version 2006.0
Page 443
The three components of the gravity must be defined. One should note that the
gravity should be parallel to one reference axis (i.e. parallel to x, y or z). Thus,
two components of the gravity must be equal to zero. In addition, it is not possible
to have a rotating gravity. Thus, it is not possible to use this module for Tilt or
centrifugal cases.
Calculation Settings
grid.param
The "grid.spacing" corresponds to the cell size of the FVM mesh. One should
select a size which is in relationship with the size of the mushy zone (in order to
have several elements through the mushy zone thickness).
The "grid.calculation.tolerance" is used in order to remove duplicate cells which
would lie at the limit in between two elements. If a regular orthogonal FEM mesh
is used and if the cell size (or grid spacing) is exactly matching the FEM mesh
(i.e. FEM elements have a size of 1 cm and FVM cells have a size of 1 mm.), it
happens that many cells are lying exactly on the element borders. The algorithm
which is used involves long loops to guarantee no double cells. The
grid.calculation.tolerance is used to check whether a cell is aligned with element
borders. If this tolerance is too large, doublon checks will be made many times
and the CPU time can rise very strongly. If this tolerance is too small, it is
possible to create "holes" in the FVM mesh. Thus, a value of 1.e-12 is
recommended. If the cellular meshing becomes very long, it is advisable to change
slightly the cell size from 1.00 to 1.00001 in order to remove these alignments. If
unstructured FEM meshes are used, this problem never happens.
In cases, such as hpdc, sometimes, the gates between the casting and the
overflows are very thin. This may lead to a lack of cells at the gate and thus, the
Version 2006.0
Page 444
overflow may become "isolated" form the casting (i.e. no cells in between the
casting and the overflow). In such case, the porosity calculation will not be
possible, as there is no way to "transmit" the pressure from the casting to the
overflow. This could also happen if very thin sections, such as flashes are
modeled. In this case, these "orphan" cells should be removed in order to allow
the calculation to run. As this procedure of "orphan search" is CPU intensive, it
can be disabled in cases where it is clear that this will not occur (for bulky parts).
Otherwise, it is highly recommended to set the "grid.search.orphan" value to ON.
calculation.param
Version 2006.0
Page 445
In the "calcosoft" format, the porosity is stored at each FVM cell. This allows a
very fine resolution of the results, but it requires the use of the specific calcosoft
post-processor (see the "Advanced Porosity Post-processing" section for more
details).
In the "ProCAST" format, the porosity results, calculated at the FVM cells are
extrapolated to the FEM nodes. This allows to view the "Advanced porosity"
results in ViewCAST.
The "store.type" allows to select which result files format should be stored.
0 : "calcosoft" format only
1 : "ProCAST" format only
2 : both formats
The "store.frequency" parameter defines the frequency of storage of the porosity
results. It is possible to store the pore fraction only (and not the pressure) with
"store.pore.fraction", however, if the pressure only storage is defined (with
"store.pressure"), the pore fraction will also be stored. These parameters will be
valid only for the "calcosoft" format results. The ProCAST format results will be
stored at the same frequency as the ProCAST temperatures, whatever values are
specified for these parameters.
In order to define a list of regions (to perform the porosity calculation in more
than one region), the region.number should be specified by a pointer (i.e. a name
starting by a "*" - see below).
Then, an additional value block should be specified with the list of regions. The
name of the value block should be the same as the one of the pointer (without the
"*"). Then, the following synthax should be used :
Version 2006.0
Page 446
Warnings
One should note that the "Advanced Porosity" module is based upon very
complex physical algorithm and thus it's mathematical handling may be quite
delicate in some cases. It is thus possible that the solution is not converging (or is
giving error with the pre-conditioner). In such case, it is advised to either : modify
the cell size, store more thermal timesteps, change (i.e. increase) the
convergence.tolerance, decrease the value the feeding.solid.fraction.cutoff or of
the permeability.solid.fraction.cutoff.
One should also take a special care to make sure that there is no remelting.
Finally, when the mushy zone is quite small (in case of large castings with low
conductivity materials), the cell size should be adapted (i.e. the cell size should be
smaller than the mushy zone length).
standard
$$$
# <-------
0.001 Pa.s
= 1
$$$
# <-------
100.e-6 m
= 100
Version 2006.0
1.E-05
1.E+09
$$$
10000.
# recommanded value
# <------0.90
value for
# 10000 = 0.1 bar /
Page 447
gas.param:
gas.nominal.concentration
CCSTP/100g (for Al)
gas.partition.coefficient
gas.equilibrium.constant.a
3 and B = 0
)
gas.equilibrium.constant.b
and B = 48822)
gas.pressure.exponent
end
$$$
# <-------
0.10 to 0.30
1.E-01
2.2e-3
0.00
0.5
gravity
101325.0
feeding.param:
mobility.limit
feeding.solid.fraction.cutoff
permeability.solid.fraction.cutoff
end
# Gravity setting
# <------- 101325.0 [Pa.s] = 1
0.80
0.98
0.98
gravity.param:
gx
gy
gz
end
# Range from 0. to 1.
# 0.98 : recommanded value
# 0.98 : recommanded value
$$$
$$$
$$$
gravity.x
gravity.y
gravity.z
1.e-12
$$$
# <------0.002 = 2 mm
# recommended value
*regions
1
2
on
on
0
1e-10
5000
value regions
data PARAMETERS
$$$
$$$
$$$
Porosity Domain 1
...
Porosity domain N
end value
Example 2 : TemplateInjection_poro.d
#------------ General information -------------porosity.param:
porosity.model
end
standard
Version 2006.0
$$$
# <-------
0.001 Pa.s
= 1
Page 448
coarsening.param:
secondary.dendrite.arm.spacing
microns
end
$$$
# <-------
100.e-6 m
= 100
1.E-05
1.E+09
$$$
10000.
# recommanded value
# <------0.90
value for
# 10000 = 0.1 bar /
$$$
# <-------
0.10 to 0.30
1.E-01
2.2e-3
0.00
0.5
injection
feeding.param:
mobility.limit
injection
feeding.solid.fraction.cutoff
permeability.solid.fraction.cutoff
end
$$$
# Injection setting
# <------- 10132500.0 [Pa.s] =
$$$
$$$
$$$
1.00
0.98
0.98
gravity.param:
gx
gy
gz
end
$$$
$$$
$$$
gravity.x
gravity.y
gravity.z
Version 2006.0
1.e-12
$$$
# <------0.002 = 2 mm
# recommended value
*regions
1
2
on
on
0
1e-10
5000
Page 449
value regions
data PARAMETERS
$$$
$$$
$$$
Porosity Domain 1
...
Porosity domain N
end value
Example 3 : TemplateContinuous_poro.d
#------------ General information -------------porosity.param:
porosity.model
end
standard
$$$
# <-------
0.001 Pa.s
= 1
$$$
# <-------
100.e-6 m
= 100
1.E-05
1.E+09
$$$
10000.
# recommanded value
# <------0.90
value for
# 10000 = 0.1 bar /
$$$
# <-------
0.10 to 0.30
1.E-01
2.2e-3
0.00
0.5
continuous
101325.0
feeding.param:
mobility.limit
feeding.solid.fraction.cutoff
permeability.solid.fraction.cutoff
end
Version 2006.0
# Continuous setting
# <------- 101325.0 [Pa.s] = 1
$$$
$$$
$$$
*plevel
1.00
0.98
0.98
Page 450
gravity.param:
gx
gy
gz
end
$$$
$$$
$$$
gravity.x
gravity.y
gravity.z
1.e-12
$$$
# <------0.002 = 2 mm
# recommended value
*regions
1
2
on
on
0
1e-10
5000
value regions
data PARAMETERS
$$$
$$$
$$$
Porosity Domain 1
...
Porosity domain N
end value
value plevel
data F(TIME)
$$$
$$$
$$$
$$$
$$$
$$$
end value
Version 2006.0
# Time level
# Time level
# Time level
Page 451
The "Run" button should be hit (in the "Advanced Porosity Solver" section) to
start the calculation. If the ProCAST Thermal calculation is not yet run, it is
possible to click on the "Execute DataCAST/ProCAST first" check box before the
"Run" button.
It is also possible to automatically launch a serie of calculation using the "Run
list" of the Manager.
If the above window does not show the "Advanced Porosity Solver" section, one
can activate it in the "Installation settings" window (by clicking in the "Advanced
Porosity module" check box, under "Modules display") :
Version 2006.0
Page 452
Version 2006.0
Page 453
Version 2006.0
Page 454
To access the "calcosoft format", one has to make a right mouse click on the
"ViewCAST" button of the manager. This will open a sub-menu and one should
select "Advanced Porosity results" (see below) :
This is automatically opening the "calcosoft" post-processor and the porosity field
is shown. One should use the tape player to visualize the last timestep.
Version 2006.0
Page 455
When a cut-off value of 1% is used, the following result is obtained (which is very
similar to the ViewCAST result above) :
Version 2006.0
Page 456
One should note that the "calcosoft" format results are more precise than the
"ProCAST" format results. This is due to the fact that the Advanced porosity
calculation is performed at cells (which are all shown in the "calcosoft" format,
whereas these cells results are "extrapolated" to the FEM nodes for the
"ProCAST" format results. If there is only a few cells which are empty (i.e. a
small macroshrinkage), this may be "diluted" in the surrounding microporosity
during the extrapolation to the FEM nodes. Thus, this small macroshrinkage may
not be visible in the ViewCAST.
Concerning the "calcosoft" format, the microporosity is shown using the yellowgreen scale, whereas the macroshrinkage is shown in orange and red. In practice
the macroshrinkage are negative values (from 0 to -100), whereas microporosity
are positive values (from 0 to 100). The macroporosity between 0 and 50% is
shown in orange, whereas the one between 50 and 100% is shown in red.
In "ProCAST" format, this distinction can not be made and both the microporosity
and the macroporosity is shown (as positive values).
Version 2006.0
Page 457
MICROSTRUCTURES
The Microstructure models were improved in version 2006.0. In particular, the
treatment of the thermal algorithm was significantly improved. This allows to
have much larger timesteps than before. Please note that due to this change the
maximum timestep should be kept to reasonable values (i.e. one should make sure
that the solidification range is crossed in several timesteps).
Concerning the prediction of the Casting strength, previously it was only a
function of the grain size. It has been improved in order to be also a function of
the ferrite-perlite fractions. As a consequence, the Casting strength will not be
only dependent upon the thermal history near the solidification, but also upon the
further cooling (i.e. it will depend upon the cooling rates between 700 and 900C).
INTRODUCTION
The microstructure module of ProCAST is now able to calculate automatically the
microstructures, based upon the composition of the alloy. This can be achieved
with the link of the module with thermodynamic databases.
Depending upon the chemical composition, the microstructure module
automatically detects the phases which will appear and the type of microstructure
which should be computed (dendritic, eutectic, nodular, ...).
For instance, if an Al-7%Si-0.3%Mg alloy is specified (A356), automatically, the
software will detect that primary dendrites will form, followed by an interdendritic
eutectic. On the other hands, if the composition of a Nodular Cast Iron (SGI) is
defined, the nodule counts, the austenite radius, the pearlite and ferrite fractions
will be computed, together with the corresponding mechanical properties (such as
hardness, yield and tensile strength). The software will also automatically detect
that if there is no Magnesium in the cast iron, the structure will be lamellar rather
than spheroidal. In the same way, if the composition is hypo-eutectic, primary
dendrites of austenite will form first before the eutectic precipitation.
The only parameters that the user may need to specify are the nucleation
parameters (see below). This is due to the fact that this is not an intrinsic property
and that it may depend upon the metal treatment. Moreover, it may be necessary
to define the growth kinetics of the eutectic phase (see below)
Version 2006.0
Page 458
Version 2006.0
Page 459
Thus, the nucleation behavior of the primary dendritic phase is fully defined by
the three above parameters.
Thus, the nucleation behavior of the eutectic phase is fully defined by the two
above parameters.
Thus, the above constant fully defines the growth characteristics of the eutectic
growth.
Default values
The following table is showing the default values which are used in the
microstructure module for the models described above and for the different alloys.
Of course, the nucleation data may change from one alloy to the other, due to
different metal treatment. These default values corresponds to values proposed in
the literature.
Version 2006.0
Page 460
Version 2006.0
Page 461
For the thermal properties, only the thermal conductivity and the density should be
defined in addition to the chemical composition. Then, the enthalpy needs to be
calculated in PreCAST, with the Thermodynamic database (based upon the
specified chemical composition), using the Lever model for Fe alloys and the
Scheil model for the other systems. Thus, at the end, the thermal conductivity, the
Version 2006.0
Page 462
density, the enthalpy, the fraction of solid, the liquidus and solidus temperatures
are defined.
Finally, the Run Parameter MICRO should be set to 1 (for the users of previous
versions, one should not anymore use values different from 1, as the selection of
the micro model is now automatic). A value of 0 will disable the microstructure
calculation.
If the nucleation or growth parameters have to be changed from the default values
(see Table below), the corresponding Run parameters can be modified (see the
Microstructure Run Parameters section for more details).
Results
Depending upon the alloy composition, the type of microstructure will be
different. As a consequence, the type of results which are computed will also be
different. All the microstructure results can be visualized in the post-processing, in
the "Contour/Micro" menu.
a) Dendritic primary phase and eutectic secondary phase
Most alloys are solidifying with a primary phase of dendrites, followed by interdendritic eutectic. In this case, the following quantities are calculated :
Version 2006.0
Page 463
The "Primary Dendrite Radius" corresponds to the primary grain size. The
"Secondary Dendrite Arm Spacing" (also called SDAS) is the distance between
the secondary dendrite arms of the primary phase.
The "Primary Solid Fraction" corresponds to the fraction of primary phase,
whereas the rest corresponds to the "Fraction of Eutectic".
The "Eutectic Grain Radius" corresponds to the radius of equiaxed eutectic grains
which are nucleating in between the dendrites of the primary phase. Finally, the
"Eutectic Inter-lamellar Spacing" is the characteristic distance of the eutectic
structure.
b) Special case of Fe-C alloys - grey iron (lamellar eutectic)
In the case of grey iron (lamellar eutectic), in addition to the quantities described
above, other quantities can be calculated :
Version 2006.0
Page 464
In the case of Nodular cast iron, nodules of graphite, surrounded by austenite are
formed, instead of eutectic grains. Thus, the "Nodule count" (which corresponds
to the density of graphite nodules), as well as the "Austenite Radius" and
"Graphite Radius" are calculated. The solid state transformations, as well as the
mechanical properties are calculated, as described above. In addition, the
"Elongation" is calculated from the microstructure results.
The Nodular cast iron (SGI) model is activated as soon as there is a non zero
amount of Mg in the chemical composition. Please note that Mg is not an element
which is present in the Computherm thermodynamic database as it will have a
negilible effect on the computed material properties. However, it should be
specified in order to trigger the SGI model.
d) Special case of Fe-C alloys - nodular cast iron (hypo-eutectic composition)
Version 2006.0
Page 465
The Nodular cast iron (SGI) model is activated as soon as there is a non zero
amount of Mg in the chemical composition. Please note that Mg is not an element
which is present in the Computherm thermodynamic database as it will have a
negilible effect on the computed material properties. However, it should be
specified in order to trigger the SGI model.
e) Special case of Fe-C alloys - steel
In the case of a steel with the composition of carbon equivalent less than 0.53 , the
following properties are obtained.
The fraction of peritectic is the solid fraction formed from the reaction of liquid
and the existing primary solid phase. The fraction of proeutectoid refers to the
fraction of proeutectoid ferrite or cementite formed from the austenite phase as a
function of time during the solid phase transformation. The carbon equivalent
value controls which type of the proeutectoid phase will form (the "carbon
equivalent" corresponds to %C + (%Si + %P)/3 ).
References
For cast iron, the mechanical properties calculations are based upon the
microstructure, according to two following papers :
Stefanescu et al, Proceedings of the 4th Decennial International Conference on
Solidification Processing, Sheffield, (July 1997), 609.
Goettsch et al, Metallurgical and Materials Transactions, 25A:5, (1994), 1063.
Version 2006.0
Page 466
EXAMPLES
In order to illustrate the application of the microstructure module on the
solidification of nodular cast iron, two calculations were performed with two
different chemical compositions.
To do so, a very simple geometry was used, as shown below. The casting is
cooled from the right with a chill, whereas the rest is in a sand mold.
This set-up produces a full range of cooling rates, as shown in the cooling curves
hereafter.
Version 2006.0
Page 467
The following figure is showing the two alloys which were used, with the
corresponding chemical compositions. The only difference between the two
calculations is the amount of Carbon (from 3.2% to 3.5%).
Version 2006.0
Page 468
The above figure is showing the different kind of results which are automatically
computed for the two alloys. One can see that for the Alloy B, there is no
"Primary phase", as the alloy is lying on the eutectic composition.
The following figures are showing the comparison of the different results. On the
left, the Alloy 1 is shown (3.2% C - Hypo-eutectic) and the Alloy 2 is shown on
the right (3.5% C - Eutectic).
Eutectic phase fraction
Nodule counts
Version 2006.0
Page 469
Austenite Radius
Ferrite fraction
Version 2006.0
Page 470
Pearlite fraction
Brinell Hardness
Elongation
Version 2006.0
Page 471
Tensile strength
Yield strength
Version 2006.0
Page 472
One can see that the SDAS is only available for Alloy 1, as there is no dendrites
(i.e. no primary phase) in Alloy 2.
Version 2006.0
Page 473
a.
b.
c.
Run Parameters/Micro:
i.
MICRO=1
ii.
Select Default Values (Gray Iron or Ductile Iron)
iii. Based on the inoculation level change the values accordingly
1.
EUNUCL (For ductile iron, default value is 2000 )
2.
EUPOWER (For ductile iron, default value is 2.5 )
3.
Run Parameters/Thermal
i.
MACROFS
ii.
PIPEFS
iii. FEEDLEN
p.dat file (for SGI only)
i.
MOLDRIG (0~1, change it according to the hardness of mold)
ii.
GRAPHITE (0~1, default=1.0)
Version 2006.0
Page 474
Steel
When the weight percent sum of all elements other than Fe is more than 5% (such
as some alloy steel or stainless steel), only equiaxed dendrite model will be
activated.
Version 2006.0
Page 475
When the melt alloy cools from an initial temperature higher than the liquidus
temperature down to a temperature slightly below it, the primary delta dendrite
phase begins to nucleate in the liquid until a recalescence occurs due to the heat
released from the growing nuclei. During the recalescence, the nucleation ceases
and the nuclei grow rapidly into dendritic grains and soon impinge on each other
at the end of recalescence. Then the growth of delta dendritic grains is replaced by
the coarsening of delta dendritic arms. When the temperature reaches to the
peritectic temperature, the peritectic transformation starts if there is still liquid
available. Otherwise, the austenite phase precipitates from the delta phase until all
of the delta phase is transformed into the austenite phase. Lastly, when the
temperature of the casting is cooled down to the alpha phase transformation
temperature, the alpha phase precipitates from the austenite phase to the eutectoid
temperature. Below the eutectoid temperature, the graphite of Fe3C phase
nucleates initially on the boundary of the austenite grains and then the coupled
growth of alpha and Fe3C phases leads to the formation of pearlite phase.
2) Carbon content is greater than 0.53% (such as Alloy B in Fig. 1)
When the melt alloy cools from an initial temperature higher than the liquidus
temperature down to a temperature slightly below it, the primary austenite
dendrite phase begins to nucleate in the liquid. The austenite phase will grow until
the end of solidification. There is no peritectic reaction here. When the
temperature of the casting is cooled down to the alpha phase(C<0.8%) or
cementite phase(C>0.8%) transformation temperature, the alpha phase or
cementite phase precipitates from the austenite phase to the eutectoid temperature.
Below the eutectoid temperature, the coupled growth of alpha and Fe3C phases
leads to the formation of pearlite phase.
The eutectoid temperature is around 723C while the eutectoid composition is
0.80% C. Slowly cooling alloys containing less than 0.80% C, hypo-eutectoid
ferrite is formed from austenite in the range 910-723C with enrichment of the
residual austenite in carbon, until at 723C the remaining austenite, now
containing 0.8% carbon transforms to pearlite, a lamellar mixture of ferrite and
iron carbide (cementite). In austenite with 0.80 to 2.06% carbon, cooling in the
temperature interval 1147C to 723C, cementite first forms progressively
depleting the austenite in carbon, until at 723C, the austenite contains 0.8%
carbon and transforms to pearlite.
Steels with less than about 0.8% carbon are thus hypo-eutectoid alloys with ferrite
and pearlite as the prime constituents, the relative volume fractions being
determined by the lever rule which states that as the carbon content is increased,
the volume percentage of pearlite increases, until it is 100% at the eutectoid
composition. Above 0.8% C, cementite becomes the hyper-eutectoid phase, and a
similar variation in volume fraction of cementite and pearlite occurs on this side of
the eutectoid composition.
Version 2006.0
Page 476
Cast Iron
The term cast iron, like the term steel, identifies a large family of ferrous alloys.
Cast irons are multicomponent ferrous alloys. They contain major (iron, carbon,
silicon), minor (<0.01%), and often alloying (>0.01%) elements.
Cast iron has higher carbon and silicon contents than steel. Because of the higher
carbon content, the structure of cast iron, as opposed to that of steel, exhibits a
rich carbon phase. Depending primarily on composition, cooling rate and melt
treatment, cast iron can solidify according to the thermodynamically metastable
Fe-Fe3C system or the stable Fe-graphite system.
When the metastable path is followed, the rich carbon phase in the eutectic is the
iron carbide; when the stable solidification path is followed, the rich carbon phase
is graphite. Referring only to the binary Fe-Fe3C or Fe-C system, cast iron can be
defined as an iron-carbon alloy with more than 2% C. Important notice is that
silicon and other alloying elements may considerably change the maximum
solubility of carbon in austenite. The formation of stable or metastable eutectic is
a function of many factors including the nucleation potential of the liquid,
chemical composition, and cooling rate. The first two factors determine the
graphitization potential of the iron. A high graphitization potential will result in
irons with graphite as the rich carbon phase, while a low graphitization potential
will result in irons with iron carbide. The metastable phase amount has both direct
and indirect effects on the properties of ductile iron castings. Increasing the
volume percent of hard, brittle carbide increases the yield strength, but reduces the
tensile strength and elongation, of ductile iron castings. Because there is no
graphite expansion for the metastable phase, the formation of carbide increases the
likelihood of internal casting porosity.
The two basic types of eutectics - the stable austenite-graphite or the metastable
austenite-iron carbide (Fe3C) - have wide differences in their mechanical
properties, such as strength, hardness, toughness, and ductility. Therefore, the
basic scope of the metallurgical processing of cast iron is to manipulate the type,
amount, and morphology of the eutectic in order to achieve the desired mechanical
properties.
The structure of the matrix is essentially determined by the cooling rate through
the eutectoid temperature range. Slow cooling rates prevalent in heavy sections
promote the transformation of ferrite.
Classification
Historically, the first classification of cast iron was based on its fracture. Two
types of iron were initially recognized:
Version 2006.0
Page 477
With the advent of metallography, and as the body of knowledge pertinent to cast
iron increased, other classifications based on microstructural features became
possible:
Graphite shape: Lamellar (flake) graphite (FG) as shown in Fig. 4.1 and 4.2,
spheroidal (nodular) graphite (SG) as shown in Fig. 5.1 and 5.2, compacted
(vermicular) graphite (CG), and temper graphite (TG); temper graphite results
from a solid-state reaction (malleabilization.)
Matrix: Ferritic, pearlitic, austenitic, martensitic, bainitic (austempered).
Version 2006.0
Page 478
Graphite is a hexagonal-close pack form of carbon that can grow in both the liquid
and solid forms of iron. In theory, in irons above the eutectic composition of
carbon, the graphite first nucleates in the liquid, and then continues to grow in the
solid. In irons below the eutectic composition, the carbon does not start to grow
until the iron reaches eutectic temperature. As seen in a micro, the larger nodules
are from growth initiated in the liquid, and the smaller nodules are from growth
that does not start until solidification temperatures are reached. The graphite
nodules continue to grow as the iron cools, so the amount of growth that occurs in
the liquid is smaller than what would be assumed by the micro.
The expansion from the graphite that grows in the liquid pushes liquid back into
the riser, and does not offset shrinkage. So in order to minimize shrinkage, it is
necessary to maximize the late formation of graphite without having to reduce the
actual amount of graphite. Understanding what happens in a non-steady state
solidification of Ductile Iron suggests a way that this can be done.
In a hypoeutectic mode of solidification, austenite forms as a solid with a lower
than average carbon content. This increases the carbon content of the remaining
liquid until it reaches the eutectic composition. Likewise, in a hypereutectic mode
of solidification, graphite nodules form in the liquid, removing carbon from the
liquid until it is reduced to the eutectic composition
Version 2006.0
Page 479
Version 2006.0
Page 480
2. Eutectic Model
This model can be applicable to both regular and irregular eutectics. In the case of
regular eutectics, growth of both phases of the eutectic structure is non-faceted in
nature. For irregular eutectic, the growth process of one of the phases is faceted.
Growth of the faceting phase requires considerably higher entropy of fusion.
Examples of faceted growth are graphite growth in stable austenite/graphite
eutectic and Silicon in Al-Si eutectic. The metastable austenite/cementite eutectic
is an example of non-faceted/non-faceted type eutectic growth. Growth of both the
stable and metastable eutectic are addressed here. Growth of the stable eutectic
usually proceeds at a higher temperature. A higher cooling rate results in the
formation of a metastable eutectic. This model assumes bulk heterogeneous
nucleation at foreign sites which are already present within melt or intentionally
added to the melt by inoculation.
The nucleation of the eutectic grains is based upon the model proposed in 1966 by
Oldfield (ASM Transaction, 598, (1966), 945). The number of nuclei is a power
law (Oldfield proposed a quadratic law) of the undercooling. The growth of the
grains is controlled by thermal undercooling at the solid/liquid interface. Solutal
undercooling is neglected here since solute diffusion during eutectic solidification
is negligible. The thermal undercooling is given by the difference between the
eutectic temperature and the actual solid/liquid interface temperature.
Version 2006.0
Page 481
This model is a special case of eutectic growth model and is applicable to cast
gray/white iron only. In cast iron, one may obtain both gray and white iron
depending on the melt composition and cooling conditions. Given a controlled
melt composition, the most important factor that will determine whether a given
region will solidify as white or gray is the cooling rate. It has been observed that
for a specific melt composition and solidification condition, there exists a
parameter called a critical cooling rate. If a region of a casting solidifies with a
cooling rate higher than the critical cooling rate, then it will be white. The reverse
is the case for gray iron.
Version 2006.0
Page 482
Version 2006.0
Page 483
Graphite precipitation:
One more micro run parameter for iron alloys solidification simulation is added
called graphite precipitation (GRAPHITE in p.dat) which tells the degree of
graphite precipitation during solidification. It varies from 0 to 1. 1 means that the
graphite expansion potential is completely considered in the simulation so the
casting will have a relatively low tendency of shrinkage. 0 means that the graphite
expansion does not occur hence there is no compensation for the shrinkage of the
liquid during solidification by graphite expansion. During the micro calculation,
the computed expansion part of the density (as a function of the phases present) is
multiplied by GRAPHITE. If GRAPHITE = 0, there will be no expansion,
whereas with a value of 1, the full expansion contribution will be taken into
account in the density. This density is only used if a porosity calculation (POROS
= 1) is made during the microstructure calculation (see the "SGI Porosity model"
section for more details). The value of GRAPHITE will not affect at all the
computed microstructure.
The default value is set to 1.0. This parameter is used to adjust the porosity
formation to the real foundry condition.
Case studies
Version 2006.0
Page 484
1. Al 4.9wt%Si
For this alloy, as the temperature cools down, the primary phase forms first. The
possibility to have and the amount of eutectic phase depends on the cooling rate.
With faster cooling rate, there is less amount of primary phase but more eutectic
phase. Fig. 6 shows the comparison of current calculation with some experiment
and other modeling results for the solidification of this alloy.
Version 2006.0
Page 485
Version 2006.0
Page 486
4) Fraction of eutectic
Version 2006.0
Page 487
Version 2006.0
Page 488
1)
Nodule counts
2)
Austenite radius
3)
Graphite radius
Version 2006.0
Page 489
4)
Pearlite Spacing
5)
Fraction of Ferrite
Version 2006.0
Page 490
6)
Fraction of Pearlite
7)
Fraction of Eutectic
Version 2006.0
Page 491
8)
9)
Tensile Strength
Version 2006.0
Page 492
10)
Yield Strength
11)
Elongation
Version 2006.0
Page 493
12)
Hardness
We can see that it is not 100% stable phase (Eutectic) every where for this casting
after solidification. On the higher cooling area (right side), the metastable phase
(ledeburite) formed. The fraction of stable phase plus the fraction of metastable
phase is 1. The hard brittle metastable phase can form at higher cooling area
when the inoculation is not good enough. The metastable phase can increase the
yield strength but reduce the tensile strength of the ductile iron castings.
Version 2006.0
Page 494
Mechanical properties of the cast grey iron part are a function of the stable and
metastable eutectic volume fractions and grain sizes. FRACTION OF EUTECTIC
gives the amount of the gray eutectic, whereas FRACTION OF METASTABLE
EUTECTIC gives the amount of the white eutectic. In most cases, the gray
structure is more desirable as it gives improved tensile strength and ductility. The
EUTECTIC GRAIN RADIUS parameter gives the gray eutectic grain radius. The
INTER LAMELLAR SPACING parameter calculates the spacing of the gray
eutectic.
1) Fraction of Primary phase
2) Fraction of Eutectic
Version 2006.0
Page 495
Version 2006.0
Page 496
6) Fraction of Ferrite
Version 2006.0
Page 497
7) Fraction of Pearlite
8) Pearlite Spacing
Version 2006.0
Page 498
9) Tensile Strength
Version 2006.0
Page 499
Version 2006.0
Page 500
resulting from this reaction. It is important to know the amount of the phase
formed through this reaction, as it usually forms as a surface layer on the primary
dendritic solid phase. As temperature cools down, pro-eutectoid ferrite or
cementite will form depends on the composition until when the temperature
reaches to the eutectoid temperature. Below the eutectoid temperature, the
pearlite forms. FRACTION OF PROEUTECTOID PHASE refers to the fraction
of proeutectoid ferrite or cementite formed from the austenite phase as a function
of time. The carbon equivalent value controls which type of the proeutectoid
phase (ferrite or cementite) will form. If the carbon content is less than the
eutectoid composition 0.8%, the proeutectoid phase would be ferrite phase,
otherwise would be cementite phase.
1)
2)
Version 2006.0
Page 501
4) Fraction of peritectic
Version 2006.0
Page 502
5) Fraction of eutectoid
6) Fraction of proeutectoid
Version 2006.0
Page 503
Version 2006.0
Page 504
"2-D"
INTRODUCTION
The current version of ProCAST does not give access anymore to 2-D modeling.
However, as it may be useful to run quick tests or for inverse modeling, "pseudo"
2-D capabilities were introduced.
The principle is that 2-D calculations are replaced by 3-D calculations on one
layer of elements. To do so, the 2-D mesh generator was modified in order to
automatically create this layer of element, in order to reproduce either a 2-D
cartesian situation or a 2-D axi-symmetric case.
On the left (see figure below), the original 2-D mesh is transformed in a "one
layer" 3-D mesh (right) for cartesian calculation (only a small thickness is
modeled).
On the left (see figure below), the original 2-D mesh is transformed in a "one
layer" 3-D mesh (right) for axisymmetric calculation (only a small sector is
modeled).
Version 2006.0
Page 505
Version 2006.0
Page 506
MESHING
To call the 2-D mesh generator, one should do a RIGHT mouse click on the
"MeshCAST" menu in the manager and the following sub-menu will appear
One can then select between 2-D or 3-D meshing. If MeshCAST-2D is selected, a
new windows opens with the 2-D meshing interface. This interface was adapted
from the 2-D mesh generator of CALCOSOFT. It allows to mix triangles and
quadrangles in the same mesh and to control quite well the size and the shape of
the mesh.
A brief description of the 2D mesh generator is made hereafter.
Firstly, the MeshCAST-2D window opens and the user can select between a new
project (i.e. a new mesh) or to load a previous project.
Version 2006.0
Page 507
Then, the drawing area should be defined. The units should be selected (1), and
the size of the drawing area must be defined (2). A grid is superimposed to the
drawing area (to help for the drawing). Its characteristics can be defined (3).
Finally, the type of problem (cartesian or axisymmetric) has to be selected (4).
The "Continue" button should be used to go to the next window (5).
Version 2006.0
Page 508
The drawing can start at this stage. The type of entity to draw should be selected
with the corresponding icon (1). Then, this entity can be drawn in the graphics
area (2). One can change between the select and deselect mode with the
corresponding button (3). The check box in-between is to activate the select or
deselect all mode.
One should notice that the intersections should be defined by the user, with the left
icon in (4). One should select the two lines which are intersecting and then press
this icon.
Once the drawing is finished, the domains definition can be done in the next
window, after pressing the "Continue" button (5).
Version 2006.0
Page 509
Before meshing, the different domains should be defined. One should first "Add"
a new domain (1) - the corresponding domain number will appear in the list (2).
Then, the contour of the domain should be selected in the graphics area (3). This
operation should be repeated for each domain. Then, one can go in the "Meshing"
window with the "Continue" button (4).
Version 2006.0
Page 510
The different domains are listed in (1). This mesh generator allows to define both
structured meshes (quadrangles) or free triangular meshes (2). Firstly, one should
select the domain to be meshed (1). Please note that structured meshes should be
meshed first. Then the type of mesh should be selected (2). If a structured mesh is
selected, one should define the four faces defining this domain (3). Then, the
number of elements along each of the four faces should be defined (3). Please note
that it is possible to have more than one line to define a face. Finally, the mesh
button can be pressed (4) to generate the mesh.
In the case of "unstructured" mesh (i.e. triangles), the operation is more simple as
one should just select the domain to be meshed (1), select the "triangle" mode (2),
define the average mesh size (3) and mesh (4).
Once all the meshes are generated, one can "Continue" to the next window.
Version 2006.0
Page 511
The last operation which can be done is the grouping of different domains
("combine"). This operation is optional. One should highlight all the domains
which should be grouped into one (1) and then the "Combine" button should be
pressed (2).
Once this is done, it is possible to save the mesh (3). At this stage, automatically,
the 2D mesh is extruded into a 3D one.
Version 2006.0
Page 512
In the case of axisymmetric cases, it is important to make sure that the axis of
symmetry is well at y = 0. Due to round-off errors, one may have slightly negative
coordinated of the nodes which are on the axis. In this case, the "Tools/Check
Axisymmetric" menu should be activated.
Version 2006.0
Page 513
CASE SETTING
After meshing, the "2-D" extruded mesh can be loaded in PreCAST.
If the "cartesian" option was selected, the mesh will look like that :
If the "axisymmetric" option was selected, the mesh will look like that :
Version 2006.0
Page 514
Such cases should be set-up exactly in the same way as 3-D cases. One should just
take care of the following points :
A symmetry boundary condition should be set on the two large faces of the model,
as shown hereafter.
Version 2006.0
Page 515
Version 2006.0
Page 516
CAFE-3D
Pre-processing
The set-up and the use of the CAFE-3D module is fully described in the
CALCOSOFT-3D manual (this manual is available in PDF format in the
ProCAST installation, in the dat/manuals/pdf directory). This section is
summarizing the specific aspects of the CAFE model related with ProCAST.
The CAFE model can be used in 3 different modes :
"post-processing" mode
weak coupling mode
full coupling mode
In all cases, the ProCAST thermal case should be set as usual in PreCAST. In
addition to the standard set-up, one should specify in PreCAST the "nucleation"
boundary conditions. To do so, in the "Boundary Conditions/Assign surface"
menu, one should "Add" nucleation BC's in the list (one entry per type of
nucleation BC).
Version 2006.0
Page 517
Then, the different surfaces of the casting should be selected and stored in the
corresponding nucleation entry (as for any other BC). Please note that no data
should be allocated to these nucleation entries, as this will be done in the separate
CAFE pre-processor.
If the "full" or "weak" coupling mode is used (for the CAFE calculation), the two
following run parameters should be set as :
THERMAL = 2
MICRO = 1024
Version 2006.0
Page 518
Version 2006.0
Page 519
Before starting the CAFE set-up (in the CALCOSOFT-3D CAFE pre-processor),
one should run first DataCAST. This will create the unformatted mesh file
(prefixg.unf), which is necessary for the loading into the CAFE pre-processor.
Then, the CAFE-3D pre-processor can be started. To launch it, one should make a
right click in the Manager and select the "CAFE pre-processor".
Please refer to the CAFE-3D manual for details about the CAFE-3D set-up.
Version 2006.0
Page 520
The CAFE-3D pre-processor will create all the necessary inputs for the
microstructure calculation. All these inputs are stored in an ASCII file called
"prefix_mica.d". This file can also be edited by hand.
Solver
To run the CAFE calculation, one should press the "ProCAST" button of the
Manager which will open the following window :
Version 2006.0
Page 521
To use the Full or weak coupling mode, the standard ProCAST executable must
be launched (one can in this case either use the Standard Run button or the one
under "CAFE Solver (full-weak coupling mode)". The CAFE will be activated as
the Run parameter MICRO=1024 is defined. The distinction between weak and
full coupling is defined by the parameter "nheat" (card 70) in the prefix_mica.d
input file (weak coupling : nheat = 0 / full coupling : nheat = 1).
Post-processing
To view the CAFE results, the CALCOSOFT-3D post-processor must be used. It
can be accessed directly from the ProCAST Manager with a right click on the
"ViewCAST" button.
Version 2006.0
Page 522
Version 2006.0
Page 523
INVERSE MODELING
INTRODUCTION
The Inverse Modeling module of Procast allows to determine, from simple and
well controlled temperature measurements, thermophysical data or boundary
conditions.
The principle of direct and inverse modeling is presented in the figure hereafter.
Direct calculation
In a standard Procast calculation (i.e. a "direct" calculation), one has the following
inputs and outputs :
Inputs :
geometry
thermophysical properties
initial conditions
boundary conditions
Outputs :
Version 2006.0
thermal history
Page 524
Inverse calculation
The principle of inverse modeling is to provide the thermal history as an input (i.e.
measured cooling curves at several locations) and to have one of the four inputs
described above as an output (i.e. unknown).
Thus, two possibilities are proposed in the Procast inverse module :
a) Thermophysical calculation
geometry
initial conditions
boundary conditions
thermal history
Outputs :
thermophysical properties
geometry
thermophysical properties
initial conditions
thermal history
Outputs :
boundary conditions
Version 2006.0
Page 525
Version 2006.0
Page 526
MODEL SET-UP
Before running any inverse calculation, first, the "direct" problem has to be set-up.
For the setting of a direct ProCAST calculation, please refer to the Pre-processing
chapter.
Once the direct problem is set-up, the file prefixd.dat contains all the information
about :
geometry (mesh)
thermophysical properties of materials
interface heat transfers
boundary conditions
initial conditions
run parameters
The inverse modeling set-up intend to specify on which data the inverse
calculation will be done and to configure the run parameters for the calculation.
PreCAST allows to set-up the inverse problem in an interactive manner (with the
"inverse" menu), in order to create the prefixid.dat file.
One have first the choice to select among a "Material Properties" problem or a
"Boundary Conditions" problem.
Material Properties
If one wants to run a "Material Properties" inverse calculation, the following panel
should be activated (with the "Activate" check box) and configured.
For each material present in the model, the user has to decide whether he would
like to run an inverse calculation on the Specific heat (Cp), the thermal
conductivity (K) and/or the latent heat (L).
Version 2006.0
Page 527
By default, all the properties are deactivated ("N" stands for No). By clicking on
the "green N", it will toggle to "Y" (for Yes). In the example below, it is intended
to run an inverse calculation on the Aluminium alloy specific heat and latent heat
(no inverse calculation on the thermal conductivity).
Please note that one should run an inverse calculation on one material domain
only.
If the property (like the specific heat) is defined in the material property file as a
constant, the inverse calculation will be done on this constant value. If the value is
defined as a temperature function, the inverse calculation will be performed for
each temperature of this function.
Important
Boundary conditions
Version 2006.0
Page 528
If the property (like the interface heat transfer coefficient) is defined in the
database as a constant, the inverse calculation will be done on this constant value.
If the value is defined as a temperature (or time) function, the inverse calculation
will be performed for each temperature (time) of this function.
Inverse settings
Finally, some setting should be defined.
The node numbers, corresponding to the measurement points must be specified. It
is important to note that the node numbers should be determined in ViewCAST,
on results (as the creation of interfaces in PreCAST, after MeshCAST is adding
new nodes and the optimization of PreCAST is renumbering the nodes).
Version 2006.0
Page 529
Moreover, the order with which the nodes are entered into the list should be the
same as the order of storage of the nodes in the prefixim.dat file.
If the measurements are in different units than C, one should specify it with the
"Measurement Units" setting.
The weighting coefficient for temperature should be kept small in order to have a
good convergence. A value of O.1 C has always given good results.
Default value : 0.1
conv
: Convergence tolerance
The convergence will be reached when the relative variation, between two
iterations, of each property will be smaller than the value, conv.
Default value : 0.05
Version 2006.0
Page 530
varb
During an iteration, the beta values will be perturbed one after the others in order
to determine the sensitivity coefficients of each property. To do so, each beta
value will be changed by a given amount corresponding to the value of varb
multiplied by the beta value. Values of varb between 0.05 - 0.2 are convenient and
correspond to a variation of 5 to 20% of the beta values
Default value : 0.1
itermax
: maximum number of iterations before the stop of
the calculation
unit code for the temperature or time "tbeta" used for temperature-dependant or
time-dependant properties
ibunit :
beta value (unknown property), initial guess the initial guess corresponds to the
value initially defined in the HEAT boundary condition of the direct calculation.
ibeta :
flag to indicate what is the type of the beta value (constant, time- or temperaturedependant) :
1 : The type of the beta value is CONSTANT
2 : The type of the beta value is TEMPERATURE dependant
3 : The type of the beta value is TIME dependant
weight :
Version 2006.0
Page 531
maximum variation of the beta value between two iterations (i.e. if relax=3, the
beta value can increase at maximum by 3 times the previous value or can be no
smaller than 1/3 of its previous value within the same iteration).
PreCAST proposes default values that may be changed by the user in the
prefixid.dat file (by editing the prefixid.dat file).
The default values are :
ibeta :
This value prompts the time or temperature value of the corresponding beta value.
This value can not be changed by the user. If ibeta = 1 (CONSTANT type). a zero
value is written.
beta :
This value corresponds to the initial guess. By default, it is the value which was in
the corresponding dataset.
weight :
By default this value is set equal to initial guess of the beta value. This value
should be of the same order of magnitude than the final beta value or larger (for a
good convergence).
betamin :
This value is the minimum allowed value for beta. It should be set to the physical
minimum that can be expected. By default, betamin is set to 1/100 of beta
betamax :
This value is the maximum allowed value for beta. It should be set to the physical
maximum that can be expected. By default, betamax is set to 100 times beta
relax :
Version 2006.0
Page 532
Version 2006.0
Page 533
INVERSE RUN
In order to run an inverse calculation, the following files should be present :
Then, the calculation can be started. To do so, a right click on the "ProCAST"
button on the manager is opening a sub-menu, which allows to start the inverse
solver.
During the execution of the calculation, several files are created, beside the usual
ProCAST files :
prefixir.dat
This file contains the evolutions of the beta values along the iterations, as well as
the evolution of the residual.
prefix.stat
This file contains information of the status of the calculation (see
below)prefixic.dat
Version 2006.0
Page 534
This file contains the calculated temperature evolutions (of the last calculated
iteration) corresponding to the locations of the measured curves.
prefix.tmp
This file is a temporary file created during the calculation. This file is needed only
during the calculation and is useless for the user. This file is automatically deleted
at the end of the calculation However, if the calculation is interrupted by the user
before completion, it should be deleted manually by the user.
prefix.list
When the calculation is run in the background mode, the printout which appears
normally on the screen is listed in this file. This file is not written in a continuous
mode, but only when the buffer of the machine is unloaded (thus, this may take a
few minutes after the beginning of the calculation before having something in this
file).
Monitoring the calculation :
As inverse calculation could be quite long and are often run in "background
mode", there are three ways to monitor the status of the calculation. The
monitoring of the calculation is a good way to check whether the calculation is
stable and whether it is converging in good conditions.
The three monitoring options are the following :
a) Status of the calculation :
the following command :
invstat
prefix
prints the status of the calculation (it is equivalent to "prostat" in Procast). The
number of iterations already calculated is printed as well as the beta values of the
last iteration. This can also be accessed from the manager with a right click on the
"Status" button.
Version 2006.0
Page 535
The residual of the current iteration is also printed. The residual is the average
temperature difference between the measured and the calculated curves (average
on all the steps and all the curves, at the given iteration). The residual will
decrease towards zero along the iterations. The monitoring of the residual is a
good way to see how the convergence is reached.
In addition, it is mentioned if the calculation is still under progress or if it is
completed (if the calculation was interrupted by the user, it will still display that
the calculation is under progress).
This command should be called in the current directory where the prefix* files are
present (this is not like "prostat" which can be called from any directory).
b) Evolution of the beta values
The beta values obtained at each iteration are stored in the file prefixir.dat Thus,
during the calculation, it is possible to visualize the evolution of each beta value
(from the first to the current iteration) and see whether the calculation is
converging or not. The format of the file prefixir.dat is the same than the one of
the measurement. Curve "i" of the file prefixir.dat correspond the ith beta value.
In addition, the residual is stored as the last curve of this file. Thus, it is possible
to graph the evolution of the residual as function of the iterations.
c) Comparison between the measurements and the calculated curves
The goal of inverse modeling is to find the best set of beta values with which the
calculated temperature curves will match the measured ones.
At the end of each iteration, the calculated curves corresponding to the location of
the measurement points are stored in a file labeled prefixic.dat This file is updated
at each iteration. Thus, it is possible to visualize during the calculation (with
ViewCAST) the superimposition of the measured curves and the calculated ones.
Of course, during the first iteration, the match will be bad, however, at the end of
the calculation, the match should be quite good.
Final result
At the end of the calculation, the final beta values correspond to the values of the
last iteration in the file prefixir.dat.
At the end of the calculation, it is advisable to check the quality of the
convergence. To do so, the file prefixir.dat can be graphed (see section
"Monitoring the calculation" above). The convergence is good if the beta values
are not changing much during the last iterations and if the residual is small (the
residual has the unit of degrees and a residual of "1" means that the average
difference between the measured and calculated curves is 1 degree).
Version 2006.0
Page 536
Version 2006.0
number
number
number
number
of beta values :
100
time-steps for the measurements : 500
of thermocouples :
20
of iterations :
100
Page 537
FILE FORMATS
Inverse settings file
All the inverse settings are stored in the file prefixid.dat. Like the file prefixp.dat
the user may change this file "by hand" with an editor and then run again inverse.
The format of this inverse settings file is described hereafter.
Definitions :
ntcunit :
node number in the mesh used for the calculation corresponding to the
measurement points (ntccode=1) (the order of the node numbers should
correspond to the order of the cooling curves in the file prefixim.dat)
imat :
Version 2006.0
Page 538
flag to indicate if the beta value will be taken into account in the inverse
calculation
1: the beta value is an unknown of the problem and will be taken into account in
the calculation (it is a constant value)
2: the beta value is an unknown of the problem and will be taken into account in
the calculation (it is a temperature-dependant value)
3: the beta value is an unknown of the problem and will be taken into account in
the calculation (it is a time-dependant value)
weight(i) :
iterations
All the parameters and variables defined above are defined in groups (or families),
which are labeled by keywords. The keywords are :
MEASUREMENTS
PARAMETERS
THERMAL_CONDUCTIVITY
SPECIFIC_HEAT
LATENT_HEAT
INTERFACE_HEAT_TRANSFER
EXTERNAL_FILM_COEFFICIENT
EXTERNAL_HEAT_FLUX
EXTERNAL_EMISSIVITY
Format :
MEASUREMENTS
ntcunit
Loop on the thermocouples (i)
ntccode(i)
if ntccode(i)=0
imat(i)
ixyzunit(i)
x(i)
y(i)
z(i)
else if ntccode(i)=1
node
endif
End Loop on the thermocouples
PARAMETERS
tau
sigma
Version 2006.0
[integer]
[integer]
[integer]
[integer]
[real]
[real]
[real]
[integer]
[real]
[real]
Page 539
conv
varb
itermax
[real]
[real]
[integer]
For all the following keywords, the structure of data is the same. Only the
necessary keywords should be specified, which correspond to the desired beta
values.
THERMAL_CONDUCTIVITY
SPECIFIC_HEAT
LATENT_HEAT
INTERFACE_HEAT_TRANSFER
EXTERNAL_FILM_COEFFICIENT
EXTERNAL_HEAT_FLUX
EXTERNAL_EMISSIVITY
Loop on the beta values (i)
ibindex(i)
[integer]
itunit(i)
[integer]
ibunit(i)
[integer]
tbeta(i)
[real]
beta(i)
[real]
ibeta(i)
[integer]
weight(i)
[real]
betamin(i)
[real]
betamax(i)
[real]
relax(i)
[real]
End Loop on the beta values
As the data are read by a "Free Format Reader", data can be placed in any fashion
(lines, columns,...), as long as the sequence of the data corresponds to the
specified one. However, each line should not exceed 256 characters. Moreover,
comments can be included in the data file after a "#" sign (all the characters
between the "#" sign and the end of the line will be ignored and can be used as
comments).
Version 2006.0
Page 540
x(1),y(1,nx),y(2,nx),...,(ny-1,nx),y(ny,nx)
Description :
nx :
number of curves
node#i :
node number corresponding to the curves. These number can be any integer
numbers (as they are not used by the inverse calculation). The only requirement is
that ny integer values are present. The reason of these values is that it corresponds
to the format which is compatible with PostCAST.
x(j) :
Version 2006.0
Page 541
INVERSE APPENDIX
Inverse model
Version 2006.0
Page 542
Version 2006.0
Page 543
Version 2006.0
Page 544
Usually, within the direct solver, it is only possible to set easily constant initial
temperatures. Thus, it is important to perform an experiment where the initial
temperatures of the different domains are as constant and well known as possible.
If the filling stage is very important (as one can loose a significant amount of
superheat during the filling), one should consider whether the design is
appropriate or if a lower average constant initial temperature can be considered.
c) temperature measurements
As it is not always easy to know well the material properties of the materials
which are used in the real process, it may be advisable to use different materials
for the inverse experiments. For instance if one wants to determine the heat
transfer coefficient at the casting-mould interface in a high pressure die casting
mould (with an aluminium alloy casting and a steel mould), it is possible to cast a
well known Al-Si7%-Mg0.3% in a block made out of a well known steel alloy.
The idea is to use similar materials where the materials properties are better
known, as the heat transfer coefficient will not depend upon slight differences in
material properties. For instance, in the example shown above, it is not important
to take a steel which is resistant at high temperatures (like it is needed for an hpdc
die), as the inverse experiments will be run only a few times (unlike the hpdc die
which will undergo many cycles, under high loads).
d2) boundary conditions (in case of material properties determination)
This point is probably the most critical in inverse modeling when one wants to
determine the thermophysical properties of a material. Unlike case (d1) above
where it is possible to use well known materials, in this case, one has to find a
method to determine the boundary conditions in the experimental set-up. Usually,
one have to run a first inverse model to determine the boundary conditions and
then to run a second inverse calculation for the material properties determination.
Version 2006.0
Page 545
Version 2006.0
Page 546
"useless" and provide not very accurate measurements, they may degrade the final
answer.
Finally, the advantage of modeling first the experiments, before running them, is
also to show how much CPU time will be needed to run one direct calculation.
Thus, it will show whether the set-up is appropriate or not for fast inverse
calculations.
Version 2006.0
Page 547
Figure A1.1
Comparison of the measured and calculated curves obtained
with 1 values of thermal conductivity, 1 specific heat values and 1 latent heat.
Version 2006.0
Page 548
Figure A1.2
Comparison of the measured and calculated curves obtained
with 3 values of thermal conductivity, 4 specific heat values and 1 latent heat.
Figure A1.3
Comparison of the measured and calculated curves obtained
with 6 values of thermal conductivity, 5 specific heat values and 1 latent heat.
Version 2006.0
Page 549
In such a case, the user has to decide the type of dependency that should be used
in the inverse calculation. To illustrate that, one can use the example of the
determination of the equivalent specific heat of an alloy (as shown in Figure
A1.4). One can see that the curve is almost constant above and below liquidus,
whereas it exhibits very strong variations between the liquidus and solidus
In such a case, the user has to define himself the temperature points at which the
unknowns will be calculated. As in the above mentioned example, these points
should be selected in between the solidus and liquidus temperatures (with a
reasonable number of points in order to describe the curve).
Above the liquidus or below the solidus, one can consider constant values or
variable values as desired.
Concerning the number of points to be selected, one should select enough points,
but not too many. Indeed, if too many points are selected, one can obtain
oscillations of the final beta values.
Figure A1.4
Equivalent specific heat obtained by inverse modeling, showing
the tabulation points (tabulated temperatures) that were chosen.
Version 2006.0
Page 550
Are the thermocouples well defined ? Is the order of the list of the
thermocouple corresponding to the order of the measured curves in the
measurement file ?
In case of time- or temperature-dependent beta values, are the tabulation
points appropriate (for instance, if one is looking for a variation between 500
and 700 and the tabulation points are chosen between 600 and 800, it might
not work).
Are they any (or enough) measurements in the range of the tabulation points?
(for instance if one looks for a temperature dependant heat transfer coefficient
between 100 and 800C and that the measurements are only between 400 and
700C, the values below 400 and above 700C will never converge).
Are the initial temperatures of the model corresponding to the ones of the
measurements (or are the initial measured temperatures constant ?).
Is the geometry of the model and the known boundary conditions well defined
?
Are the material properties of the different materials used in the measurements
well known ?
Version 2006.0
Page 551
INPUT-OUTPUT FILES
This section is listing the different files which are created and used by the
different modules of ProCAST.
MeshCAST files
PREFIX.gmrst
PREFIX.sm
PREFIX.stlsm
PREFIX.igs
PREFIX.iges
PREFIX.stp
PREFIX.step
PREFIX.xmt_txt
PREFIX.x_t
PREFIXd.dat
PREFIX_pre_sh.sm
PREFIX.patran
PREFIX.out
PREFIX.ideas
PREFIX.unv
PREFIX.nastran
PREFIX.ansys
PREFIX.stl
PREFIX.bstl
PREFIX.mesh
PREFIX_sub.sm
PREFIX_sub_act.sm
PREFIX.elem
PREFIX.node
PREFIX.wrk
merge
PREFIX.ceg
merge
usr_cmnds
*.gif
mesh.print
PREFIX.psm
IO
IO
O
I
I
I
I
I
I
I
I
I
IO
I
I
I
I
I
I
IO
O
O
O
O
O
Binary
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
IO
ASCII
PreCAST files
PREFIXd.dat
IO
ASCII
General model setup (including the
geometry), also used for Restarts
PREFIXp.dat
IO
ASCII
Run Paramerters file, also used for
Restarts
PREFIXg0.dat
IO
ASCII
Cut-off values per domain for Thixo casting
bc.db
I
ASCII
Boundary conditions database
matl.db
I
ASCII
Material properties database
stress.db
I
ASCII
Stress properties database
intf.db
I
ASCII
Interface heat transfer database
proc.db
I
ASCII
Process database
CT_Al.db
I
ASCII
Al Thermodynamic database
CT_Fe.db
I
ASCII
Fe Thermodynamic database
CT_Ni.db
I
ASCII
Ni Thermodynamic database
CT_Mg.db
I
ASCII
Mg Thermodynamic database
Version 2006.0
Page 552
CT_Ti.db
I
PREFIXt.unf
I
condition
PREFIX.rnm
I
PREFIXd.unf
I
extraction
PREFIXid.dat
IO
PREFIX.out
I
PREFIX.unv
I
PREFIX.mesh
I
predefined_%d_p.dat
default_p.dat
ASCII
Binary
Ti Thermodynamic database
Temperature results for extracting initial
Binary
Binary
Node renumbering
Time step information, used for IC
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
Inverse data
PATRAN mesh neutral file
IDEAS mesh universal file
Meshcast file
Pre-defined p.dat files
Default values of the p.dat file
Version 2006.0
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Page 553
PREFIXtv.unf
PREFIXx.unf
PREFIXy.unf
PREFIXz.unf
PREFIXu.unf
PREFIXv.unf
PREFIXw.unf
PREFIX.vf
PREFIX.view
PREFIX.serr
PREFIXp.out
O
O
O
O
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
ASCII
ASCII
ASCII
ASCII
Turbulent viscosity
X displacements
Y displacements
Z displacements
X velocity
Y velocity
Z Velocity
View factors after symmetrization
View factor groups
View factor row sum errors
ProCAST run information file
Binary
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
O
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Micromodel,
Micromodel,
Micromodel,
Micromodel,
Micromodel,
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Micromodel
Average penetration, penalty number
Pressure relaxation info
Radiation face remapping
Highest filled element (FREESURF=2)
Shrinkage data, obsolete
Slave-master data
Load residual
Thermal load
Last thermal load
Tri-to-quad (Radiation)
Turbulence data, obsolete
Thixocasting viscosity
Temperature gradient at the solidus
Filling BC shut-off times
Version 2006.0
O
O
O
Binary
Binary
Binary
Page 554
PREFIXs3.unf
PREFIXss.unf
PREFIXsp.unf
PREFIXsx.unf
PREFIXsy.unf
PREFIXsz.unf
PREFIXsxy.unf
PREFIXsyz.unf
PREFIXszx.unf
PREFIXepn.unf
PREFIXac.unf
PREFIXac.ntl
PREFIXdas.unf
PREFIXdas.log
PREFIXdas.ntl
PREFIXm.unf
PREFIXm.log
PREFIXm.ntl
PREFIXi.unf
PREFIXi.log
PREFIXi.ntl
PREFIXl.unf
PREFIXl.log
PREFIXl.ntl
PREFIXr.ntl
PREFIXg.ntl
PREFIXf.ntl
PREFIXe.ntl
PREFIXs.ntl
PREFIXt.ntl
PREFIXt.log
PREFIXq.ntl
PREFIXq.log
PREFIXp.ntl
PREFIXp.log
PREFIXv.ntl
PREFIXv.log
PREFIXd.ntl
PREFIXd.log
PREFIX.tt
PREFIX.lv
PREFIX.clip
*.gif
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
Binary
ASCII
ASCII
Binary
ASCII
ASCII
Binary
ASCII
ASCII
Binary
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
ASCII
Miscellaneous
(Some files are created for internal purposes)
PREFIX.cm
PREFIX.mtx
bem.out
PREFIXc.out
debug.out
dbg.out
mem_usage
poros.out
PREFIX.rhs
PREFIX.lhs
PREFIX.sol
PREFIXrf.out
PREFIX.rhs_v
sconv
PREFIXf.out
PREFIX.phs
Version 2006.0
Page 555
PREFIX_rho.dat
PREFIX_k.dat
PREFIX_mu.dat
PREFIX_tdd.dat
PREFIX.err
PREFIXt.out
Version 2006.0
Page 556
Version 2006.0
If the content of the d.dat file was not modified, one should not run
DataCAST and it is possible to run directly ProCAST. Typically,
this happens if the content of the p.dat file only was modified. If
PreCAST is loaded, the d.dat file is not modified only if the Run
Parameters are changed. If anything else is changed in PreCAST, it
is necessary to run first DataCAST and then ProCAST.
When the d.dat file is modified, DataCAST has first to be run and
then ProCAST should be launched.
Page 557
CONVERGENCE PROBLEMS
If a calculation does not converge or has problem to converge (this can be seen in
the p.out file - please refer to the Troubleshooting section for more details), the
following operations can be tried :
The Thermal model does not converge
check the thermal material properties (in particular if an enthalpy
curve is defined, check that the curve is strictly in an ascending
order)
modify the solver parameters from TDMA to CGSQ (with the
CGSQ Run parameter)
Increase the convergence thermal Run parameter CONVT from 1
to 2 (maximum 5)
Modify the mesh in order to have a better quality
The fluid flow model does not converge (velocities of pressures)
modify the solver parameters from TDMA to CGSQ (with the
CGSQ Run parameter)
Increase the convergence flow Run parameter CONVV from 0.05
to 0.1 (maximum 0.5)
Modify the mesh in order to have a better quality
Version 2006.0
Page 558
STRESS CALCULATIONS
Stress calculations might be rather long (in terms of CPU time), as the six
components of the stress tensor have to be computed, in addition to
Temperature, Pressure and the three components of the velocity.
Stress calculations are starting as soon as the fraction of solid is larger than a
critical fraction defined by the Run Parameter CRITFS. By default, CRITFS =
0.5.
Version 2006.0
Page 559
The figure below illustrates the change of the resulting interface heat transfer
coefficient, as a function of the air gap width.
Version 2006.0
Page 560
Version 2006.0
Page 561
STRESS VISUALIZATION
Stresses are rather complex to view, as it corresponds to a tensor (with 6
components).
ProCAST offers different possibilities to visualize these results, in addition to
each of the components of the stress tensor.
Effective Stress
There are several ways to evaluate the local stress state with a single value. The
formula below shows how the Effective Stress in ViewCAST is calculated (e.g. in
one dimension, the effective stress corresponds to the stress itself). To simplify,
one could consider that it corresponds to some kind of "average" value of the
stress. As shown in the equation below, the Effective Stress is always positive and
thus, it does not allow to determine whether the state of stress is in compression or
in tension. To determine that, the other viewing possibilities should be used.
The stress components which are in the square root correspond to the principal
stresses.
Version 2006.0
Page 562
The shear stresses components are highlighted in red in the figure above. The
"maximum shear stress" is defined by the equation below (where the stress
components corresponds to the maximum and the minimum normal stresses).
Principal Stresses
Version 2006.0
Page 563
Considering the stress state of a given unit volume, it is possible to find a unit
cube with a different orientation, where all the shear stresses are zero. Thus, only
the normal stresses are remaining and they are called principal stresses. The
principal stress 1 is the larger value and the principal stress 3 is the smaller value.
Positive values show a tensile state, whereas negative values show a compressive
state.
Version 2006.0
Page 564
TUTORIALS
GENERAL INTRODUCTION
ProCAST is a modular system and allows the coupling of various modules.
Flow Chart
Several steps are necessary in ProCAST in order
Version 2006.0
Page 565
PreCAST is the pre-processor of ProCAST and allows the user to read a finite
element model, to assign material properties to the different model
components, to set the interface and boundary conditions and to define the
initial conditions.
DataCAST compiles the settings and provides the necessary files for the
calculation.
ViewCAST is the post-processor and allows the user to visualize the results as
ProSTAT reports the status of the calculation.
The user is guided through these different operation with the help of the file
manager (see below):
Version 2006.0
Page 566
MeshCAST
Meshcast is the fully automatic 2D and 3D tetrahedral mesh generator of
ProCAST. It provides facilities for CAD reading, geometry quality checks and
repairs. An automatic shell generator is included in MeshCAST and is especially
dedicated to investment casting.
Meshcast is interfaced with various CAD file formats. In addition to ProCAST
files, MeshCAST can read the following
Version 2006.0
Page 567
Usually the user will read a geometry and check its quality (the geometry will be
eventually repaired, simplified or improved). The surface mesh will be generated
and the solid mesh will be generated based on the surface mesh. The quality of the
solid mesh is checked and the solid mesh can be optimized.
The solid mesh is then ready to be read in PreCAST, the pre-processor of
ProCAST.
PreCAST
PreCAST is the pre-processor of ProCAST and allows the user to read a finite
element model, to assign material properties to the different model components, to
set the interface and boundary conditions and to define the initial conditions.
Version 2006.0
Page 568
ProCAST
DataCAST compiles the settings and provides the necessary files for the
calculation.
ProCAST runs the analysis.
ViewCAST
ViewCAST is the post-processor and allows the user to visualize and analyse the
results.
Version 2006.0
Page 569
Version 2006.0
Page 570
RESTART
The goal of this tutorial is to explain how to continue or restart a ProCAST
calculation.
The reason to restart a calculation is multiple:
It is important to remind that the *p.dat file contains run parameters information
as the *d.dat file contains calculation data as the material properties, the mesh
information, the boundary, interface and initial conditions.
A. NATURAL STOP:
Version 2006.0
Page 571
B. HARDWARE PROBLEM:
Version 2006.0
Page 572
AS A SUMMARY:
Version 2006.0
Page 573
Version 2006.0
Page 574
BOUNDARY CONDITIONS
In order to assign the boundary conditions, i.e. all conditions applied to the
external side of the model, the user will proceeded as follow:
Version 2006.0
Page 575
Use (3) the selection/deselection tools in order to select the area of the model
where the corresponding Boundary conditions will be applied. Do not forget to
Store (4) the selection.
Version 2006.0
Page 576
Please note that once you pressed the Store button (4), the surface of the selected
area is displayed (5). This is useful in order to know the filling area and to
calculate the right filling velocity. In the velocity Boundary condition type, you
have a velocity calculator tool based on the filling type and the inlet area:
Assign (6) then the right Boundary condition entry from the database (7).
Version 2006.0
Page 577
The database entry is then indicated in the top right table (DB Entry) of the preprocessor interface.
The Boundary condition database (7) can be checked and enriched. The read
button allows the user to check the content of a Boundary condition. The Add->
button allows the addition of a Boundary condition among the available list. It is
also possible to copy and modify an existing Boundary condition. The use of the
Del button will delete the selected entry. It is possible to sort the Boundary
conditions per Boundary type.
Version 2006.0
Page 578
Please note that some boundary conditions can be constant or time and/or
temperature dependent. In this case one has to be careful as the constant multiplies
the content of the table (if defined) and both tables multiplies each other.
Version 2006.0
Page 579
Version 2006.0
Page 580
SYMMETRY
In order to optimize the calculation cpu time, one should take advantage of the
rotational or mirror symmetries of the geometry to be modeled.
Version 2006.0
Page 581
The axis of symmetry is defined with two points as the coordinates of three points
are necessary to define the plane of symmetry. In order to get the coordinates of
the symmetry, activate the 'Get Co-ord' button and click on the axis of symmetry
nodes or on the nodes of the symmetry plane. The cursor must be in the
corresponding X1, X2 or X3 coordinate input box.
Do not forget to press Apply once the axis or th plane of symmetry is defined.
Version 2006.0
Page 582
TIME STEPS
Version 2006.0
Page 583
Version 2006.0
Page 584
MATERIAL PROPERTIES
The goal of this tutorial is to help the user with the definition of material
properties which are required for the different material types:
mould
sand mould in case of lost foam process
filter
casting for Newtonian flow
casting for non-Newtonian flow (semi-solid infiltration)
exothermic
In case the data are tabulated, a linear interpolation is applied between the defined
point and a horizontal (constant) extrapolation is applied (except for the enthalpy).
Version 2006.0
Page 585
Casting material
In order to perform a heat (with liquid to solid phase change) and fluid analysis in
a casting material, the following material properties are required (two
possibilities):
A.
Thermal:
Thermal conductivity [W/mK]
Density [Kg/m3]
Specific heat [J/KgK]
Latent heat [J/Kg]
Fraction of solid [-]
Solidus and Liquidus temperatures [C]
Fluid:
Viscosity [centipoise]
B.
Thermal:
Version 2006.0
Page 586
Please note that in order to predict porosity, the density of the material must be
temperature dependent as shown below. The density change between solidus and
liquidus temperature will represent the shrinkage fraction during solidification.
Version 2006.0
Page 587
In case the fluid flow behavior is shear rate dependent (non-Newtonian), the user
can describe the viscosity as a function shear rate and temperature with the help of
the Carreau Yasuda relationship:
Version 2006.0
Page 588
For the use of the Power Cut-off viscosity relationship, please see the Thixo
casting section.
Exothermic material
An exothermic material is defined with the following thermal data:
Thermal:
Thermal conductivity [W/mK]
Density [Kg/m3]
Specific heat [J/KgK]
Version 2006.0
Page 589
In order to take into account the energy released by the exothermic reaction, an
exothermic model is available in the material interface.
Select the Exothermic menu and enter:
The burnt fraction as a function of time
The ignition temperature
The exothermic energy
Version 2006.0
Page 590
Filter material
In order to perform a heat and fluid flow analysis in a filter material, the following
material properties are required:
Thermal:
Thermal conductivity [W/mK]
Density [Kg/m3]
Specific heat [J/KgK]
Fluid -> Filter:
Void fraction [-]
Specific Area [1/cm]
Pressure drop (optional)
Mold material
In order to perform a heat analysis in a mold material, the following material
properties are required:
Thermal:
Thermal conductivity [W/mK]
Density [Kg/m3]
Version 2006.0
Page 591
In case of a sand mold for the Lost Foam process, the permeability of sand must
be defined under the Fluid section. The following material properties are thus
required:
Thermal:
Thermal conductivity [W/mK]
Density [Kg/m3]
Specific heat [J/KgK]
Fluid -> Fluid Permeability:
Permeability [cm2]
Thermodynamic databases
Version 2006.0
Page 592
In case of a binary alloy, it is not too difficult to extract the solidification path
from the corresponding phase diagram using a Lever or a Scheil relationship. But
most of the time, industrial casting alloys are composed of multi-element.
Let's compare the solidification behavior of two slightly different alloys in order
to illustrate the importance of having the right data.
Version 2006.0
Page 593
Version 2006.0
Page 594
The minimum of the Gibb's free energy curve (or the linear combination of two
curves) at temperature T will provide the stable phase diagram.
Version 2006.0
Page 595
Version 2006.0
Page 596
(4) click then on apply and (5) select the Scheil or Lever rule.
The available base systems are Al, Fe, Ni, Ti and Mg. The available alloying
elements are described in the "Databases/Thermodynamic Databases" section.
Version 2006.0
Page 597
CYCLING MODELLING
Version 2006.0
Page 598
In PreCAST there are two techniques to model cycling: A. the Standard method
and B. the Die Combo technique. This appears in the Interfaces menu when one
want to Add-> a new interface condition. The Standard or Die Combo technique
choice is here proposed.
The first technique consists of the modeling of the different cycling sequences
with the help of time dependent interface heat transfer coefficients and heat
boundary conditions.
Version 2006.0
Page 599
This classical technique has the advantage of offering all flexibility and is able to
model any situation. However in case of process optimization, it can be quite
tedious to change all tabulated interface heat transfer coefficients and heat
boundary conditions if the cycling timing has to be changed.
The Die Combo technique facilitate the cycling modeling through the use of
dedicated interfaces.
Version 2006.0
Page 600
The different cycle times are directly entered in the Run Parameter Cycles menu.
And the interface heat transfer coefficient and heat boundary conditions are
directly entered in the Die Combo interface.
Version 2006.0
Page 601
Using the Browse button, look for the Cycling (thermal only) calculation and set
the Step number at which the temperature distribution will be extracted.
Version 2006.0
Page 602
Version 2006.0
Page 603
VIRTUAL MOLD
The virtual mold should be used in order to save cpu time. The thermal aspect of
the mold is taken into account (heat diffusion, local heating, heat saturation) but as
the mold is not meshed, the user cannot visualize the temperature distribution in
the mold.
Please note that a zero velocity Boundary condition should be set to the surface of
the casting in case the virtual mold is used.
The virtual mold is defined in the Geometry menu.
Enter the dimensions of the virtual mold. Please make sure that the virtual mold
will include the complete casting geometry. As the external boundary of the
virtual mold is considered as adiabatic (no heat exchange), the size of the virtual
mold should be set large enough. Using the Check Geom option you can have the
external dimension of the geometry. This helps to define the size of the virtual
mold.
Version 2006.0
Page 604
The calculated thermal depth can be displayed using the Show Depth button and
the scale of the thermal depth can be adjusted using the Set Scale button. It is also
possible to remove the virtual mold.
Version 2006.0
Page 605
The mold material is now displayed in the Material menu and material properties
can be assigned to the virtual mold.
Version 2006.0
Page 606
An interface heat transfer coefficient has to be defined between the casting and the
mold.
Version 2006.0
Page 607
Version 2006.0
Page 608
PROCESS TEMPLATES
The goal of these tutorials is to help the user with the typical settings and run
parameters. The coupling of Thermal and Fluid flow modules allows the user to
model any process, i.e. gravity casting, sand casting, high pressure die casting,
low pressure die casting, investment casting ... But each process requires some
specific settings.
Basically the user will set in Precast:
GEOMETRY: a solid mesh
MATERIALS: a casting and mold material
INTERFACE: an interface heat transfer coefficient between casting and mold
BOUNDARY CONDITIONS: a heat and pouring conditions
PROCESS: a gravity vector and moving solids
INITIAL CONDITIONS: initial temperatures and the casting cavity will be set
as empty
RUN PARAMETERS: general, heat, cycling, radiation, flow and turbulence
Please note that the user is able to use predefined templates for the run parameters,
i.e. the parameters for the process of interest can be defined only once at the
beginning.
Version 2006.0
Page 609
HPDC-Cycling
The goal of this tutorial is to show you step by step how to setup a high pressure
die casting model in ProCAST.
The geometry used for this tutorial is composed of a:
casting
top die
bottom die
insert
meshed cooling channels
unmeshed cooling channels
Version 2006.0
Page 610
The goal here is to run first a thermal analysis with several cycles in order to
predict the number of necessary cycles to reach steady state. In a second step, the
temperature distribution of the mold components will be used as initial
temperature for a fluid flow or stress analysis.
In this tutorials both options for the cooling channels are considered: meshed and
not meshed channels. The goal is to illustrate the set-up of both options.
The interfaces will be treated with the Diecombo technique.
Version 2006.0
Page 611
Enter first the case name under the 'Case' field of the file manager and press the
PreCAST menu.
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
MeshCAST file (.mesh) with the corresponding prefix. And if there is no
MeshCAST file, you will have to use the Open file menu of PreCAST in order to
look for the right file.
Version 2006.0
Page 612
Version 2006.0
Page 613
Assign material properties to the casting, mold and inserts. In this case,
Aluminium is assigned to the casting material, Steel H13 to the top and bottom
dies and to the insert and water is assigned to the meshed cooling channels.
Version 2006.0
Page 614
Please note that the temperature of the casting material will be re-initialized before
each cycle as the temperature of the mold components at cycle N-1 will be taken
Version 2006.0
Page 615
into account for cycle N. It is therefore important to set the right Material type for
each component.
Version 2006.0
Page 616
Version 2006.0
Page 617
Version 2006.0
Page 618
Version 2006.0
Page 619
Version 2006.0
Page 620
Version 2006.0
Page 621
Please note that the casting will be 'attached until ejection' to the top die.
Version 2006.0
Page 622
Version 2006.0
Page 623
Air cooling condition will be assigned to the external surface of the mould
Version 2006.0
Page 624
Version 2006.0
Page 625
temperature of 200C. The inner surface of the non-meshed channels are selected
using the propagate selection tool.
Version 2006.0
Page 626
Version 2006.0
Page 627
Version 2006.0
Page 628
Version 2006.0
Page 629
Version 2006.0
Page 630
Version 2006.0
Page 631
Version 2006.0
Page 632
Version 2006.0
Page 633
10
11
Version 2006.0
Page 634
Version 2006.0
Page 635
HPDC-Flow
The goal of this tutorial is to show you step by step how to setup a high pressure
die casting model in ProCAST.
The geometry used for this tutorial is composed of a:
casting
top die
bottom die
insert
meshed cooling channels
unmeshed cooling channels
The goal here was first to run a thermal analysis with several cycles in order to
predict the number of necessary cycles to reach steady state. In a second step, the
temperature distribution of the mold components will be used as initial
temperature for a fluid flow or stress analysis.
In this tutorials we are going to extract the die temperature distribution of HPDCCycling in order to use it as initial temperature.
Version 2006.0
Page 636
Create a new calculation directory and copy from your hpdc-cycling directory the
hpdcd.dat and hpdcp.dat files.
Version 2006.0
Page 637
Version 2006.0
Page 638
Version 2006.0
Page 639
Version 2006.0
Page 640
The temperature distribution can be displayed in order to check that the extract
procedure worked.
Version 2006.0
Page 641
RUN PARAMETERS: General and Thermal are not modified, The number of
Cycles is set to 1 and Flow is set to 3. The frequency for saving the velocity
results (FLOW) is set equal to TFREQ (frequency for saving the temperature
results).
Version 2006.0
Page 642
Version 2006.0
Page 643
Version 2006.0
Page 644
Version 2006.0
Page 645
Version 2006.0
Page 646
LPDC
The goal of this tutorial is to show step by step how to setup a low pressure die
casting model in ProCAST.
Version 2006.0
Page 647
Version 2006.0
Page 648
Enter first the case name under the 'Case' field of the file manager and press the
PreCAST menu.
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
MeshCAST file (.mesh) with the corresponding prefix. And if there is no
MeshCAST file, you will have to use the Open file menu of PreCAST in order to
look for the right file.
Version 2006.0
Page 649
Version 2006.0
Page 650
Version 2006.0
Page 651
Version 2006.0
Page 652
Version 2006.0
Page 653
Assign material properties to the casting, mold and cores. In this case, CuAluminium Bronze is assigned to the casting material, Steel_H13 to both parts of
the dies and Sand_Silica to the core domains.
Version 2006.0
Page 654
Once the interfaces are created, assign a heat transfer value to each interface.
Typical values are 500 W/m2K between casting and cores and 1000 W/m2K
between casting and steel dies.
Version 2006.0
Page 655
Version 2006.0
Page 656
A natural air cooling heat Boundary condition will be applied all around the dies
and a pressure Boundary condition together with a temperature Boundary
condition will be imposed at inlet of the gating system.
Please note that the simulation model might have a different height compared to
the current casting if the full filling column was not taken into account in the
simulation model. In such a case the pressure ramp should be shifted accordingly.
Please also note that the pressure which is applied to the inlet for the filling
sequence will be the difference between the reference pressure which will be
defined in the run parameters and this pressure BC. In this case where a total
pressure of 0.2 bar is applied, a zero reference pressure will be defined in the
FLOW run parameters.
Version 2006.0
Page 657
Version 2006.0
Page 658
In the Preferences menu, select the LPDC filling option. The right and dedicated
run parameters will be automatically set.
Version 2006.0
Page 659
Version 2006.0
Page 660
Version 2006.0
Page 661
10
Version 2006.0
Page 662
11
12
Version 2006.0
Page 663
Version 2006.0
Page 664
Version 2006.0
Page 665
Gravity-Sand
This tutorial will guide you step by step in the ProCAST set-up of a sand casting
simulation.
The geometry used for this tutorial is composed of a:
casting
sand box
cores
Version 2006.0
Page 666
Please enter first the case name under the 'Case' field of the file manager and press
the PreCAST menu.
Version 2006.0
Page 667
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
MeshCAST file (.mesh) with the corresponding prefix. And if there is no
MeshCAST file, you will have to use the Open file menu of PreCAST in order to
look for the right file.
Version 2006.0
Page 668
Version 2006.0
Page 669
Assign material properties to the casting, mold and cores. In this case,
Ductile Iron is assigned to the casting material, Sand silica to the mold and
Steel_H13 to the core domains.
Version 2006.0
Page 670
Version 2006.0
Page 671
Version 2006.0
Page 672
Once the interfaces are created, assign a heat transfer value to each interface.
Typical values are 500 W/m2K between casting and sand and 1000 W/m2K
between casting and steel cores.
Version 2006.0
Page 673
A natural air cooling heat Boundary condition will be applied all around the sand
box, a velocity Boundary condition will be applied and a temperature Boundary
condition will be imposed to some nodes of the top surface of the down sprue.
Version 2006.0
Page 674
Version 2006.0
Page 675
Version 2006.0
Page 676
Version 2006.0
Page 677
In the General run parameters panel, DTMAX is set to 5 seconds in this case.
Version 2006.0
Page 678
Version 2006.0
Page 679
Version 2006.0
Page 680
Do not forget to press Apply in order to validate the Run Parameters selection.
Version 2006.0
Page 681
10
11
Version 2006.0
Page 682
Position the model using the picture manipulation tools (zoom, drag, rotate, ...).
Version 2006.0
Page 683
Version 2006.0
Page 684
Version 2006.0
Page 685
Version 2006.0
Page 686
Isolated pockets of liquid are displayed using the Fraction of Solid cut-off option.
Investment
The goal of this tutorial is to show you step by step how to setup an investment
casting model in ProCAST.
The geometry used for this tutorial is composed of a:
casting
a shell mold
an enclosure
Version 2006.0
Page 687
Remark: It is possible to model the preheating of the shell mold in the furnace
using the same model as the one for fluid flow modeling. The temperature
distribution of the shell mold is then used as initial condition for the fluid flow
calculation using the extract procedure. To do so:
Version 2006.0
Page 688
Please enter first the case name under the 'Case' field of the file manager and press
the PreCAST menu.
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
MeshCAST file (.mesh) with the corresponding prefix. And if there is no
MeshCAST file, you will have to use the Open file menu of PreCAST in order to
look for the right file.
If this Case name is not entered, you will have to go in the File menu of PreCAST
and open the right file.
Version 2006.0
Page 689
Version 2006.0
Page 690
Version 2006.0
Page 691
Two mirrors of symmetry will be considered here. The Get Co-ord button is
useful to pick the node coordinates on the planes of symmetry.
You can check that the symmetry was well applied if the symmetry is set in the
Boundary conditions.
Version 2006.0
Page 692
Version 2006.0
Page 693
Inconel 718 and a shell material are applied to the casting and shell mold.
Version 2006.0
Page 694
Version 2006.0
Page 695
Version 2006.0
Page 696
Remarks:
Version 2006.0
Page 697
Version 2006.0
Page 698
Check that the normals of the enclosure are all pointing inside the enclosure. In
case it is not the case, use the Align or Reverse options.
Version 2006.0
Page 699
Version 2006.0
Page 700
In the Preferences menu, select the Gravity filling option. The right and dedicated
run parameters will be automatically set.
Version 2006.0
Page 701
Version 2006.0
Page 702
Version 2006.0
Page 703
Version 2006.0
Page 704
10
Version 2006.0
Page 705
11
Useful tip:
Version 2006.0
Page 706
12
Version 2006.0
Page 707
Version 2006.0
Page 708
Version 2006.0
Page 709
Prediction of pockets of liquid using the fraction of solid field and the cut-off
option
Filter
This tutorial will guide you step by step in the ProCAST set-up of a sand casting
simulation with a filter material.
The geometry used for this tutorial is composed of a:
casting
filter
chill
A virtual mold will be generated for the sand mold.
Version 2006.0
Page 710
Please enter first the case name under the 'Case' field of the file manager and press
the PreCAST menu.
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
Version 2006.0
Page 711
Version 2006.0
Page 712
Symmetry definition
Version 2006.0
Page 713
In order to define the coordinates of the two points which are defining the axis of
symmetry, press the 'Get Co-ord' button and then click with the mouse on a mesh
points located on the axis of symmetry.
You can check that the symmetry boundary conditions are well set. Go in the
Boundary conditions menu and check that you have the symmetry well defined.
Version 2006.0
Page 714
Generate a virtual mold larger than the model. The virtual mold has to be large
enough as the external boundary of the virtual mold is considered as adiabatic.
Version 2006.0
Page 715
Version 2006.0
Page 716
Version 2006.0
Page 717
In addition to the thermal conductivity, the density and the specific heat, a filter is
defined by its Void Fraction and Surface Area (it is also possible to define the
Pressure Drop properties - optional).
Version 2006.0
Page 718
Assign then heat transfer coefficients between the chill and the casting and
between all components and the virtual mold. Please note that no heat transfer
Version 2006.0
Page 719
coefficient is defined between the casting and the filter. This will be defined in the
Assign Volume Boundary Conditions menu.
Version 2006.0
Page 720
A zero velocity Boundary condition has to be applied all around the casting as no
coincident meshed mold is considered. A velocity Boundary condition and an
imposed temperature are applied to the top of the gating system.
Version 2006.0
Page 721
The interface heat transfer coefficient is defined in the Assign Volume -> Filter
Heat condition.
Version 2006.0
Page 722
Version 2006.0
Page 723
10
Version 2006.0
Page 724
In the Preferences menu, select the Gravity filling option. The right and dedicated
run parameters will be automatically set.
In the General run parameters panel, DTMAX is set to 5 seconds and TSTOP is
set to1000C in this case. This means that the calculation will be stopped as soon
as the casting domain temperature is below 1000C.
Version 2006.0
Page 725
Version 2006.0
Page 726
Version 2006.0
Page 727
Version 2006.0
Page 728
11
Version 2006.0
Page 729
12
13
Version 2006.0
Page 730
Version 2006.0
Page 731
Version 2006.0
Page 732
Version 2006.0
Page 733
Stress
This tutorial will guide you step by step in the ProCAST set-up of a stress model.
The geometry used for this tutorial is composed of a:
casting
die
Version 2006.0
Page 734
Please enter first the case name under the 'Case' field of the file manager and press
the PreCAST menu.
PreCAST will load first a ProCAST file (d.dat) if it is present in the current
directory. If PreCAST does not find any ProCAST file, it will look for a
MeshCAST file (.mesh) with the corresponding prefix. And if there is no
MeshCAST file, you will have to use the Open file menu of PreCAST in order to
look for the right file.
Version 2006.0
Page 735
Version 2006.0
Page 736
Version 2006.0
Page 737
Version 2006.0
Page 738
Assign the Interfaces between the components. Please note that the mesh
was already optimized in MeshCAST and thus the interfaces are already created
(they are already set as COINC).
Version 2006.0
Page 739
Version 2006.0
Page 740
A natural air cooling heat Boundary condition is assigned all around the die and a
zero displacement Boundary condition is applied to the top and bottom surface of
the die.
Version 2006.0
Page 741
Version 2006.0
Page 742
Version 2006.0
Page 743
Version 2006.0
Page 744
Version 2006.0
Page 745
Version 2006.0
Page 746
10
11
Version 2006.0
Page 747
Temperature results:
Version 2006.0
Page 748
Version 2006.0
Page 749
Y-Displacements:
Version 2006.0
Page 750
Non-Coincident model
This tutorial will guide you step by step in the ProCAST set-up of a noncoincident model.
We can use non-coincident meshes in ProCAST. This facilitates the assembly of
the different model components but you need to be careful as in such a case the
interfaces between the components are defined according to a tolerance. If all
components are well positioned and the mesh densities are not too different, the
thermal or flow predicted results will be good. But one should thus run a thermal
calculation first and check if the temperature distribution is correct. If the
interfaces are not good, i.e. not within the tolerance (1/10 of the smallest element
length), then one won't have heat exchange at some interfaces. It is then possible
to change this tolerance value (see the Interfaces assignment section for more
details). The above test procedure has then to be run again.
If you run a fluid flow analysis with a non-coincident model, do not forget to
assign a zero velocity boundary condition all around the casting.
The geometry used for this tutorial is composed of a:
casting
Version 2006.0
Page 751
sand mold
Version 2006.0
Page 752
Version 2006.0
Page 753
Version 2006.0
Page 754
Version 2006.0
Page 755
Version 2006.0
Page 756
In case of a non-coincident model one has to turn the coincident interface type in
non-coincident and add the other non-coincident interfaces. Please note that the
casting is the Master and the Mold is the Slave.
Version 2006.0
Page 757
Version 2006.0
Page 758
Version 2006.0
Page 759
Version 2006.0
Page 760
Version 2006.0
Page 761
Version 2006.0
Page 762
Velocity: zero velocity all around the casting (this necessary in case of a nocoincident mold, in case of virtual mold or in case there is no mold)
Velocity: -y velocity at the top of the downsprue.
Temperature: Pouring temperature of the metal assigned at the top of the
downsprue.
Version 2006.0
Page 763
Version 2006.0
Page 764
Version 2006.0
Page 765
Version 2006.0
Page 766
Version 2006.0
Page 767
Version 2006.0
Page 768
10
Version 2006.0
Page 769
11
Version 2006.0
Page 770