Home

SIMULATION OF RESIDUAL STRESSES IN CASTINGS

1. 159 OPTIONS WINDOW 22 a 160 STORING DATA DEFINITIONS WINDOW ese ese sese ese esee ese tese renun 160 SOLIDIFICATION DEFINITIONS 2 esee ese 161 STRESS SIMULATION OPTIONS WINDOW 0 161 FAST POSTPROCESSING PREPARATION 1 eere teste ene 162 163 POSTPROCESSOR MAIN INTERFACE 164 RESULTS TAB SELECTED IN THE POSTPROCESSOR S CONTROL PANEL WINDOW 164 TEMPERATURE FIELD RESULT DISPLAYED IN THE POSTPROCESSOR S MAIN WINDOW 165 CUT VIEW SETTING WITH THE SLICE FUNCTIONALITY 166 CUT VIEW DISPLAYED IN THE MAIN WINDOW 166 COOLING CURVE SELECTED FOR DISPLAY IN THE CONTROL PANEL WINDOW 167 Table of Contents FIGURE 10 78 COOLING CURVE DISPLAY IN THE POSTPROCESSOR MAIN 2 0 22 168 FIGURE 10 79 EXPORTING THE CURVES FROM THE CURVE S OPTIONS 000 169 FIGURE 10 80 IMPORTING AN RPT FILE INTO EXCEL TEXT IMPORT WIZARD STEP OF 3 WINDOW 171 FIGURE 10 81 I
2. 33 FIGURE 4 10 ABAQUS VS MAGMA VON MISES CURVES FOR THE CYLINDER MODEL 34 FIGURE 4 11 ABAQUS VS MAGMA MAXIMUM PRINCIPAL STRESSES FOR THE CYLINDER MODEL 34 FIGURE 4 12 ABAQUS VS MAGMA MINIMUM PRINCIPAL STRESSES FOR THE CYLINDER MODEL 35 FIGURE 4 13 ABAQUS TOP AND MAGMASOFT BOTTOM COLOR SPECTRUMS FOR THE MISES RESULTS OF THE VCINDER 36 FIGURE 4 14 ABAQUS AND MAGMASOFT BOTTOM COLOR SPECTRUMS FOR THE RESIDUAL MAXIMUM PRINCIPAL STRESSES OF THE CYLINDER 6 4 4 400 00 24 FIGURE 4 15 ABAQUS TOP AND MAGMASOFT BOTTOM COLOR SPECTRUMS FOR THE RESIDUAL MIN PRINCIPAL STRESSES OF THE CYLINDER MODEL 0 38 FIGURE 5 1 FRONT AND TOP VIEW OF THE ORIGINAL 40400000000 serene 40 FIGURE 5 2 BOTTOM VIEW OF THE ORIGINAL MODEL eee 41 FIGURE 5 3 THE HUB MOLD PART UNITS METERS 41 FIGURE 5 4 LOCATION OF THE COOLING AND STRESS POINTS OF THE PART 42 FIGURE 5 5 MESH OF THE HUB MODEL USED IN 6 1 esee esses eene 43 FIGURE 5 6 MAGMASOFT MESH OF THE HUB 2000 0 0 00 000 hee eee ee ee sese eei 44 FIGURE 5 7 CONSTRAINING THE RIGID BODY TRANSLATIONS IN X Y AND Z
3. bete de siad 154 OE 162 10 163 102 3 1 Resah Visuali zaio 163 165 TIE CUES scien 167 10 2 3 2 Results Preparation for COmlparison eere e 169 10 2 92 PIEXDOILIDO ota rb oe ita i 169 10 3 RESULTS COMPARISON APPROACH o eorr epe etapa anu dig ates Rant 170 10 3 1 Thermal Results Comparison 171 10 3 1 1 Combining the Abaqus thermal 5 171 I0 5 LT T Loading the Abagus tpt files into Excel error c e ves Rcx epis 171 10 3 1 1 2 Combining the Before and After Shake Out Abaqus rpt 173 10 33 12 Loading the Magma txt file into Excel oce eerte 173 10 3 13 560009 up the Matlab te 175 10 3 be Plot ne the 177 T SUR the comparison IMAGE 177 10 3 2 Stress results Comparison approach rente 177 10 3 2 1 I Loading the Abaqus rpt files into 20 177 10 3 22 Loadi
4. E aM et 16 2 TENE ABRERATPSDE NIINCE 17 FIGURE 2 3 STRESS TRAIN CURVES DIFFERENT TEMPERATURES LINEAR HARDENING APPROACH 17 FIGURE 2 4 ONE DIMENSIONAL FRICTIONAL DEVICE REPRESENTING IDEAL PLASTICITY 19 FIGURE 3 1 STEPS SEQUENCE FOR RESIDUAL STRESS 616 1 25 FIGURE 4 1 THE CYLINDER PART UNITS 8 1 6 esta ean 26 FIGURE 4 2 THE MOLD PART UNITS METERS vtae eet intus Geni ouod Pal etae tet 26 FIGURE 4 3 HYPERMESH MESH USED IN ABAQUS see tese esset esee sese een 2 FIGURE 4 WIAGMA SOBT MESH 21 FIGURE 4 5 CONSTRAINING THE RIGID BODY TRANSLATIONS IN X Y AND Z IN THE CYLINDER 29 FIGURE 4 6 CONSTRAINING THE ROTATIONS IN Y AND Z IN A SINGLE NODE SELECTION IN RED FOR THE CYLINDER MODE 30 FIGURE 4 7 CONSTRAINING THE ROTATIONS X IN A SINGLE SELECTION RED FOR THE CYLINDER MODE T 31 FIGURE 4 8 ABAQUS VS MAGMASOFT COOLING CURVES FOR THE CYLINDER MODEL 32 FIGURE 4 9 ABAQUS TOP AND MAGMASOFT BOTTOM THERMAL COLOR SPECTRUMS OF THE LAST STEP AFTER SHAKE OUT OF THE CYLINDER MODEL 2
5. acai EAD 58 FIGURE 5 21 MAGMASOFT COLOR SPECTRUMS FOR THE RESIDUAL MIN PRINCIPAL STRESSES OF THE ce 59 FIGURE 6 1 FRONT AND TOP VIEW OF THE OPTIMIZED HUB 404400000000 61 VI Table of Contents FIGURE 6 2 BOTTOM VIEW OF THE OPTIMIZED MODEL ecce eene 62 FIGURE 6 3 LOCATION OF COOLING AND STRESS POINTS 5 AND 6 FOR THE OPTIMIZED 62 FIGURE 6 4 MESH OF THE OPTIMIZED HUB MODEL USED IN 4 3 64 FIGURE 6 5 MAGMA MESH OF THE OPTIMIZED HUB 040 40 0 00000000 esee serenus 65 FIGURE 6 6 CONSTRAINING THE RIGID BODY TRANSLATIONS IN X Y AND Z IN THE OPTIMIZED HUB 67 FIGURE 6 7 CONSTRAINING THE ROTATIONS IN THE Y AND Z AXES IN THE OPTIMIZED HUB 68 FIGURE 6 8 CONSTRAINING THE ROTATION IN THE X AXIS IN THE OPTIMIZED HUB 69 FIGURE 6 9 ABAQUS VS MAGMASOFT COOLING CURVES FOR THE OPTIMIZED HUB MODEL 70 FIGURE 6 10 ABAQUS THERMAL COLOR SPECTRUMS FOR THE LAST STEP AFTER SHAKE OUT OF THE OPTIMIZED HOB MODE Liere 71 FIGURE 6 11 MAGMASOFT THERMAL COLOR SPECTRUMS FOR THE LAST STEP AFTER SHAKE OUT OF THE OPTIMIZED HUB MODE TT 72 FIGURE 6 12 VON MISE
6. 2 Fa E a n SIS a a ng TEE AT ATTE TA Ta Tm T I m y TD MC TT LT CI Pepe AX ALTES E UIST sr b he ca A T ET E Figure 6 19 Abaqus color spectrums for the residual Min Principal stresses of the Optimized Hub 79 Optimized Hub Results Empty 39 0 23 3 7 6 81 23 8 39 5 55 2 70 9 86 5 102 2 117 9 133 6 149 3 165 0 180 7 MPa Empty 39 0 23 3 7 6 81 23 8 39 5 55 2 70 8 86 5 102 2 117 9 133 6 149 3 165 0 180 7 Y Figure 6 20 Magmasoft color spectrums for the residual Min Principal stresses of the Optimized Hub 80 Optimized Hub Results 6 9 Simulation time of the Optimized Hub Abaqus 99hrs 32min 119hrs 57min 81 Original and Optimized Hub Comparison Original and Optimized Hub Comparison 7 1 Mises Figure 7 1 Original Hub top and Optimized Hub bottom Mises comparison Top View 82 Original and Optimized Hub Comparison 42 2 0 0 000e 00 et 2 20 amp 8e 07 0 000 00 0 0 Figure 7 2 Original Hub top and Optimized Hub bottom Mises comparison Bottom view Original and Optimized Hub Comparison 2 208 07 0 000e 00 Figure 7 3 Original H
7. M VV V OOM M NN BAVA Me sayy ARAN NO cj 22 5 4 7 TATE AAA AII DORR DT GMO SOE mL XO orc 2 402e 08 2 186 03 1 970e 08 1 755 08 1 539e 08 1 323e 08 1 107 08 8 910 07 6 752 07 4 593 07 2 434 07 2 753e 06 AW DIC a OVATE a m K AANA A S a a NY 7 BAD RINT S ritas VEN AY y PS SI nate JUN Nat 74 AA 2 7 7 rAFAVAY V v 4 ow L a A TATATA Fara 1 Ys nys El ay 4 I EDUCAT 7 gt e ava 737 7 f Ae A Taveras ware ve 28251 XN A f SIS NIS ODE aye 42 SF arata L a Figure 6 15 Abaqus color spectrums for the Mises results of the Optimized Hub 75 Optimized Hub Results Mises MPa Empty 157 7 146 4 135 2 123 9 112 7 101 4 90 2 78 9 67 7 56 4 45 2 33 9 22 7 11 4 0 2 C Mises MPa Empty 157 7 146 4 135 2 123 3 112 7 101 4 30 2 78 9 67 7 56 4 45 2 33 9 22 7 11 4 0
8. i Film interaction propert Conv HTC hal v Definition Property Reference Sink temperature 20 Figure 10 32 Cast Ambient Convection interaction ri cu J m Radiation between the casting and the ambient Right click the Interaction collector of the 5 model gt Create gt Name it Cast Ambient Radiation gt Step After Shake Out Type Surface radiation to ambient Continue Select the whole external surface of the casting gt Done gt Set the Edit Interaction window as follow Edit Interaction Mame Cast Ambient Radiation Type Surface radiation to ambient Step After Shake Out Heat transfer Surface Cast Surf Edit Region Emissivity D 76 Ambient temperature 20 I nstantaneous Ambient temperature amplitude Note The absolute zero temperature and Stefan Bolkzmann constant must be specified in the Edit Model Attributes dialog Figure 10 33 Cast Ambient Radiation interaction 130 Appendix 11 Boundary Conditions No mechanical boundary conditions are needed to be specified in the thermal problem 12 Predefined Field Requests Remains unchanged 13 Job creation Expand the Analysis item in the Model tree Right click the Jobs collector gt Create gt Name it ASO RI gt Source Model gt In the displayed models list select 5 gt Continue gt Write a short description of the analysis if desired gt Job Type Fu
9. entities must belong to a collector If we have not created one before the import of the geometry Hypermesh create it automatically which means that all our surfaces get automatically grouped in one collector when we first import the part This collector gets the name of the imported file without the extension which is replaced prt if is not originally a prt file In the case of this simple model we will keep the surfaces of the mold in one collector and will create two more collectors to separate the mesh of the external surfaces and the mesh of the cavity surfaces 99 Appendix Create a collector as follow Organize menu gt Collectors gt Check the create option gt Click the arrow for the collector type and select Component gt Name it Ext Mesh gt Assign a color if desired Create In the same way create a collector named Cavity Mesh and another one called Surface The existent collectors can be seen inside the Components item in the Model Browser tab If the Model Browser is not displayed go to the View menu and select Model Browser Rename the firstly created collector the one with the name of the step file from ProEngineer as Ext Surfaces Right click the collector named MOLD SURFACE PRT gt Rename gt Name it Ext Surfaces gt Return The collectors list should look as x Model utiy E E Assembly Hierarchy amp Components 4 oo bf
10. Figure 10 9 Automesh panel is divided in two for display purposes 101 Appendix By clicking the button a list of selection criterias is displayed To take advantage of the collector choose by collector then check the check box of the Ext Surfaces collector and press the Dye button Similarly mesh the surfaces in the Cavity_Surfaces collector with an element size of 0 01 Remember to Make Current the Cavity Mesh collector and to select elems to currentcomp in the Automesh panel so the elements get automatically collected in the appropriate collector The surface mesh for the mold is now ready Save the file File menu gt Save Note Hypermesh do not recognize CTRL S so use the Save option in the File menu The model should look as Model BES mo 57 Entities SP Assembly Hierarchy Ea Components 4 n Lawity Mesh 1 D m I Surface 2 pg gm 1 4 Mesh 3 E L w 4 Ext Surface Figure 10 10 Mold Surf Mesh hm model Create a Save As copy of the file and name it Cylinder Surface Mesh hm File menu gt Save As 2 File name Cylinder Surf Mesh hm gt Save 102 Appendix Notice that when you make a Save As copy the copy opens immediately closing the file that originate it That means you should be now seeing the Cylinder Surf Mesh hm model in which we will continue working now We want to hav
11. VC US SIS ET Bere ae Sta el pe ear T BU ME 4 OB RENS spl Ei TTT ke Rr Tho Figure 6 17 Abaqus color spectrums for the residual Max Principal stresses of the Optimized Hub 77 Optimized Hub Results MPa Empty 116 0 105 0 94 1 83 1 724 61 1 50 1 39 2 28 2 17 2 48 15 7 26 7 MaxPrinci 116 0 105 0 94 1 83 1 72 1 61 1 50 1 39 2 28 2 17 2 6 2 4 8 15 7 26 7 37 7 Figure 6 18 Magmasoft color spectrums for the residual Max Principal stresses of the Optimized Hub 78 Optimized Hub Results Min Principal C3 3 C3 ar SEG YR T noe rar eee TiN Ge FO Fav arate S PC VEU VV ATITEA y RN AAAS s ANY SNR NIS DADOS 5 SAM A S BA eu LL e vata Unc Evi AU nn A NAVAS AANA Nye A TATATA TATATATA S SIS RR Vi Ok AN XS 5 DUAE LS UE K zl
12. gt CAST gt CHANGE Undo Point Material Uni Gri Ctrl Point Special Figure 10 54 Location of the Material button in the Preprocessor interface 150 Appendix 3 Creating the mold and assigning a material type The material type that happens to be selected before the creation of a new feature Is the one to be assigned to it Therefore to not need to change the material type after creating the mold box the right material type sand mold for us must be selected first Material gt There is more than one way to define the mold box see for example the begin box and the set cube commands in the Magmasoft Online Help documentation we do it with the set cube command defining a body diagonal by introducing the coordinates of two opposite corners via the keyboard Set Cube 1 21 x2 y2 z2 2 Return Example P cube 0 100 100 100 gt Return This results in a cube with a corner in the origin Magmasoft automates the boolean operation that subtract the casting part from the mold box creating the cavity but the decision of which volume must be subtracted from which one is made by the order of the volumes in the volumes list In this automatic cavity creation a volume that appear later in the list is removed from the volume that appear earlier in the list Since our first volume was the Magma_Cylinder it appears first in the list but we need the cylinder to be remo
13. Figure 6 6 Constraining the rigid body translations in X Y and Z in the Optimized Hub By fixing this node in the space three rigid body translations are constrained Consequently the part would shrink toward this node Now the remaining task is to constraint the three degrees of freedom corresponding to the rigid body rotations 67 Optimized Hub Results Rotations in the Y and Z axes A node at the same Z level of the fully constrained one and also in 0 is fixed in Y and Z See Figure 6 7 The node is free to move in the X direction so it can follow a correct shrinking trajectory Edit Boundary Condition YZ Mame Type Displacement Ratation Step Stress 1 Static General Region Picked C5Y5 Transform T BCS CS5YS E 2 Distribution Uniform 1 ET Cus L2 NENNEN us a unt radians CI URZ radians ug3 radians Amplitude Ramp w Note displacement value will be maintained in subsequent steps Figure 6 7 Constraining the rotations in the Y and Z axes in the Optimized Hub 68 Optimized Hub Results Rotation in the X axis A node in 0 with the same X coordinate that the totally fixed node but in a different Z coordinate in this case in the bottom flat surface of the lower inner ring of the cylindrical section is constrained in X and See Figure 6 8 In this
14. ag Figure 10 24 Selection option tools gt Select the whole geometry gt Done gt Section CAST SEC gt OK gt Done 117 Appendix Similarly the section assignment for the Mold is performed as follow Expand the Mold part in the Model tree Right click the Sections Assignment collector gt Create gt Select the whole geometry gt Done gt Section MOLD SEC gt gt Done 5 Mesh Element Type Assign the Heat Transfer element type family to both parts A DC3D4 element type will be automatically selected For the Cylinder Right click the Mesh item under the Cylinder part in the Model tree gt Switch Context gt Mesh menu gt Element Type gt Select the whole geometry Done Element Library Standard Geometric Order Linear gt Family Heat Transfer gt gt Done For the Mold Right click the Mesh item under the Mold part in the Model tree gt Switch Context gt Mesh menu gt Element Type gt Select the whole geometry gt Done gt Element Library Standard gt Geometric Order Linear gt Family Heat Transfer gt OK gt Done 118 Appendix So far the Model tree should look as Madel Results Model Database w w E Models 1 gt E 1 650 a Parts 2 E CYLINDER H Hh Features 1 id Sets Surfaces m Skins Stringers Section Assignments 1 CAST SEC Solid Homogeneous B Composite Layups Engineering Features Mesh M
15. O utput history frequenc y 0 End Step After Shake out PARTS name Cast Ele ment type DC 304 Section Section 1 PIC KEDSET2 Solid Section elset PickedSet2 matenal CASI MAT 12 End Part 189 Appendix ASSEMBLY Assembly name Assembly Instance name 2Cast 1 part2Cast End Instance Surface type LEM ENT na me C AST EXT SURF End Assembly MATERIALS Matenal name CASI MAT C onductivity Density Latent Heat Latent Heat INTERAC TION PROPERTIES Film Property name Conv HIC PHYSICAL CONSTANTS Physic al Consta nts absolute zero 273 15 stefan boltzmann 5 6 7e 08 PREDEFINED FIELDS Cast Initial Type Temperature 190 Appendix Conditions type TEM PERATURE file 29 d okument 5 ac e R3 0db step l inc 506 STEP Before Shake Out Step Before Shake Out extrapolation PARABOLIC inc 43200 Heat Transfer end PERIOD deltmx 1 10 43200 le 12 43200 INTERAC TIONS Interaction SURFFILM 1 ofilm CAST EXT SUFF 20 Conv HTC Interaction SURFRADIA TE 1 Sra diate AST EXT SURF R 20 0 76 OUTPUT REQUESTS Resta rt write frequenc y 0 FIELD OUTPUT F Output 1 O utput field Node Outpu
16. 1 1500 1500 2000 Specific Heat 6 6 1 816 98 820 101 858 127 993 327 1074 527 1123 727 1166 927 1201 1127 1230 1327 1333 33 2000 The graphics of the mold thermal material data can be found in Density Figure 10 21 Conductivity Figure 10 22 Specific Heat Figure 10 23 181 Appendix 10 4 2 Stress Material Data CAST there is no mold in our stress analysis ABAQUS MAGMASOFT Expansion Coefficient Expansion Expansion Coef c eme IC Coane 1 00E 05 20 1 07E 05 20 1 07E 05 200 1 07E 05 199 1 23E 05 400 1 37E 05 201 1 33E 05 600 1 37E 05 399 1 33E 05 1120 1 52 05 401 1 33E 05 1160 1 52 05 599 7 67 06 2000 1 33E 05 601 1 33E 05 1159 1 00E 10 1161 1 00E 10 2000 Young s Modulus 126800000000 20 126800 20 112600000000 200 112600 200 108400000000 400 108400 400 103 00000000 600 103700 600 915 9000000 1120 91579 1120 500000000 1160 500 1160 500000000 2000 900 Poisson s Ratio 200 026 60 0 49 Hardening Coefficient ONLY FOR MAGMA 182 Appendix Plasticity Data ONLY FOR ABAQUS Yield Stress Pa Plastic Strain u Temp C 183 1 429590E 08 0 000000E 00 2 00E 01 1 623300 08 0 000139366 2 00 01 1 793920 08 0 000297 755 2 00 01 1 930600E 08 0 00048416 2 00E 01 2 043150E 08 0 00069044 2 00E 01 2 144730E 08 0 000905715 2 00E 01 2 232710 08 0 00113216 2 00 01 2 309010 08 0 00136818 2 00 01 2 374720E 08 0 001616
17. However is not mandatory The codes Write a skeleton code as in Figure 10 86 1 HMAGHASOFT Thermal Results Magma Time E S 38 Temperature B g 5 ABAQUS Thermal Results 10 Shihaqus Time ii cC 12 13 54 14 Temperature D 15 17 Plot Setting 16 figure 19 plot A B h 20 hold on xlabeli Time 3 23 wvlabeli i Temperature 24 titlel Thermal Results Comparison legend Magqmasott Abagus Figure 10 86 Template code for the thermal comparison 175 Appendix In Fzgure 10 86 we presented our template code for the thermal comparison As you can see all the vectors A B C and D are empty so they have to be populated with the respective information from Excel In the Plot Setting section in Figure 10 86 the first line of code figure 1 was written to assign a name 717 in this case to the graphic to be created This name will not be plotted With the second line plot A B b we specify that we want to plot the A vector in the X axis against the B vector in the Y axis and that we want the color of this curve to be blue b The third line hold on is the one that produce the combination of the curves by including in the same graphic all plots specified after hold on and before a hold off For more informatio
18. TEKNISKA HOGSKOLAN HOGSKOLAN I JONKOPING SIMULATION OF RESIDUAL STRESSES IN CASTINGS Ruben Lora Echavarria Jayesh Vasant Namjoshi THESIS WORK 2007 MACHINE TECHNOLOGY TEKNISKA HOGSKOLAN HOGSKOLAN I JONKOPING SIMULATION OF RESIDUAL STRESSES IN CASTINGS Ruben Lora Echavarria Jayesh Vasant Namjoshi This thesis work is performed at the Jonkoping Institute of Technology within the subject area of Machine Technology The work is part of the university s Master of Science with a Mayor in Mechanical Engineering Specializing in Product Development and Industrial Design The authors are responsible for the given opinions conclusions and results Supervisor Niclas Stromberg Credit points 30 points Date 2 1 2008 Archive number Postal Address Visiting Address Telephone Box 1026 Gjuterigatan 5 036 10 10 00 551 11 J nk ping Abstract Abstract This work presents a study and implementation of the simulation of residual stresses in castings The objects of study are a cast iron truck Hub part provided by the company Volvo 3P and an optimized version of the Hub resulting from the application of a topology optimization process The models are solved through an uncoupled thermo mechanical solidification analysis performed both in the FE commercial software Abaqus and the FD commercial software Magmasoft and the results are compared First a thermal analysis is carried out where the casting 1
19. 2 Procedure type General gt Heat transfer gt Continue 2 Time period seconds 28800 gt Incrementation Type Automatic 2 Maximum number of increments 28800 Initial Increment size 10 Minimum Increment size 1E 12 gt Maximum Increment size 28800 Max allowable temperature change per increment 10 gt Max allowable emissivity change per increment 0 1 gt 8 Predefined Fields definition We assume a homogeneous initial temperature of 1400 C for the casting and 20 C for the mold Cylinder initial temperature Right click the Predefined Fields collector gt Create gt Name it Cast Initial Temp gt Step Initial gt Category Other gt Type Temperature gt Continue 2 Select the whole geometry of the Cylinder gt Done gt Distribution Direct specification gt Section variation Constant through region gt Magnitude 1400 Mold initial temperature Right click the Predefined Fields collector gt Create gt Name it Mold Initial Temp gt Step Initial gt Category Other gt Type Temperature gt Continue gt Select the whole geometry of the Mold gt Done gt Distribution Direct specification gt Section variation Constant through region gt Magnitude 20 120 Appendix 9 Interaction Properties definition The interaction property to be defined describes the heat transfer coefficient HTC between the Cylinder and the Mold We used a constant HTC of 1000 in Magmasoft to
20. I BSO gt Continue gt Write a short description of the analysis if desired gt Job Type Full analysis gt OK Now the Job is ready to be submitted for calculation in the solver Note that the results obtained after calculation of this I BSO model are needed for the setup and further run of the second part of the thermal analysis which now follows For details about how to run the analysis refer to section 10 1 2 125 Appendix 10 1 1 3 1 2 After Shake Out model The After Shake Out model is created from a modified duplicate of the Before Shake Out one In this copy we will remove the Mold from the Assembly module to represent the shake out As long as is not in the assembly it will not affect the simulation but to avoid confusion and to simplify the inp file the Mold and the related information material section sets etc will be removed from the entire model In this model the initial temperature field of the casting will be read from the thermal history stored in the BSO RI odb file of the Before Shake Out analysis The cylinder will be cooled down by means of a convective and a radiate interaction with the ambient applied to the whole external surface of the cylinder The same 13 steps as in the Before Shake Out model will be followed and just the differences will be detailed Copying the model Right click the I BSO item in the model tree gt Copy Model gt Name it 5 gt Removing th
21. In Magmasoft this condition is defined automatically so the user has no participation in the setting 28 Cylinder Results 4 3 2 Mechanical boundary conditions 4 3 2 1 Stress analysis step The user does not participate in the definition of boundary conditions for the stress analysis in Magmasoft It is an automatic procedure Therefore we only present our Abaqus approach The task 1 to restrain the rigid body translations and rotations in X Y and 7 but allow the body to deform naturally to shrink basically In the cylinder model the 6 degrees of freedom has been constrained as follow Translations in X Y and Z In the flat face of the Cylinder lying in Z 0 a node Y 0 is constrained in X Y 7 See Figure 4 5 Notice that in the picture X 1 the horizontal axis Y the vertical axis and the Z axis is perpendicular to the paper Edit Boundary Condition Mame 2 Displacement Rotation Step Initial Region Picked CSY Global U1 U2 U3 CIURI 2 Note The displacement value will maintained in subsequent steps Figure 4 5 Constraining the rigid body translations in X Y and Z in the Cylinder Three rigid body translations have been constrained by fixing the point in the space As a result the part would shrink toward the point However the body could still pivot in our fixed node so now the rotations have to be constrained 29 Cylinder
22. Initial gt Category Mechanical gt Type Displacement Rotation gt Continue gt Select a node as the one with red marks in Figure 41 gt Done gt Check the check boxes for U1 and U2 2 OK Figure 10 41 Semi fixed node aligned in z with the totally constrained one 138 Appendix 12 Predefined Field Requests The Stress components and invariants the Equivalent plastic strain and the Translations and rotations will be requested to be written into the output database The existent Nodal Thermal History field output request must be deleted from this model Note No History Output Request 1s necessary Expand the Field Output Requests collector gt Right click the Nodal Thermal History item gt Delete gt Yes Right click the Field Output Requests collector gt Create gt Name it III Stress Output gt Step Before Shake Out gt Continue gt Set the Edit Field Output Request window as Note When we select the step to which the output request will be applied the request gets automatically propagated to the following steps Edit Field Output Request Mame ITI Stress Qurpuer Step Stress BSO Procedure Static General Domain Whole model Frequency ni Timing Output at exact times Output Variables gt Select From list below Preselected defaults All Edit variables 5 PEEQ U m Stresses m Strains E Displacement velocity Acceleration Forces Reactio
23. Principal Avg 75 IE ess MPa Empty 18 56 14 98 11 41 f 83 4 25 0 6 2 9 6 49 10 06 13 64 17 22 20 80 24 38 27 95 31 53 Figure 4 15 Abaqus top and Magmasoft bottom color spectrums for the residual Min Principal stresses of the Cylinder model 38 Cylinder Results 4 8 Simulation time for the Cylinder Ph in 36min 39 Original Hub Results 5 Original Hub Results 5 1 Geometry Figure 5 1 Front and Top view of the Original Hub model Original Hub Results Figure 5 2 Bottom view of the Original Hub model The mold of the original Hub is just a box with the Hub cavity in its center The external dimensions are 0 7x0 7x0 65 See Figure 5 3 Figure 5 3 The Hub Mold part Units Meters 41 Original Hub Results 5 2 Thermal and stress curves points placement Figure 5 4 Location of the cooling and stress points of the Hub part 5 3 Mesh Magmasoft 42 Original Hub Results Figure 5 5 Mesh of the Hub model used in Abaqus 43 Original Hub Results Figure 5 6 Magmasoft mesh of the Hub model 44 Original Hub Results 5 4 Boundary Conditions 5 4 1 Thermal boundary conditions 5 4 1 1 Before Shake out Conduction Between the external surface of the casting and the surface of the mold cavity For details on how to set this type of boundary condition in Abaqus refer to section 10 1 1 3 1 1 un
24. Select the Job Module in the context bar eens gt Jobs menu gt Submit gt Select the job to be submitted The status of the Job is always presented next to the Job name in the Analysis tree For example if the calculation is running in the solver the message running will appear between parentheses as follow ii Analysis Jobs 3 Figure 10 43 Job status 140 Appendix To monitor the progress of an analysis job Expand the Analysis item in the Models tree Expand the Jobs collector Right click the job that is running Monitor Or Select the Job Module in the context bar Jobs menu Monitor Select the job to be monitored 10 1 3 Post Processing 10 1 3 1 Results Visualization In this step of the process we will obtain a graphical representation of the results through colored spectrums applied to the model and curves which values can be written to text files for further comparisons The results are presented into the Visualization module and read from the odb file They can be read at any moment of the calculation since each time an increment is completed Abaqus write the results to the odb being accessed and an update can be performed First we must load our odb file into the Visualization module 10 1 3 1 1 Loading the Output Data Base Expand the Analysis item in the Models tree Expand the Jobs collector Right click the job you want to see results from e g BSO R1 Results
25. The settings and cut selection is memorized and it can be activated or deactivated with the Activate Deactivate View tool 2 in the toolbox View Cut Manager Create Show Mame C gd Plane Edit Rename s s amp s s s zPlane Delete Options Dismiss Motion of Selected cut 0 1 Sensitivity Figure 10 47 View Cut Manager window 10 1 3 1 3 Removing a part from the viewport When we have more than one part in a model as in the I BSO where we have the Mold and the Cylinder we may face the necessity of remove one part from the viewport of the Visualization module to have a better view of the results of another way to achieve it is 143 Appendix Expand the Output Databases collector in the Results tree gt Expand the odb item of your analysis Job e g 1 4 gt Expand the Instances collector gt Right click the part you want to remove e g MOLD 1 gt Remove This 15 a Boolean operation applied just to the viewport which means that the part is not deleted is just not displayed If is necessary to display the part again the same previous procedure will do it by selecting Add instead of Remove at the end This procedure is a shortcut of the capabilities of the Display Group option for more information see Abaqus CAE User s Manual 10 1 3 1 4 Creating X Y
26. Time Seconds x 10 Figure 10 91 Thermal results comparison of the Cylinder model with and without symmetry 200 10 8 1 4 Stress results 25 Mises Pascal 0 5 1000 800 600 Temperature Celsius 400 200 Appendix Time Second 201 MAGMA MAGMA Symmetry ABAOQUS ABAQUS Symmetry T 1 MAGMA MAGMA Symmetry ABAQUS ABAQUS Symmetry Man F 6 7 8 x 10 Appendix Min Principal Pascal 2 E MAGMA Symmetry ABAQUS ABAQU S Symmetry 1 I 1 2 3 4 6 7 8 Time Seconds X 10 Figure 10 94 Minimum Principal Stress comparison of the Cylinder model with and without symmetry 202 Appendix 10 8 2 Original Hub 10 8 2 1 Geometry A half of the Original Hub has been used for the symmetric analysis as shown in Figure 10 95 Accordingly half of the mold has also been used Figure 10 95 Half of the Original Hub geometry as used in the symmetry analysis 10 8 2 2 Simulation time 7hrs 38min 48hrs 56min No Symmetry 52hrs 6min IOhrs 50min G2hrs 56min 203 Appendix 10 8 2 3 Stress results x10 Misesi Pascal 2 li 71 V 1 j 0 MAGMA MAGMA Symmetry ABAQUS ABAQOUS Symmetry 10 5000 10000 15000 Time Seconds Figure 10 96 Mises comparison of the Original Hub model with and without symmetry 10 Max Princ
27. dip Mesh 8r gb Cavity Surfaces 10 Ig gm dip Ent Mesh Ext Surfaces gm Figure 10 8 Collectors of the Mold Surf Mesh hm model in Hypermesh To move an entity into a collector we proceed as follow Organize menu gt Entities or Shift F11 gt Click the button and specify the type of entity to select from the list gt Select the desired entity or entities gt Click the dest button and select the collector to which you want to assign the entity gt Move Accordingly move the corresponding entities to each one of the collectors previously created The collectors can be hidden or displayed using the Display panel which can be accessed from the View menu Shortcut d key View menu gt Display gt Collectors For details on how to use the Display panel refer to Hyperworks 2006 100 Appendix Notice that for the imported step geometry of our mold Hypermesh do not distinguish between the external surfaces and the surfaces of the cavity even if they are isolated one from the other Instead it considers them as one single surface which represents an obstacle to separate the external from the cavity ones into different collectors To go around this problem the surfaces are copied not moved into another collector then the fist collector is hidden next the surfaces that are not wanted to be in the new collector e g the external ones are deleted Finally the opp
28. postprocessor 2 Press i button the Magmasoft main interface gt geometry The fist time that you choose Postprocessor gt On geometry for the current project version Magmasoft initiates a geometry conversion process to the ACIS 9 format and the ACIS R converter window appears Here you must enter parameters for the control of the conversion process The size of the geometry defines the duration of this process For our case set the ACIS R converter window as in Figure 10 71 gt OK 2 Next ACIS R converter Total amount of available memory Maximum volume size for subtracts default is calculated fram the available memory 35000 Maximum number af facets per volume rdefault is calculated fram the available memori Save model to sat file Extensive STL check Polygon reduction Planar OK Skip Quit Figure 10 71 ACIS converter window For details about this conversion parameters see the MAGMA Online Help documentation 163 Appendix The next time that the Postprocessor get opened it will jump to the Postprocessor main interface For our project it opens by default as Figure 10 72 Control Panel Curves 7 Ser Slice X Ray Vector Dist Anim Material Scales Rotate Images Views Light Results X Ray Clipping Animation Distortion ProcRot Curves Vector Slicil Tracer Print Mesl Project Cylinder_26 Version v
29. 1 16E 03 1 000000E 06 0 000000E 00 2 00E 03 1 050000E 06 0 0179 2 00E 03 1 050000E 06 0 99 2 00E 03 Yield Stress ONLY FOR MAGMA The Abaqus data presented next left was extracted from the Abaqus plasticity data 1 42959E 02 1 19744E 02 9 39556E 01 6 81353E 01 1 00000E 00 1 00000E 00 1 00000E 00 142959000 119744000 93955600 68135300 1000000 1000000 1000000 The graphics of some of the stress material data can be found in Young s Modulus Figure 10 34 Poisson s Ratio Figure 10 35 Expansion Coefficient Figure 10 36 Plasticity Figure 10 37 185 Appendix 10 5 Keywords of the Abaqus input files Next as a reference we present the keywords of our Abaqus input inp files corresponding to the Cylinder simulations Before Shakeout PARTS name Cast Ele ment type 2DC 304 Section Section 1 PIC KEDSET2 Solid Section elset PickedSet2 matenal CASI MAT i End Part Part name Mold Ele ment type DC 304 Section Section 2 PIC KEDSET2 Solid Section elset PickedSet2 matenal MOLD MAT End Part ASSEM BLY Assembly name Assembly Instance name 2Cast 1 part2C ast End Instance 186 Appendix Instance name Mold 1 part Mold End Instance Surface type LEM ENT na me C AST EXT SURF Surface type LEM ENT na me MO LD INT SURF Surface type LEM ENT na me M O LD EXT
30. 10 40 FIGURE 10 41 FIGURE 10 42 FIGURE 10 43 FIGURE 10 44 FIGURE 10 45 FIGURE 10 46 FIGURE 10 47 FIGURE 10 48 FIGURE 10 49 FIGURE 10 50 FIGURE 10 51 FIGURE 10 52 FIGURE 10 53 FIGURE 10 54 FIGURE 10 55 FIGURE 10 56 FIGURE 10 57 FIGURE 10 58 FIGURE 10 59 FIGURE 10 60 FIGURE 10 61 FIGURE 10 62 WINDOW FIGURE 10 63 FIGURE 10 64 FIGURE 10 65 FIGURE 10 66 FIGURE 10 67 FIGURE 10 68 FIGURE 10 69 FIGURE 10 70 FIGURE 10 71 FIGURE 10 72 FIGURE 10 73 FIGURE 10 74 FIGURE 10 75 FIGURE 10 76 FIGURE 10 77 Table of Contents CONDUCTIVITY MATERIAL DATA CURVE FOR THE CYLINDER PART 114 SPECIFIC HEAT MATERIAL DATA CURVE FOR THE CYLINDER PART 115 DENSITY MATERIAL DATA CURVE FOR THE 115 CONDUCTIVITY MATERIAL DATA CURVE FOR THE 116 SPECIFIC HEAT MATERIAL DATA CURVE FOR THE MOLD PART ecce 116 SBEECTIONOP TON TOOLS 117 MODEL TREE AFTER COMPLETING THE FIRST 5 STEPS OF THE SETUP 119 HTC CONDUCTION INTERACTION PROPERTY BETWEEN THE CAST AND THE 121 CONDUCTIVE INTERACTION BETWEEN THE CAST AND THE MOLD cerne 123 CONVECTIVE INTERACTION BETWEEN THE MOLD AND THE 123 RADIATION INTERACTION BETWEEN THE MOLD AND THE AMBIENT 124 F
31. 66 64 1 Thermal boundary conditions 66 OAPI Berre Shake OU teca tuas ca Ret Ma a trai 66 CA2 66 6 4 2 Mechanical boundary conditions 2 67 64 2 I Stress Analysis SE Pran ees e ee eerte teet e ple ieu aes 67 6 5 COOLING CURVES FOR THE OPTIMIZED 70 6 6 THERMAL COLOR SPECTRUM S 53 iiec EP MR OS 71 6 7 STRESS CURVES FOR THE OPTIMIZED essere sese nee 79 6 6 STRESS COLOR SPECTRUMS obe inodo E 75 6 9 SIMULATION TIME OF THE OPTIMIZED 81 EU ML M Bebe 82 TA MAMOM PRINCIPATIZS Mx telae 85 T3 MINIMUM PRINCIPALE 88 8 Conclusions and CISCUSSIONS ccccccccsccccccccccccccccccsccsccsccsccscccces 9 Referernce8s 2 Te UE UI OL 94 e pA PP bond Ud 94 Process eie metet Ee 95 95 POM WZ Mesh 97 the Geometry 98 DD Errori iR Cle
32. 8 a 2 ll 0 1 10 5000 10000 15000 Time Seconds Figure 5 14 Maximum Principal stresses from PNTO of the Original Hub 52 Original Hub Results 1 x MAGMA ABAQUS or 1 1 Min Principali Pascal tn 0 5000 10000 15000 Time Seconds Figure 5 15 Minimum Principal stresses from PNTO of the Original Hub 53 Original Hub Results 5 8 Stress color spectrums Figure 5 16 Abaqus color spectrums for the Mises results of the Original Hub 54 Original Hub Results Mises MPa Empty 191 8 178 1 164 4 150 7 137 0 123 3 109 6 Figure 5 17 Magmasoft color spectrums for the Mises results of the Original Hub 55 Original Hub Results Figure 5 16 Abaqus color spectrums for the residual Max Principal stresses of the Original Hub 56 Original Hub Results MaxPrincipalStress MPa Empty 91 84 81 57 71 30 61 03 50 77 40 50 30 23 19 96 9 69 0 57 10 84 21 11 31 38 41 65 51 92 Y MaxPrincipalStress a MPa Empty 91 84 81 57 71 30 61 03 50 77 40 50 30 23 19 96 9 69 0 57 10 84 21 11 31 38 41 65 51 92 A Figure 5 19 Magmasoft color spectrums for the residual Max Principal stresses of the Original Hub 57 Original Hub Results Figure 5 20 Abaqus color spectrums f
33. Alloy maqma cUHSET 1400 00 Sand Hold magma COLDBOX 20 00 ok prev cancel select data expand hide parameters help Figure 10 58 The material definitions window database request selection property dataset database project group Cast Alloy ok cancel MAGMAdAata help Figure 10 59 Database request window 155 Appendix The material data can be edited from the MAGMAdata window for the project database that automatically appears after importing the materials If you can not see the MAGMAdata window it can be accessed from the button in the Magmasoft main interface and then select Project from the Database menu To edit the casting material Select GJL 150 from the MAGMAdata window as in Figure 10 60 EN From the Edit menu select the parameter to be edited and input the correct data gt Data menu gt Save gt Close gt Database menu gt Quit For details about the correct material data see the Appendix section 10 4 d d Germany DIN 1691 185 66 15 Japan JIS 65501 76 Fo 15 ASTM 848 03 BB Iso 185 Grade 150 Figure 10 60 MAGMAdata window for the project database 156 Appendix Notice that Magmasoft tells us the units being used for each material data See for example Figure 10 61 Figure 10 61 Default Young s Modulus
34. BSO and after shake out II ASO 110 Appendix 10 1 1 3 1 1 Before Shake Out model 1 Importing the models Open Abaqus CAE gt Create Model Database An empty model is automatically created Expand the Model tree left side of the Abaqus CAE user interface under the Model tab Right click the empty model Model 1 gt Rename gt Rename it as I BSO Now we will import the mesh of the cylinder and the mesh of the mold created with Hypermesh in section 10 1 1 2 as two independent new models First the inp file of the cylinder File menu Import gt Model Cylinder inp gt Ok Now the inp file of the mold File menu Import gt Model gt Mold inp gt Ok These models just contain a Part Copy the Part object from the Cylinder model to the model I BSO where the simulation will be set as follow Model menu Copy Objects gt From model Cylinder gt To model I BSO 2 Click the arrow next to the Parts object category gt Check the box next the part name Part I Apply With the Copy Objects dialog box still open copy the Part object from the Mold model to I BSO From model Mold To model I BSO Click the arrow next to the Parts object category gt Check the box next the part name Part I gt Ok Confirm that the parts are in the model I BSO and rename them as Cylinder and Mold respectively Delete the two imported models 111 Appendix The Model tree should look like Model
35. Continue gt Select a node as in Figure 10 38 Done gt Check the check boxes for U1 U2 and U3 gt Figure 10 38 Totally constrained node Boundary Conditions Stress Analysis The Edit Boundary Condition window should look as Edit Boundary Condition Mame Ll 123 Displacement Riotation Step Initial Region Picked Edit Region cays Figure 10 39 Edit Boundary Conditions window for a fully constrained node 137 Appendix In the same face another node aligned with the first one in the direction of one of the axis parallels to the face in this case the x axis will be constrained in the other two directions that are not aligned allowing contraction In our study case the node will be free in X and fixed in Y and 7 Right click the BCs collector of the III STRESS model gt Create gt Name U 23 gt Step Initial gt Category Mechanical gt Type Displacement Rotation gt Continue gt Select a node as the one with red marks in Figure 10 40 gt Done gt Check the check boxes for U2 and U3 gt OK Figure 10 40 Semi fixed node red aligned in the x direction with the totally fixed one In the opposite end face a node aligned with the first one in the axis of the length in this case the Z axis will be constrained in the other two axes Right click the BCs collector of the III STRESS model gt Create 2 Name it U 12 Step
36. IN THE OPTIMIZED 46 FIGURE 5 8 CONSTRAINING THE ROTATIONS IN THE X AND Z AXES IN THE ORIGINAL 47 FIGURE 5 9 CONSTRAINING THE ROTATION IN THE X AXIS IN THE ORIGINAL 48 FIGURE 5 10 ABAQUS VS MAGMASOFT COOLING CURVES FOR THE HUB 2 49 FIGURE 5 11 ABAQUS THERMAL COLOR SPECTRUMS FOR THE LAST STEP AFTER SHAKE OUT OF THE ORIGNAL HOD MODED emm 50 FIGURE 5 12 MAGMASOFT THERMAL COLOR SPECTRUMS FOR THE LAST STEP AFTER SHAKE OUT OF THE ORIGNAL HUB MODEL a a a 51 FIGURE 5 13 VON MISES CURVES FROM OF THE ORIGINAL 52 FIGURE 5 14 MAXIMUM PRINCIPAL STRESSES FROM PNTO OF THE ORIGINAL HUB 52 FIGURE 5 15 MINIMUM PRINCIPAL STRESSES FROM PNTO OF THE ORIGINAL 53 FIGURE 5 16 ABAQUS COLOR SPECTRUMS FOR THE MISES RESULTS OF THE ORIGINAL HUB 54 FIGURE 5 17 MAGMASOFT COLOR SPECTRUMS FOR THE MISES RESULTS OF THE ORIGINAL HUB 55 FIGURE 5 18 ABAQUS COLOR SPECTRUMS FOR THE RESIDUAL MAX PRINCIPAL STRESSES OF THE e eru EE DD MM Lt EIE EE 56 FIGURE 5 19 MAGMASOFT COLOR SPECTRUMS FOR THE RESIDUAL MAX PRINCIPAL STRESSES OF THE ORIGINAL HU D arerin 57 FIGURE 5 20 ABAQUS COLOR SPECTRUMS FOR THE RESIDUAL MIN PRINCIPAL STRESSES OF THE ORIGINAL HUD es pe
37. Images views Print Results N Styles AL Axis x Options x EX Solidification J Temperature a Stress Figure 10 77 Cooling curve selected for display in the Control Panel window see the stress curves Select Curves from the Postprocessor s Control Panel window From the Results group list expand the Stress item gt From the list displayed under Stress select all che results you want to see for the to be selected stress point From the Curves list select the curve corresponding to the stress point that you are interested in 167 Appendix Temperature Curve 3 T iD i Cc a Le Time s Figure 10 78 Cooling curve display in the Postprocessor main window To see the position of the cooling and stress points in the Postprocessor Select Results from the Postprocessor s Control Panel window gt From the Results list expand the Geometry item gt Double click Cooling points or Stress points as needed gt The points will appear in the main window Notice that the points are displayed over the actual mesh of the part 168 Appendix 10 2 3 2 Results Preparation for Comparison The approach consist in export the thermal and the stress curves from the Magmasoft Postprocessor to text files txt from where the X Y data of the curves can be obtained in a table form at that can be used in a software as Matlab where plots of the Abaqus
38. MODEL WITH AND WITHOUT ARR 202 FIGURE 10 95 HALF OF THE ORIGINAL GEOMETRY AS USED IN THE SYMMETRY ANALYSIS 203 FIGURE 10 96 MISES COMPARISON OF THE ORIGINAL HUB MODEL WITH AND WITHOUT SYMMETRY 204 FIGURE 10 97 MAXIMUM PRINCIPAL STRESS COMPARISON OF THE ORIGINAL HUB MODEL WITH AND NEITHOUT SYM VIELE EA 204 FIGURE 10 98 MINIMUM PRINCIPAL STRESS COMPARISON OF THE ORIGINAL HUB MODEL WITH AND WITHOUTSYNIMETRY 25 c EU EE 205 FIGURE 10 99 HALF OF THE OPTIMIZED HUB GEOMETRY AS USED IN THE SYMMETRY ANALYSIS 206 FIGURE 10 100 THERMAL RESULTS COMPARISON OF THE OPTIMIZED HUB MODEL WITH AND WITHOUT trc s red S Ra E LE aM Mt M nM Lc M te 207 FIGURE 10 101 MISES COMPARISON OF THE OPTIMIZED HUB MODEL WITH AND WITHOUT SYMMETRY 208 FIGURE 10 102 MAXIMUM PRINCIPAL STRESS COMPARISON OF THE OPTIMIZED HUB MODEL WITH AND RA EN A TOC aw ase 208 FIGURE 10 103 MINIMUM PRINCIPAL STRESS COMPARISON OF THE OPTIMIZED HUB MODEL WITH AND AITHOUTISYMMEBEBR nete e etai 209 Introduction 1 Introduction During the solidification process of castings residual stresses are developed due to temperature gradients between different parts of the casting mechanical constraints imposed by the mold du
39. Poisson s Ratio d Hardening Coefficient Only in Magmasoft e Plasticity Only in Abaqus i Yield Stress Plastic Strain 9 Define the initial boundary condition a Initial temperature of the casting as in the thermal analysis 10 Load the nodal thermal history generated in the thermal simulation as a predefined temperature field Only in Abaqus 11 Define the mechanical boundary conditions Only in Abaqus a Constrain the rigid body translations and rotations in X Y and Z but allow the body to deform 24 Implementation 3 2 Procedure The simulation procedure steps are explained in details in the Appendix sections 10 1 and 10 2 according to the following diagram Pre Processing Geometry Definition Mesh Generation Simulation Setup Solidification Calculation Solution of the governing differential equations Stress strain analysis Post Processing Results Visualization Results Preparation for Comparison Results Comparison Figure 3 1 Steps sequence for residual stress analysis In sections 10 1 and 10 2 to focus on the methods and to minimize geometry related problems we go through the whole process using a simple geometry specifically a cylinder The Results Comparison method is treated apart in the Appendix section 10 3 All the needed steps to perform our simulations in Abaqus and Magmasoft are presented in a sequence format so the reader can use them as a step by step guide of
40. a residual stress simulation The casting material for all the models is grey iron and our mold material is based in the Coldbox sand material defined in the Magmasoft database for details on the material data refer to the Appendix section 10 4 25 Cylinder Results 4 Cylinder Results 4 1 Geometry Figure 4 1 The Cylinder part Units Meters Figure 4 2 The Mold part Units Meters 26 Cylinder Results 4 2 Mesh Notice that Magmasoft uses the Control Volume Finite Difference Method so for each element there is only one node which is positioned in the center of the element This justifies the fact of having the same number of nodes as elements in Magmasoft We tried to match the number of nodes for the casting in both softwares Still it is difficult to control the number of elements assigned to the casting and the mold in Magmasoft therefore the difference For details about how to mesh the parts in Abaqus and Magmasoft refer to sections 10 1 1 2 and 10 2 1 2 respectively Figure 4 3 Hypermesh mesh used in Abaqus E 3 Figure 4 4 Magmasoft mesh 27 Cylinder Results 4 3 Boundary Conditions 4 3 1 Thermal boundary conditions 4 3 1 1 Before Shake out Conduction Between the external surface of the casting and the surface of the mold cavity For details on how to set this type of boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 9 Interaction
41. and Magma curves are combined to perform the actual comparison 10 2 3 2 1 Exporting a curve With the curve you want to export displayed in the main window do as follow Go to the Options tab as in Figure 10 79 gt On the Spread sheet file subdivision of the Options tab assign a filename to the curve gt Press the Write File button The curve is exported as a text file txt to the folder of the current project version Control Panel Curves Images Views Print Results Styles Axis Options 54 rid None Legend Position right Orientation top right Ho of curves shown 20 to hide legend Explore data Mode single value Interpolate no 1 Spread sheet file Filename curves txt Write File Figure 10 79 Exporting the curves from the Curve s Options tab 169 Appendix 10 3 Results Comparison Approach The comparison consists in combining the plots of the result curves from the different models We do so in Matlab but first the results are loaded 1n Excel where we adjust all the data to have the same units and solve other problems explained in this section We compare results generated from the same point in the geometry of the models involved Matlab do not recognize the report rpt files generated by Abaqus and therefore we make then pass by Excel However it does recognize the text txt files from Magmasoft but we c
42. condition If we assume perfect thermal contact the heat leaving one body must be equal to that entering the other In which case for a point P in the contact surface T P t T P t 2 19 915 2 20 The subscripts 1 and 2 refer to the two bodies 10 Theoretical Background 2 1 4 The Heat Conduction Equation 2 1 4 1 1 D transient time dependent heat conduction equation OT O 0T Sep 50 2 21 Ot 2 227 Where is the internal generation of heat per unit time per unit volume present within the body 2 1 4 2 1 D steady state heat conduction equation Since the steady state is independent of time is defined as ILL m 0 2 22 Ox C 2 1 4 3 The 3 transient Heat Conduction Equation For casting processes it represents the basis of all heat conduction calculations Is the general form of the heat conduction equation and 15 as follows 2 RUE pre 2 23 gt Ot Ox Ox 02 If we consider no Q gen and replace C by its equivalent value we get H OT OT Qro T _ e er o uer a er OT Ot OxV Ox 02 02 Which is the same equation presented in section 1 1 as the classical heat equation p 2 25 11 Theoretical Background 2 1 5 Numerical solutions The purpose of the numerical solution of partial differential equations is to determin
43. heat of the particular substance 2 1 2 5 Thermal diffusivity Is the ratio of the thermal conductivity to the volumetric heat capacity of the material k 2 13 Where represents the volumetric heat capacity Mediums with high thermal diffusivity reach thermal equilibrium rapidly with their surroundings due to their capacity of fast heat transfer compared with their mass Theoretical Background 2 1 3 Initial and Boundary conditions Initial conditions and boundary conditions are needed together with the heat conduction equation to fully define a transient thermal problem If the given problem is in steady state there is no necessity to define initial conditions The initial conditions represent the initial temperature distribution throughout the body In casting processes the initial condition 1 assumed to be constant throughout the mould is also assumed to be constant for the melt in the mould filling simulation where the temperature will be a superheating temperature For the solidification simulation the initial condition 1 given by the temperature field immediately after filling For simplicity when there is no interest in the mould filling simulation a constant temperature throughout the melt after filling can be assumed and the superheating temperature of the melt can be used as initial condition for the solidification simulation 2 Next five types of boundary conditions relevant
44. history is requested to be written in the output database odb file Right click the Field Output Requests collector gt Create gt Name it Nodal Thermal History gt Step Before Shake Out gt Continue 2 Set the Edit Field Output Request window as in Figure 10 30 124 Appendix Edit Field Output Request Marne Modal Thermal History Step Before Shake Out Procedure Heat transfer Domain Whole madel v Timing Output at exact times Output Variables gt Select From list below Preselected defaults All Edit variables b Displacement velocity Acceleration Energy E Thermal MT Madal temperature F TEMF Element temperature Mote Error indicators are not available when Domain is Whole Model or Interaction Output for rebar Output at shell beam and layered section points gt Use defaults Specify Include local coordinate directions when available Figure 10 30 Field Output Request of the Nodal Thermal History Note Is not necessary to create any History Output Request 13 Job creation When a Job is created an input file inp for the FE solver is written This file will not be read until the Job is submitted for calculation To create the Job do as follow Expand the Analysis item in the Model tree gt Right click the Jobs collector gt Create gt Name it BSO R1 gt Source Model 2 In the displayed models list select
45. in the form of waves or moving subatomic particles Among the radiation types we are specifically interested in the Thermal radiation Thermal radiation 15 heat transfer by the emission of electromagnetic waves from the surface of an object due to temperature differences which carry energy away from the emitting object The basic relationship governing radiation from hot objects 15 called the Stefan Boltzmann law 4 p 2 3 Where is the coefficient of emissivity 21 for ideal radiator is the Stefan Boltzmann constant of proportionality 5 669E 8 W m2K4 A is the radiating surface area is the temperature of the radiator and is the temperature of the surroundings The three of the previously mentioned heat transport mechanisms can be expressed by the model law that state that a flux is proportional to a difference in driving potential divided by a resistance in our case Theoretical Background 2 4 Being the Thermal Resistance Fu for each one of them as follow con Ax R kA 2 5 1 iA 2 6 1 2 7 hA 2 7 Where P is ug cmm 2 8 Theoretical Background 2 1 2 Material properties 2 1 2 1 Thermal conductivity k Is the ability of a material to conduct heat It is defined as the quantity of heat transmitted during a period of time A through a thickness L in a direction normal to a surface of area due to a temperature dif
46. into Excel Open Excel gt File menu gt Open 2 Files of type All files gt Browse the Abaqus rpt file for the stress result Open gt In the Text Import Wizard Step 1 of 3 window select Fixed width Next In the Text Import Wizard Step 2 of 3 window adjust the vertical separation line to properly divide the two data columns Next In the Text Import Wizard Step 3 of 3 window select Colum data format General Confirm the columns are correctly separated gt Press Finish 10 3 2 2 Loading the Magma txt file into Excel Open Excel gt File menu gt Open gt Files of type Text files gt Browse the Magma txt file gt Open gt In the Text Import Wizard Step 1 of 3 window select Delimited gt Next In the Text Import Wizard Step 2 of 3 window check the Tab check gt Next gt In the Text Import Wizard Step 3 of 3 window select Colum data format General Confirm the columns are correctly separated gt Press 177 Appendix 10 3 2 3 Modifying the units of the Magma XY data Our stress results from Magmasoft are given in Mega Pascals MPa while the Abaqus stress results are given in Pascals Pa We choose to adapt the Magma units to the Abaqus ones consequently we have to multiply each value of the stress column of the Magma txt file by a coefficient of 1 00 06 Go to Excel gt Create a column for the coefficient containing the value 1 00 06 in as many cells as
47. is P Qo And y 2 0 is determined by the following consistency condition yf o a 0 if f o a 0 2 41 20 Theoretical Background 2 2 3 5 J2 Plasticty model The yield function which defines the elastic range of the material behavior and when plasticity begins 1s governed by the second invariant of the deviatoric stress tensor This is known as flow theory and is given as 88 2 42 The flow theory makes the yield function independent of hydrostatic pressure 2 2 3 5 1 J2 Plasticity model Constitutive Laws The equilibrium equation of a quasi static mechanical problem 15 divo 0 2 43 Where o is the stress tensor The total strain is governed by 1 s Wu 2 44 Where u represents the displacement The constitutive law is e 2 45 Where D D T is the temperature dependent elastic tensor The thermal strain is of the form al T T 2 46 Where a a T is the thermal expansion parameter 21 Theoretical Background The yield function is defined as f c T Y lt 0 2 47 Where is the linear hardening parameter The flow rule is given by 2 48 0 Where 1 the plastic multiplier and which value is govern by the Karush Kuhn Tucker conditionsy gt 0 T X0 and y oc T 0 dt 2 49 22 Implementation 3 Implementation 3 1 Process Summary Next we present a general list of w
48. strain relationship Is given by c E e e 2 32 The yield condition is defined from the assumption that the absolute value of the stress the frictional device cannot be greater than o 7 0 f o lo 0 lt 0 2 33 If ffo lt 0 the is zero and the instantaneous response of the device is elastic If ffo 0 the frictional device slip with constant slip rate in the direction of the applied stress The following expression describes the flow rule OO Where represents the slip rate and 15 70 19 Theoretical Background The conditions that the stresses must be admissible and the plastic flow can take place just on the yield surface are known as Kuhn Tucker complimentary conditions and mathematically look like gt 0 lt 0 0 2 35 A final condition known as consistency condition must be stated that is 0 if f o 0 2 36 As mentioned before this mathematical model corresponds to ideal plasticity The constitutive model to account for isotropic linear hardening effects has the following differences A hardening law Oey 2 37 The Yield condition changes to le lo Ka 0 where 20 gt 0 K20 2 38 Here Kis the plastic modulus and is a function of the amount of plastic flow slip known as an internal hardening variable The Kuhn Tucker complementary conditions are now y20 f o a lt 0 0 2 39 The flow rule
49. stress values exist in the file gt Create another column for the stress with new units where the value of each cell would be equal to the value in the same row of the original data multiplied by the coefficient e g in Figure 10 88 the value of the cell F6 would be B6 D6 Ensure that the first numerical data from each column is in the same row B F Curve 1 2 Time Mises 3 5 Coefficient Mew Mises 4 1 0 000000 5 0 0 0 000000 0 002 0 000141 1 00 906 141 486000 6 36077 1 63607 70 000000 200 10 5995 1 10599600 096000 9 127192 1 127 19200 000000 10 400 13 4409 1 1 5440900 000000 11 500 13 478 1347 9000 000000 12 BOO 13 2322 1 00 06 13232200 000000 13 00 12 867 1 12857000 14 BUD 12 4483 1 UD E Ib 12448300 DB 000 15 900 11 995 11993000 000000 16 11 5349 1 00 906 11524900 000000 17 1100 11 0946 1 11094600 000000 18 1200 10 5815 1 00 E416 10689 1600 000000 19 15300 10 2935 1 10293300 000000 20 1400 9 93637 1 00 06 g93637 0 000000 21 1500 9 59507 9595070 000000 22 1600 9 28254 1 00 906 9202540 000000 23 1700 8 89214 099140 000060 24 1900 8 71725 1 of 1724250 00000 25 1900 8 465273 1 00 462 30 000000 25 2000 8 22375 1 00 06 0223r SU 000000 27 2100 7
50. stresses are calculated using an uncoupled thermo mechanical solidification analysis A thermal analysis is performed first and then the thermal history is read into a quasi static mechanical analysis to calculate the residual stresses using a J2 plasticity model An academic problem is set using a simple geometry to implement and explain the procedure Then residual stresses are calculated on the truck Hub part provided by Volvo and finally the same simulation is performed on a topologically optimized version of the mentioned part Introduction 1 1 Background The thermal analysis The governing equation for the thermal analysis is the classical heat equation 1 1 Where p k are temperature dependent and represent density enthalpy and thermal conductivity respectively Tis the temperature and 2 is the time The stress analysis The equilibrium equation for the residual stress analysis is diva 0 1 2 Where o is the stress tensor The yield surface equation for the J plasticity model reads f o T 3J he o lt 0 1 3 Where J 1 the second invariant of the deviatoric stress tensor is the temperature dependent hardening parameter is the equivalent plastic strain and is the temperature dependent yield strength Introduction 1 2 Purpose and aims 1 Compare the residual stress development of parts subjected and not subjected to topology optimization processes 2 P
51. use a similar HTC in Abaqus the interaction is defined through a clearance dependent data set as in Figure 10 26 Right click the Interaction Properties collector gt Create gt Name it CAST MOLD CONTACT INTERACTION PROPERTY gt Contact Continue Thermal menu Thermal Conductance 2 Check Use only clearance dependency data gt Fill in appropriate data see the Appendix section 2 OK The Edit Contact Property window may look as Ell dit Contact Property Mame CAST MOLD COPRTACT INTERACTIORN PROPERTY Contact Property Options Thermal Conductance Geometric Properties Mechanical Thermal Thermal Conductance Use only clearance dependency data Use only pressure dependency data Use both clearance and pressure dependency data Clearance Dependency Use temperature dependent data Use mass flow rate dependent data Standard only Number of field variables Conductivity Clearance 1000 0 Figure 10 26 HTC Conduction interaction property between the cast and the mold 10 Interactions definition The conduction between the casting and the mold is described by means of a surface to surface contact interaction using the previously created property in the step 9 The interaction is assigned to the Initial step 21 Appendix Another approach consist in simulate perfect conduction by means of a TIE Constraint but the contact interaction produce better comparisons b
52. 0 00 gt Solid 711 19h 43min 20s 100 00 lt gt Solid 712 19h 45min 00s 100 00 a lt 2 Solid 713 19h 46min 40s 100 00 lt 2 Solid 714 19h 48min 20s 100 00 Q Solid 715 19h 50min 00s 100 00 P lt 2 Solid 716 19h 51min 40s 100 00 mae Solid 717 19h 53min 20s 100 00 1 lt 2 Solid 718 19h 55min 00s 100 00 a Solid 719 19h 56min 40s 100 00 X Solid 720 19h 58min 20s 100 00 z lt gt Solid 721 20h 00min 005 100 00 Stress E pe ET Solid_721 t 20h 00min 00s P 100 00 Apply User Results Reload Figure 10 74 Temperature field result displayed in the Postprocessor s main window Similarly to see the stress results Select Results from the Postprocessor s Control Panel window From the Results list double click Stress Under the Stress item double click Stress gt From the list of stress results that appear double click the one that correspond to the stress type and time that you are interested in 10 2 3 1 1 Creating a cut view Select Slice from the Postprocessor s Control Panel window gt Check the check box for the Activate deactivate slicing dialog Choose a Slice direction The cut is displayed in the main window so move the slide bar to adjust the cut The slide bar can be moved with the mouse or the left and right arrow keys of the keyboard The cut view remain 15 applied to all the results that you display in the main window until the Activat
53. 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 4 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 6 00 02 Appendix 1 618970 08 0 0105471 6 00 02 1 627120 08 0 0108175 6 00E 02 1 637110 08 0 0111259 6 00 02 1 644790E 08 0 0113966 6 00E 02 1 653270E 08 0 0117063 6 00E 02 1 660060E 08 0 011991 1 6 00E 02 1 668310E 08 0 0122878 6 00E 02 1 675400E 08 0 0125722 6 00 02 1 683160 08 0 0128692 6 00 02 1 689570 08 0 0131675 6 00 02 1 696200 08 0 0134655 6 00 02 1 702530 08 0 0137505 6 00 02 1 709090E 08 0 0140486 6 00E 02 1 715030 08 0 0143472 6 00 02 1 720500 08 0 0146329 6 00 02 1 726920 08 0 0149442 6 00 02 1 731860E 08 0 0152172 6 00 02 1 737160E 08 0 0155295 6 00E 02 1 741900E 08 0 015829 6 00E 02 1 745880E 08 0 0161028 6 00E 02 1 833170 08 0 988303 6 00 02 1 000000E 06 0 000000E 00 1 12E 03 1 050000E 06 0 019989 1 12E 03 1 050000E 06 0 99 1 12E 03 1 000000E 06 0 000000E 00 1 16E 03 1 050000E 06 0 0179 1 16E 03 1 050000E 06 0 99
54. 00 02 2 399900 08 2 435650 08 2 473230 08 2 504700 08 2 534860 08 2 565580 08 2 590850 08 2 618460 08 2 642040 08 2 664980 08 2 688010 08 2 707800 08 2 727910 08 2 746250 08 2 761730 08 2 899820 08 6 813530 07 7 328360 07 7 829640 07 8 290950 07 8 762440 07 9 253260 07 9 694570 07 1 008660E 08 1 045960E 08 1 084920E 08 1 118360 08 1 151410E 08 1 177690E 08 1 208230E 08 1 235960E 08 1 263380 08 1 286670 08 1 311810 08 1 337120 08 1 357820 08 1 377000 08 1 394580 08 1 413970 08 1 432950 08 1 449730 08 1 463780 08 1 478850 08 1 494900 08 1 508070 08 1 521220E 08 1 533070 08 1 546070 08 1 557300 08 1 568860E 08 1 579110E 08 1 589350E 08 1 600560E 08 1 610410E 08 Appendix 0 00380089 0 00407479 0 00434701 0 00461697 0 00488039 0 00516621 0 0054415 0 00571469 0 00598377 0 00626864 0 00655337 0 00684093 0 00711289 0 00740163 0 00769286 0 987431 0 000000E 00 0 000257529 0 000503261 0 000765946 0 00101438 0 00127428 0 00153873 0 00179445 0 00206517 0 00232106 0 00259532 0 00286988 0 00313745 0 00341427 0 00368045 0 00397337 0 0042436 0 00451206 0 00479363 0 0050794 0 00535333 0 00562869 0 00591563 0 00621616 0 00649218 0 0067707 0 00706152 0 00735141 0 00763067 0 00792319 0 00820362 0 00849623 0 00877717 0 00907105 0 00937932 0 00964791 0 00995523 0 0102506 184 4 00 02 4
55. 01 Directory 260 19 Cast Alloy 03 01 Sand Mold GJL 150 View Geometry Apply Figure 10 72 Postprocessor main interface see the thermal results Select Results from the Postprocessor s Control Panel window From the Results list double click Solidification gt double click Temperature From the list of thermal results that appear double click the one that correspond to the time that you are interested in Curves XYZ Scr Control Panel X Ray Vector Dist Anim Scales Rotate Images Views Light Results X Ray Clipping Animation Distortion ProcRot Vector Slicing Tracer Print Mesh Result selection Group wr Switch to Browser Mode Deselect Percen EAS Geometry Solidification at stress Figure 10 73 Results tab selected in the Postprocessor s Control Panel window 164 Appendix Material Scales Rotate Images Views Light X Ray Clipping Animation Distortion ProcRot Curves Vector Slicing Tracer Print Mesh Result selection Group Results Solidi fication Temperature Switch to Browser Mode Reus Tme Peren 2 Solid 704 19h 31min 40s 100 00 be lt gt Solid 705 19h 33min 20s 100 00 lt 2 Solid 706 19h 35min 00s 100 00 Se Solid_707 19h 36min 40s 100 00 lt a Solid 708 19h 38min 20s 100 00 Q2 Solid 709 19h 40min 00s 100 00 X Solid 710 19h 41min 40s 10
56. 2 X Figure 6 16 Magmasoft color spectrums for the Mises results of the Optimized Hub 76 Optimized Hub Results ees mutat P PLU SS ee a ena te 0 Am DP nh Tur i V atu fam LESS E 1 26 78 0 val ahs O64e 07 ar ea e Ta va 2 ales zy i ran FATA Le YaA OATS ini IRIS OA n Qt Po ROGA mof dg MM od A VANS A CAUSA SINN EE J TIN vii 4 ALA 0 IN NA SAEZ RRA RIS d SORA iE m US N QT ITE 4 F T Fal Xa wa Fd Ta ts LA Lr Y a Y TA Es Thy Y AFAFA Lr oer ti Lan A 1 E Farag a 7 e nd coo Fi A un P Deor NEA e TS SPESE UNE AGORA Tac ur tke 3 eR eer relate atn pun UPS TS S D EY E MBA SEC VES Crus a eral vs TES NS T CR VAYA 1 ATAT Fi Fi m Tati TA Jj dn Ps Sas Pid ES S PIS Esos
57. 86 2 00E 01 2 434780E 08 0 00186615 2 00E 01 2 489900E 08 0 00212343 2 00E 01 2 537010 08 0 00237927 2 00 01 2 580740 08 0 00264977 2 00E 01 2 618640E 08 0 00291306 2 00E 01 2 749570 08 0 987724 2 00 01 1 197440 08 0 000000 00 2 00 02 1 372210 08 0 000150232 2 00E 02 1 526430E 08 0 000318051 2 00E 02 1 661990E 08 0 000501695 2 00E 02 1 779330E 08 0 000705482 2 00E 02 1 877650 08 0 000916054 2 00 02 1 964240 08 0 0011459 2 00 02 2 042710 08 0 00137793 2 00E 02 2 109960E 08 0 00162412 2 00E 02 2 168660E 08 0 0018 285 2 00 02 2 224920 08 0 00212359 2 00E 02 2 273760 08 0 00238524 2 00 02 2 315650 08 0 00264341 2 00 02 2 357800 08 0 00291061 2 00 02 2 394380 08 0 00317783 2 00 02 2 427480 08 0 00345259 2 00 02 2 458020 08 0 0037248 2 00 02 2 483570 08 0 00400585 2 00 02 2 607750 08 0 987776 2 00 02 9 395560 07 0 000000 00 4 00 02 1 109670 08 0 000147702 4 00 02 1 257960 08 0 00031501 4 00 02 1 391810 08 0 000502729 4 00 02 1 513970 08 0 000693089 4 00 02 1 615460 08 0 000901 722 4 00 02 1 719130E 08 0 00110838 4 00E 02 1 807870E 08 0 00132821 4 00E 02 1 896110E 08 0 00155609 4 00E 02 1 973380E 08 0 00177833 4 00E 02 2 040020E 08 0 00201 76 4 00 02 2 108630 08 0 00227036 4 00 02 2 166270 08 0 00250988 4 00 02 2 217540 08 0 00276262 4 00 02 2 270110 08 0 00302181 4 00 02 2 316240 08 0 00328666 4 00 02 2 359570 08 0 00353864 4
58. 9947 2984 700 000000 20 JA 77 7988 1 00 fe SoU 000000 29 2300 7 57117 1 UD E Ib 75421170 00 0006 20 2400 2467 fo 46 0 000000 Figure 10 88 Changing Magmasoft stress curve results from MPa to Pa 178 Appendix 10 3 2 4 Setting the Matlab M File Proceed as in section 10 3 1 3 10 3 2 5 Plotting the comparison Proceed as in section 10 3 1 4 10 3 2 6 Exporting the comparison image Proceed as in section 10 3 1 5 179 Appendix 10 4 Material Data 10 4 1 Thermal Material Data Note Apart from the Latent Heat which just differs in the units the rest of the thermal material data is common between Abaqus and Magmasoft CASTING Densit Specific Heat 7100 1 450 1 7074 5 100 467 30 7049 1001 200 506 100 7023 7998 300 563 200 6998 6001 400 621 300 6850 6001 1000 663 400 6814 1160 741 500 6882 1173 851 600 6813 8799 1255 1036 700 6745 25 1355 1100 725 6310 1802 2000 744 810 744 900 Conductivity 804 1000 830 1100 1 844 1160 100 740 1173 200 747 1200 300 778 1300 400 813 1400 500 854 1500 1160 871 1600 1173 872 1700 2000 872 2000 General Parameters ABAQUS MAGMASOFT Latent Heat J Kg 230000 Latent Heat KJ Kg Liquidus Temp C Liquidus Temp C Solidus Temp C Solidus Temp C The graphics some of the casting thermal material data can be found in Density Figure 10 18 Conductivity Figure 10 19 Specific Heat figure 10 20 180 Appendix
59. AZS IN CONG Zi FFAG deos CE DRY 5 6 i Mask include _ comp El Cavity Mesh mask C unmask Figure 10 13 A masked view of the Cylinder model where inner elements can be seen The Mold Open the Mold Surf Mesh hm model File menu gt Save As Name it Mold Volume Mesh hm Save Now the active file should be the Volume Mesh hm model 105 Appendix The option tetra mesh is used here instead of volume tetra as in the Cylinder volume mesh Volume tetra just allows the selection of one single closed volume identified by the surface that enclose it and while meshing it ignore other closed volumes that could be inside the selected one This work for the cylinder but for the mold we want to mesh a volume between two closed volumes i e the cavity volume and the external walls of the mold In this case the tetra mesh option is used which identify closed volumes by the surface elements that enclose them and allow multiple selections This option meshes the volume in between the cavity and the mold external walls I our case we can select all the elements but if it would be necessary the surface elements be selected by collectors or individually For the mold also create a new collector called Volume Mesh as explained in section 10 1 1 2 3 to store the volume mesh Activate it Make Current Mesh menu gt Tetramesh gt Select tetra mesh as the me
60. Browser gt Press the button in the Export Abaqus deck window gt Browse a destination folder and name the file Cylinder inp Save For the Export option in the Export Abaqus deck window select all cal gt Ok The inp file of the Cylinder mesh is now exported and ready to be used in Abaqus Save the file File menu Save 109 Appendix 10 1 1 3 Abaqus Simulation Setup Here a step by step procedure to setup and run first the thermal simulation and then the stress simulation in Abaqus CAE V6 7 is presented and commented Setup Overview 1 Importing the models 2 Materials definition 3 Sections definition 4 Sections assignment 5 Mesh element type 6 Assembly 7 Steps definition 8 Predefined Fields definition 9 Interaction Properties definition 10 Interactions definition 11 Boundary Conditions 12 Predefined Field Requests 13 ob creation 10 1 1 3 1 The Thermal Simulation We will include a shake out process in our thermal simulation which means that before completion of the cooling the casting is removed from the mold and is left to cool down until room temperature exposed to the ambient This implies that the mold must be present in the simulation corresponding to the before shake out BSO period and must not be in the simulation of the after shake out ASO period Therefore in this steps guide the thermal simulation will be found divided in two models before shake out to be called I
61. C Al gen E cca Dn Luo Ma nens 8 2 8 24 3 Initial and 9 21 3 Prescribed bounddry Teripe abl 9 2 1 3 2 Perfectly insulated adiabatic boundary 9 2 1 3 3 Convecion boundary cOBIIH n de 10 2 1 S d Radiation boundary condon zu eed ida 10 2 1 3 5 Internal boundary two solids bodies in contact 10 2 14 The HeatConducttonBguattOB amp 11 2 1 4 1 1 D transient time dependent heat conduction equation 11 2 1 4 2 1 D steady state heat conduction 11 2 1 4 3 The 3 D transient Heat Conduction 11 Dh DIN 12 2 L5 I Fimite element Method eoe ettet oe t e Eee a 12 2 1 5 2 Tine Cire ne Ce MENO eere tte aetema 13 2D LE SIRESS ANALYS 14 2 2 WAR ESIC 14 PIO IA E ene Ree 14 2 2 25 14 e 15 15 2 2 5 1 Ideal PIastiety o repe 15 17 23 3 3 Temperature Dependent SHE SS ee
62. Casting Processes Polyteknisk Forlag Kgs Lyngby ISBN 8750209698 MAGMA Online Help documentation MAGMASOFT version 4 4 MAGMA GmbH Aachen Germany Magmasoft 2000 MAGMASOFT version 4 4 MAGMAstress Module manual MAGMA GmbH Aachen Germany MATLAB 2006 MATLAB Help documentation version R2006a The Mathworks Inc USA 92 References Simo J C Hughes T J R 1998 Computational Inelasticity Springer New York Wikipedia free encyclopedia 2008 http en wikipedia org Acc 21 01 2008 93 Appendix 1 0 Appendix 10 1 Abaqus Implementation The same general steps presented Figure 3 1 will now be followed for the Abaqus implementation of our residual stress analysis For reference purposes the figure is presented again Pre Processing Geometry Definition Mesh Generation Simulation Setup Solidification Calculation Solution of the governing differential equations Stress strain analysis Post Processing Results Visualization Results Preparation for Comparison Results Comparison Figure 10 1 Steps sequence for the residual stress analysis 94 Appendix 10 1 1 Pre Processing 10 1 1 1 Geometry Definition A cylinder of length 0 4m diameter 0 25m and 0 02m radius at each end will be used for the detailed description of the entire process see Fzgure 10 2 Figure 10 2 The Cylinder part Units Meters The mold for the cylinder is simply a box with the cylinder
63. Curves We plot the nodal history of an analysis variable by using the XY Data tool To plot the thermal history of a node in our cylinder from the BSO R1 analysis for example proceed as follow Press the Create XY Data button from the tool box gt Source ODB field output gt Open the Variables tab gt Position Unique Nodal gt Check the checkbox for NT11 Nodal temperature Open the Elements Nodes tab Select the node s from which you want to see results using one of the listed methods gt Plot The XY Data create option can also be accessed from Tools menu XY Data Create To confirm that the displacement boundary conditions where satisfied in the stress analysis STRESS R1 we plot the displacement results in the fixed directions of the nodes where we applied the constraints as follows Press the Create XY Data button from the tool box Source ODB field output Open the Variables tab Position Unique Nodal Click the arrow next to U Spatial displacement gt Select the degrees of freedom you want to check in your node s Open the Elements Nodes tab gt Select the node s from which you want to see results using one of the listed methods Plot 144 Appendix XY Data from ODB Field Output Steps Frames Note x Y Data will be extracted From the active steps Frames Active Steps Frames variables Elements Mades variables Position Unique Modal Click checkboxes or edit Ehe identi
64. Data Setup File Name g idokumentiNT11 B50 RiLrpt Append to file Mutouk Format Layout 2 Single table For all x data Interpolate between X values iF necessary Separate table For each XY data Page width characters gt No limit Specify Number af significant digits Murmber Farmat Data write xY data Column totals Column minimax Figure 10 50 Report XY Data window 147 Appendix 10 2 Magmasoft Implementation The same general steps presented in Figure 3 1 that where followed for the Abaqus implementation will be followed for the Magmasoft one For reference purposes the figure is presented again Pre Processing Geometry Definition Mesh Generation Simulation Setup Solidification Calculation Solution of the governing differential equations Stress strain analysis Post Processing Results Visualization Results Preparation for Comparison Results Comparison Figure 10 51 Steps sequence for the residual stress analysis 10 2 1 Pre Processing 10 2 1 1 Geometry Definition The mesh to be used will be generated in Magmasoft and not in Hypermesh as we did in the Abaqus implementation In this case the model to be imported is the model of the casting The mold is created directly in the Magmasoft preprocessor A material must be assigned to the geometries during the Geometry Definition step However the details about the material data are to be edited later
65. ED HUB BOTTOM MAXIMUM PRINCIPAL STRESS COMPARISON VIE 86 FIGURE 7 6 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MAXIMUM PRINCIPAL STRESS COMPARISON INCLINED VIEW 87 FIGURE 7 7 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MINIMUM PRINCIPAL STRESS COMPARISONS TOP VIEW Vr 88 FIGURE 7 8 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MINIMUM PRINCIPAL STRESS COMPARISON antes 99 FIGURE 7 9 ORIGINAL HUB TOP AND OPTIMIZED BOTTOM MINIMUM PRINCIPAL STRESS COMPARISON INCLINED VIEW 90 FIGURE 10 1 STEPS SEQUENCE FOR THE RESIDUAL STRESS 516 0 94 FIGURE 10 2 THE CYLINDER PART UNITS 8 2 0 1 6000 95 FIGURE 10 3 THE CYLINDER MOLD PART UNITS 6 446400 95 FIGURE 10 4 STL EXPORT WINDOW 1 1 emen ee 97 FIGURE 10 5 USER PROFILES WINDOW 1 1 heme nennen 98 FIGURE 10 6 WIREFRAME APPEARANCE OF THE STL GEOMETRY OF THE MOLD IN 98 FIGURE 10 7 GENERAL APPEARANCE OF T
66. FOR THE CYLINDER cssccsssccsscccsscscesccseccsscsccscossccsscsccsccssccscsesseonscssccecsecenscs 39 5 Original Hub 440 5 40 5 2 THERMAL AND STRESS CURVES POINTS 210 ee 42 42 Sd BOUNDARY C ONDINON S 45 Ihe rmalboundaty conditlOBS p eo Pet 45 SA LA Ho aud vente bd DUI MD 45 3 41 D2 ATC SHAK IU ET 45 5 4 2 Mechanical boundary OS 46 54 2 I Stress amalysis SLC D odo a o eai 46 5 5 COOLING CURVES FOR THE 49 S OSTHERMAI COLORASPEGTRUMS 50 5 7 STRESS CURVES FOR THE ORIGINAL 00 0 ehem 52 5 8 STRESS COLOR SPECTRUM e FU ua Lau 54 5 9 SIMULATION TIME OF THE ORIGINAL 60 LAE IDEE A ERA 61 6 2 THERMAL AND STRESS CURVES POINTS 62 63 04 BOUNDARY CONDITOR S
67. HE AUTO CLEANUP PANEL 0 2 2727 1 99 FIGURE 10 8 COLLECTORS OF THE MOLD SURF MESH HM MODEL IN 100 FIGURE 10 9 AUTOMESH PANEL IS DIVIDED IN TWO FOR DISPLAY PURPOSES eese 101 FIGURE 10 10 MOLD SURE 2 102 FIGURE 10 11 CYLINDER SURF MESH HM MODEL 103 FIGURE 10 12 TETRAMESH PANEL SETTING FOR THE CYLINDER VOLUME MESH HM MODEL 104 FIGURE 10 13 MASKED VIEW OF THE CYLINDER MODEL WHERE INNER ELEMENTS CAN BE SEEN 105 FIGURE 10 14 TETRAMESH PANEL SETTING FOR THE MOLD VOLUME MESH HM MODEL 106 FIGURE 10 15 A MASKED VIEW OF THE MOLD MODEL WHERE INNER ELEMENTS CAN BE 107 FIGURE 10 16 UTILITY BROWSER APPEARANCE FOR THE ABAQUS USER 108 FIGURE 10 17 IMPORTED CAD FILES IN 6 8 112 FIGURE 10 18 DENSITY MATERIAL DATA CURVE FOR THE CYLINDER 200000 114 VII FIGURE 10 19 FIGURE 10 20 FIGURE 10 21 FIGURE 10 22 FIGURE 10 23 FIGURE 10 24 FIGURE 10 25 FIGURE 10 26 FIGURE 10 27 FIGURE 10 28 FIGURE 10 29 FIGURE 10 30 FIGURE 10 31 FIGURE 10 32 FIGURE 10 33 FIGURE 10 34 FIGURE 10 35 FIGURE 10 36 FIGURE 10 37 FIGURE 10 38 FIGURE 10 39 FIGURE
68. IELD OUTPUT REQUEST OF THE NODAL THERMAL HISTORY 22022 125 CONVECTIVE INTERACTION PROPERTY BETWEEN THE CASTING AND THE AMBIENT 129 CAST AMBIENT CONVECTION 000 00 000 00000 ee eee eser serene 130 CAST AMBIENT RADIATION 020 2 ene teser se eee 130 YOUNG MODULUS MATERIAL DATA CURVE FOR THE CYLINDER 132 POISSON S RATIO MATERIAL DATA CURVE FOR THE CYLINDER PART 133 THERMAL EXPANSION COEFFICIENT MATERIAL DATA CURVE FOR THE CYLINDER 133 PLASTICITY MATERIAL DATA CURVE FOR THE CYLINDER PART 00 134 TOTALLY CONSTRAINED NODE BOUNDARY CONDITIONS STRESS ANALYSIS 137 EDIT BOUNDARY CONDITIONS WINDOW FOR A FULLY CONSTRAINED NODE 137 SEMI FIXED NODE RED ALIGNED IN THE X DIRECTION WITH THE TOTALLY FIXED ONE 138 SEMI FIXED NODE ALIGNED IN Z WITH THE TOTALLY CONSTRAINED ONE 138 FIELD OUTPUT REQUEST CONFIGURATION FOR THE III STRESS MODEL 139 JOB STATUS t co 140 JXNZODBAN THE RESULTS TREP Ei tnc ni Me seta 142 PLOT CONTOURS ON DEFORMED SHAPE BUTTON SELECTED eene 142 VIEW CUT MANAGER BUTTON 143 VIEW CUT MANAGER WINDOW eee T E A A A A A aT 143 VARIABLES TAB IN THE XY DATA FROM ODB OUTP
69. MPORTING AN RPT FILE INTO EXCEL TEXT IMPORT WIZARD STEP 2 OF 3 WINDOW 172 FIGURE 10 82 IMPORTING AN RPT FILE INTO EXCEL TEXT IMPORT WIZARD STEP 3 OF 3 WINDOW 72 FIGURE 10 83 IMPORTING AN TXT FILE INTO EXCEL TEXT IMPORT WIZARD STEP 1 3 WINDOW 173 FIGURE 10 84 IMPORTING AN TXT FILE INTO EXCEL TEXT IMPORT WIZARD STEP 2 OF 3 WINDOW 1 74 FIGURE 10 85 IMPORTING AN TXT FILE INTO EXCEL TEXT IMPORT WIZARD STEP 3 OF 3 WINDOW 1 74 FIGURE 10 86 TEMPLATE CODE FOR THE THERMAL 175 FIGURE 10 87 EXAMPLE OF A POPULATED 2 0 6 0000 0 176 FIGURE 10 88 CHANGING MAGMASOFT STRESS CURVE RESULTS FROM MPA 178 FIGURE 10 89 COMPARISON BETWEEN CONDUCTION INTERACTION WITH 1000 TIE CONSTRAINT AND MAGMASOFT 000 otn Sai eo o teu 198 FIGURE 10 90 1 87 OF THE CYLINDER GEOMETRY AS USED IN THE SYMMETRY ANALYSIS 199 FIGURE 10 91 THERMAL RESULTS COMPARISON OF THE CYLINDER MODEL WITH AND WITHOUT SYMMETRI Ait at ILL 200 FIGURE 10 92 MISES COMPARISON THE CYLINDER MODEL WITH AND WITHOUT SYMMETRY 201 FIGURE 10 93 MAXIMUM PRINCIPAL STRESS COMPARISON OF THE CYLINDER MODEL WITH AND WITHOUT 201 FIGURE 10 94 MINIMUM PRINCIPAL STRESS COMPARISON OF THE CYLINDER
70. Magmasoft T 20 199 201 399 401 ggg 601 1159 1161 2000 1 Relative Temperature change used in Magmasoft DTM Delta T Magmasoft 196 Appendix SRelative Thermal Strain used in Magmasoft SM Strain Magmasoft for i 1 length T 1 AlphaM i DIM 2 end SM B Absolute Thermal Strain to used Abaqus SA Strain Abaqus 1 5 1 for 1 2 length T CUL esM FC i end SA C SAbsolute Reference Temperature to be used Abaqus RT Reference Temperature Global Temperature for Abaqus considering the Reference Temperature TA Temperature Abaqus end TA D Thermal Expansion Coefficient for Abaqus AlphaA 1 0 for 1 2 length T E i SA i TA i end AlphaA E 197 Appendix 10 7 Conduction Interaction Vs Tie Constraint As mentioned in the step 10 of Section 10 1 1 3 1 we could have simulated the conductive heat transfer between the casting and the mold in Abaqus by means of a Tie Constrain which represents perfect conduction or by a Contact Interaction property where a value for the heat transfer coefficient must be defined We compare simulations of the Cylinder model before shake out using a HTC of 1000 in Magmasoft a Contact Interaction with an equivalent HTC of 1000 in Abaqus and using a Tie Constraint in Abaqus Finally we decide that the Contact Interaction approach will produce better comparisons with the 1000 simulation in Magmas
71. OLD d Features 1 ib Sets Surfaces m Skins Stringers zi 7 Section Assignments 1 MOLD SEC 50119 Homogeneous Ex Composite Layups ES fi Engineering Features Mesh zi Fre Materials 2 CAST MAT MOLD MAT Sections z CAST SEC MOLD SEC Figure 10 25 Model tree after completing the first 5 steps of the setup 6 Assembly Since the parts where made with respect to the same coordinate system when they get inserted in the assembly they will be automatically in the correct relative position and no further positioning operations will be needed Note If instead of importing the parts they would have been created in Abaqus is in the Assembly module where a Boolean operation should be performed to produce the mold cavity Expand the Assembly item in the Model tree Right click the Instances collector Create Select both parts holding the Shift key 2 OK 119 Appendix 7 Steps definition An Initial Step is created by default The initial conditions and the contact interaction between the casting and the mold will later be defined on it However a General Heat transfer step must be added the before shake out step Here the mold ambient interactions are described and the output data that we are interested in is specified The duration of this step define the time that the casting remains in the mold prior shake out Right click the Steps collector gt Create gt Name it Before Shake Out
72. Properties Definition and 10 Interactions Definitions For details on how to set this boundary condition in Magmasoft refer to section 10 2 1 3 instruction for Figure 10 63 Convection Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition 15 defined automatically so the user has no participation in the setting Radiation Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting 4 3 1 2 After shake out Convection Between the external surface of the casting and the ambient A temperature dependent convective heat transfer coefficient property was defined in Abaqus For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 2 under 9 Interaction Properties Definition and 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting Radiation Between the external surface of the casting and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 2 under 10 Interactions Definitions
73. Results Models 1 E 1850 LE Parts 23 c CYLINDER E MOLD Figure 10 17 Imported CAD files in Abaqus 2 Materials definition The material properties to be defined for the casting part CYLINDER in the I BSO model are Density Conductivity Specific Heat Latent Heat Consta nt The properties to be defined for the mold part MOLD are Density Conductivity Specific Heat The values of our material data can be found in the Appendix section 10 4 However in this section we present the curves of the previously mentioned temperature dependent material properties Abaqus CAE does not use specific units but the units must be self consistent throughout the model which means that derived units of the chosen system can be expressed in terms of the fundamental units without conversion factors see Abaqus 2007 112 Appendix An example of a self consistent set of units is the International System of units 51 which fundamental units are length meters m mass in kilograms kg time in seconds s temperature in degrees Kelvin and electric current in Amperes Derived units as Newton Joule J or Coulomb C must be expressed in terms of the fundamental ones The geometries for this model where created in meters so the materials properties values must be consequent with it Creating the material for the Cylinder part Right click the Materials container in the Mo
74. Results Rotations in the Y and Z axis A node also lying in Z 0 and Y 0 but in the opposite side of the part with respect to the fixed node is constrained in Y and Z See Figure 4 6 Edit Boundary Condition Mame YZ Displacement Rotation Step Initial Region Picked CSY Global Note The displacement value will be maintained in subsequent steps Figure 4 6 Constraining the rotations in Y and Z in a single node selection in red for the Cylinder model The constraint in Y avoids the rotation in Z and the constraint Z avoids the rotation in Y The reason why the node is left free to move in X is because that is the correct contraction direction toward the totally fixed node 30 Cylinder Results Rotation in the X axis A node aligned in the Z axis with the totally fixed one that is with the same X coordinate and 0 lying in the opposite flat face of the Cylinder 15 constrained in X and Y See Figure 4 7 In this way the displacement of that node can just happen in Z ensuring that the length axes of the body will remain parallel to the Z axis Edit Boundary Condition Mame XY Displacement Rotation Step Stress 1 Static General Region Picked CSYS Global Distribution Uniform ut 4 0 UR1 radians un2 radians un3 radians Amplitude Ramp M
75. S CURVES FROM PNT5 OF THE OPTIMIZED 400 73 FIGURE 6 13 MAXIMUM PRINCIPAL STRESSES FROM PNT5 OF THE OPTIMIZED 73 FIGURE 6 14 MINIMUM PRINCIPAL STRESSES FROM PNT5 OF THE OPTIMIZED 74 FIGURE 6 15 ABAQUS COLOR SPECTRUMS FOR THE MISES RESULTS OF THE OPTIMIZED HUB 75 FIGURE 6 16 MAGMASOFT COLOR SPECTRUMS FOR THE MISES RESULTS OF THE OPTIMIZED 76 FIGURE 6 17 ABAQUS COLOR SPECTRUMS FOR THE RESIDUAL MAX PRINCIPAL STRESSES OF THE OPTIMIZED LIUB 77 FIGURE 6 18 MAGMASOFT COLOR SPECTRUMS FOR THE RESIDUAL PRINCIPAL STRESSES OF THE OP TIEN bei 78 FIGURE 6 19 ABAQUS COLOR SPECTRUMS FOR THE RESIDUAL MIN PRINCIPAL STRESSES OF THE OPTIMIZED etes op EI M DM M De 79 FIGURE 6 20 MAGMASOFT COLOR SPECTRUMS FOR THE RESIDUAL MIN PRINCIPAL STRESSES OF THE OPTIMIZED HUB cR E 80 FIGURE 7 1 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MISES COMPARISON TOP VIEW 82 FIGURE 7 2 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MISES COMPARISON BOTTOM VIEW ubt Emi x E M am M MEE MS 83 FIGURE 7 3 ORIGINAL HUB TOP AND OPTIMIZED HUB BOTTOM MISES COMPARISON INCLINED VIEW incu E A ML iL CM T Mn c EE MR 84 FIGURE 7 4 ORIGINAL TOP AND OPTIMIZED HUB BOTTOM MAXIMUM PRINCIPAL STRESS CONIPARIDON TOP VIEW N S 85 FIGURE 7 5 ORIGINAL TOP AND OPTIMIZ
76. SURF End Assembly MATERIALS Matenal name CASI MAT C onductivity Density Latent Heat Specific Heat Matenal name MOLD MAT C ond uc tivity Density Specific Heat INTERAC TION PROPERTIES Surface Interaction name CASI MOLD C ONTAC T INTERAC TIO N PRO PERTY T Gap Conductance 1000 0 0 1000 PHYSICAL CONSTANTS 187 Appendix Physic al Constants absolute zero 273 15 stefan boltzmann 5 6 7e 08 PREDEFINED FIELDS Name Field 1 Type Temperature Initial Conditions type TEM PERA TURE _PickedSet48 1400 Name Field 2 Type Temperature Initial Conditions type TEM PERA TURE _PickedSet49 20 INTERAC TIONS Interaction CASIT MOLD C ONTACT INTERAC TIO N PRO PERTY 1 Contact Parr intera ctionZC AST MOLD CONTAC T INTERAC TON PRO PERTY MOLD IN T SURF CAST EXT SURF STEP Before Shake Out Step Before Shake Out extrapolation PARABOLIC inc 28800 Heat Transfer end PERIOD deltmx 10 10 28800 le 12 28800 INTERAC TIONS Interaction SURFFILM 1 188 Appendix ofilm MOLD EXT SURF F 20 20 Interaction SURFRA DIA TE 1 Sra diate MOLD EXT SURF R 20 0 76 OUTPUT REQUESTS Resta rt write 0 FIELD OUTPUT F Output 1 O utput field Node Output
77. Select the Visualization Module in the context bar gt File menu gt Open gt Browse and select the odb file from which you want to see results gt OK 141 Appendix The odb is placed in the Results tree under the Output Databases collector Model Resulks Output Databases 11 E 1 BSO R1 ndb Spectrums 71 xwPlats x Data Paths Display Groups 1 Movies Images Figure 10 44 An odb in the Results tree They are different types of plots that allow us to see results for details see Abaqus CAE User s Manual We chose the Plot Contours on Deformed Shape option This option represent the values of our selected analysis variable as colored faces It can be selected from the toolbox EDT c ws Figure 10 45 Plot Contours on Deformed Shape button selected Or from the menu Plot menu Contour On Deformed Shape 142 Appendix 10 1 3 1 2 Cut Sections We can display a cut section of our model using the View Cut Manager tool from the toolbox View Cut Manager Figure 10 46 View Cut Manager button selected Or from the menu Tools menu gt View Cut gt Manager From the View Cut Manager window see Figure 10 47 we can select one of the tree default planar cut sections or we can create our own for details see the Abaqus CAE User s Manual As we adjust the cut it is applied in real time to the model
78. UT 0 20 145 ELEMENTS NODES TAB IN THE XY DATA FROM ODB OUTPUT 146 REPORT AY DATA WINDOW mos oe aoe 147 STEPS SEQUENCE FOR THE RESIDUAL STRESS 6516 2 148 TYPICAL APPEARANCE OF THE MAGMASOFT MAIN 0 0 00 149 MAGMA CYLINDER STL FILE IMPORTED INTO THE MAGMASOFT PREPROCESSOR 150 LOCATION OF THE MATERIAL BUTTON IN THE PREPROCESSOR 150 ENTITY SELECTIONS WINDOWS WITH THE VOLUMES SELECTED PRIOR ORGANIZING 152 MAGMASOFT MESH GENERATION ene sees sites seen 153 PROCESS MODE WINDOW sereni a E E R 154 THE MATERIAL DEFINITIONS 0 6000 155 gt DATABASE ET Ee tae ko ete E AE SEEN ume Ea 155 MAGMADATA WINDOW FOR THE PROJECT DATABASE 3 3 156 DEFAULT YOUNG S MODULUS FOR THE GJL 150 MATERIAL 157 GJL 150 MATERIAL SELECTED FROM THE PROJECT DATABASE IN THE DATABASE REQUEST DT MERC REPRE HP 157 HEAT TRANSFER DEFINITIONS WINDOW 6 ene ee sese tesis 158 OPTIONS 159 SHAKE OUT DEFINITIONS
79. a rt write 0 FIELD OUTPUT F Output 1 O utput field Node Output NI U Element Output directions YES EE PE PEEQ S THE O utput history frequenc y 0 End Step STEP Stress 2 Step name Stress 2 inc 243200 Sta tic 0 01 43200 1e 11 43200 194 Appendix PREDEFINED FIELDS ASO Temperature Temperature ABAQUS 9 fg R6 odb bstep 1 binc estep 1 einc 99 OUTPUT REQUESTS Resta rt write 0 FIELD OUTPUT F Output 1 O utput field Node Output NI U Element Output directions YES EE PE PEEQ S THE O utput history frequenc y 0 End Step 195 Appendix 10 6 Thermal Expansion Coefficient calculation Magmasoft Abaqus Magmasoft uses a local definition of the thermal expansion coefficient while Abaqus uses a global definition Taking this into account if the thermal expansion coefficient material data from the Magma database will be used in Abaqus a conversion of the data must be performed We carry the conversion in Matlab as follow Adapting the Magmasoft Thermal Expansion Coefficient data to Abaqus Thermal Expansion Coefficient from Magmasoft AlphaM AlphaM 1 07E 05 s0 7E 05 s3 1H 05 ge 09 22 05 12 s00E 10 00E 10 Mm RPP EP PP PE s6Global Temperature from
80. an c 99 Organizing b eee ees 99 A NTeshinesthe 101 EO che 1 2 5 Mesh The Volume da C tete tod ides Geb test Eee Ere cet tenete dtr 104 IV Table of Contents 10 1 1 2 6 Exporting the meshes 108 I0 T 1 5 AbaqussimiulattOn Se CUD 110 TO Tie henna lS awaits a bete UM atta ole eal 110 IQ TT EI BeloreShake Out tmodel 2i io de netus 111 2 A Wer Shake Outb mode sai EE Ee Eee 126 131 Gc Ie EUM Ec TE 140 10 153 POSEDEOCOSSLIHP cus oe en eta say 141 TIO 3 D Results VasualizabloH c doo tee 141 IO T3 A T Eboading the Output 141 143 10 1 3 1 3 Removing a part from the 143 TOAST d Crea nos XX CURVES 144 1071 32 Results Preparation for Comparison UIS eoe eet Rast eesti 145 Savine ACNE 146 2 ex OTL WAG CUE 146 102 MIAGMASOET MENTATION 148 10 2 ag ced ba 01 Sich M cU 148 PO TAG COME Deb 148 10 2 1 2 Mesh eener une 153 10 21 53 Ma masolt Simulation Setup secco dte eterno nites
81. baqus vs Magma Maximum Principal stresses for the Cylinder model 34 Cylinder Results 3X10 MAGMA ABAQUS 2 1 4 amp at 3 4 0 1 2 3 6 7 8 4 Time Seconds Figure 4 12 Abaqus vs Magma Minimum Principal stresses for the Cylinder model The stress results presented in Figure 4 10 Figure 4 11 and Figure 4 12 were obtained from the central point of the geometry of the whole Cylinder Comments These results show that the Cylinder model develop more stresses in tension than in compression 35 Cylinder Results 4 7 Stress color spectrums 5 Mises Avg 7596 3 384 07 3 131 07 2 878 07 2 625 07 2 372 07 2 118 07 1 865 07 1 612e 07 1 359 07 1 105 07 8 522 06 5 990 06 3 457 06 Mises MPa Empty 43 38 40 37 37 35 34 34 31 33 28 31 25 30 22 29 19 27 16 26 13 25 10 23 7 22 4 21 1 19 Figure 4 13 Abaqus top and Magmasoft bottom color spectrums for the Mises results of the Cylinder model 36 Cylinder Results 5 Principal Aug 75595 06 S MaxPrincipalStress MPa Empty 59 50 54 62 49 73 44 85 39 96 35 08 30 19 25 30 20 42 15 53 10 65 5 76 0 88 4 01 Figure 4 14 Abaqus top and Magmasoft bottom color spectrums for the residual Maximum Principal stresses of the Cylinder model Cylinder Results 5 Min
82. cavity in the center its general dimensions in meters 0 65 x 0 65 x 0 80 see Figure 10 3 Figure 10 3 The Cylinder Mold part Units Meters 95 Appendix Both Abaqus and Magmasoft have CAD capabilities for us to create the cylinder and its mold but instead we decide to use the specialized CAD software ProEngineer Wildfire 3 to create and edit these geometries not only because we receive the other models the Hub and the optimized Hub as ProEngineer part files but also because 1 a good practice to use the best characteristics of every software at hand if the effectiveness is not compromised The cylinder is a solid part created by a simple revolve feature The mold general shape is an extrusion of a rectangular section To create the cavity of the mold the cylinder is removed form the mold with a Boolean operation In ProEngineer is as follow Note For simplicity create the cylinder and the mold in the same relative position with respect to their coordinate systems so if you make the coordinate systems coincide when assembling the parts they will get in the correct position without needing any further adjustment With the mold part active go to Insert gt Shared Data gt Merge Inheritance gt Select the Remove material button gt Select the Open a model which geometry will be copied button Open the cylinder Select its coordinate system then the mold coordinate system gt Press the button Now the cylin
83. cooled down from a super heated temperature to room temperature The thermal history obtained is then used as an external force to calculate the residual stresses by means of a quasi static mechanical analysis using a J2 plasticity model The simulation procedures are explained through a simplified model of the Hub and then applied to the geometries of interest results comparison between the original Hub and its optimized version 1 also presented The theoretical base is given in this work as well as detailed implementation procedures The results shows that the part subjected to the topology optimization process develop less residual stresses than its original version Key Words Key Words Residual stresses l hermo mechanical Quasi static Castings FEM FDM Numerical simulation Solidification Thermal analysis Stress analysis Meshing Abaqus Magmasoft Hypermesh Matlab ProEngineer Table of Contents Table of Contents J 66966 GROIN 2 3 3 2 Theoretical background 26 6 eee e eee eee e eee e eee 4 ZI THE THERMAL PAIN AGW SIS E 4 deat UT ANS erene E accu edd eme 4 4 5 5 Nl Thermal COMIC aeter east 7 2 12 DOS E E 7 7 DM DAT Aven TA
84. d later through a modification on the Predefined Fields Two new general static steps will be created to define the analysis of the stresses before and after shake out In each one of them the corresponding odb file from the thermal analysis must be read into The After Shake Out step must be deleted Expand the Steps collector of the III STRESS model gt Right click the After Shake Out step gt Delete gt Yes Stress step before shake out Right click the Steps collector gt Create gt Name it Stress BSO gt Procedure type General gt From the list select Static General gt Continue Time period seconds 28800 the same as in the I BSO model gt Incrementation Type Automatic Maximum number of increments 28800 Initial Increment size 10 Minimum Increment size 1E 12 Maximum Increment size 10000 gt OK Stress step after shake out Right click the Steps collector gt Create gt Name it Stress ASO gt Procedure type General From the list select Static General Continue 135 Appendix Time period seconds 43200 the same as in the 5 model 2 Incrementation Type Automatic 2 Maximum number of increments 43200 gt Initial Increment size 10 Minimum Increment size 1E 12 Maximum Increment size 10000 gt 2 OK 8 Predefined Fields definition The Cast Initial Temp field will be modified to be constant and of the same value as in the I BSO model Expand th
85. del tree Create 2 Name it CAST MAT gt Under the General material editor menu select Density Under the Thermal material editor menu select Conductivity Specific Heat and Latent Heat Fill in the appropriate data For reference see the Appendix section 10 4 Creating the material for the Mold part Right click the Materials container in the Model tree gt Create gt Name it MOLD MAT gt Select Density Conductivity and Specific Heat gt Fill in the right data 2 OK 113 Appendix The curves for the Cylinder temperature dependent material data are 7100 7000 6900 6800 6700 Density Kg m 6600 6500 6400 900 1000 1500 2000 Temperature Celsius Figure 10 18 Density material data curve for the Cylinder part 94 Conductivity W mK 4 I 4 P gt e 38 500 1000 1500 2000 Temperature Celsius Figure 10 19 Conductivity material data curve for the Cylinder part 114 Appendix 1100 1000 900 800 700 Speific Heat J KgK 600 900 500 1000 1500 2000 Temperature Celsius e Figure 10 20 Specific Heat material data curve for the Cylinder part Our curves for the MOLD material data are 1501 1500 8 1500 6 1500 4 1500 2 1500 1499 8 Density Kg m 1499 6 1499 4 1499 2 t 900 1000 1500 2000 Temperature Celsius Figure 10 21 Density material data curve for the Mold part 115 Specif
86. der 9 Interaction Properties Definition and 10 Interactions Definitions For details on how to set this boundary condition in Magmasoft refer to section 10 2 1 3 instruction for Figure 10 63 Convection Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition 15 defined automatically so the user has no participation in the setting Radiation Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting 5 4 1 2 After shake out Convection Between the external surface of the casting and the ambient A temperature dependent convective heat transfer coefficient property was defined in Abaqus For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 2 under 9 Interaction Properties Definition and 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting Radiation Between the external surface of the casting and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 2 under 10 Interac
87. der and the mold are ready to be exported For the implementation of the analysis in Abaqus the meshes are created in Hypermesh In this case the Mold will be exported as a step stp file Just the surfaces and the coordinate system must be saved to the stl In Hypermesh the mesh for the Cylinder will be obtained from the Mold one so is not necessary to export the Cylinder Exporting the mold as a step file File menu gt Save Copy gt Step gt Name it Mold Surface gt OK 2 Check the box for the Surfaces gt Ensure the other check boxes are unchecked Press the button Select the Mold coordinate system 2 OK For the implementation in Magmasoft the mold is not necessary as a external CAD file so just the Cylinder will be exported This time is an STL stl file File menu gt Save a Copy gt STL gt Name it Magma Cylinder gt OK gt Press the button Select the Mold coordinate system Set the rest of the window as Figure 10 4 96 Appendix Export S TL Coordinate System CSYS FA ESYS Format Binary ASCII Allow negative values Deviation Control Chord Height 0 000200 Angle Control 0 500000 File name haama Lylinder Figure 10 4 STL export window ProEngineer 10 1 1 2 Mesh Generation The mesh of the parts is created in Hypermesh 8 0 and exported as an Abaqus inp file containing just nodes and volume elements The mold surface is imported and me
88. during the Simulation step 148 Appendix Overview 1 Importing the model 2 Assigning a material type to the Cylinder 3 Creating the mold and assigning a material type 5 Placing cooling curves points and stress curves points 1 Importing the model Create a new Project Project menu Create Project Project Mode Shape Casting Batch Production gt Name it Magma Cylinder gt Press the Return key gt When running in Windows Magmasoft usually don t recognize the Enter key of the numeric key pad so use the Return key instead Since the mesh will be generated Magmasoft the stl file created ProEngineer Is directly imported into the Magmasoft preprocessor The model to be imported is the Magma Cylinder stl created in ProEngineer Cylinder 26 version 01 shape casting batch production 05101 2008 MAGMASOFT project preprocessor enmeshment simulation export postprocessor database info help Figure 10 52 Typical appearance of the Magmasoft main interface Click the button on the main window gt File menu gt LOAD SLA gt Browse the Magma_Cylinder stl model gt Open gt Ok 149 Appendix ereatina deseire display list nee Figure 10 53 Magma Cylinder stl file imported into the Magmasoft preprocessor 2 Assigning a material type to the Cylinder Select menu gt Volume gt Select the volume SLA Magma Cylinder from the list gt Return gt
89. e deactivate slicing dialog check box get unchecked again 165 Appendix Control Panel curves X Ray Vector Dist Anim Material Scales Rotate Images Views Light Results X Ray 0 Animation Distortion ProcRot Curves Vector Slicing Tracer Print Mesh v Activate deactivate slicing dialog Slice direction CX Y CZ 33 20 Improve display quality Figure 10 75 Cut view setting with the Slice functionality Project View Support Help Project Cylinder_26 Version v01 Directory CAMAGMAsoft Cylinder Inn DATES MPa Empty 113 58 1 11 48 9 28 4 87 2 6 0 4 1 43 3 93 6 13 MinPrincipalStress 136 t 3h 45min 00 100 00 Figure 10 76 Cut view displayed in the main window 166 Appendix 10 2 3 1 2 The curves In the Postprocessor we can just see the results history curves of the cooling and stress points which positions where defined in the Preprocessor module Therefore no creation of curves from new points in the part is possible in the Postprocessor To see the cooling curves Select Curves from the Postprocessor s Control Panel window From the Results group list select Solidification From the Curves list select the curve corresponding to the cooling point that you are interested in The curve will be displayed in the main window as in Fzgure 10 78 Control Panel Curves
90. e 17 2 2 3 2 Pfasuceity Materia models 18 2 27 34 The de I E 18 2 2 3 4 2 Illustration for a Simple Mathematical 18 22 35 23PI3SDe V mode cose teret 21 2 2 3 5 J2 Plasticity model Constitutive 55 21 21 211 002 0 060400000000000000000000000000 21 3 S SS PROCESS 23 SEHR 25 4 Cylinder Results ta eee edu eos 20 AA E TEE 26 A EE 24 4 BOUNDARY C ONDION S 28 A Thermal bponndary COPIGRIONS 28 OS M E 28 234 2 ATter sDake QUU oe drm teo tob ton e 28 Table of Contents 4 3 2 Mechanical boundary conditions 29 2 3 2 Stress analysis SiE 29 Zu COOLING CURVES a vetet ao E dos adus 22 a 252 33 STRBSS 34 lSLIBBSSCODORSPBECTRUMS 220 36 4 8 SIMULATION TIME
91. e Mold and related information from the model Part Expand the new II ASO model in the model tree Expand its Parts collector gt Right click the MOLD part gt Delete gt Yes Material Expand the Materials collector of the II ASO model gt Right click MOLD MAT gt Delete gt Yes Section Expand the Sections collector of the II ASO model gt Right click MOLD SEC 2 Delete gt Yes Assembly instance Expand the Assembly item of the model gt Expand the Instances collector 2 Right click MOLD 1 gt Delete gt Yes 126 Appendix Interaction Expand the Interactions collector of the model gt Right click CAST MOLD CONTACT INTERACTION gt Delete Yes Interaction Property Expand the Interactions Properties collector of the 5 model gt Right click CAST MOLD CONTACT INTERACTION PROPERTY gt Delete gt Yes Predefined Fields Expand the Predefined Fields collector of the 5 model Right click Mold Initial Temp Delete Yes 1 Importing the models The part file of the Cylinder 1s copied together with the rest of the model therefore there is no need to import or create any part file 2 Materials definition No modifications needed 3 Sections definition No modifications needed 4 Sections assignment No modifications needed 5 Mesh element type No modifications needed 127 Appendix 6 Assembly After deletions of the MOLD instance no more changes are needed
92. e Predefined Fields collector of the III STRESS model gt Right click Cast Initial Temperature gt Edit gt Distribution Direct specification gt Magnitude 1400 gt OK 9 Interaction Properties definition There are no interaction properties involved in this analysis Consequently the Conv HTC interaction property must be deleted Expand the Interaction Properties collector of the III STRESS model gt Right click Conv HTC 2 Delete 2 OK 10 Interactions definition There are no interaction properties involved in this analysis Therefore the Cast Ambient Convection and the Cast Ambient Radiation interactions must be deleted Expand the Interactions collector of the III STRESS model gt Right click Cast Ambient Convection 2 Delete gt OK Expand the Interactions collector of the III STRESS model gt Right click Cast Ambient Radiation 2 Delete gt OK 136 Appendix 11 Boundary Conditions The Cylinder must be constrained in such a way that no rigid body motions could happen Then the translations and the rotations in X Y and Z ust be constrained However it must be able to shrink To do so three nodes will be fixed First in one of the flat end faces of the cylinder a node will be totally constrained Right click the BCs collector of the III STRESS model gt Create 2 Name U 123 gt Step Initial gt Category Mechanical gt Type Displacement Rotation gt
93. e the same mesh in the interface between the casting and the mold Therefore we will keep the mesh that correspond to the cavity mesh of the mold and use it as the surface mesh of the casting Since we are now working in the Cylinder Surf Mesh hm the external surface of the mold and the corresponding elements are not needed and must be deleted Right click the Mesh collector in the Model Browser gt Delete gt Ok Right click the Suefaces collector in the Model Browser gt Delete gt Ok If you are not using collectors delete the surfaces using the Delete option in the Edit menu shortcut F2 check the delete associated elems check box to delete surfaces and elements at once The cylinder surface mesh is ready now It should look as Utility amp Entities AA Assembly Hierarchy E Components 21 1 4 Mesh i g 4 Surfaces Figure 10 11 Cylinder Surf Mesh hm model Save the file File menu Save 103 Appendix Note Remember that Hypermesh do not recognize CTRL S so use the Save option in the File menu The next step is to create a volume mesh from each one of the surface meshes 10 1 1 2 5 Meshing the volumes The Cylinder Create a Save As copy of the Cylinder Surface Mesh hm file and name it Cylinder Volume Mesh hm File menu gt Save As gt File name Cylinder Volume Mesh hm gt Save The model displayed
94. e the value of the dependent variable at various predefined points nodal points The values of the dependent variable will always be the primary unknowns The resulting equation systems the primary unknowns are written so that for each nodal point in the calculation domain there is an equation for every dependent variable These equations are referred to as discretization equations The calculation domain is divided into sub domains called cells elements or control volumes with the intention of identify the dependent variable in a smaller area as a function of the values in the nodal points In this way different profiles can be applied to each sub domain allowing more suitable sub domains for the actual problem The most commonly used numerical methods in casting simulations are the Finite Differential Method the Finite Element Method FEM and the Finite Volume Method FVM The differences between them are mainly in the profile assumptions for the cells elements or control volumes and in the methods of deriving the discretization equations Nevertheless they have also much in common for instance the all need a geometry definition which describe the calculation domain appropriate material data definition of initial and boundary conditions they all use solvers of linear algebraic equations to perform the calculations and they all use a postprocessor to present the results According to Abaqus 2007 Analysis user s manua
95. eometry of the mold in Hypermesh 98 Appendix Save the model as Mold_ Surf Mesh hm File menu gt Save gt File name Mold_ Surf Mesh hm gt Save 10 1 1 2 2 Geometry Cleanup They may be small gaps between surfaces These gaps turn the geometry into a not closed volume representing a problem to create the volume mesh that we are aiming for To repair this gaps we used the Auto Cleanup option Geometry menu gt Auto Cleanup gt Click suis gt Select All gt Set the target element size 0 001 gt _autocleanup gt This value is modified with the button For details on the effect of those parameter see Hyperworks 2006 Auto Geometry Cleanup include comp CYLINDER 14 target elem size parameter 0 004 Topology cleanup parameters use current parameters __ editparameters Elements quality criteria wsecurentcrtera TEER Figure 10 7 General appearance of the Auto Cleanup panel 10 1 1 2 3 Organizing the model Hypermesh group the entities of the models into collectors this collectors allow us to handle the collected data as a unit Is convenient to separate the surfaces to be meshed into logical groups e g external surfaces and cavity surfaces because it does not only simplify the visualization of the model since collectors can be hidden or displayed but also the selection of large groups of surfaces as usually casted industrial parts have
96. erimentally determined as the slope of the stress strain curve obtained during tensile tests carried out on samples of a given material 2 2 2 1 Elastic strain According to the Hook s law within the elastic limit of a material the stress is proportional to the strain This strain is known as elastic strain 64 and is expressed as g 6 2 26 14 Theoretical Background 2 2 2 2 Thermal strain When a metal body is heated or cooled it expand or contract if it s free to deform The amount of deformation that the body undergoes is then proportional to the rise or fall down of the temperature This give place to the mathematical expression for the thermal strain g a T dT 2 27 This thermal deformation can be a contraction or an expansion can result in deformation only or stresses only if the body is free to contract or is totally constrained respectively or it can result in a combination of deformation and stresses which is the most common situation in reality for castings 2 2 3 Plasticity When the load applied to a material produce a deformation over the elastic limit Is said that the material is in the plastic region where any experimented deformation is permanent The transition from elastic to plastic behavior is called yield and the stress that correspond to this transition is called yield stress At elevated temperatures the metals undergo such irreversible deformations and in casting pr
97. ers as Data preview me lt Figure 10 64 Importing an txt file into Excel Text Import Wizard step 2 of 3 window gt In the Text Import Wizard Step 3 of 3 window select Colum data format General see Figure 10 85 Confirm the columns are correctly separated gt Press This screen lets you select each calumn and set Column data Format the Data Format 2 General General converts numeric values to numbers date Text values En dates and all remaining values to text O Date MDY s Do not import column skip Data preview Figure 10 85 Importing an txt file into Excel Text Import Wizard step 3 of 3 window 174 Appendix 10 3 1 3 Setting up the Matlab M File In a Matlab M File we will create two vectors per curve to be compared containing each one of them the data from one of the columns variables of the XY data for the curve Then the codes to plot the curves in the same graphic will be written We will always use the time in the X axis of the graphic Create the M File Open Matlab gt File menu gt New M File Note We recommend to comment the M Files simply write 96 at the beginning of a comment line so it get ignored in the code for ease of understanding for any reader Also is useful to separate the vectors in cells Cell menu Insert cell divider to simplify the finding of the variables later on you can use the Go function in the Go menu
98. ertinent conversion from the Magmasoft data was performed for details see the Appendix section 10 6 Pla stic ity The curves for the used material data are Young s Modulus Pascal 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Temperature Celsius Figure 10 34 Young Modulus material data curve for the Cylinder part 132 Appendix 0 5 0 45 0 4 Poissons Ratio 0 3 0 25 e 500 1000 1500 2000 Temperature Celsius Figure 10 35 Poisson s Ratio material data curve for the Cylinder part 1 3 Expansion Coefficient 1 C e co 0 8 0 900 1000 1500 2000 Temperature Celsius Figure 10 36 Thermal Expansion Coefficient material data curve for the Cylinder part 133 Appendix Stress Pascal 0 01 0 02 0 03 0 04 0 05 0 06 0 07 Strain Figure 10 37 Plasticity material data curve for the Cylinder part 3 Sections definition No modifications needed 4 Sections assignment No modifications needed 5 Mesh element type The element type must be changed to 3D STRESS Expand the Parts collector in the III STERSS model Expand the CYLINDER part item gt Right click the Mesh item gt Switch Context gt Mesh menu gt Element Type gt Select the whole geometry gt Done gt Element Library Standard Geometric Order Linear Family 3D STRESS gt OK 2 Done 134 Appendix 6 Assembly No modifications needed 7 Steps definition The Initial step will be altere
99. etween Abaqus and Magmasoft by allowing us to set the same value of heat transfer coefficient for details see the Appendix section 10 7 The contact interaction give the possibility of run the Before and After Shake Out simulations as consecutive steps in the same model by suppressing the interaction in the After Shake Out step which can not be done with the tie constraint However this alternative will have the mold present in the after shake out step even if is not interacting with the casting which means that the solver will continue calculating the cooling of the mold consuming unnecessarily precious processing capacity of the computer and slowing down the calculation Therefore even with the contact interaction we choose to run the Before and After Shake Out steps in different models where the thermal history in the output database of the Before Shake Out is read as initial temperature field in the After Shake Out step In the latest case the mold is totally removed from the After Shake Out simulation A convective interaction between the mold and the ambient is defined through a Surface film condition interaction The interaction is assigned to the Before Shake Out step A radiation interaction between the mold and the ambient is defined through a Surface radiation to ambient interaction The interaction is assigned to the Before Shake Out step Conduction Cylinder Mold Right click the Interaction collecto
100. ference A T under steady state conditions and when the heat transfer is dependent only on the temperature gradient E E W mK 2 9 Ax AT 2 1 2 2 Density Indicate the mass per unit volume of a material pec Kg m 2 10 Density is a temperature and pressure dependent material property In solids and liquids is just slightly affected by these factors but in gases 1s strongly dependent in both of them 2 1 2 3 Specific Heat cv and In general is the measure of the heat energy required to increase the temperature of a unit quantity of a substance by a defined temperature step For example how much heat must be added to increase the temperature of one gram of water by one Celsius degree The specific heat is defined at constant pressure J Kg K and a constant volume C J m K In gases C and C have important differences but since for most solids and liquids C and C are equal in casting processes for simplicity we call specific heat C H 2 J Kg K 2 11 Where 1s the enthalpy per unit mass Theoretical Background 2 1 2 4 Latent Heat L Is the amount of energy in the form of heat that is released or absorbed by a substance during a change of phase solid liquid or gas Kg 2 12 Where Q is the amount of energy needed to change the phase of the substance 77 is the mass of the substance and L correspond to the specific latent
101. fiers shown next to Edit below Equivalent plastic strain b L 5 Stress components THE Thermal strain components w Spatial displacement Magnitude Edit LI LH LI LE LI LES Section point Figure 10 48 Variables tab in the XY Data from ODB Output window 10 1 3 2 Results Preparation for Comparison When we need to compare Abaqus results with Magmasoft results the approach consist in export the curves of the Abaqus analysis results to Abaqus Report files rpt from where the X Y data of the curves can be obtained in a table form at that can be used in a software as Matlab where the actual comparison of the data from both sources can performed If instead the results to compare belong to one or more Abaqus odb files the comparison can be made in the same Visualization module through the XY Data Manager In section 10 1 3 1 we saw how to plot a curve We can create an X Y data report of the curve being displayed in the viewport but also instead of plotting the curves we can save them to the Data Manager without needing to see them and then select from a list which curve s you want to export to the report file If you want to compare curves from two different odb files is mandatory to save them first prior to the combination of the results Note The saved curves just remain in memory during the session 145 Appendix 10 1 3 2 1 Saving a curve The setup of t
102. for the GJL 150 material The database request window should now look like database request Figure 10 62 GJL 150 material selected from the project database in the database request window Press the ok button in the database request window 157 Appendix Similarly assign and modify the material for the mold as follow In the material definitions window select the Sand Mold material class shown in Figure 10 589 ENT gt Database Project gt Group Sand Mold gt MEE gt COLDBOX from the MAGMAdata window shown Figure 10 60 5 Fiom the Edit menu select the parameter to be edited and input the correct data gt Data menu gt Save gt Close gt Database menu gt Quit gt ok ok the material definitions window Let s now select the conductive heat transfer coefficient between the casting and the mold From the heat transfer definitions window select the Cast Alloy material class as in Figure 10 63 _select data gt Database MAGMA gt Group Constant gt Select C1000 0 from the list gt ok gt Press ok in the heat transfer definitions window heat transfer definitions selection boundary Cast Alloy Sand Hold database magma file name 1000 0 constant list material class material class database file name material group id material group id database file name Cast Alloy Sand Hold
103. for the modeling of casting processes are introduced together with their mathematical representation 2 1 3 1 Prescribed boundary temperature T P t T P t 2 14 Where P is a position on the surface in 1 D described just by the x value t is time and denotes prescribed 2 1 3 2 Perfectly insulated adiabatic boundary An adiabatic boundary has no heat flux across 0 2 15 Where 7 is the outward pointing normal to the surface at point Another way to define this boundary condition is setting the heat transfer coefficient h to zero Newton s law Theoretical Background 2 1 3 3 Convection boundary condition The heat flux across the bounding surface is proportional to the difference between the temperatures of the surface 7 P t and the surrounding t cooling medium It is defined by the Newton s convective law of cooling OT 2 16 Pur h T t T P t n As mentioned for equation 2 2 is the convective heat transfer coefficient 2 1 3 4 Radiation boundary condition When a boundary surface receives heat by radiation the following expression applies K P 0 ur T0 2 17 n Where Hu ec T 2 18 Which for simplicity assumes constant This is used when the time step of the analysis is so small that the temperatures may be assumed constant during the time step 2 1 3 5 Internal boundary two solids bodies in contact
104. hat have to be set to perform a residual stress simulation in a problem like ours The following summary corresponds to an uncoupled thermo mechanical analysis as described in the Introduction chapter of this work The Thermal simulation 1 Mesh the part 2 Define the material properties a CASTING 1 Density ll Conductivity ili Specific Heat iv Latent Heat v Liquidus Temperature vi Solidus Temperature b SAND MOLD 1 Density ll Conductivity ili Specific Heat 3 Define the initial boundary conditions a Initial temperature of the casting b Initial temperature of the mold 4 Define the shake out event By time or temperature 23 Implementation 5 Define the interactive boundary conditions before shake out Only in Abaqus a Conduction Between the external surface of the casting and the surface of the mold cavity b Convection Between the external surface of the mold and the ambient c Radiation Between the external surface of the mold and the ambient 6 Define the interactive boundary conditions after shake out Only in Abaqus a Convection Between the external surface of the casting and the ambient b Radiation Between the external surface of the casting and the ambient The Stress simulation 7 Use the same mesh used in the thermal simulation for the casting the mold is not present in our stress analysis 8 Define the material properties a Expansion Coefficient b Young s Modulus c
105. he curve must be carried out as explained in section 10 1 3 1 but at the end we just press the Save button from the Data from ODB Field Output window see Figure 10 49 Is not relevant if the curve have been plotted or not XY Data from ODB Field Output Steps Frames Note XY Data will be extracted From the active steps Frames Active Steps Frames PTEN Variables Elements Modes Selection Method Edit Selection Add Selection Delete Selection Pick From viewport Nodes selec ted Node labels sets Internal seks L Highlight items in viewport Figure 10 49 Elements Nodes tab in the XY Data from ODB Output window 10 1 3 2 2 Exporting the curves Report menu gt gt Select from the list one or more curves to export to the rpt file gt Switch to the Setup tab gt Assign a mane gt Select a destination folder with the Select button gt Set desired options gt Press Apply to create the file and continue exporting results or OK to create it and finish Note The Append to file button allows writing more than one result to the same rpt file When exporting a result to an existent report file if Append to file is checked the result will follow the previous one in the rpt if is not checked the result will overwrite the previous one 146 Appendix Report XY Data
106. hin a medium or between mediums It takes place in gases liquids and solids In conduction there is no flow of any of the material mediums The governing equation for conduction 1s called the Fourier s law of heat conduction and it express that the heat flow per unit area 1 proportional to the normal temperature gradient where the proportionality constant is the thermal conductivity cp 2 1 4 2 1 Where q 1s the heat flux perpendicular to a surface of area W A 1s the surface area through which the heat flow occurs m is the thermal conductivity W mK T is the temperature C and x is the perpendicular distance to the surface traveled by the heat flux Theoretical Background 2 1 1 2 Convection Is the heat transfer by mass motion of a fluid when the heated fluid moves away from the heat source It combines conduction with the effect of a current of fluid that moves its heated particles to cooler areas and replace them by cooler ones The flow can be either due to buoyancy forces natural convection or due to artificially induced currents forced convection The equation that represents convection comes from the Newton s law of cooling and is of the form 4 hA T T 2 2 Where is the convective heat transfer coefficient W m K is the temperature of the cooling fluid and 1 the temperature of the surface of the body 2 1 1 3 Radiation In general radiation is energy
107. hose to load them in Excel also just to store all the results in the xls format At this point we assume that all the curves to be combined and or compared have been exported as suggested in this report from their sources Overview Thermal results comparison approach 1 Combining the Abaqusthemal results 1 1 Loading the Abaqus rpt files into Excel 1 2 Combining the Before and After Shake Out rpt files 2 Loading the Magma txt file into Excel 3 Setting the Matlab M File 4 Plotting the compa nson 5 Exporting the comparison image Stress results comparison approach 1 Loading the Abaqus rpt file into Excel 2 Loading the Magma txt file into Excel 3 Modifying the units of the Magma XY data 4 Setting the Matlab M File 5 Plotting the compa nson 6 Exporting the comparison image 170 Appendix 10 3 1 Thermal Results Comparison Approach 10 3 1 1 Combining the Abaqus thermal results As we mention in section 10 1 3 2 if all the results to be compared are from Abaqus we could compare them in the Abaqus Visualization module But there is a drawback for this approach from our Abaqus implementation procedure and that is that the After Shake Out simulation results are apart in a different odb file from the Before Shake Out simulation results That means that when plotting curves from the After Shake Out odb file the time axis will start from zero instead as from the last time of the Before Shake Out simulation So even if we can crea
108. ic Heat J KgK Mold Conductivity W mK Appendix 1 6 1 5 gt EE NO e Oo 0 7 9 6 500 1000 1500 2000 Temperature Celsius Figure 10 22 Conductivity material data curve for the Mold part 1400 1300 1200 1100 1000 900 800 700 900 1000 1500 2000 Temperature Celsius Figure 10 23 Specific Heat material data curve for the Mold part 116 Appendix 3 Section definition Create two Solid Homogeneous sections with the corresponding material for the Cylinder and the Mold respectively Cylinder section Right click Sections in the Model tree gt Create gt Name it CAST SEC gt Category Solid gt Type Homogeneous gt Continue gt Material CAST MAT gt Mold section Right click Sections in the Model tree gt Create gt Name it MOLD SEC gt Category Solid gt Type Homogeneous gt Continue gt Material MOLD MAT OK 4 Section assignment Assign the respective section to the whole geometry of the Cylinder and the Mold Cylinder section assignment Expand the Cylinder part in the Model tree gt Right click the Sections Assignment collector gt Create E If the Selection option tools not displayed press 42 and ensure that the T Select from all entities option is selected since it allow to select elements from both outside and inside a part x Select From
109. ic style all the curves results are presented from Matlab 10 3 1 2 Loading the Magma txt file into Excel Open Microsoft Excel gt File menu gt Open gt Files of type Text files gt Browse the Magma txt file gt Open 2 In the Text Import Wizard Step 1 of 3 window select Delimited see Figure 10 83 gt Next gt The Text Wizard has determined that your data is Delimited IF this is correct choose Next or choose the data type that best describes your data Original data type Choose the File type that best describes your data Delimited Characters such as commas or tabs separate each Field CO Fixed width Fields are aligned in columns with spaces between each Field File origin 932 Japanese ShiFE JI5 Preview of File a Master Thesis CURVES AMD REPORTS Magma curve z Exk Start import row CURVE ZABAQUS CURVE 2 3 urve 1 PI CASTING 1 344 1 E imeTemperatureTimeTemperature 5 cxe 3 Figure 10 83 Importing an txt file into Excel Text Import Wizard step I of 3 window gt In the Text Import Wizard Step 2 of 3 window check the Tab check box see Figure 10 84 gt Next gt 173 Appendix This screen lets vau set the delimiters your data contains You can see how your kext is affected in the preview below Delimiters Tab Semicolon Comma Space Other Text qualifier Treat consecutive delimit
110. in the assembly 7 Steps definition The Initial step will be altered later through a modification to the Predefined Fields The Before Shake Out step will be removed and the After Shake Out step will be created Expand the Steps collector of the H ASO model gt Right click the Before Shake Out step Delete gt Yes Right click the Steps collector gt Create gt Name it After Shake Out 2 Procedure type General gt Heat transfer gt Continue 2 Time period seconds 43200 gt Incrementation Type Automatic 2 Maximum number of increments 43200 Initial Increment size 10 Minimum Increment size 1E 12 Maximum Increment size 43200 Max allowable temperature change per increment 1 Max allowable emissivity change per increment 0 1 gt 8 Predefined Fields definition The Cast Initial Temp field will be modified to be read from the odb file of the BSO RI analysis Expand the Predefined Fields collector of the II ASO model gt Right click Cast Initial Temperature gt Edit gt Distribution From results or output database file gt Press Select button for the File name gt Browse and select the BSO R1 odb file gt OK gt Step 1 gt Increment insert the number of the last increment of the BSO RI analysis gt Interpolation Compatible gt OK 128 Appendix Note To find the number of the last increment of the BSO R1 analysis go to the folder where the analysis was saved a
111. ipal Pascal MAGMA Symmetry ABAQUS ABAOUS Symmetry 0 5000 10000 15000 Time Seconds Figure 10 97 Maximum Principal Stress comparison of the Original Hub model with and without symmetry 204 Appendix MAGMA MAGMA Symmetry ABAQUS ABAQUS Symmetry 0 5 Min Principal Pascal gt 1 5 1 2 5 3 5 5000 10000 15000 Time Seconds Figure 10 98 Minimum Principal Stress comparison of the Original Hub model with and without symmetry 205 Appendix 10 8 3 Optimized Hub 10 8 3 1 Geometry A half of the Hub has been used for the symmetric analysis as shown in Figure 10 100 Accordingly half of the mold has also been used Figure 10 99 Half of the Optimized Hub geometry as used in the symmetry analysis 10 8 3 2 Simulation time 206 Abaqus Appendix 10 8 3 3 Thermal results 1400 MAGMA Symmetry ABAQUS ABAQUS Symmetry 1200 4 1000 4 gt 800r e E qam D 600 E S 400 200 0 0 5000 10000 15000 Time Second Figure 10 100 Thermal results comparison of the Optimized Hub model with and without symmetry 207 Appendix 10 8 3 4 Stress results 10 3 5 2 5 Mises Pascal PJ 0 5 MAGMA MAGMA Symmetry ABAOLUS ABAQUS Symmetry 5000 10000 15000 Time Seconds Figure 10 101 M
112. is now Cylinder Volume Mesh hm in which a volume mesh with tetrahedral elements will now be created Create a new collector called Volume Mesh as explained in section 10 1 1 2 3 to store the volume mesh Activate it Make Current Mesh menu gt Tetramesh gt Select volume tetra as the mesh type gt Set the enclosed volume option to surfs Select the surfaces of the cylinder gt set the rest of the panel as in Figure 10 12 mesh gt Tetramesh tetra mesh enclosed volume tetra remesh w 14 C mesh volume tetra match existing mesh 7D trias include Cavity Mesh use curvature use proximity element size 0 01 z reum Figure 10 12 Tetramesh panel setting for the Cylinder Volume Mesh hm model By clicking one surface Hypermesh automatically select all the surfaces that together with the selected one enclose a volume 104 Appendix The volume mesh for the Cylinder is ready Save the file File menu Save Optional To see the appearance of the elements inside the cylinder the Mask option can be used It allows us to hide selected elements from the display For details about how to use the Mask option refer to the Hypermesh help documentation doe Ar EUREN 429 y OA AES SN SIN e vd 4 INK EATS 37 i A N
113. ises comparison of the Optimized Hub model with and without symmetry x10 3 9 25 Max Principal Pascal T 0 5 o4 MAGMA MAGMA Symmelry ABAOUS Symmetry CC 5000 10000 15000 Time Seconds Figure 10 102 Maximum Principal Stress comparison of the Optimized Hub model with and without symmetry 208 Appendix MAGMA MAGMA Symmetry ABAOUS ABAQUS Symmetry Min Principal Pascal RI L 0 5000 10000 15000 Time Seconds Figure 10 103 Minimum Principal Stress comparison of the Optimized Hub model with and without symmetry 209
114. l version 7 1 ABAQUS Inc Providence RI USA Becker A A 2004 some of the main features of the FE and the FD methods are 2 1 5 1 Finite element method 1 The solution domain is divided into a grid of finite segments or elements 2 The governing partial differential equations are solved for each element in mesh 3 The elements are assembled together and the continuity requirements and equilibrium conditions are satisfied with adjacent elements 4 A unique solution can be obtained to the whole system of linear algebraic equations once the boundary conditions are satisfied 5 The solution matrix is populated with relatively few non zero coefficients 6 The FE method is suitable for the analysis of complex geometries and is not difficult to modify the element size in particular regions 12 Theoretical Background 2 1 5 2 Finite difference method 1 The solution domain is divided into a grid of cells or elements 2 The derivatives in the governing partial differential equations are converted into finite difference equations 3 These finite difference approximations are applied to each interior point so that the displacement of each node 15 a function of the displacements at the other nodes connected to it 4 A unique solution can be obtained to the whole system of linear algebraic equations once the boundary conditions are satisfied 5 Ihe solution matrix is banded 6 The FD method is not suitable for very co
115. le 3 Hardening law 2 2 3 4 1 The Yield Surface The definition of a yield surface is very useful in the multi dimensional formulation of the plasticity theory Is a general way to define the yield criterion by means of a yield function f for the material The yield surface encompasses the elastic region of the material behavior which means that the state of the stresses while inside the surface is elastic Since the yield function 15 defined to be zero in the plastic state the yield point 1s reached when the stresses reach the surface and outside the surface the material behavior becomes plastic 2 2 3 4 2 Illustration for a Simple Mathematical Model The classical rate independent ideal plasticity model can be illustrated with a simple model of a one dimensional mechanical device This device exhibit the notion of irreversible response and consists of a spring with a spring constant E m o gt 0 A and coulomb friction element with constant 7 This device is assumed to have unit length and unit area initially 18 Theoretical Background E Figure 2 4 One dimensional frictional device representing ideal plasticity represents the stress or force applied to the model o is the flow stress of the friction device and represents the total strain or change in length The total strain is decomposed into elastic and plastic strain p esp pe 2 31 By equilibrium conditions and using eq 2 31 the elastic stress
116. lick the Simulation button in the Magmasoft main interface gt Set the process mode window as Figure 10 57 Ok process mode permanent mold calculate batch production calculate stress sand mold calculate filling calculate solidification calculate stress prepare fast postprocessing ok cancel help Figure 10 57 Process mode window The next step consists in assign a material to the casting and the mold Since our material data differ from the one in the Magmasoft database we will copy a material for the casting specifically GL 150 and for the mold COLDBOX from the Magma database to the Project database where they will be modified and from where they will be used In the material definitions window select the Cast Alloy material class as in Figure 10 589 _select data gt Database Project gt Group Cast Alloy gt MAGMAdata see Figure 10 59 Import menu gt From MAGMA gt From the Choice list select GL 150 gt 000 gt Choice gt Lent Now the materials are in the Project database 154 Appendix material definitions selection material Cast Alloy T initial 1400 000 database Tdliquidus 0 000 file name T solidus 0 000 list material class database file name initial temp material group id database file name initial temp Cast
117. ll analysis 2 OK 10 1 1 3 2 The Stress Simulation The residual stress analysis is performed in a single model The model is a modified copy of the After Shake Out model to utilize the same part file and other unchanged data like the section and the material Note that is a duplicate of the 5 model and not of the II BSO That is because we are not interested in the mold part in this analysis neither The stresses before and after shake out are solved in two consecutive steps The nodal thermal history is used as an external force to solve stress calculations Therefore the corresponding temperature field must be read into each step of the analysis In this model the initial temperature field assigned to the Initial step must be the same as the equivalent in the I BSO model In the step for the stresses before shake out the thermal history from the BSO RI odb must be read As well the thermal history from the ASO R1 odb must be read into the step for the stresses after shake out 131 Appendix Copying the model Right click the II ASO item in the model tree gt Copy Model gt Name it III STRESS gt 1 Import the CAD files Since 15 a copy of the II ASO the model already contains the part file 2 Materials definition The material properties to be defined in stress analysis for the Cylinder are Ela stic ity Temperature dependent Young s Modulus and Poisson s Ratio Thermal Expansion Coefficient a p
118. magma C1000 0 ok prev cancel select data expand hide parameters help Figure 10 63 Heat transfer definitions window 158 Appendix Set the Options window as in Figure 10 64 gt Thixocasting Pressurize parameters Tilt parameters sand Permeability ilie Coating parameters Shake Out parameters Guenching a parameters cancel prev Figure 10 64 Options window Select Sand Mold from the identifier list in the shake out definitions options window gt P shake out definitions selection identifier Sand Hold controlled by time control value open 28800 000 identifier list material class controlled hy material group id controlled Sand Hold 1 time Figure 10 65 Shake out definitions window 159 Appendix Controlled by Time gt Control value open 28800 gt Ok gt Ok gt Ok shake out options selection controlled time control value open 28800 000 s identifier list ok cancel help Figure 10 66 Shake out options window In the solidification definitions window select edit gt Select delete all gt Yes gt In the storing data definitions window set the input data option as Une gt To save results each X number of time steps the request must be written in the input field i
119. me the Magma txt tile nto bibe nbn esed 177 10 3 2 3 Modifying the units of the Magma XY 96 2 2925 178 10 32 4d Sctung cou o stan Dao 179 10 32 the compariSOD e euni eon alec 179 10 3 2 6 Exportine the comparison Image 179 O UP uod MP c 180 10 4 1 Fhetinal 6666 180 10 4 2501685 IM ate Ba 666 182 10 5 KEYWORDS OF THE ABAQUS INPUT FILES ese tese 186 10 6 THERMAL EXPANSION COEFFICIENT CALCULATION MAGMASOFT 8 196 10 7 CONDUCTION INTERACTION VS TIE CONSTRAINT 198 10 8 RESULTS COMPARISON WITH AND WITHOUT SYMMETRY 240 4 199 re ee rine E E TS 199 TO S A VGC OMG LEY deir eor rte tree E SA eee 199 10 5 Non T 200 IOS Thermal results et et 200 ROSSI ASU CSS Te SUNS coe oi imd eiim 201 1O S2 Guia M 203 203 10 52 Simul 203 204 10 5 S Optimized HU otis Fa E MEE 206 10 8 5 cate 206 10 5 3 2S 206 te LI LEM E 207 10 5 9 2S IT6SS tate t amt 208 Table of Contents Figures Table
120. mentation of residual stress analysis during the design of castings can lead to important improvements on the mechanical behavior of the final parts on aspects as crucial as fatigue life Therefore we strongly recommend the use of this type of numerical simulations as part of the design routine of casted parts As well the difference in the residual stress development of parts that has and has not undergone topology optimization procedures suggest the benefits of the inclusion of shape optimization in the design process 91 References 9 References Abaqus 2007 Abaqus CAE User s Manual version 6 7 1 ABAQUS Inc Providence RI USA Abaqus 2007 Analysis user s manual version 6 7 1 ABAQUS Inc Providence RI USA Becker A A 2004 An introductory guide to finite element analysis London Professional Engineering ISBN 1860584101 Chandra U Ahmed A 2002 Modelling for casting and solidification processing Marcel Dekker New York Gustafsson E Stromberg N 2006 Optimization of Casting by using response surface methodology SweCast AB Jonkoping Sweden Gustafsson E Hofwing M Stromberg N 2007 Simulation and Measurement of Residual Stresses in a Stress Lattice SweCast AB amp Department of Mechanical Engineering University of Jonkoping Hyperworks 2006 Hypermesh User s guide from Hyperworks 8 051 Altair Engineering Inc Troy MI USA Jesper Hattel 2005 Fundamentals of Numerical Modeling of
121. mplex geometries and is difficult to change the element size in particular regions 6 The FD method is not as popular for stress analysis problems as for heat transfer and fluid flow problems The approximation quality of the FE method 1s better but it comes with a greatest computer calculation time price 13 Theoretical Background 2 2 The Stress Analysis 2 2 1 Residual Stresses Residual stresses are tensions or compressions that exist in the bulk of a material without applying an external load In a casting process while cooling residual stresses are induced due to temperature gradients across the whole casting mechanical constraints given by the mold during the shrinkage of the metal and volumetric change and transformation plasticity related to the solid state phase transformation Hence residual stresses are a function of the shape of the casting and the cooling rate of the casting process Compressive residual stresses are desirable in a component as they improve the fatigue life and reduce the stress corrosion cracking tendency since they also offer resistance to crack propagation 2 2 2 Elasticity Within a certain limit when a load is applied a component undergoes a deformation that is recovered when the load is released This behavior of a material is known as elasticity and its limit is known as the elastic limit The measure of the elastic behavior of a material is known as Young s modulus This modulus is exp
122. n about Matlab commands see the MATLAB 2006 Populating the vectors Go to Excel gt Select all the numerical values from the time column in the Magma data imported from the txt file 2 Copy CTRL C gt Go to Matlab gt Paste them CTRL V in the A vector between the signs In the same way copy paste the information of the temperature column into the vector Similarly populate the two Abaqus results vectors and D Only for illustration purposes in Figure 10 87 we present the A vector populated with values from 0 60 in steps of 10 for you to have an idea about how do a populated vector look like 5 MAGMASOFT Thermal Results Magma Time 0 10 30 40 50 1 4 si 0 5 d 60 Figure 10 87 Example of a populated vector 176 Appendix 10 3 1 4 Plotting the comparison Once all the vectors contain their respective information and the codes for the plot have been written simply do as follow j Press the Evaluate entire file button 10 3 1 5 Exporting the comparison image A new window per figure appear with the vectors plotted on it On the window corresponding to the Figure you want to export File menu gt Save as Browse a destination for the file 2 Name it In the Save as type option choose an appropriate format Save 10 3 2 Stress results comparison approach 10 3 2 1 1 Loading the Abaqus rpt files
123. n the form From time in seconds To time in seconds X increment see Figure 10 67 gt Press the Return key gt Ok storing data definitions select result qroups data list input data alf materials 0 72000 100 insert delete delete all cancel help o 1 s s s s s s s s s Li Figure 10 67 Storing data definitions window 160 Appendix We are not interested in the feeding simulation then is necessary to ensure that calculate feeding is switched to NO in the solidification definitions window as in Figure 10 68 solidification definitions use solver solver d stop simulation time stop value 12000 000 5 calculate feeding x Yes x no H criterion temperature 1161 300 d C criterion temperature 2 1175 000 d C storing data edit ok prev cancel help Figure 10 68 Solidification definitions window Set all the options to NO in the stress simulation options window as in Figure 10 69 gt Ok stress simulation options user defined material selection x Yes no user defined stress inputfoutput x Yes no calculate hottear criterion x Yes wz no calculate machining x Yes no ok prev cancel help Figure 10 69 Stress simulation options window 161 Appendix In the fast postprocessing preparation window check the
124. nd open the BSO R1 sta file in a text editor The number in the last line of the column second from left to right corresponds to the last increment 9 Interaction Properties definition convective interaction property will be defined with temperature dependent data for the heat transfer coefficient between the casting and the ambient Right click the Interaction Properties collector of the 5 model gt Create gt Name it Conv HTC gt Film condition gt Continue gt Check the Use temperature dependent data check gt in appropriate data OK Our interaction property data looks like Edit Interaction Property Mame Type Film condition of Field variables Data 1 2 3 4 5 T B 9 BD Figure 10 31 Convective interaction property between the casting and the ambient 129 Appendix 10 Interactions definition Convection between the casting and the ambient Right click the Interaction collector of the 5 model gt Create 2 Name it Cast Ambient Convection Step After Shake Out Type Surface film condition Continue Select the whole external surface of the casting Done Set the Edit Interaction window as follow Edit Interaction Mame cCast Ambient Convection Surface Film condition Step After Shake Out Heat transfer Surface Cast Surf Edit Region
125. ndition in Abaqus refer to section 10 1 1 3 1 2 under 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting 66 Optimized Hub Results 6 4 2 Mechanical boundary conditions 6 4 2 1 Stress analysis step We must mention again that the user does not participate directly in the definition of boundary conditions for the stress analysis in Magmasoft Therefore we present just our Abaqus approach As we have established in the Cylinder and Original Hub results the task here also Is to constrain the rigid body translations and rotations in X Y and Z but allow the body to deform In the Optimized Hub model the 6 degrees of freedom has been constrained as follow Translations in X Y and Z In the flat surface of the upper inner ring of the cylindrical section a node in 0 is constrained in X Y and 7 See Figure 6 6 Notice that in the picture X is the horizontal axis Y the vertical axis and the Z axis is perpendicular to the paper L Edit Boundary Condition AT Marne Type Displacement Rotation Step Stress 1 Static General Region Picked CSYS Transform__T BCS CSYS XS Distribution Uniform ces fo LES 7 Junt radians Clurz radians radians Amplitude Ramp Mote displacement value will be maintained in subsequent steps
126. ns Contact F Energy Failure Fracture Mote Error indicators are nok available when Domain is Whole Model or Interaction Output for rebar Qutput at shell beam and layered section points 5 Use defaults Specify Include local coordinate directions when available Figure 10 42 Field output request configuration for the III STRESS model 139 Appendix 13 Job Creation Expand the Analysis item in the Models tree Right click the Jobs collector gt Create gt Name it STRESS R1 gt Source Model gt In the displayed models list select III STRESS gt Continue gt Write a short description of the analysis if desired gt Job Type Full analysis 2 OK Or Select the Job Module in the context bar Jobs menu gt Create gt Name it STRESS R1 gt Source Model gt In the displayed models list select III STRESS gt Continue gt Write a short description of the analysis if desired gt Job Type Full analysis 2 OK 10 1 2 Calculation When submitting a Job for analysis an input inp file is automatically written The only participation of the user in this phase 1 in the actual submitting of the Job since the calculation itself is performed automatically by the solver To submit a Job for analysis Expand the Analysis item in the Models tree gt Expand the Jobs collector gt Right click the job to be submitted e g BSO R1 gt Submit Or
127. ocesses which occurs over a large temperature range this inelastic or plastic behavior becomes important In plasticity is not possible to define the stresses as functions of the strains on total form so Hooke s law which is a total constitutive law does not apply Instead it Is possible to express the changes in stresses as changes in strains which is known as an incremental constitutive law 2 2 3 1 Ideal Plasticity Ideal Plasticity is the simplest approximation to the inelastic behavior of a material It assumes that the yield stress is constant independently of the mechanical strain see Figure 2 1 Notice that we said mechanical strain and not total strain The mechanical strain is equal to the elastic strain plus the plastic strain 15 Theoretical Background mech el g pg 2 28 And the total strain the one see from outside the component is the sum of the mechanical and the thermal strains and therefore the strain for the thermo elasto plastic case total mech E E e g e g 2 29 The importance of this differentiation and the fact that ideal plasticity is defined with respect to the mechanical strain is because even if the total strain of a fully constrained component subjected to a thermal gradient 15 zero the mechanical strain is not and is this strain the one to be used in the stress strain plot mech Figure 2 1 Ideal Plasticity If the event happens at constant tem
128. odified in this step Note The displacement value will be maintained in subsequent steps Figure 4 7 Constraining the rotations in X in a single node selection in red for the Cylinder model 3l Cylinder Results 4 4 Cooling curves The following results presented in Fzgure 4 8 were obtained from the central point of the geometry of the whole Cylinder 1400 1 T ABAQUS 1200 1000 v 800 4 3 ao o E 600 400 200 0 0 1 2 3 4 5 6 7 8 Time Second x 10 Figure 4 6 Abaqus vs Magmasoft cooling curves for the Cylinder model 22 Cylinder Results 4 5 Thermal color spectrums 4 243e 01 Temperature Empty 35 42 35 39 35 36 35 34 35 31 35 28 35 26 35 23 35 21 35 18 35 15 35 13 35 10 35 07 35 05 Figure 4 9 Abaqus top and Magmasoft bottom thermal color spectrums of the last step after shake out of the Cylinder model 33 Cylinder Results 4 6 Stress curves 2 5 10 T T T MAGMA ABAQUS 2 1 5 2 8 1 0 5 0 0 1 2 3 5 6 7 8 Time Seconds x 10 Figure 4 10 Abaqus vs Magma Von Mises curves for the Cylinder model x 10 5 MAGMA ABAQUS 4 3 2f 2 e a f 1 o c a gt __ AL 2 3 0 1 2 4 6 7 8 Time Seconds x 10 Figure 4 11 A
129. oft In Fzgure 8 1 the comparison of these simulations is presented Conduction Interaction Tie Constraint Magma 1000 Temperature C 7005 0 5 1 1 9 2 2 5 3 Time seconds x 10 Figure 10 89 Comparison between Conduction Interaction with HTC 1000 Tie Constraint and Magmasoft 1000 198 Appendix 10 8 Results comparison with and without symmetry If the problem in hand is symmetric is very convenient from a time saving point of view to perform the analysis in the smaller part that represent the whole geometry when appropriate boundary conditions are used We carry out our simulation procedure in the using symmetry conditions for the Cylinder the Original Hub and the Optimized Hub Next we present a comparison between the stress results obtained from the whole geometry and from the symmetric geometries The comparison includes CPU simulation time comparison 10 8 1 Cylinder 10 8 1 1 Geometry A 1 8 of the cylinder has been used for the symmetric analysis as shown in Figure 10 90 Accordingly a 1 8 of the mold has also been used Figure 10 90 1 8 of the Cylinder geometry as used in the symmetry analysis 199 Appendix 10 8 1 2 Simulation time No Symmetry 7hrs 12min 7hrs 36 min Abaqus 10 8 1 3 Thermal results MAGMA MAGMA Symmetry ABAQUS ABAQUS Symmetry 1400 1200 1000 600 Temperature Celsius 400 200 4 5 6 7 8
130. om corresponding to the rigid body rotations 46 Original Hub Results Rotations in the X and Z axis A node also in the top flat surface of the Original Hub with the same Z coordinate and in X 0 is constrained in X Z See Figure 5 8 Edit Boundary Condition Mame 2 Displacement Rotation Step Initial Region Picked CSY Global ui uz CIURI C ur2 uR3 Note The displacement value will be maintained in subsequent steps Figure 5 6 Constraining the rotations in the X and Z axes in the Original Hub constraint in X avoids the rotation in the Z axis and the constraint in 7 avoids the rotation in the X axis Notice that the node is free to move in the Y direction so the part is still able to deform normally 47 Original Hub Results Rotation in the Y axis node in 0 and vertically aligned with the totally fixed node same Y coordinate but in a different Z coordinate in this case in the flat surface of the end of the cylindrical section of the hub is constrained in X and Y See Figure 5 9 In this way the rotation in the X axis 15 restrained and all vertical axes of the part are fixed to remain parallel to the Z axes dit Boundary Condition Mame Type Displacement Rotation Step Initial Region Picked CSY Global ut uz Jus uR2 uR3 Note The displacement value will be maintained in sub
131. or the residual Min Principal stresses of the Original Hub 58 Original Hub Results ubi pad MPa Empty 42 6 22 5 2 4 174 37 7 57 8 779 98 0 118 1 138 1 158 2 178 3 198 4 218 5 238 5 Y MPa Empty 42 6 22 5 2 4 Figure 5 21 Magmasoft color spectrums for the residual Min Principal stresses of the Original Hub 59 Original Hub Results 5 9 Simulation time of the Original Hub 52hrs 6min 62hrs 56min 60 Optimized Hub Results 6 Optimized Hub Results 6 1 Geometry Figure 6 1 Front and Top view of the Optimized Hub model 61 Optimized Hub Results Figure 6 2 Bottom view of the Optimized Hub model The Optimized Hub mold has the same dimensions as the Original Hub mold as defined in section 5 1 6 2 Thermal and stress curves points placement Figure 6 3 Location of cooling and stress points 5 and 6 for the Optimized Hub 62 Optimized Hub Results 6 3 Mesh Magmasoft 63 Optimized Hub Results Figure 6 4 Mesh of the Optimized Hub model used in Abaqus 64 Optimized Hub Results Figure 6 5 Magma mesh of the Optimized Hub model 65 Optimized Hub Results 6 4 Boundary Conditions 6 4 1 Thermal boundary conditions 6 4 1 1 Before Shake out Conduction Between the external surface of the casting and the surface of the mold cavity For details on how to set this ty
132. osite 15 done in the first collector so the new collector is hidden and the first collector is displayed there the surfaces to be deleted are the equivalent ones to those left to remain in the new collector e g the cavity ones 10 1 1 2 4 Meshing the surfaces By default Hypermesh try to adapt the sizes of the elements in the surface being meshed to that of the surrounding meshes in the proximity of the union between the surfaces Taking that into account a surface needed to keep a constant element size is better to be meshed first because the mesh that adapt to the surrounding ones don t keep a constant element size unless you assign the same size to the elements of all surfaces involved The surface meshes were created using the Automesh function shortcut F12 key The external surface of our mold was meshed with the following procedure Right click the Mesh collector in the Model Browser gt Make Current gt Mesh menu Automesh gt Select surfs surfaces as the entity type gt Select the surfaces to mesh gt set the rest of the window as in Figure 10 9 gt Mesh gt Return Mesh Include 14 Blementsize 0 0480 size and bias mesh type trias C Oloptimize C edge deviation C suface deviation C rigid body mesh interactive comp GO Ext Mesh elems to current comp size control skew control break connectivity previous settings link opposite edges AI
133. pe of boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 9 Interaction Properties Definition and 10 Interactions Definitions For details on how to set this boundary condition in Magmasoft refer to section 10 2 1 3 instruction for Figure 10 63 Convection Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition 15 defined automatically so the user has no participation in the setting Radiation Between the external surface of the mold and the ambient For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 1 under 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting 6 4 1 2 After shake out Convection Between the external surface of the casting and the ambient A temperature dependent convective heat transfer coefficient property was defined in Abaqus For details on how to set this boundary condition in Abaqus refer to section 10 1 1 3 1 2 under 9 Interaction Properties Definition and 10 Interactions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting Radiation Between the external surface of the casting and the ambient For details on how to set this boundary co
134. perature the mechanical strain 15 equal to the total strain 16 Theoretical Background 2 2 3 2 Hardening better approximation to the real behavior of metallic materials is this concept of hardening because most of them instead of behave ideal plastically exhibit Increasing stress with increasing plastic deformation which the hardening approach assumes to have a linear relation Hence is known as strain hardening or linear hardening Mech E Figure 2 2 Linear hardening 2 2 3 3 Temperature Dependent Yield Stress The yield stress is also a function of the temperature and this dependence is very important in casting processes for the presence of plastic deformations T1 T2 T3 T4 Mech gr Figure 2 3 Stress train curves at different temperatures linear hardening approach 17 Theoretical Background Taking the temperature dependence into consideration the two main dependences of yield stress for the plasticity in a casting process are 1 Temperature 2 Plastic strain This can be expressed mathematically as 5 56 T gP 2 30 2 2 3 4 Plasticity Material models The residual stresses are calculated by a quasi static rate independent elasto plastic analysis and the majority of the plasticity models use incremental theories where the mechanical strain rate is divided into an elastic part and a plastic part Those incremental plasticity models are expressed in terms of 1 Yield function 2 Flow ru
135. properties as material section and interaction This last process would be equivalent to copy the nodes and elements from the inp file generated in Hypermesh and replace the nodes and elements directly on an inp file generated by Abaqus containing the rest of the necessary keywords The details about our approach to import the models into Abaqus will be discussed in section 10 1 1 3 Exporting the Mold mesh Confirm you are working in the Mold Volume Mesh hm model If the User Profile has been set to Abaqus as illustrated in Figure 10 5 the Utility browser should look like in Figure 10 16 Model Conversion Mastan To Abagus FromMastran Import Iptions Import Export Cleanup Hierarchy Tools Component Browser otep Manager Contact Manager Ahagqus GeomiMash Disp LIAModel Figure 10 16 Utility browser appearance for the Abaqus User Profile 108 Appendix Procedure Click the Export button in the Utility Browser gt Press the button in the Export Abaqus deck window gt Browse a destination folder and name the file Mold inp gt Save gt For the Export option in the Export Abaqus deck window select all 24 gt Ok The inp file of the Mold mesh is now exported and ready to be used in Abaqus Save the file File menu Save Exporting the Cylinder mesh Open the Cylinder Volume Mesh hm model gt Click the Export button in the Utility
136. r gt Create gt Name it CAST MOLD CONTACT INTERACTION gt Step Initial gt Surface to surface contact gt Continue gt Select the external surface of the Cylinder for the Master surface If the Selection option tools are not displayed press and ensure that the Select from exterior entities option is activated gt Done gt Slave type Surface gt Select the surface of the cavity of the Mold gt Done gt Define the Edit Interaction window as in Figure 10 27 127 Appendix Edit Interaction Mame CAST MOLD CONT ACT INTERACTION PROPERTY 1 Surface bo surface contact Standard Step Initial Master surface CAST EXT SURF Edit Region a Slave surface MOLD INT SURF edit Region Sliding Formulation 2 Finite sliding Small sliding Discretization method to surface bal Degree of smoothing Far master surFace Use supplementary contact points 2 Selectively Never C Always Switch Slave Made SurFace Adjustment Clearance gt adjustment Adjust only to remove overclosure Specify tolerance For adjustment zone Adjust slave nodes in set Contact interaction property CAST MOLD CONTACT INTERACTION PROPERTY Options Contact controls DeFaulE bal Figure 10 27 Conductive interaction between the cast and the mold Convection Mold Ambient Right click the Interac
137. reak line click at Ehe desired position To DELETE a break line double click on the line To MOVE a break line click and drag it Data preview 10 20 30 40 3l 60 ee ee ee ee ee ee ee spo uL 11 PI CAST 1 2715 1 l 4E 03 lt Figure 10 81 Importing an rpt file into Excel Text Import Wizard step 2 of 3 window gt In the Text Import Wizard Step 3 of 3 window select Colum data format General see Figure 10 82 Confirm the columns are correctly separated gt Press This screen lets vau select each column and set Column data format the Data Format General general converts numeric values to numbers date Text values to dates and all remaining values to text Date MOY Do nat import column skip Data preview gt Figure 10 82 Importing an rpt file into Excel Text Import Wizard step 3 of 3 window 1 2 Appendix 10 3 1 1 2 Combining the Before and After Shake Out Abaqus rpt files Add the last value of the time column of the Before Shake Out XY data to each value of the time column of the After Shake Out XY data Copy the After Shake Out XY data after the Before Shake Out one in their respective columns Now all the time and temperature data 1 together in one column each from where we will create the corresponding vectors in Matlab We could combine the data and plot it in Abaqus but to maintain the same graph
138. resent a methodology to perform numerical simulations of residual stresses 3 Compare solutions obtained from the FE solver Abaqus and the FD solver Magmasoft 1 3 Delimits For the purpose of our work a general understanding of the Finite Element and the Finite Difference formulations are sufficient This work does not present the mathematical details of the FE or the FD method No details about the topology optimization process are intended to be provided in this work The topologically optimized version of the part provided by Volvo 3P Was given Theoretical Background 2 Theoretical background 2 1 The Thermal Analysis This chapter aims to provide basic information related with the simulation of solidification in castings about heat transfer mechanisms material properties boundary conditions the heat conduction equation and the numerical methods 2 1 1 Heat transference When a system is at a different temperature than its surroundings the Nature tries to reach thermal equilibrium To do so as the second law of thermodynamics explains the thermal energy always moves from the system of higher temperature to the system of lower temperature This transfer of thermal energy occurs due to one or a combination of the three basic heat transport mechanisms Conduction Convection and Radiation 2 1 1 1 Conduction Is the transference of heat through direct molecular communication i e by physical contact of the particles wit
139. ring shrinkage of the cast metal and volumetric change and transformation plasticity associated with the solid state phase transformation according to Chandra U Ahmed A 2002 Since the residual stresses can increase or decrease the fatigue life of a component an interest on its consideration during the design process has grown in the industry of casted parts Scientific information supporting the validity of such interest 1 offered in Gustafsson E Hofwing M Stromberg N 2007 Considerable differences when residual stresses are included or not in shape optimization processes of castings has been presented in Chandra U Ahmed A 2002 Modelling for casting and solidification processing Marcel Dekker New York Gustafsson E Stromberg N 2006 Differences between results of residual stresses obtained from the commercial softwares Abaqus and Magmasoft are also presented in Gustafsson E Hofwing M Stromberg N 2007 This work presents a comparison of residual stress development between parts that has and has not undergone topology optimization processes As well we provide a detailed procedure to carry out residual stress simulations both in Abaqus and Magmasoft including the steps for the geometry preparation mesh generation and results comparison using ProEngineer Hypermesh and Matlab respectively The results obtained from the two solvers are also compared and the theoretical fundamentals are given The residual
140. sequent steps Figure 5 9 Constraining the rotation in the X axis in the Original Hub 48 Original Hub Results 5 5 Cooling curves for the Original Hub 1400 MAGMA ABAQUS 1200 1000 2 5 800 3 c 5 5 2 600 p 400 200 0 ca 2 0 5000 10000 15000 Time Seconds Figure 5 10 Abaqus vs Magmasoft cooling curves for the Hub model The results presented in Figure 5 10 were obtained from PNTO which location is shown is Figure 5 4 49 Original Hub Results 5 6 Thermal color spectrums Figure 5 11 Abaqus thermal color spectrums for the last step after shake out of the Original Hub model 50 Original Hub Results Temperature TC Empty 20 94 20 94 20 93 20 93 20 92 20 92 20 92 20 91 20 91 20 90 20 90 20 89 20 89 20 89 20 88 Temperature re Empty 20 94 20 94 20 93 20 93 20 92 20 92 20 92 20 91 20 91 20 90 20 90 20 83 20 69 20 83 20 88 Figure 5 12 Magmasoft thermal color spectrums for the last step after shake out of the Original Hub model 51 Original Hub Results 5 7 Stress curves for the Original Hub Using PNTO as reference point as shown Figure 5 4 x10 e i gt MAGMA ABAQUS 5 4 T Mises Pascal Cd 2 i iH gue 0 5000 10000 15000 Time Seconds Figure 5 13 Von Mises curves from PNT0 of the Original Hub x10 8 E MAGMA ABAQUS 4
141. sh type gt Set the float trias quads option to elems gt Click gt Select All gt set the rest of the panel as Figure 10 14 gt mesh gt reum Tetramesh tetra mesh float trnias quads tetra remesh 14 mesh C volume tetra fixed trias quads Compas split quads include comp sValume tetra mesh normally Figure 10 14 Tetramesh panel setting for the Mold Volume Mesh hm model The volume mesh for the mold is ready Save the file File menu Save 106 Appendix Optional The Mask option can be used again to see the appearance of the inner mesh of the model MV NDS VAN WA AWS lt amp KANA e mA unu P Pn en Ti PU n T a E TUR wr CETT D A m PI d e n ANUS m p e AER a 2 OA A Figure 10 15 A masked view of the Mold model where inner elements can be seen 107 Appendix 10 1 1 2 6 Exporting the meshes to Abaqus The volume meshes for the Mold and the Cylinder were exported as separate inp files that just contain a collection of nodes and elements These inp files were then imported into Abaqus CAE as new models but since the model just contains the geometry the part was copied to a previously prepared model that includes
142. shed in Hypermesh as a surface mesh produce matching nodes in the interface between the casting and the mold the surface mesh for the mold is duplicated to another file In that duplicate file the external surfaces of the mold are deleted remaining just the surface of the cavity already meshed which will be used as the surface of the cylinder From each file a volume mesh is generated and saved to an inp file Note This version of Hypermesh doesn t have the undo option so if you may need to go back to certain point in your work progress you can create a save as of the model in the desired point Start Hypermesh and select the Abaqus User profile as the profile option Start Hypermesh gt Preferences menu gt User Profiles gt Application Hypermesh Check the Abaqus option 2 OK Appendix User Profiles Customize user interface Application Hyper esh t Default HuperMesh OptiStruct Radioss ti Abagus f Acan C C LsDyna t Maduma C Nastan C Pamerash Pamcrash2G2004 gt t Pemas C CFD Always show at start up Figure 10 5 User Profiles window 10 1 1 2 1 Importing the Geometry File menu gt Import gt Geometry gt Step gt Browse the Mold_Surface stl part gt Open 8 iique Han DRY 1 Geometry include comp El Mold Ext Mesh Figure 10 6 Wireframe appearance of the stl g
143. solidififcation stress add on mesh and criteria fill solid options as in Figure 10 70 Ok fast postprocessing preparation result preparation filling temperature filling entrapped air filling pressure filling velocity fill criteria material trace solidification criteria fill solid x ray range show all above 1160 000 d fraction liquid 4 fraction solid stress add mesh interpolated unprepared results only new conversion Figure 10 70 Fast postprocessing preparation window 10 2 2 Calculation The calculation is performed automatically by the solver is just start the simulation what is needed from the user OK T T After selecting ok in the fast postprocessing preparation window the online job simulation control window appears To begin the calculations of the analysis press the start button dismiss Once the calculation is finished press dismiss 162 Appendix 10 2 3 Post Processing 10 2 3 1 Results Visualization The graphical representations of the results are accessed through the Postprocessor module Concerning our interest they can be colored spectrums applied to the model or curves of the history of a variable in a point specified in the Preprocessor This curves values can also be written to text files for further comparisons To access the post processing module
144. t O utput history frequenc y 0 End Step 19 Appendix Stress Analysis PARTS Part name CAST Ele ment type C 304 Section Cast Sect Solid Section elset PickedSet2 matenal CASI MAT 1 End Part ASSEM BLY Assembly name Assembly Instance name CASI 1 part2C AST End Instance End Assembly MATERIALS Matenal name CASI MAT Density Ela stic Expansion Pla stic 192 Appendix PHYSICAL CONSTANTS Physic al Consta nts absolute zero 273 15 stefan boltzmann 5 6 7e 08 BOUNDARY CONDITIONS Name BC 6 Type Disolacement Rotation Boundary _PickedSet74 2 2 _PickedSet74 3 3 Name BC 7 Type Displacement Rotation Boundary PickedSet76 1 1 PickedSet76 2 2 Name BC 9 Type Displacement Rotation Boundary PickedSet81 1 1 PickedSet81 2 2 _PickedSet81 3 3 PREDEFINED FIELDS Name Cast Initial Temp Type Temperature Initial Conditions type TEM PERA TURE _PickedSet49 1400 STEP Stress 1 193 Appendix Step name Stress 1 inc 228800 Sta tic 0 01 28800 1e 11 28800 PREDEFINED FIELDS Name BSO Type Temperature Temperature file C Frog HJ ABAQUS 5 ace R3 odb bstep 1 binc 1 einc 50 OUTPUT REQUESTS Rest
145. te a graphic that shows both curves in the Visualization module they will not be shown one after the other but instead superposed To create a continuous curve of the Abaqus thermal results the rpt files are imported into Excel and then combined 10 3 1 1 1 Loading the Abaqus rpt files into Excel Open Microsoft Excel gt File menu gt Open 2 Files of type files gt Browse the Abaqus rpt file gt Open gt In the Text Import Wizard Step 1 of 3 window select Fixed width see Figure 10 80 gt Next The Text Wizard has determined that your data is Delimited IF this is correct choose Next or choose the data type that best describes your data Original data type Choose the File type that best describes your data Delimited Characters such as commas or tabs separate each Field Fixed width Fields are aligned in columns with spaces between each field Start import at row File origin 437 OEM United States 4 Preview File a i Master Thesis 5 ace R5 rpt gt NTll PI CAST 1 N 2715 1 1 4E 03 z Figure 10 80 Importing an rpt file into Excel Text Import Wizard step 1 of 3 window gt In the Text Import Wizard Step 2 of 3 window adjust the vertical separation line to properly divide the two data columns see Figure 10 81 gt Next gt 171 Appendix This screen lets vau set Field widths column breaks Lines with arrows signify a column break To CREATE a b
146. thermal color spectrums for the last step after shake out of the Optimized Hub model 72 Optimized Hub Results 6 7 Stress curves for the Optimized Hub Using 5 as reference point as shown Figure 6 3 x10 e MAGMA ABAQUS 5 4 T Mises Pascal Cd 2 4 5000 10000 15000 Time Seconds Figure 6 12 Von Mises curves from PNT of the Optimized Hub x 10 MAGMA ABAOLUS 4 1 7 11 D e A 22 1 J 4 E 10 5000 10000 15000 Time Seconds Figure 6 13 Maximum Principal stresses from PNT5 of the Optimized Hub 73 Optimized Hub Results Min Principal Pascal 5000 10000 15000 Time Seconds Figure 6 14 Minimum Principal stresses from PNT5 of the Optimized Hub 74 Optimized Hub Results 6 8 Stress color spectrum 1 970 08 1 755e 08 1 539e 08 1 323 08 1 107e 08 5 bU 8 910 07 6 752 07 gt 4 593 07 2 434 07 2 753 06 UNI VN DEVIS A VIN yO gt L INCUN VE NOCT AE OC v J A AVI ORG VN SANCII SINK TAVAT TATATATA d T inm 2005 TAVA Xv ca RAN 8 Ze A V A n Vx Y 8 PON ANS
147. tion collector gt Create gt Name it MOLD AMBIENT CONVECTION 3 Step Before Shake Out gt Type Surface film condition gt Continue Select the external surface of the Mold gt Done gt Define the Edit Interaction window as Edit Interaction Mame MOLD 4MBIENT CONVECTION Type Surface Film condition Step Before Shake Qut Heat transfer Surface MOLD EXT SURF Film coefficient Film coefficient amplitude Sink amplitude Instantaneous Figure 10 26 Convective interaction between the mold and the ambient 123 Appendix Radiation Mold Ambient Right click the Interaction collector gt Create gt Name it MOLD AMBIENT RADIATION gt Step Before Shake Out gt Type Surface radiation to ambient Continue Select the external surface of the Mold Done 2 Define the Edit Interaction window as Mame MOLD AMBIENT PR ADTIATIGOR Type Surface radiation to ambient Step Before Shake Out Heat transfer Surface MOLD EXT SURF edit Region Ambient temperature lo sd Ambient temperature amplitude Instantaneous hal Mote The absolute zero temperature and Stefan Boltzmann constant must be specified in the Edit Model Attributes dialog Figure 10 29 Radiation interaction between the mold and the ambient 11 Boundary Conditions No mechanical boundary conditions are needed to be specified in the thermal problem 12 Predefined Field Requests In this step the nodal thermal
148. tions Definitions In Magmasoft this condition is defined automatically so the user has no participation in the setting 45 Original Hub Results 5 4 2 Mechanical boundary conditions 5 4 2 1 Stress analysis step As mentioned in the Cylinder model results the user does not participate directly in the definition of boundary conditions for the stress analysis in Magmasoft It is an automatic procedure Therefore just the Abaqus approach is presented The task 1 to constrain the rigid body translations and rotations in X Y and Z but allow the body to deform In the Original Hub model the 6 degrees of freedom has been constrained as follow Translations in X Y and Z In the top flat surface of the Original Hub a node in 0 is constrained in X Y and Z See Figure 5 7 Notice that in the picture X is the horizontal axis Y the vertical axis and the Z axis is perpendicular to the paper L Edit Boundary Condition Ed Mame Displacement Ratatian Step Initial Region Picked csvs Global ut uz un1 un2 un3 Note The displacement value will be maintained in subsequent steps Figure 5 7 Constraining the rigid body translations in X Y and Z in the Optimized Hub By fixing this node in the space three rigid body translations are constrained Consequently the part would shrink toward this node Now the remaining task is to constraint the three degrees of freed
149. ub top and Optimized Hub bottom Mises comparison Inclined view 84 Original and Optimized Hub Comparison 7 2 Maximum Principal Stress 5 Principal Avg 75 Figure 7 4 Original Hub top and Optimized Hub bottom Maximum Principal Stress comparison Top view 85 Original and Optimized Hub Comparison 5 Princip a 9 TET Figure 7 5 Original Hub top and Optimized Hub bottom Maximum Principal Stress comparison Bottom view Original and Optimized Hub Comparison 5 Principal 5 Principal 55 Figure 7 6 Original Hub top and Optimized Hub bottom Maximum Principal Stress comparison Inclined view 97 Original and Optimized Hub Comparison 7 3 Minimum Principal Stress 5 Min Principal Avg 7596 m Min Principal 596 Figure 7 7 Original Hub top and Optimized Hub bottom Minimum Principal Stress comparison Top view 98 Original and Optimized Hub Comparison 5 Min Principal Avg 75 S Min Principal Avg 7595 Figure 7 8 Original Hub top and Optimized Hub bottom Minimum Principal Stress comparison Bottom view Original and Optimized Hub Comparison 5 Min Principal Aug 7556 Figure 7 9 Original Hub top and Optimized Hub bottom Minimum Principal Stress comparison Inclined view 90 Conclusions and discussions 8 Conclusions and discussions The imple
150. ved from the mold and not the other way around Therefore the mold volume must be placed before the casting in the list of volumes Select menu gt Volume gt Select the volume corresponding to the mold gt Click the upper entity selector button above the Move Before and Move After buttons Select the volume corresponding to the casting gt Click the lower entity selector button at this point it must look like Figure 10 55 gt Hove Before ___ Return _ 151 Appendix Select Volume Hagma_Cyl inder Move Before Move After Hagma Lulinder Return Next Level Figure 10 55 Entity Selections windows with the volumes selected prior organizing Save your work File menu gt Save Active gt Name it M Cylinder gt Save 152 Appendix 10 2 1 2 Mesh Generation First exit the Preprocessor File menu Exit Click the button in the Magmasoft main interface gt Choose the automatic method gt Assign the desired number of elements for the whole model casting and mold gt generate gt dismiss gt gt gt gt 62 5 ae Figure 10 56 Magmasoft mesh generation window Note Since Magmasoft use a Control Volume Finite Difference formulation where there is just one node in the center of each element the number of elements are equivalent to the number of nodes in Abaqus 153 Appendix 10 2 1 3 Magmasoft Simulation Setup C
151. way the rotation in the X axis is restrained and all vertical axes of the part are fixed to remain parallel to the Z axes lI Edit Boundary Condition 3 Mame xY Displacement Rotation Step X Stress 1 Static General Region Picked CSY Transform T BCS CSYS Distribution UniForm Ut 0 02 0 us UR1 982 ni radians radians Amplitude Ramp v Note The displacement value will be maintained in subsequent steps Figure 6 6 Constraining the rotation in the X axis in the Optimized Hub 69 Optimized Hub Results 6 5 Cooling curves for the Optimized Hub MAGMA ABAQUS 1000 4 800 600 Temperature Celsius 400 200r 1 E 0 5000 10000 15000 Time Seconds Figure 6 9 Abaqus vs Magmasoft cooling curves for the Optimized Hub model The reference point for the curves in Figure 6 9 is PNT5 which location is shown in Figure 6 3 70 Optimized Hub Results 6 6 Thermal color spectrums Figure 6 10 Abaqus thermal color spectrums for the last step after shake out of the Optimized Hub model 71 Optimized Hub Results i n Empty 20 49 20 49 20 49 20 49 20 48 20 48 20 48 20 48 20 48 20 48 20 47 20 47 20 47 20 47 20 47 eure Empty 20 49 20 49 20 49 20 49 20 48 20 48 20 48 20 48 20 48 20 48 20 47 20 47 20 47 20 47 20 47 Figure 6 11 Magmasoft

SIMULATION OF RESIDUAL STRESSES IN CASTINGS

Contents

Download Pdf Manuals

Related Search

Related Contents