ProCAST User Manual
PROCAST USER MANUAL VERSION 2006.0
Revised version (April 2006) - CL/PRCA/06/02/00/A
Version 2006.0
Page 1
ProCAST User Manual
PROCAST USER MANUAL
1
INTRODUCTION
7
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
ProCAST User Manual
FLUID FLOW & FILLING RADIATION STRESS DATABASES MATERIAL DATABASE MATERIAL PROPERTIES THERMODYNAMIC DATABASES Calculation of Stress Properties Databases limitations Influence of alloying elements INTERFACE DATABASE BOUNDARY CONDITIONS DATABASE PROCESS DATABASE STRESS DATABASE STRESS MODELS AND PROPERTIES Digitized Hardening Plastic and Viscoplastic properties determination RUN PARAMETERS GENERAL RUN PARAMETERS THERMAL RUN PARAMETERS CYCLING RUN PARAMETERS RADIATION RUN PARAMETERS FLOW RUN PARAMETERS TURBULENCE RUN PARAMETERS STRESS RUN PARAMETERS MICRO RUN PARAMETERS PRE-DEFINED RUN PARAMETERS RUN PARAMETERS RECOMMENDATIONS POROSITY MODELS POROS=1 POROS=4 POROS=8 DENSITY DEFINITION ACTIVE FEEDING SGI POROSITY MODEL VIRTUAL MOLD FILTERS EXOTHERMIC CYCLING LOST FOAM THIXO CASTING CENTRIFUGAL CASTING MULTIPLE MESHES AND NON-COINCIDENT MESHES GEOMETRY MANIPULATION MESH OPTIMIZATION USER FUNCTIONS USER FUNCTIONS TEMPLATES External heat transfer coefficient Function External temperature Function Emissivity Function Heat flux Function Interface heat transfer coefficient Function Mass Source Flow Rate Function
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
ProCAST User Manual
Mass Source Vector Function Translation Vector Function Imposed Velocity Vector Function Solid Transport Velocity Vector Function External Function
319 320 321 322 322
RUN OF THE CALCULATION
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
ProCAST User Manual
PARALLEL VERSION INSTALLATION HARDWARE AND OS LAM/MPI AND MPICH COPYRIGHTS
425 428 430
ADVANCED POROSITY CALCULATIONS
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
ProCAST User Manual
TIPS & TRAPS
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
ProCAST User Manual
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
ProCAST User Manual
Mold filling and Fluid flow calculation
Thermomechanical calculation
Version 2006.0
Page 8
ProCAST User Manual
SOFTWARE ORGANIZATION The software is organized around a Manager, which calls the different modules : • • • •
MeshCAST : the mesh generator PreCAST : the pre-processor, coupled with databases DataCAST / ProCAST : the solvers ViewCAST : the post-processor and data export unit
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
ProCAST User Manual
Thermal module • • • • • • • •
Heat conduction (Fourier equation) Latent heat release during solidification Cycling in die casting Sleeves (insulating and exothermic) Non-coincident meshes Solidification time Secondary Dendrite Arm spacing Porosity indicator
Radiation module • • • • •
Net radiation method Full view factors capabilities Mirror and rotational symmetries Relative motion of materials Solid or surface enclosures
Fluid flow 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 • • • • • • •
Elastic, Elastic-plastic, Elasto-visco-plastic Rigid or vacant materials Automatic calculation of the air gap heat transfer Contact algorithm between the different materials Contact pressure Die Fatigue prediction Hot Tearing indicator
The features linked to specific processes (e.g. cycling, tilt, ...) are embedded in the corresponding physical modules.
Version 2006.0
Page 10
ProCAST User Manual
USER MANUAL PRESENTATION After a "Getting started" chapter, which presents briefly the set-up of a simple thermal case, the different modules of the software, starting by the "Software Manager" are presented. The "Pre-processing" chapter is probably the most important as it describes the setting up of a case, from the FEM mesh to the run of the calculation. After an introduction on common features, this chapter is divided according to the "Physical modules", i.e. Thermal, Fluid Flow & Filling, Radiation and Stress. Then, the Databases, Run Parameters and Advanced features are presented. After the Run of the calculation chapter, three chapters are dedicated to the Results viewing, the Results analysis and the Results exports. Finally, Tips & Traps and Tutorials will illustrate the use of the software. MeshCAST, the mesh generator, is presented in a separate manual.
Version 2006.0
Page 11
ProCAST User Manual
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 • •
• • • •
In the "Copy" panel, the possibility to access directly to the "Template" directory of the installation was added (see "Software Manager/File Manager" section for more details). The access to the CAFE pre-processor and post-processor were introduced in the Manager, as well as to the Advanced Porosity post-processor (see the "Software Manager/Module calls" and "Software Manager/Advanced module calls" sections for more details). A bug in the "Copy" functionnality was corrected. The possibility to launch DMP calculations on Linux in Batch was introduced (see the "Parallel Solver/Use of the Parallel solver" section for more details). The FlexLM libraries were upgraded from 9.2 CRO to 10.1.3 STL. A bug in the status window (wrong update of the number of cycles) was corrected.
MeshCAST •
• • •
• • • • •
Version 2006.0
A link between GEOMESH and MeshCAST was created in order to speed-up the repair and surface meshing process (especially to account for the more efficient readers of GEOMESH) (see the MeshCAST manual for more details). Surface mesh Assembly (see the MeshCAST manual for more details). Boolean Assembly (see the MeshCAST manual for more details). The Surface mesh algorithm was improved in order to have a better quality surface mesh. Especially, cylinders defined by only one surface can now be meshed automatically. Thin cylinders (e.g. cooling chanels) are not collapsed anymore when the mesh size is coarse. The Autofix is now fully automatic (no more need to specify a length) A functionnality to remove automatically filets was introduced (see the MeshCAST manual for more details). The Layered shell mesher was re-introduced in MeshCAST (see the MeshCAST manual for more details). The format of the volume mesh has been adapted to allow the handling of files with more than 10 millions elements. When a Volume mesh is loaded, new buttons "Select All" and "Deselect All" were introduced in order to select/deselect automatically all the domains.
Page 12
ProCAST User Manual
• •
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
Time-dependant Inlet boundary conditions were introduced (see the "Boundary Conditions Database" section for more details). In the case of shot piston, it is possible now to define the position of the piston in three different ways : position vs time, velocity vs time, velocity vs position (see the "Process Database" section for more details). Wall BC and Velocity BC are now exclusive (i.e. if a Wall BC with a "Select All" is set, the previously defined Velocity BC's will be preserved). Now, it is not anymore possible to set BC's on Periodic BC faces (the Periodic BC faces are "protected", like Symmetry faces). A new "Pick" function has been introduced, to pick the coordinates and the node number of a given point selected interactively on the model (see the "Geometry Manipulation" section for more details). Stress properties (Youngs modulus, Poisson Ratio, Expansion coefficient) are now automatically calculated, based upon the Thermodynamic databases (see "Calculation of Stress Properties" section for more details). When defining the interfaces, the "Wireframe" mode has been reactivated (in order to see a domain which is inside an other one). 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 Microstructure modelling, the default values of the Run parameters were removed (see the "Micro Run Parameters" section for more details). The mesh can be now optimized also in the case of Periodic BC. If a non-coincident mesh is used, a warning prevents the definition of a virtual mould. New User functions have been introduced (see "User Functions" section for more details). A problem in the initialization of CORE materials during cycling has been corrected. A problem of Extract of temperatures with the prefix_t.unf file larger then 2 GB has been corrected. Enclosures can now again be defined by QUAD elements A new Run Parameter (GATEFS) was introduced for the handling of the third stage pressure in HPDC (see the "Thermal Run Parameters" and "Active Feeding" sections for more details). The set-up of Centrifugal cases was changed for more accuracy and the RELVEL Run Parameter was introduced (see the "Centrifugal casting" and the "Flow Run Parameters" sections for more details). Due to the changes in PreCAST and DataCAST, it is advised to reload the d.dat and p.dat files generated on previous versions and save them before rerunning the cases with the version 2006.0.
Page 13
ProCAST User Manual
•
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).
Thermodynamic databases • • •
Stress properties (Youngs modulus, Poisson Ratio, Expansion coefficient) are now automatically calculated, based upon the Thermodynamic databases (see "Calculation of Stress Properties" section for more details). The latest available Computherm databases are included in this version. The material properties calculated with these databases will be more accurate than the previous ones. 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.
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
ProCAST User Manual
• • •
• • •
The pressures calculated by the "Advanced Porosity module" can now be visualized in ViewCAST (see the "Advanced Porosity Post-processing" section for more details). In order to better visualize pockets of air, a new display mode of the free surface ("Foreground") was added (see the "Results Viewing/Display Parameters" section for more details). Stress results can be viewed also when the directory is in Read-only mode (on Windows only). This is especially useful to view stress cases stored on a CD or a DVD. The stress results are computed and stored in a temporary file on the local disk of the computer (see the "Results Viewing/Field selection" section for more details). If more than one symmetry is defined in the model, it can be automatically retrieved (it was previously working only with one symmetry). In the case of a Tilt casting, it was mandatory to have the p.dat file present in the directory to view the case. This obligation was removed. In the XYPlot window, when nodes were selected interactively, the selection window (with the cursors) was disappearing each time the model had to be rotated or moved. This limitation is now removed.
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".
Version 2006.0
Page 75
ProCAST User Manual
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
ProCAST User Manual
Once the "C:\flexlm\pam_lmd.lic" file is specified, the "Next>" button can be pressed in order to finish the FlexLM installation.
When the "Finish" button is pressed, the software is unlocked and all the modules can be used.
Version 2006.0
Page 77
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
The reading of multiple meshes (for non-coincident meshes or for radiation calculation) is described in the "Advanced features" section of the Pre-processor.
Version 2006.0
Page 83
ProCAST User Manual
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
ProCAST User Manual
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 :
Neg-Jac (negative Jacobian) and Neg-Area correspond to problems in the mesh. These buttons allow to locate where these problems are in the mesh in order to give indication where to modify the mesh in MeshCAST. Volumes gives access to the volume of each material domain, whereas Min-Max indicates the dimensions of the model.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
For non-coincident meshes, it is possible to have access to the non-coincident tolerances with a right click on the NCOINC label. A window will open with the two tolerances.
Boundary conditions assignment
Version 2006.0
Page 95
ProCAST User Manual
After the definition of the interfaces, the Boundary conditions should be specified. This is done in the "Boundary Conditions" menu.
Three types of boundary conditions can be applied : •
• •
Surface boundary conditions ("Assign Surface"), which correspond to all the conditions applied to the outside of the model or the outside of a given material domain. This is the most commonly used type of boundary conditions. Volume conditions ("Assign Volume"), which corresponds to conditions which are applied in a whole volume (e.g. volumetric heat or mass source). Boundary conditions assigned to Enclosures ("Assign Enclosure") in case of radiation problem (see the "Radiation" section for more details).
Assign Surface The principles of Boundary conditions definition are described in the following figure.
Version 2006.0
Page 96
ProCAST User Manual
Firstly, the desired boundary conditions should be "Added" (1) in the upper list (2). With the "Add ->" button, one has the following choices :
The boundary conditions corresponding to Thermal problems are "Symmetry", "Periodic", "Temperature" and "Heat". Then the "location" where the boundary condition should be applied on the geometry should be specified. The selection tools (3) allow to "paint" the desired area on the geometry (4). The values to be assigned to the boundary conditions should be selected in the database (5) and then they are assigned to the corresponding boundary condition (6). See the "boundary conditions database" section for more details about the different type of boundary conditions, as well as the database management. It is possible to have a quick access to the database entries in the following way. When one makes a right click on the boundary condition entry in the upper list (2), the lower list is pointing on this particular boundary condition entry in the lower list (5). The "Selection tools" allow to "Select", "Deselect", "Propagate", "Clip", "Copy" and "Paste" faces or nodes on the geometry.
Version 2006.0
Page 97
ProCAST User Manual
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.
Selection of individual faces or nodes.
Deselection of individual faces or nodes.
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.
Definition of the propagation angle.
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
ProCAST User Manual
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.
Button to clip a model. This allows to perform a selection inside a model, where is it not accessible from the outside.
Button to "Backtrack" the clip.
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.
Paste of selection. The copied selections (above button) can be pasted on a different boundary condition (i.e. on the active boundary condition when the Paste button is pressed). Please note that one can not copy the node selection of a Temperature BC to a Heat BC (as faces are expected).
Assign Volume In the "Assign Volume" menu, the following entities can be assigned :
Version 2006.0
Page 99
ProCAST User Manual
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".
Process conditions assignment For Thermal problems (as well as for flow), the gravity vector should be defined in the "Process" menu.
Gravity
Version 2006.0
Page 100
ProCAST User Manual
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
ProCAST User Manual
Initial conditions assignment The initial temperature of each material should be defined in the initial condition menu.
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
ProCAST User Manual
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
ProCAST User Manual
Run parameters assignment The following Run Parameters should be specified :
Version 2006.0
Page 104
ProCAST User Manual
The thermal module should be activated with THERMAL = 1 or 2. The storage frequency of the thermal results (TFREQ) should be specified, as well as the activation of the porosity model (POROS) . Please refer to the "Thermal Run Parameters" section for the full description of all the Run Parameters.
Version 2006.0
Page 105
ProCAST User Manual
FLUID FLOW & FILLING Fluid flow model The fluid flow module allows to perform mold filling calculation (free surface) as well as fluid flow computation, by the resolution of the Navier-Stokes equation. The typical results which can be obtained are the following : • • • • • •
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
ProCAST User Manual
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
ProCAST User Manual
Boundary conditions menu Concerning the boundary condition assignments, one has to specify the velocity, pressure, wall and/or inlet BC. See the "boundary conditions database" section for more details about the different type of boundary conditions, as well as the database management. See the "Thermal/Boundary Conditions Assignment" section for more details about how to assign boundary conditions. Please note that when a filling-flow calculation is performed on a model where there is no mold, it is necessary to set a zero-velocity boundary condition around the whole mesh (i.e. at all the location where a mold would be in contact with the casting). A zero velocity boundary condition can also be replaced by a WALL boundary condition, which has the same effect.
Process menu For fluid flow problems, the gravity has to be defined (see the "Thermal/Process Conditions Assignment" section for more details).
Initial conditions menu Nothing should be specified in this menu for fluid flow problems.
Run parameters menu The following Run Parameters should be specified :
Version 2006.0
Page 108
ProCAST User Manual
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
ProCAST User Manual
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 : • •
Effect of radiative heat transfer Shadowing effect
Radiation can be either treated as a simple radiative flux (described in the "Thermal/Boundary conditions" section), or with a complex radiation algorithm which takes into account reflexions, obstructions and shadowing effects. The setup of a case with such complex radiation (called hereafter "Radiation with view factors") is explained in this section. As Radiation with view factors involves the calculation of the interaction of the components (casting and mold) with the environment (furnace, castshop, ...), it is necessary to include the environment into the model. This is done with an "Enclosure". If the casting is put into a furnace, the Enclosure is the furnace itself (or the inner skin of the furnace). However, it the casting (and mold) is sitting on the floor of the castshop, one should set an "artificial" enclosure which will surround the casting and which will have the same effect as the environment. An enclosure can be either a solid (represented by a solid 3-D FEM mesh, as the casting, mold, etc...), or by a closed surface (represented by a closed FEM surface mesh). The figure below shows a casting within a solid enclosure (for symmetry reasons, only a sector is modeled).
Version 2006.0
Page 110
ProCAST User Manual
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
ProCAST User Manual
File/Multiple meshes menu When a model contains an enclosure (i.e. a surface mesh which will act as an enclosure), there are two ways to load the corresponding meshes : a) the enclosure is built into MeshCAST, together with the solid mesh. In this case, the enclosure should be "tagged" as an enclosure. b) the enclosure is built separately in MeshCAST, as a conventional surface mesh (without and "enclosure tag"). Then, the solid mesh and surfaces meshes should be loaded into PreCAST using the "File/Multiple meshes" menu. Please note that in previous versions, the surface meshes of the enclosures had to be "tagged" as enclosures. This is not anymore needed.
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
ProCAST User Manual
The full geometry of the two concentric cylinders.
The following figures illustrate the different possibility of simplification by symmetry. •
One mirror (M1)
Two examples of simplified geometry by one mirror.
•
Version 2006.0
Two orthogonal mirrors (M1 and M2)
Page 113
ProCAST User Manual
Two examples of simplified geometry by two orthogonal mirrors.
•
A simple rotation of n sectors (R)
A simplified geometry by a rotation of 8 sectors.
•
Version 2006.0
A rotation of n sectors (R) associated with a mirror (M1)
Page 114
ProCAST User Manual
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)
A simplified geometry by a rotation with 4 sectors and two orthogonal mirrors.
The symmetries are defined in the Geometry/Symmetry menu :
Version 2006.0
Page 115
ProCAST User Manual
In case of a rotation, the axis of rotation should be defined by the coordinates of two points and the number of sectors. Then, the "Rotational" check box should be checked. One should be careful that the mesh should be defined so that the selected number of sectors will not create overlaps of the mesh (i.e. to specify 7 sectors for a mesh which corresponds to a sector of 60°). For mirror symmetries, each plane of symmetry should be defined by the coordinates of three points (which should not be co-linear). Then the "Mirror-1" and "Mirror-2" (if applicable) check box should be checked. If two mirror symmetry are used, one should be careful to make sure that the two planes are orthogonal. In the case of mirror symmetry with a rotation, the axis of rotation should be either perpendicular or parallel to the mirror plane(s). In case of parallel plane(s), the axis of rotation should be within the mirror plane(s). Then, the "Apply" button should be used to validate the symmetry definition. To disable a symmetry, just uncheck the corresponding check box. The "Get Co-ord" button allows to pick nodes of the FEM mesh for an interactive definition of the mirror planes or rotational axis. To use it, first click in the X coordinate box of the point which should be defined interactively, then click on "Get Co-ord" and finally, click on the desired node on the geometry. The corresponding coordinates will fill automatically the corresponding fields. Repeat it for the other points. Please note that one should click very close to the desired node (otherwise, it may be possible that a node "behind" is selected).
Materials and Interface menu
Version 2006.0
Page 116
ProCAST User Manual
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.
Boundary conditions menu When the ambiance or the furnace is modeled with an enclosure, one should specify the temperature and the emissivity of the enclosure in the "Assign Enclosure" menu.
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
ProCAST User Manual
View icon : Viewing of the face orientation
Reverse icon : Reversal of the face orientations
Align icon : Automatic alignment of the face orientations To view the face orientations, one should first select the desired "Enclosure set" and then press the "View" icon. The arrows are drawn for the selected set as shown in the figure hereafter.
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
ProCAST User Manual
Moreover, it allows to define the gravity vector (see the "Thermal/Process Conditions Assignment" section for more details). When "Assign Enclosure" is opened, the following panel appears on the upper right corner of the window. Each enclosure set is displayed and three database entries are possible. The first column corresponds to a "Translation", the second one to "Rotation" and the third one to "Revolution". If a "*" appears, it means that no motion is defined for this set. Otherwise, the number indicates the corresponding database entry.
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
ProCAST User Manual
Run parameters menu The following Run Parameters should be specified :
Version 2006.0
Page 120
ProCAST User Manual
The radiation module is automatically activated when Heat Boundary conditions with View Factors ON are set. Thus, no specific Run Parameter should be activated at this stage. There is however three Run parameters which are important to set if the enclosure is moving. In this case, the update of the view factors will be triggered by the VFTIME or VFDISP parameters. If VFDISP is activated, the Enclosure ID on which the motion will be recorded should be specified. Please refer to the "Radiation Run Parameters" section for the full description of all the Run Parameters.
Version 2006.0
Page 121
ProCAST User Manual
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
These different models are presented in the Databases section.
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
ProCAST User Manual
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
ProCAST User Manual
Interface menu Nothing specific should be defined at the level of the interfaces.
Boundary conditions menu The user can specify three types of stress boundary conditions, as shown in the figure below.
Version 2006.0
Page 124
ProCAST User Manual
The specification of the displacements is very important in order to guarantee that the model will not rotate or move in an unexpected fashion, under the effect of stresses. The figure below is summarizing typical displacement constraints.
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
ProCAST User Manual
One should note that a symmetry is also inducing displacement constraints as shown in the figure below.
It is necessary to specify displacements on the casting domain when the mold is set to "Vacant". However, when the mold is set to "Rigid", it is not necessary to specify displacements on the mold, except if one wants to model the Elastic spring-back when the mold is opened. In the case of die casting, it is recommended to set a zero vertical displacement at all the top and bottom points of the upper and lower dies (black triangles on the figure below), as well as one lateral zero displacement on each side of each die (blue triangle on the figure below).
Version 2006.0
Page 126
ProCAST User Manual
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
ProCAST User Manual
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.
Process Initial conditions menu Nothing should be specified in these menus for stress problems.
Run parameters menu The following Run Parameters should be specified :
Version 2006.0
Page 128
ProCAST User Manual
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.
Advanced stress features Elastic Springback
The stress module of ProCAST allows to simulate the Elastic Springback occurring when a mold is opened. In the figure below (left picture), the mold in blue is closed and it induces some stresses on the "T-bar". When the mold is
Version 2006.0
Page 129
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Constant or by a Table (function of temperature), by clicking on the corresponding radio button. When a Constant is selected, the user can choose the units and enter the constant material property in the white field below "Enter". When a Table is selected, the user can input the temperature-dependant property in the white field on the bottom left of the screen. Each time a new line is filled, the graph on the right is updated. In this case, the units for the temperature and the property can be selected. Above the table, the buttons allow to erase the whole table, erase only the selected line, import or export the table. For imports, the table should be in the form of a text file, with X and Y values on the same line, separated by at least a blank. The number of lines should not exceed 100. The Export format is the same. If a "Table" is defined and the user decides to switch to a "Constant" (or viceversa), a warning will be prompted saying that the table will be lost. A confirmation will then be asked to the user. Of course, the data will be totally lost only when the material is "Stored". All the properties which can be defined as temperature-dependant are organized in the same way. The only thing which changes is the hierarchy of tabs, as shown below for the Fluid/Viscosity/Carreau-Yasuda/Zero Viscosity definition.
Version 2006.0
Page 134
ProCAST User Manual
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
ProCAST User Manual
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
These properties can be either constant or temperature dependant. With phase change (solidification)
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
The fraction of solid curve must be temperature-dependant. It should start at 0.0 at high temperature and increase to 1.0 towards the low temperatures. The fraction of solid should be a strictly descending curve and it should be strictly defined between 0.0 and 1.0. If it is not the case, a warning will be issued. If there is a isothermal transformation (e.g. eutectic plateau), it should be "spread" over an interval of one degree. The latent heat, liquidus and solidus temperatures are defined by constants. Please note that the liquidus and solidus temperatures should be consistent with the fraction of solid curve (no consistency checks are performed). The liquidus and solidus temperatures are used for the porosity models and for the calculation of the permeability of the mushy zone in the case of flow calculations. ProCAST offers an alternative in the definition of the phase change. Instead of defining the specific heat, and the latent heat, one can define the corresponding enthalpy curve. The enthalpy as a function of temperature, H(T), is defined as follows :
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
ProCAST User Manual
specific heat Latent heat
or enthalpy
If this is the case, the following message is displayed :
and the user has to select which definition is preferred. In previous versions (v4.x.x and v3.x.x), there was no check to prevent both definitions. Thus, if a model which was created in a previous versions is loaded into PreCAST v2006.0, the following warning will be displayed (during the load in PreCAST) :
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.
Fluid flow problems
Version 2006.0
Page 137
ProCAST User Manual
For fluid flow problems, it is mandatory to define the viscosity. Then, optionnal definitions are available, such as "Surface Tension", "Permeability" and "Filter". Viscosity
Several viscosity models are available in ProCAST : Newtonian Carreau-Yasuda Power-cutoff
The Carreau-Yasuda model corresponds to Non-Newtonian flow, where the viscosity depends upon the shear rate (see the equation below) :
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 :
The Power-cutoff is used in the case of Thixocasting. Surface tension
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
ProCAST User Manual
Permeability
For a casting material, the permeability is defined by a Karman-Cozeny model, modified by Beckermann at low fraction of solid. The user has also the ability to define its own permeability table, as a function of temperature. In this case, a high permeability corresponds to a "free flow", whereas a low value corresponds to "no flow". For "casting" materials, the permeability is applied only in between the solidus and the liquidus temperatures. For mold materials (in the case of lost foam), a permeability should be defined. In this case, one can define a constant or a temperature dependant permeability. For Filter materials, if the Permeability is defined, it will override the default permeability calculated from the Filter tab. Filter
Filters are characterized by the following properties : Void fraction Surface area Pressure Drop
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
ProCAST User Manual
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
ProCAST User Manual
and the following equation :
where v, ∆P and ∆x are measured values which can be made in a simple experiment (or provided by the Filter supplier). Please note that the Flow rate, v, corresponds to the velocity used in the experiment and not the velocity of your casting model. If both the "Pressure Drop" and "Permeability" tabs are defined, the "Permeability" values are ignored (and replaced by the ones obtained from the above equation). c) Specified permeability
It is also possible to define a given permeability value in the "Permeability tab". In this case, the "Void fraction" is not used for the automatic Karman-Cozeny relationship. See the "Filters" section for more details about the settings of cases with Filters.
Thermodynamic Databases
Version 2006.0
Page 141
ProCAST User Manual
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) :
CompuTherm LLC databases Al database: Ag Al B C Cr Cu Fe Ge Hf Mg Mn Ni Sc Si Sn Sr Ti V Zn Zr Fe database: Al C Co Cr Cu Fe Mg Mn Mo N Nb Ni P S Si Ti V W Ni database: Al B C Co Cr Fe Hf Mo N Nb Ni Re Si Ta Ti W Zr Ti database: Al B C Cr Cu Fe H Mo N Nb Ni O Si Sn Ta Ti V Zr Mg database: Ag Al Ca Ce Cu Fe Gd Li Mg Mn Nd Sc Si Sr Y Zn Zr
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
ProCAST User Manual
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
ProCAST User Manual
"Scheil" and "Lever" correspond to two different microsegregation models. In the case of "Lever", the Lever Rule is applied, which corresponds to a complete mixing of the solute in the solid (i.e. very good diffusion in the solid). On the other hand, the Scheil model corresponds to no diffusion at all in the solid phase (both model consider complete mixing or infinite diffusion in the liquid). The main difference between the two models is the shape of the fraction of solid curve at the end of solidification, as well as the solidus temperature (see figure below).
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
ProCAST User Manual
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
ProCAST User Manual
Density
Enthalpy
Version 2006.0
Page 146
ProCAST User Manual
Fraction of solid
Version 2006.0
Page 147
ProCAST User Manual
Liquidus-Solidus
Viscosity
Version 2006.0
Page 148
ProCAST User Manual
Please note that when a thermodynamic database is used, as the enthalpy is calculated, the specific heat and the latent heat should not be defined (as they are contained in the enthalpy). During the Thermodynamic database calculation, a file named "prefix.phs" is created. It contains for each temperature the phase fractions, as well as the composition of each phase. This information is not needed for a ProCAST calculation, but it can be interesting for other purposes (e.g. growth kinetics calculations).
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
ProCAST User Manual
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.
Calculation of Stress Properties Beside the thermal properties, it is possible to calculate automatically some Stress properties. At this stage, the Young's modulus, the Poisson's ratio and the Thermal expansion coefficient can be calculated based upon the phases obtained from the thermodynamic databases. When the Properties calculation is started in PreCAST with either the Scheil or the Lever model, the following window appears :
Version 2006.0
Page 150
ProCAST User Manual
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
ProCAST User Manual
Young's modulus
Version 2006.0
Page 152
ProCAST User Manual
Poisson's ratio
Thermal expansion coefficient
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
Influence of alloying elements The goal of this section is to illustrate the influence of alloying elements on properties and to show why properties obtain with thermodynamic databases may differ from experimental data. An AlSi9Cu3Fe will be considered to illustrate this example. The usual average chemical composition of such alloy is the following :
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
ProCAST User Manual
Starting from AlSi9Cu3Mg0.3, the other alloying elements are added progressively. The effect on the solid fraction curve is shown in the following figures (please note that the Temperature scale is changing from one graph to the next one):
Version 2006.0
Page 157
ProCAST User Manual
AlSi9Cu3Mg0.3
Version 2006.0
Page 158
ProCAST User Manual
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
ProCAST User Manual
AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15 (Cr is rising the liquidus temperature from 612 to 640°C)
Version 2006.0
Page 160
ProCAST User Manual
AlSi9Cu3Mg0.3Fe1.3Mn0.55Ni0.55Zn1.2Cr0.15Ti0.15 (When Ti is added, the Al3Ti phase appears, with a very high liquidus temperature above 760°C) The following figure are showing the solid fraction curves for all the alloys together.
Version 2006.0
Page 161
ProCAST User Manual
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
ProCAST User Manual
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 588°C 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
ProCAST User Manual
Interface Database The "Interface" database contains two types of entries : Interface Die Combo
"Interface" corresponds to "standard" interface heat transfer coefficients between two different domains, where a "coincident" or "non-coincident" interface has been defined (not applicable in case of perfect contact - EQUIV). "Die Combo" is a specific entry to define a "composite" type of heat transfer in the case of die casting. It allows to define automatically the sequence of closed and opened dies.
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
ProCAST User Manual
tables, please refer to the "Boundary Conditions Database" section). It is also possible to specify that the interface heat transfer coefficient will be defined by a User Function (6). Finally, the data are saved with the "Store" button (7).
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
ProCAST User Manual
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
ProCAST User Manual
Boundary Conditions Database The "Boundary Conditions" database contains the following types of entries in the "Assign Surface" menu : Temperature Heat Velocity Pressure Inlet Turbulence Vent Inject Displacement Point load Surface load
in the "Assign Volume" menu : Volumetric Heat Momentum Source Mass Source
and in the "Assign Enclosure" menu : Emissivity
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
ProCAST User Manual
The following panel corresponds to the "Temperature" database entry.
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
ProCAST User Manual
In order to define a Time-dependant function, one should click on the "Time" button (see 2 in the figure below). This automatically opens a table, as well as a graph. The values of the table can be entered in (3). It is possible to manage the table (4), with "Erase" operations, as well as "Import" or "Export". For the import, the data should be stored in a text file, with the X and Y values on the same line. One can have up to 100 lines (i.e. 100 points). The Export will also create an text file. For both the Import and the Export, it is possible to Browse in order to find the right location of the imported or exported file. Finally, when the table is well defined, one should SAVE it (5), before Storing (6) the database entry.
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
ProCAST User Manual
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
ProCAST User Manual
In the case of "Complex" radiation calculations, one should include "View Factors" (see the "Radiation" section for more details). In order to active the "Radiation" model, the "View Factor" button (see item 4 in the figure above) should be turned ON.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
The position of the domain is defined as a function of time. This position is relative to the original position at step 0. The translation vector is defined by X, Y and Z which is multiplied by the Time function (in order to have the domain moving, it is necessary to define a time function. Otherwise, the domain will remain fixed - no motion). When x(t) is defined, it is possible to define the translation vector with user functions.
Version 2006.0
Page 182
ProCAST User Manual
v(t) - Position vs Time
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
ProCAST User Manual
v(x) - Position vs Position
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
The "speed" of revolution can be defined either as a constant or as a function of time, with the following panel.
Stress Database To access the Stress database, one should to in the Material/Stress menu
Version 2006.0
Page 187
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
All the properties can be defined as Constants or as Temperature dependant tables. The standard input panel is used as for the material database (see the "Material Database" section for more details) :
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
ProCAST User Manual
this material (as the fraction of solid will not be defined, it will be considered as "zero" and no stress will be calculated).
Stress Models and Properties Warning
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.
Linear Elastic model The Elastic model is mainly characterized by the Young's modulus. It corresponds to the slope of the initial part of the stress-strain curve.
Version 2006.0
Page 191
ProCAST User Manual
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.
The thermal expansion can be defined by two different ways : "Thermal Strain" "Secant" thermal expansion coefficient
In the case a Strain curve as a function of Temperature is measured, one can directly input such a curve in the "Thermal Expansion/Strain" tab.
Version 2006.0
Page 192
ProCAST User Manual
However, in the equation (see above), it is the "Secant" thermal expansion coefficient which is used. Thus, one can define "Secant" thermal expansion coefficient directly.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Linear hardening is defined as follows :
whereas Non-linear hardening is defined as :
Version 2006.0
Page 196
ProCAST User Manual
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
ProCAST User Manual
Expansion coefficients) and the Elasto-Plastic model (i.e. the Young's modulus and the Hardening coefficient) should also be defined.
In order to account for the visco-plasticity, three models are available : a) Perzyna b) Norton c) Strain Hardening Creep Perzyna model
This model allows to describe the secondary (steady-state) creep with threshold.
Version 2006.0
Page 198
ProCAST User Manual
Norton model
This model allows to describe the secondary (steady-state) creep with no threshold.
Version 2006.0
Page 199
ProCAST User Manual
As there is no threshold value, the data for the Yield Stress and the Hardening will be ignored.
Strain Hardening Creep model
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
ProCAST User Manual
For the three models, one have (for the plastic model) :
or
Moreover, the value of the normalization stress ( σ* ) is defined by the user (usually a value of 1 is recommended, in the same unit system as the measured stresses). One should note that the value of η and n will depend upon the selected value for σ*. One should be careful that viscoplastic models which have been set in versions prior to 2005.0 should be modified in PreCAST in order to add the normalization stress (σ*). Otherwise, the solver will not work. Further details on the Perzyna law can be found in : "Numerical Modelling in Materials Science and Engineering", M. Bellet, M. Rappaz and M. Deville, Spinger, 2003, pp. 306-310.
Summary
Version 2006.0
Page 201
ProCAST User Manual
With the three available models (elastic / elasto-plastic / elasto-viscoplastic), it is possible to cover the whole range of behavior occurring in a solidifying alloy. The figure below shows how one can "go" from one model to the other, by setting the appropriate temperature-dependant material properties.
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 0.001°C. 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.
Temperature-dependant stress data The three graphs hereafter are showing how it is recommended to define the Elastic Young's modulus, the Yield stress and the Hardening coefficient (in the case of Linear Hardening), over the whole temperature range. These
Version 2006.0
Page 202
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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).
Plastic and Viscoplastic properties determination In order to determine the plastic and viscoplastic properties (σo, H, η, n) at different temperatures, including the transition, two methods are suggested (for the Perzyna model with linear hardening) : • •
Version 2006.0
Tensile test method Creep test method
Page 206
ProCAST User Manual
Tensile test method In the "tensile test method", one should perform controlled tensile tests. These tensile tests should be performed for different temperatures using various strain rates. To do so, the tensile test should be done at constant RAM speed (in order to pull at specified strain rates). The stress and strain should be recorded during these tests. This is usually obtained with a Gleeble machine. The figure below shows the results of such measurements : at low temperature (T1 - green curve), the curve is independent from the strain rate. Thus, the behavior is elastic-plastic. At high temperature (T3 - red curves), one can see that the stress level is depending upon the strain rate and that there is no hardening (viscoplasticity only), as the curve is horizontal at a given stress level. At "intermediate" temperature (T2 - blue curves), one has both plasticity (with hardening) and viscoplasticity (as the curve depends upon the strain rate).
Version 2006.0
Page 207
ProCAST User Manual
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 (σo) and the Hardening coefficient (H). The value of the viscous parameter (η) 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
ProCAST User Manual
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 (σy) is equal to the Yield stress σo.
Version 2006.0
Page 209
ProCAST User Manual
In order to obtain the viscoplastic parameters (η and n), one should first determine the value of the Yield stress (σo), 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 (σo) 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 (σ- σY)/σ* (which is the stress level of the plateau minus the Flow stress, σY, which is equal to σo in this case, as H = 0) and dε/dt should be plotted in a log-log graph. The value of η and n can be deduced from this graph (see below).
Version 2006.0
Page 210
ProCAST User Manual
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 (η and n), one should first determine the value of the Yield stress (σo), 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 (σo) 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 (σ- σY)/σ* (which is the extrapolated stress level at zero strain minus the Flow stress, σY. In this case, σY is equal to σo, as εpl = 0 - because the extrapolation is performed a zero strain) and dε/dt should be plotted in a log-log graph. The value of η and n can be deduced from this graph (see previous figure).
Version 2006.0
Page 211
ProCAST User Manual
Creep test method In the "creep test method", one should perform controlled creep tests (plastic deformation as a function of time for a given fixed load). The creep tests should be performed for different temperatures using various loads (stress). The figure below shows the results of such measurements : at low temperature (T1 - green curves), the curves are independent from time (plateau). Thus, the behavior is elastic-plastic. At high temperature (T3 - red curves), one can see that the plastic strain is increasing linearly with time (viscoplasticity only). At "intermediate" temperature (T2 - blue curves), one has both plasticity (with hardening) and viscoplasticity.
Version 2006.0
Page 212
ProCAST User Manual
Low Temperature
At low temperature (usually T1 < Tm/3), the following curves are measured at different loads.
Version 2006.0
Page 213
ProCAST User Manual
From these curves, one can get the Yield stress (σo) and the Hardening coefficient (H). To do so, the values (shown in the graph below) of ε and σ (at the different loads) should be plotted in a σ-ε graph. A linear or Power Law Hardening can be fitted on those points.
Version 2006.0
Page 214
ProCAST User Manual
Beside the above, the value of the viscous parameter (η) 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, dε/dt (see the red curves in the figure below).
Version 2006.0
Page 215
ProCAST User Manual
In order to obtain the viscoplastic parameters (η and n), one should first determine the value of the Yield stress (σo), 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 (σo) is zero. Then, for each measured creep curve (shown above), the value of strain rate (dε/dt) should be plotted as a function of (σ- σY)/σ* (which is the stress level of the plateau minus the Flow stress, σY, which is equal to σo in this case, as H = 0) in a log-log graph (see below). The value of η and n can be deduced from this graph (see below).
Version 2006.0
Page 216
ProCAST User Manual
Beside the above values, as there is no hardening, the value of H should be set to zero. Therefore the threshold (σy) is equal to the Yield stress σo. 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
ProCAST User Manual
From these curves, one can get the Yield stress (σo) and the Hardening coefficient (H) as described above in the "Low Temperature" section. To do so, the values (shown in the graph below) of ε and σ (at the different loads) should be plotted in a σ-ε graph. A linear or Power Law Hardening can be fitted on those points. In order to obtain the viscoplastic parameters (η and n), one should first determine the value of the Yield stress (σo), 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 (σo) is zero. Then, for each measured creep curve (see blue curves in the zoom below), the values of the INITIAL strain rate (dε/dt) should be plotted as a function of (σσY)/σ* (which is the load, σi, minus the Flow stress, σY. In this case, σY is equal to σo, as εpl = 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
ProCAST User Manual
Version 2006.0
Page 219
ProCAST User Manual
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
ProCAST User Manual
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).
General Run Parameters The Run Parameters which are used by any module are grouped in the "General" tab. It is also possible to defined the units which will be used for the display of the results (it will not affect the results themselves).
Version 2006.0
Page 221
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Enter an integer value.
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
ProCAST User Manual
64 = nodal turbulence intensities, 128 = nodal turbulence dissipation rates, 1024 = nodal displacements, 8192 = surface heat fluxes, and 32768 = nodal magnetic potentials
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
ProCAST User Manual
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.
Thermal Run Parameters The Thermal model activation, the Temperature results storage frequency, as well as the Porosity model and parameters are the main parameters to be defined in this tab.
Version 2006.0
Page 227
ProCAST User Manual
Version 2006.0
Page 228
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Enter a floating (real) value.
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
ProCAST User Manual
0 - automatic density calculation, by extrapolation of the densities at liquidus and solidus, weighted by the fraction of solid. 1 - the density table, which is defined in the database is used.
Cycling Run Parameters When cycling is modeled, one should define the corresponding Run Parameters in this tab (number of cycle and cycling time). In addition, when "Die Combo" interface definitions are used, additional parameters should be set.
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
ProCAST User Manual
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}
Radiation Run Parameters When a calculation with Radiation, including View Factors is run, the following parameters should be defined. If the "View factor OFF" option is selected, it is not needed to define these Run parameters.
Version 2006.0
Page 234
ProCAST User Manual
Version 2006.0
Page 235
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
Flow Run Parameters When fluid flow and/or filling models are activated, the "Flow" Run Parameters should be defined. The main parameters are present in the "Standard" tab. In the "Advanced 1" tab, one should mainly define the WSHEAR and the WALLF parameters. In most cases, all the other parameters can remain as the default values.
Version 2006.0
Page 238
ProCAST User Manual
Version 2006.0
Page 239
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
example, if the pressure boundary condition drives the flow at a gauge of 1 atmosphere, the boundary condition is set to 2 atm. PREF should be set to 1 atm. 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}
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
ProCAST User Manual
a real value sets the time Enter a floating (real) value. Choose the time units from: {sec | min}.
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.
Turbulence Run Parameters
Version 2006.0
Page 246
ProCAST User Manual
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
ProCAST User Manual
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
Stress Run Parameters
Version 2006.0
Page 248
ProCAST User Manual
Version 2006.0
Page 249
ProCAST User Manual
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
ProCAST User Manual
CRACK = 3 activates the cracking model, with a feedback on the stress properties. In this case, the hardening properties are changed, according to the cracking model.
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
ProCAST User Manual
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.
Micro Run Parameters
Version 2006.0
Page 252
ProCAST User Manual
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
ProCAST User Manual
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.
The default values correspond to the following table :
Version 2006.0
Page 254
ProCAST User Manual
Pre-defined Run Parameters ProCAST allows to define "Pre-defined sets" of Run Parameters in the Preferences.
Version 2006.0
Page 255
ProCAST User Manual
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
ProCAST User Manual
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) :
Moreover, in the "dat/pref" directory, there is also a file named "default_p.dat". This file contains the "default" values of the Run Parameters which will appear when a new case is created. The user can change the content of this file (which has exactly the same structure as a normal p.dat file) at his convenience.
Run Parameters Recommendations ProCAST provides the access to many Run parameters, in order to allow the treatment of all kind of situations.
Version 2006.0
Page 257
ProCAST User Manual
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.
For all processes 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
POROS = 1 (this model is now recommended for all processes) MACROFS = 0.7 FEEDLEN = X (The value FEEDLEN depends upon the size of the mushy zone and thus, the size of the casting. A value ranging from a few millimeters to a few centimeters is recommended. This should be calibrated with experiments. A value of 0 is not advised as this will produce a uniform microporosity throughout the part, beside the macroshrinkage)
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
High pressure die casting (HPDC) Timestep handling
Version 2006.0
Page 258
ProCAST User Manual
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
Low pressure die casting (LPDC) Timestep handling
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
WSHEAR = 0 (never use WSHEAR = 2 for LPDC) FREESFOPT = 1 WALLF = 0.8 LVSURF = 1.0 PINLET = 1 for a pressure filling or PINLET = 0 for a velocity/inlet 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
ProCAST User Manual
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
ProCAST User Manual
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.
To disable the porosity and piping calculations 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.
Version 2006.0
Page 261
ProCAST User Manual
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.
When a "region" is cooling down, if the density is increasing with decreasing temperature (the usual case for most alloys), some shrinkage occurs. At each timestep, the accumulated shrinkage occurring at all the nodes which have a solid fraction equal to or lower than MACROFS, plus those nodes between the
Version 2006.0
Page 262
ProCAST User Manual
MACROFS and MACROFS + FEEDLEN isosurfaces, is computed. This shrinkage is then distributed according to the following scenarios: 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, piping occurs and the free surface of the casting (usually the riser) goes down by the amount corresponding to the shrinkage (For display purposes, in the pipe, the shrinkage porosity value is set to 1 and FVOL is set to a value below 0.5, so that it will exhibit piping, i.e. "empty nodes").
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
ProCAST User Manual
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
ProCAST User Manual
During the same time, microporosity is computed in the following way: a) microporosity can appear only in the zone where the fraction of solid is in between MACROFS and 1. b) within this zone, two situations may occur : b1) there is still some mushy zone (or liquid) below MACROFS. In this case, microporosity can occur only at a distance higher than the value of FEEDLEN from the MACROFS isosurface (Zone A in the figure below). This means that, if high gradients are present, the distance between the MACROFS and solidus isosurface is smaller than FEEDLEN, no microporosity occurs (Zone B in the figure below). The amount of microporosity is equal to the density change between the local fraction of solid and 1. b2) there is no more mushy zone below MACROFS. In this case, the FEEDLEN parameter is not active anymore and there could be microporosity in the whole region between MACROFS and 1. This is due to the fact that as there is no more "open liquid" to feed the shrinkage, local microporosity has to occur in order to compensate the local shrinkage. The amount of microporosity is equal to the density change between the local fraction of solid and 1.
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
ProCAST User Manual
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
ProCAST User Manual
In version 3.2.0, this Run parameters was called MOBILE (MOBILE still exists in version 2004.0 for other purposes. Thus, it should not be used anymore for porosity settings).
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
ProCAST User Manual
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
ProCAST User Manual
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).
SGI Porosity model Instead of shrinking during solidification, some alloys do also exhibit some expansion. The most well known material which exhibits this behavior is the Nodular Cast Iron, also called Spheroidal Graphite Iron (SGI). Grey Iron may also exhibit such behavior. The volume change as a function of temperature is shown in the figure hereafter for different cast irons.
Version 2006.0
Page 269
ProCAST User Manual
POROS model in case of expansion
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
ProCAST User Manual
At the same time, microporosity is computed in the following way: a) Microporosity can appear or disappear (partially or fully) only in the zone where the fraction of solid is in between MACROFS (+FEEDLEN) and 1. b)
Within this zone, two situations may occur: b1) There is still some mushy zone (or liquid) below MACROFS. In this case, microporosity can occur or disappear (partially or fully) only at a distance greater than the value of FEEDLEN from the MACROFS isosurface (Zone A). This means that, if high gradients are present, the distance between the MACROFS and solidus isosurface is smaller than FEEDLEN, no microporosity occurs or disappears (Zone B in the figure below). The amount of microporosity formation or disappearance is equal to the density change between the local fraction of solid and 1. b2) There is no more mushy zone below MACROFS. In this case, the FEEDLEN parameter is not active anymore and there could be microporosity in the whole region between MACROFS and 1 for shrinkage. This is due to the fact that as there is no more "open liquid" to feed the shrinkage, local microporosity has to occur in order to compensate the local shrinkage. The amount of microporosity is equal to the density change between the local fraction of solid and 1. If the density change between the local fraction of solid and 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.
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
ProCAST User Manual
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
ProCAST User Manual
Coupled Microstructure - Porosity calculation
When a microstructure calculation is performed in the case of SGI, the density is automatically calculated at each location (i.e. each node) of the casting and this local density is used "on-line" in the porosity calculation (POROS = 1 model only). The following figures are showing the temperature at different locations of a casting and the corresponding density curves (with the same colors). As one can see in the bottom graph, the fast cooling zones exhibit only shrinkage (red curve), whereas the low cooling rate zones exhibit a very important expansion (blue curve). In between, one have mixed behaviors.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Version 2006.0
Page 281
ProCAST User Manual
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.
Once the model is loaded, a Filter material should be assigned to the corresponding material domain (see the "Databases/Material Properties" section for the description of the filter properties definition).
Version 2006.0
Page 282
ProCAST User Manual
Then, the "FILTER" type should be assigned to this material domain. In the Interface menu, the interfaces between the casting and the filter material (on both sides) should be kept as "EQUIV" (i.e. no interface should be created between the casting material and the filter).
Version 2006.0
Page 283
ProCAST User Manual
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
ProCAST User Manual
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) :
The material properties of the Exothermic sleeve must be defined as follows :
Version 2006.0
Page 285
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
1
Page 288
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Version 2006.0
Page 292
ProCAST User Manual
The Run Parameters should be configured as a regular Gravity filling problem with an inlet pressure. Additional Run parameters should be set in the p.dat file for Lost Foam :
Version 2006.0
Page 293
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
correspond to the shear rate encountered in domain 4 (which should be a higher value than domain 2 as the section is smaller). In domains 1, 2 and 4, it is not necessary to set a cut-off value (the default value of 1 can be used), as the effective shear rates will correspond well to the reality. To determine the shear rate in a gate, one can view the "Non-Newtonian Shear Rate" in ViewCAST.
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
ProCAST User Manual
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
ProCAST User Manual
MULTIPLE MESHES AND NON-COINCIDENT MESHES ProCAST has the capability to handle non-coincident meshes (see figure below).
If the different domains are meshed separately, they can be loaded in PreCAST using the "File/Multiple Meshes" menu.
Then, a window opens, which allows to load the different meshes.
Version 2006.0
Page 299
ProCAST User Manual
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
ProCAST User Manual
defined, one can "Apply" the whole selection (5), which will create the appropriate interfaces.
When a non-coincident interface is defined, it is possible to change the default tolerances used for the detection of the nodes on the opposite side. To do so, one should make a right click on the NCOINC label and the following window will appear :
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
ProCAST User Manual
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
ProCAST User Manual
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).
Restore the X-Y orientation of the model (Z-axis perpendicular to the screen)
Interactive zoom (the model is enlarged when the cursor is moved towards the bottom of the screen and it is reduced when the cursor is moved towards the top of the screen)
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)
Drag of the model (interactive move of the model on the screen)
Version 2006.0
Page 303
ProCAST User Manual
Material (or domain) selection
Wire frame display mode
Hidden wire frame display mode
Version 2006.0
Page 304
ProCAST User Manual
Hidden display mode with the mesh
Version 2006.0
Page 305
ProCAST User Manual
Shaded display mode
Activation of the enclosure viewing (for radiation models)
Version 2006.0
Page 306
ProCAST User Manual
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
ProCAST User Manual
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 : • • •
User Functions to define a given condition, instead of a constant or a table External Functions which can be called from the User Functions, in order to get the values of some fields at any node. External Computation Function which can be called during the execution of the calculation (e.g. for coupling with external softwares).
User Functions description Currently, the following User Functions are available : • • • • •
Interface BC : interface heat transfer coefficient Heat BC : external heat transfer coefficient Heat BC : external temperature Heat BC : emissivity Heat BC : heat flux
• • •
Imposed velocity (X, Y and Z components) Solid transport velocity (X, Y and Z components) Translation vector (X, Y and Z components)
• •
Mass source position (X, Y and Z components) Mass source flow rate
Each of these functions have the following arguments : • • • • • •
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
ProCAST User Manual
allows to define for instance a heat transfer coefficient as a function of a "control temperature", corresponding to a given location, somewhere else ("remote control"). User Function Templates are available in the next section.
External Functions description From any User Function, it is possible to get the value of some fields (e.g. Temperature, fraction of solid, velocity) at any node, specified by its node number. To do so, the following functions are available : • • • • •
usertemp1(node#) for temperature userfs1(node#) for the fraction of solid uservx1(node#) for the X-component of the velocity uservy1(node#) for the Y-component of the velocity uservz1(node#) for the Z-component of the velocity
In addition, it is possible to get the node number by giving the coordinates, using the following function : •
nodNum (xin, yin, zin, domain#, xout, yout, zout)
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.
External Computation Function description Beside User Functions for the definition of flexible conditions, ProCAST is providing an "External User Function", which is called at the following times : • • • •
at the beginning of the calculation at the beginning of each timestep at the end of each timestep at the end of the calculation
This allows to perform a number of operations, as defined by the user, at these different moments. As for the User Functions, the "internal functions" allow to retrieve the values of the main fields (e.g. temperature, fraction of solid, velocities, ...) at any location of the model. Such External calls can be used for instance to couple ProCAST with other softwares. The Template of the External Function is available in the next section.
Version 2006.0
Page 309
ProCAST User Manual
Language and Compiler All the User Functions and External Functions should be written in C language, based upon the provided Templates. In order to be able to use User Functions, a C compiler and linker should be available on the machine. Specific compiler and linker are needed (only the compiler/linker which have been used for the creation of the executables are guaranteed to work with the software). There is no guarantee that the software will work with other compilers/linkers. Windows
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
Intel C compiler (icc), version 8.1 SGI
C compiler version 7.3.1 IBM
C compiler version 6
Version 2006.0
Page 310
ProCAST User Manual
SUN
C compiler version 5.4 HP (HP-UX 11.11)
C compiler version B11.11.06 HP (HP-UX 11.22)
C compiler version A.05.50
Use of User Functions Once the User Functions are defined (i.e. programmed), they should be placed in the current working directory (i.e. next to the standard input files). Then, when the "procast" executable is launched (either from the Manager, or "manually" in a Command window), the User Functions are automatically compiled and linked. A local DLL is created, as well as a local executable. Then, this executable is automatically launched and the calculation starts. This DLL and local executables are automatically deleted at the end of the execution (if the calculation is stopped manually or if it crashes, these files will remain in the local directory).
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
ProCAST User Manual
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
ProCAST User Manual
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
User Functions Templates The following section is presenting the templates of all the User and External Functions. The meaning of the arguments are described in the comments of the functions. • • • • •
External heat transfer coefficient Function (convehtransfer.c) External temperature Function (texternal.c) Emissivity Function (emissivity.c) Heat flux Function (heatflux.c) Interface heat transfer coefficient Function (interhtransfer.c)
• •
Mass Source Flow Rate Function (masssourceflowrate.c) X-component Mass Source Vector Function (xmasssource.c) - same for Y and Z X-component Translation Vector Function (xtranslation.c) - same for Y and Z X-component Imposed Velocity Vector Function (vximposed.c) - same for Y and Z X-component Solid Transport Velocity Vector Function (vxsolidtransport.c) same for Y and Z
• • • •
External Function (externalcompute.c)
External heat transfer coefficient Function This function is called : convehtransfer.c
It is used to define a convective heat transfer coefficient on the outside of a domain, to be used in conjunction with an external temperature.
Version 2006.0
Page 313
ProCAST User Manual
#include #include #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* );
convective heat transfer coefficient (applied on external surfaces)
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 ------------- */
/* ------------ Do not forget to remove the call to exit ------------ * * ------------ hereafter before running the calculation ------------ */ printf("---> exit in C user function convehtransfer exit in C user function texternal exit in C user function emissivity exit in C user function heatflux exit in C user function interhtransfer exit in C user function masssourceflowrate exit in C user function xmasssource exit in C user function xtranslation exit in C user function vximposed exit in C user function vxsolidtransport 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
ProCAST User Manual
on the PC and are automatically used (even in a latter post-processing session). These results are stored in a directory named : C:\Documents and Settings\USER\Local Settings\Temp\Prefix . One should just be careful that if two cases having the same prefix are visualized from a CD, the data will be overwritten. Please note that the fields which are shown in the menus correspond only to the available result files in the working directory. For instance, if results are computed in the "Action" menu - such as Niyama (see the Result analysis chapter for more details), the corresponding item will be added in the menu.
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
ProCAST User Manual
DISPLAY TYPES The type of display can be selected either with the following icons
or with the Picture Menu
Snapshot - 3D view of the selected field (e.g. Temperature), displayed on the surface of the selected domains. For filling the free surface is viewed in 3D. This mode can be accessed directly with the F5 key. Slice - 2D cuts (shown within the 3D model). One or several slices can be selected, either along the X, Y and/or Z planes or along any plane. This mode can be accessed directly with the F6 key. Scan - Scan of 2D slices along either the X, Y or Z axis. This mode can be accessed directly with the F7 key. CutOff - X-Ray view of the inside of a model. The model is made partially transparent (according to a criterion such as above a given temperature) in order to see features which are inside the model, such as pockets of liquid, surrounded by a solid shell (which is made transparent). This mode can be accessed directly with the F8 key. The selected mode can be viewed by a red square around the corresponding icon.
Version 2006.0
Page 341
ProCAST User Manual
Examples of Snapshots (during filling and after), of slice and CutOff (X-Ray) view
Example of Scan slices
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
The figure hereafter is illustrating these different possibilities.
Version 2006.0
Page 349
ProCAST User Manual
Shrinkage and Advanced Porosity display When the Shrinkage Porosity or Advanced Porosity contours are selected for the first time, automatically, the "cut-off" mode is activated, with the definition "above 0.01" (which corresponds to a porosity value higher than 1%). Of course, it is always possible to go back to the normal snapshot mode, by clicking on the snapshot icon
Version 2006.0
.
Page 350
ProCAST User Manual
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
ProCAST User Manual
The Display parameters are defined in the Parameters menu :
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
ProCAST User Manual
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
ProCAST User Manual
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 :
The following figure shows an example of a shaded enclosure.
Version 2006.0
Page 354
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
Slice Data, Scan Data, CutOff Data These options are described in the "Display types" section.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
"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
ProCAST User Manual
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
ProCAST User Manual
In a Stress calculation, it is possible to superimpose the Undeformed geometry (in wireframe mode) to the deformed geometry. The extent of the part deformation can be adjusted with the "Displacement Mag." option. If the wireframe has "too many lines", it is possible to adjust that with the "Feature Angle" option.
Version 2006.0
Page 363
ProCAST User Manual
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
ProCAST User Manual
The scale can be defined either automatically, or semi-automatically, via the "Min-Max" or "Semi-Auto" settings. It is also possible to change each value independently (with the Manual field), by clicking on each value to be changed.
Version 2006.0
Page 365
ProCAST User Manual
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
ProCAST User Manual
It is possible to deactivate this display by unchecking the "Display Tsol-Tliq" in the above window. When a scale was defined other than "Automatically", the minimum and maximum values are stored for each field (the storage is done upon exit of the model if the "Store the last view and Exit" option is selected. When the model is loaded the next time, these settings will be automatically retrieved. The values are indicated in the "Stored" field (the values can not be modified interactively, use the Min-Max fields above). The values are stored in a local file called prefix.scale (this file can be copied in an other case if one wants to use the same values for this new model). Finally, the color scale can be change, by clicking on the color map itself. The following panel opens and the colors can be changed. It is also possible to store and load a different color scale.
Version 2006.0
Page 367
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
Three types of selections are available : • • •
Interval - selection of the curves every N nodes Nodes - selection node by node External - loading of a set of external curves (e.g. measurements)
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
ProCAST User Manual
Nodes When "Nodes" is selected, the following node selection window appears :
Version 2006.0
Page 371
ProCAST User Manual
The node numbers can be entered manually, or an interactive selection can be made with the
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
ProCAST User Manual
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
ProCAST User Manual
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.
X-Y Plot settings The display of the X-Y plots can be tuned in the X-Y Plot settings window. For the settings of the time scale, the calculated minimum and maximum values are shown in blue (1). The vertical scale can be adjusted automatically with the "Auto" check box (2).
Version 2006.0
Page 374
ProCAST User Manual
Version 2006.0
Page 375
ProCAST User Manual
GEOMETRY MANIPULATION The geometry can be manipulated with the following icons :
Opens the Rotate definition window (see below)
Restores the original orientation of the model (with the Z-axis perpendicular to the screen)
Interactive zoom (zoom up when the mouse is move towards the bottom of the screen)
Auto-scale (or zoom out)
Interactive center of the model (the location which is clicked is moved towards the center of the screen)
Interactive drag of the model
Rotation When the Rotate icons is activated, the following Rotate window is opened :
Version 2006.0
Page 376
ProCAST User Manual
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).
Automatic Rotation in the X-direction
, 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 key is pressed in the same time, the rotation direction is reversed. When the 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
ProCAST User Manual
Beside the zoom icon (which allows to increase or decrease the size of the model), the F2 and F3 keys can be used to increase, respectively decrease the size of the model (zoom in and zoom out). Moreover, with the center mouse key, it is possible to zoom interactively a specific zone of the model with a rubber band box. Once the center key is pressed, it defines one corner of the rubber band box and the opposite corner is defined by the movement of the mouse until the key is released. Then, the zoomed model is automatically displayed. It is possible to zoom the model several time using this technique. To zoom out, use the Auto-scale icon
.
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
ProCAST User Manual
Parameters/StoreView/View-X. To retrieve (or Restore) a given view, use the "Restore" option (one can also use the Ctrl-... key where ... is the number of the view (1 to 6)). With the "Replace" option, the current view will replace the one which was previously stored. The "Reset view" (or Ctrl+0) will automatically reset the orientation of the model in the isometric view with the gravity pointing downwards. These views are stored in a file named prefix.lv (only if the option "Store the last View and Exit" is selected upon exit of ViewCAST). Please note that in View-1, it is always the last view which is active on the screen upon exit which is stored (which will replace any previously stored one). Please note that this *.lv file is not compatible with versions before 2006.0.
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
SO : Solid (displays the selected field values - e.g. Temperature) IN : Invisible (hide the selected material) WI : Wireframe (displays the selected material with wireframe only) HM : Hidden mesh (displays the FEM mesh in hidden mode) SH : Shaded material
Page 379
ProCAST User Manual
•
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
ProCAST User Manual
When the Apply button is pressed, the corresponding offsets are applied, leading to a picture as follows :
Display node and element numbers The node numbers of the surface of the mesh can be viewed with the "Ctrl-N" keys. When "Shift-Ctrl-N" keys are pressed, a small window is opening :
Version 2006.0
Page 381
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
These settings will be used for the next ViewCAST launch's.
Version 2006.0
Page 384
ProCAST User Manual
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 10°C 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
ProCAST User Manual
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
ProCAST User Manual
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
and the Mapping factor "M" will be calculated as :
If the following parameters are used : a b c d
= 1.0 = 0.0 = 1.0 = -0.5
the Mapping factor "M" corresponds to the Nyiama criterion :
For the calculation of the Nyiama criterion, suggested values are : • L Upper Temp = Tliquidus + 2 • L Lower Temp = Tsolidus • R,G Temp = Tsolidus + 0.1 * (Tliquidus - Tsolidus)
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
ProCAST User Manual
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 :
These constants A and B should be calibrated with experiments. The solidus temperature should be entered in the "Solidus" field (see below) and the temperature corresponding to a critical solid fraction (at which feeding is not anymore possible - typically 60-80%) should be entered in the "Critical" field.
Version 2006.0
Page 388
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
SDAS (secondary dendrite arm spacing)
The secondary dendrite arm spacing (also called SDAS or λ2 ) 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.
The SDAS is calculated in ViewCAST according to the following parameters.
Version 2006.0
Page 392
ProCAST User Manual
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
ProCAST User Manual
POROSITY To analyse the porosity in a casting, several options are available : • • • • •
Temperature field with the cut-off option Fraction of solid field with the cut-off option Shrinkage porosity field Niyama criteria for critical cooling rates Specific RGL criteria
Temperature field and Fraction of solid field The porosity being primarily due to enclosed pockets of liquid, one can well observe them by looking at the Temperature or Solid fraction fields in cut-off mode (in order to visualize inside the casting during the solidification). For temperatures, it is recommended to select a cut-off value above the solidus temperature, whereas cut-off values of solid fractions below 70% shall be used.
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.
Specific RGL criteria One has the possibility to define a customized criterion function, with the RGL option (see the Criterion Functions section for more details).
Version 2006.0
Page 394
ProCAST User Manual
FATIGUE LIFE INDICATOR In the case of die casting, when a stress calculation is run, ProCAST calculates automatically the fatigue life in the dies. The model is based upon a "strain-driven" approach and a power law relationship, which corresponds to low cycle fatigue. During the calculation, the accumulated plastic strain (if any) is recorded at each location of the die. Then, the life of the die is calculated as a power law of this plastic strain. If this life is shorter than the previously calculated life at this given location, this latter value is stored. If there is no plastic strain, then, the elastic strain is recorded and the life of the die is calculated with an other power law relationship. Again, if the computed life is shorter as the previously calculated (at this same location), this latter value will be used. The parameters of the model correspond to measured values for a typical die alloy (H13 steel), published by the Society of Automotive Engineers. One should note that the values are measured at room temperature and thus, the effect of temperature is only taken into account through the temperature-dependant stress properties (i.e. Youngs modulus, Yield stress, Hardening, ...). Thus, one should consider that ProCAST calculates a "Fatigue life INDICATOR" and that it does not correspond to absolute values of number of cycles. No data are required for this model and it can be used as an indicator for different steel grades generally used for molds. One should notice that it should be mainly used to compare different designs with the same mold material (e.g. to compare the effect of cooling channels or of cycling time). One should not compare the influence of different steels on the "Fatigue life indicator".
Version 2006.0
Page 395
ProCAST User Manual
HOT TEARING INDICATOR During a stress calculation, ProCAST calculates the susceptibility to hot tearing. The hot tearing indicator allows to model cracks which are forming during the solidification (i.e. it corresponds to openings in between dendrites which are not yet totally solidified and which are opening because of a tensile stress). Hot tearing are formed only above the solidus temperature. The hot tearing indicator is a "strain-driven" model based upon the total strain which occurs during the solidification. The model is computing the elastic and the plastic strain at a given node when the fraction of solid is between CRITFS (usually 50%) and 99%. There is no parameter for this model (except CRITFS which is a general stress Run Parameter). As the amount of plastic strain will strongly depend upon the stress properties in the mushy zone, the hot tearing indicator should be used to compare different designs with the same alloy. One should not compare hot tearing indicator levels for different alloys.
Version 2006.0
Page 396
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
Radiation faces correspond to a "surface mesh" or radiative faces, like an enclosure.
Version 2006.0
Page 401
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
The format of this "Neutral file" is the following : Loop on the Timesteps Timestep Number_of_nodes Loop on the Number_of_nodes node_number, x, y, z, s1, s2, s3, s4, s5, s6 End_loop End_loop
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
ProCAST User Manual
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.
DMP : Distributed Memory Processing
Distributed Memory Processing is an architecture where each CPU accesses its own memory. Data are shared between processors through message passing. The send and receive operations require actions by the processors on both ends of the communication. The Message-passing Interface, or MPI, is the software standard that has been developed by an industry/government consortium. Such a configuration (see picture below) requires a very fast network between the CPU's.
Version 2006.0
Page 408
ProCAST User Manual
SMP/DMP combined architecture
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.
ProCAST Parallel architecture
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
ProCAST User Manual
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
ProCAST User Manual
HOW DOES PROCAST PARALLEL WORKS ? In order to distribute the calculation between the N processors, the geometry (model) is split in N sub-domains. This partitioning is done fully automatically by the software. It is done in order to balance the load between the processors as evenly as possible (i.e. almost the same number of nodes in each sub-domain) and in such a way that the amount of communication between the processors is minimized (i.e. the least number of common nodes between the sub-domains). This principle is illustrated in the figure hereafter.
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
In order to prevent such detrimental behavior, one should rather have a partitioning like the one on the right (in the figure below), rather than the one on the left.
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
ProCAST User Manual
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
ProCAST User Manual
USE OF THE PARALLEL SOLVER The Parallel solver is fully compatible with the Scalar solver. This means that the input and output files are strictly identical (except two additional Run parameters for the Parallel solver called NFFWID and NFFSAF). Thus, the Pre- and Post-Processing should be made in the Scalar version, as usual. The two "Parallel" Run Parameters (NFFWID and NFFSAF) should be added manually in the p.dat file. If they are not specified, the default values of NFFWID = 10 and NFFSAF = 2 are used. Recommended values are NFFWID = 10 to 20 and NFFSAF = 2 to 5 (see the "How does ProCAST Parallel works" section for more details about the meaning of these Run Parameters). Once a case is set-up in PreCAST, the standard "DataCAST" should be called (it is not anymore necessary to use a specific DMP DataCAST executable). The Manager allows to have a direct access to the Parallel solver. To do so, one should first activate the Parallel Processing Preferences in the "Software Manager/Software configuration" window. The interactive launch of the Parallel solver is valid only for Linux and Windows at this stage. For other UNIX platforms, the run should be launched using manual commands in a Command window (see the end of this section for more details) Once the Manager is configured as decribed above, the "ProCAST" button is opening the following window for Linux :
Version 2006.0
Page 416
ProCAST User Manual
and for Windows :
Version 2006.0
Page 417
ProCAST User Manual
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
ProCAST User Manual
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)
mpirun : MPI executable -options(1) : LAM MPI options to speed-up communications -ssi rpi usysv
CPU_selection : selection of processors (2 possibilities) -np N selection of N processors ci,j-k list of processors procastparallel : ProCAST Parallel executable prefix : case name -options(2) : ProCAST Parallel options none should be specified at this stage To run the case in the background, one can use the following command : mpirun -options(1) CPU_selection procastparallel prefix -options(2) > runlog.txt &
For Windows : mpiexec -options(1) CPU_selection procastDMP.exe prefix
mpiexec : MPI executable -options(1) : Mapping options When the calculation data are on a different drive (e.g. the M: drive), one should specify it with the "-map option" under "Option(1)". The synthax is the following : "-map m:\\Disk1\Data" This means that the \\Disk1\Data directory is mounted as the "M" drive.
Version 2006.0
Page 419
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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
When a cluster of processors is used (e.g. Linux cluster of 4 machines, having each 2 CPU's), they should be connected through a dedicated fast network, such as a 1 Gigabit/s Ethernet or a Myrinet switch. Please note that the ProCAST Parallel installation (with the libraries) may be different if a 1 Gigabit/s Ethernet or a Myrinet switch is used. A special care should be taken upon the type of cables which are used. The cable should be designed for fast communication (for 1 Gigabit/s Ethernet, Category 6 cables - or above - should be used). Concerning the operating system, it is important to make such that the installation of each node is the same in order to optimize the performances. SSH
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
ProCAST User Manual
Disk mounting procedure
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
ProCAST User Manual
PARALLEL VERSION INSTALLATION As a pre-requisite for the DMP installation, the scalar version must be installed first and the $ProCAST20060 environment variable must be set. This section describes the necessary configuration on the following platforms : • •
Linux Windows
Linux LAMHOME environment variable
The "lam-7.1.1" directory can be placed anywhere (but it should be accessible from all nodes in the same way). It is advised to put it in the $ProCAST20060 directory (sideways to the bin directory). The LAMHOME environment variable should be set to point to the /..../lam-7.1.1 directory. The directory $LAMHOME/bin directory should be included in the PATH (for ALL nodes) : set path=($LAMHOME/bin $path)
LD_LIBRARY_PATH
The environment variable LD_LIBRARY_PATH should contain the path of the libraries delivered with the software (i.e. the one contained in the libDMP directory of the installation). To do so, one should specify in the .cshrc (or .tcshrc) file of ALL nodes the following : setenv LD_LIBRARY_PATH /......../libDMP:$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
ProCAST User Manual
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
Windows (XP and 2000 Professional only) The Windows DMP version is designed to work for a multiple CPU's Windows machine. The supported OS are Windows XP and Windows 2000. Files and directories installation
When the Windows_DMP.exe InstallShield will automatically install the ProCAST DMP solver in the bin directory of the installation, as well as the MPICH2 environnement. Please note that the installation of MPICH2 requires the prior installation of the patch "Microsoft .Net Framework version 1.1". If it is not installed, the following message will appear. Click Yes and load the Patch.
You can also search for this patch in the "Windowsupdate" web site of Microsoft :
Version 2006.0
Page 426
ProCAST User Manual
Version 2006.0
Page 427
ProCAST User Manual
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".
Linux (32 bits) The requirement for Linux is the level of the Linux kernel and of the glibc. One should have levels higher or equal to : • •
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
With these requirements, the ProCAST parallel version is technically compatible with any Linux operating system (e.g. RedHat, Suse, Mandrake, ...). However, the version is guaranteed (i.e. tested) only on Redhat EL3 (RedHat Entreprise Linux 3). The following versions of libraries and compilers were used : LAM-7.1.1 Petsc-2.3 Parametis-3.1 gcc-3.1 compiler(for LAM, Petsc and Parametis) icc 8.1.026 compiler (for ProCAST)
Linux (64 bits) The requirement for Linux is the level of the Linux kernel and of the glibc. One should have levels higher or equal to : • •
Version 2006.0
kernel : glibc :
>= 2.4.21 >= 2.2.1
Page 428
ProCAST User Manual
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
With these requirements, the ProCAST parallel version is technically compatible with any Linux operating system (e.g. RedHat, Suse, Mandrake, ...). However, the version is guaranteed (i.e. tested) only on Redhat EL3 (RedHat Entreprise Linux 3). The following versions of libraries and compilers were used : LAM-7.1.1 Petsc-2.3 Parametis-3.1 gcc-3.1 compiler(for LAM, Petsc and Parametis) icc 8.1.026 compiler (for ProCAST)
Version 2006.0
Page 429
ProCAST User Manual
LAM/MPI AND MPICH COPYRIGHTS Copyright and Software License for LAM/MPI : Copyright (c) 2001-2004 The Trustees of Indiana University. All rights reserved. Copyright (c) 1998-2001 University of Notre Dame. All rights reserved. Copyright (c) 1994-1998 The Ohio State University. All rights reserved. Indiana University has the exclusive rights to license this product under the following license. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. All redistributions of source code must retain the above copyright notice, the list of authors in the original source code, this list of conditions and the disclaimer listed in this license; 2. All redistributions in binary form must reproduce the above copyright notice, this list of conditions and the disclaimer listed in this license in the documentation and/or other materials provided with the distribution; 3. Any documentation included with all redistributions must include the following acknowledgement: "This product includes software developed at the Ohio Supercomputer Center at The Ohio State University, the University of Notre Dame and the Pervasive Technology Labs at Indiana University with original ideas contributed from Cornell University. For technical information contact Andrew Lumsdaine at the Pervasive Technology Labs at Indiana University. For administrative and license questions contact the Advanced Research and Technology Institute at 1100 Waterway Blvd. Indianapolis, Indiana 46202, phone 317-274-5905, fax 317-274-5902." Alternatively, this acknowledgement may appear in the software itself, and wherever such thirdparty acknowledgments normally appear. 4. The name "LAM" or "LAM/MPI" shall not be used to endorse or promote products derived from this software without prior written permission from Indiana University. For written permission, please contact Indiana University Advanced Research & Technology Institute. 5. Products derived from this software may not be called "LAM" or "LAM/MPI", nor may "LAM" or "LAM/MPI" appear in their name, without prior written permission of Indiana University Advanced Research & Technology Institute. Indiana University provides no reassurances that the source code provided does not infringe the patent or any other intellectual property rights of any other entity. Indiana University disclaims any liability to any recipient for claims brought by any other entity based on infringement of intellectual property rights or otherwise. LICENSEE UNDERSTANDS THAT SOFTWARE IS PROVIDED "AS IS" FOR WHICH NO WARRANTIES AS TO CAPABILITIES OR ACCURACY ARE MADE. INDIANA UNIVERSITY GIVES NO WARRANTIES AND MAKES NO REPRESENTATION THAT SOFTWARE IS FREE OF INFRINGEMENT
Version 2006.0
Page 430
ProCAST User Manual
OF THIRD PARTY PATENT, COPYRIGHT, OR OTHER PROPRIETARY RIGHTS. INDIANA UNIVERSITY MAKES NO WARRANTIES THAT SOFTWARE IS FREE FROM "BUGS", "VIRUSES", "TROJAN HORSES", "TRAP DOORS", "WORMS", OR OTHER HARMFUL CODE. LICENSEE ASSUMES THE ENTIRE RISK AS TO THE PERFORMANCE OF SOFTWARE AND/OR ASSOCIATED MATERIALS, AND TO THE PERFORMANCE AND VALIDITY OF INFORMATION GENERATED USING SOFTWARE. Indiana University has the exclusive rights to license this product under this license.
MPICH Copyright Notice: Copyright (c) 1993 University of Chicago Copyright (c) 1993 Mississippi State University Permission is hereby granted to use, reproduce, prepare derivative works, and to redistribute to others. This software was authored by: Argonne National Laboratory Group W. Gropp: (630) 252-4318; FAX: (630) 252-5986; e-mail:
[email protected] E. Lusk: (630) 252-7852; FAX: (630) 252-5986; e-mail:
[email protected] Mathematics and Computer Science Division Argonne National Laboratory, Argonne IL 60439 Mississippi State Group N. Doss: (601) 325-2565; FAX: (601) 325-7692; e-mail:
[email protected] A. Skjellum:(601) 325-8435; FAX: (601) 325-8997; e-mail:
[email protected] Mississippi State University, Computer Science Department & NSF Engineering Research Center for Computational Field Simulation P.O. Box 6176, Mississippi State MS 39762 GOVERNMENT LICENSE Portions of this material resulted from work developed under a U.S. Government Contract and are subject to the following license: the Government is granted for itself and others acting on its behalf a paid-up, nonexclusive, irrevocable worldwide license in this computer software to reproduce, prepare derivative works, and perform publicly and display publicly. DISCLAIMER This computer code material was prepared, in part, as an account of work sponsored by an agency of the United States Government. Neither the United States, nor the University of Chicago, nor Mississippi State University, nor any of
Version 2006.0
Page 431
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
The principle of the porosity calculation is the following : a) a thermal calculation is performed first b) the temperature and fraction of solid results are then processed in the Advanced Porosity module in order to calculate the macro- and micro-porosity, as well as the pipe shrinkage. c) to do so, a finer grid is automatically superimposed to calculate the pressure drop in the mushy zone (solution of the D'Arcy equation) and the nucleation and growth of pores are computed. d) the calculation of the nucleation and growth of the pores is made according to the a gas model. The Sieverts law, presented in the paper mentioned above, is generalized as a Power law in the software. In order to achieve a porosity calculation, the appropriate settings should be defined for the thermal calculation and then the porosity input file (named "prefix_poro.d") should be configured. These settings are described in the "Advanced Porosity Pre-Processing" section.
Version 2006.0
Page 434
ProCAST User Manual
ADVANCED POROSITY PRE-PROCESSING Configuration of the thermal calculation
The configuration of the thermal model in ProCAST should be made in the standard way. The only precaution is to set the case in order to prevent piping (i.e. PIPEFS = 0). Please note that there are no restriction to activate other models during the thermal calculation, such as fluid flow or stress. However, it is important to remember that the Advanced Porosity model does not handle remelting. If this occurs, it is not guaranteed that the solution will converge. Configuration of the porosity model
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
ProCAST User Manual
In the description hereafter, each block is presented according to the following format.
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
ProCAST User Manual
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
An other important parameter of the calculation of the D'Arcy equation is the permeability (K) of the mushy zone, which is described by the following equation :
Version 2006.0
Page 437
ProCAST User Manual
Beside the fraction of solid (gs), the permeability is a function of the secondary dendrite arm spacing (λ2), 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
b) A constant secondary dendrite arm spacing value is specified
Gas and Bubble Properties nucleation.growth.param
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
ProCAST User Manual
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
The gas properties are defined in the "gas.param" block. Firstly, the concentration of dissolved gas in the liquid metal should be specified (gas.nominal.concentration). The units are [ccSTP/100g], which means cubic centimeters of gas (at normal conditions of pressure and temperature [i.e. STP conditions - 20°C - 1 bar], per 100 grams of liquid metal. The gas.partition coefficient corresponds to the ratio of the solubility limits of the gas in the solid and in the liquid (like a partition coefficient in a phase diagram). This value should be defined between 0 and 1. In the "Standard" model, the gas concentration [gas] in the liquid at equilibrium is described by the following powerlaw : where Pp is the pressure in the pore, n is the "gas.pressure.exponent" and A(T) is a defined as :
Where Ao corresponds to the "gas.equilibrium.constant.a" and B to the "gas.equilibrium.constant.b". For di-atomic gases (such as Hydrogen [H2]), n is equal to 0.5 (i.e. 1/2).
Version 2006.0
Page 439
ProCAST User Manual
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
ProCAST User Manual
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
ProCAST User Manual
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.
The value of p.level can be defined either by a constant or by a function of time (with a value block F(TIME)). feeding.param
Version 2006.0
Page 442
ProCAST User Manual
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
ProCAST User Manual
automatically a value of 0.98 is set. If nothing is specified, a value of 0.98 is used. This value has to be decreased in the case of bad convergence the model. gravity.param
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
ProCAST User Manual
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
Finally, a few calculation parameters should be defined. The "region.number" (or domain number) in which the porosity calculation should be done must be specified at this time. Please note that it is possible to run a porosity calculation in more than one domain. In this case, the domain list should be specified in a value block (see below). One should however make sure that all the domains are connected. The "enthalpy.model" should always be set to zero (it is there for compatibility reasons). The "convergence.tolerance" is the value of the tolerance which is used in the pressure solver. The recommended value is 1e-10. If a calculation is not converging, it is advised to increase this value (1e-9 to 1e-8). The "maximum.iterations" corresponds to the number of iterations that are run before the calculation stops (if the convergence.tolerance is not reached before). It is advised to leave this value as such. Concerning the porosity results storage, two formats are available : a) "calcosoft" format b) "ProCAST" format
Version 2006.0
Page 445
ProCAST User Manual
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
ProCAST User Manual
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).
Examples prefix_poro.d input file Example 1 : TemplateGravity_poro.d #------------ General information -------------porosity.param: porosity.model end
standard
# It should always be 'standard'
#------------ Material Properties -------------viscosity.param: constant.viscosity centipoise end coarsening.param: secondary.dendrite.arm.spacing microns end
$$$
#
