4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Element Sizes in Crash Calculation Dipl.-Ing Ing. Udo Jankowski, Dr.-Ing Martin Müller M Bechtel,, Dipl.-Ing. Manfred Sans,, Tecosim GmbH 4. LS-DYNA Forum, 20.-21. 21. Oktober 2005, Bamberg Investigation on mesh density in LS Dyna Agenda Tecosim-best partner for simulation Introduction Analysis of Results/ Conclusion Outlook I - I - 33
900 800 700 600 500 400 300 200 100 0 1 2 3 4 5 6 7 8 CPU # standard scale_c scale_c2 scale_x scale_y Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 Locations & References Leonberg Rüsselsheim Köln Basildon (UK) Audi AG Adam OPEL AG Claas Daimler Chrysler Daihatsu Fiat FORD General Motors HONDA ISUZU KIA John Deere Jaguar Landrover Nissan PORSCHE AG Toyota Coventry (UK) AMG Autoliv Bayer AG Bentler Bertone Bosch/ Blaupunkt Degussa-Hüls AG Dynamit Nobel EADS Faurecia Getrag Hella KG Johnson Controls Karmann Lear Magna Thyssenkrupp TRW Automotive Mahle MAN Mannesmann/Sachs Siemens VDO Igenie Office Wagon Automotive Tokio (Japan) CAE Portfolio NVH / Durability CRASH Safety Powertrain CAE Portfolio Seats Multi Body Systems (MBS) CPU time CPU time [s] CFD Optimization Software dev I - I - 34
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Introduction Any sufficiently advanced technology is indistinguishable from magic. Profiles of the future (1961) by Arthur C. Clarke (2003) Introduction Why do we simulate? Cost effective Fast Proven method We cannot test! I - I - 35
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 Introduction Number of Elements for a FE Crashmodel BIW 2500000 2000000 1500000 1000000 # of Elements 500000 0 1999 2000 2001 2002 2003 2004 Introduction How should a mesh look like? TEC ODM Mesh Element size 12 mm TEC ODM Mesh Element size 4 mm TEC ODM Mesh Element size 6 mm I - I - 36
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Simple box crash experiment: Box section 50 mm x 80 x 500mm, t= 1.0mm, mild steel Varied parameters: average edge length 15/10/5/2,5mm mesh orientation 0deg/ 25deg different mesh/ integration method: Belytschko-Tsay/ Fully Integration Varied number of spotwelds With and without mapping or stamping data Renumbering and move in space Objective: Is the result depending on the element length? Is the result depending on the element orientation? How does mapping influence/stabalise the results? How do small changes in the input influence the results? v10_1: 10mm mesh, 0deg max internal energy 6766Nmm max displacement 278mm I - I - 37
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 V10_1n: 10mm mesh, 0 deg, full integration max internal energy 6499Nmm max displacement 241mm v10_2: 10mm mesh, 25deg max internal energy 6439Nmm max displacement 295mm I - I - 38
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung V10_2n: 10mm mesh, 25 deg, full integration max internal energy 6272Nmm max displacement 286mm v10_3: 10mm triangle mesh max internal energy 6428Nmm max displacement 215mm I - I - 39
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v10_4: 10mm mesh, 25deg, more spotweld max internal energy 6560Nmm max displacement 265mm V10_4n: 10mm mesh, 25 deg, more spotweld, full integration max internal energy 6741Nmm max displacement 295mm I - I - 40
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung v10_5: 10mm mesh, 0 deg, more spotweld max internal energy 6565Nmm max displacement 262mm v10_6: 10mm mesh, 0deg mapped stamping data MpCCI max internal energy 6658Nmm max displacement 235mm I - I - 41
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v10_7: 10mm mesh, 25deg mapped stamping data MpCCI max internal energy 6350Nmm max displacement 243mm v10_6n: 10mm mesh, 0deg mapped stamping data DYNAIN max internal energy 6468Nmm max displacement 250mm I - I - 42
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung v10_7n: 10mm mesh, 25deg mapped stamping data DYNAIN max internal energy 6376Nmm max displacement 233mm Analysis: comparison of Displacement 10mm meshing 3,50E+02 3,00E+02 2,50E+02 2,00E+02 1,50E+02 1,00E+02 5,00E+01 0,00E+00 0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00 55,00 60,00 65,00 70,00 75,00 80,00 85,00 90,00 95,00 100,00 Displacement [mm] 10_1 max 279,79 10_2 max 294,56 10_3 max 215,16 10_4 max 265,30 10_5 max 262,47 10_6 max 227,10 10_6n max 250,63 10_7 max 302,08 10_7n max 232,63 Time [ms] I - I - 43
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v5_1: 5mm mesh, 0deg max internal energy 6398Nmm max displacement 296mm V5_1n: 5mm mesh, 0 deg, full integration max internal energy 6481Nmm max displacement 277mm I - I - 44
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung v5_5: 5 mm mesh, 0 deg, more spotweld max internal energy 6466Nmm max displacement 295mm v5_2: 5mm mesh, 25deg max internal energy 6363Nmm max displacement 278mm I - I - 45
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v5_3: 5mm triangle mesh max internal energy 6425Nmm max displacement 265mm V5_4: 5mm mesh, 25deg, more spotweld max internal energy 6327Nmm max displacement 277mm I - I - 46
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung v5_6: 5mm mesh, 0deg mapped stamping data MpCCI max internal energy 6401Nmm max displacement 268mm v5_7: 5mm mesh, 25deg mapped stamping data MpCCI max internal energy 6593Nmm max displacement 278mm I - I - 47
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v5_6n: 5mm mesh, 0deg mapped stamping data DYNAIN max internal energy 6346Nmm max displacement 268mm v5_7n: 5mm mesh, 25deg mapped stamping data DYNAIN max internal energy 6323Nmm max displacement 278mm I - I - 48
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Analysis: comparison of Displacement 5mm meshing 3,50E+02 3,00E+02 Displacement [mm] 2,50E+02 2,00E+02 1,50E+02 1,00E+02 5,00E+01 0,00E+00 0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00 55,00 Time [ms] 60,00 65,00 70,00 75,00 80,00 85,00 90,00 95,00 5_1 max 295,75 5_1n max 276,58 5_2 max 278,37 5_3 max 264,94 5_4 max 276,87 5_5 max 295,21 5_6 max 269,77 5_6n max 267,51 5_7 max 276,48 5_7n max 277,82 100,00 v2.5_1: 2.5mm mesh, 0deg max internal energy 6398Nmm max displacement 296mm I - I - 49
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 V2-5_1n: 2.5 mm mesh, 0 deg, full integration max internal energy 6439Nmm max displacement 291mm V2-5_5: 2.5 mm mesh, 0deg, more spotweld max internal energy 6423Nmm max displacement 289mm I - I - 50
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung V2-5_2n: 2.5 mm mesh, 25 deg max internal energy 6471Nmm max displacement 287mm v2.5_3: 2.5mm triangle mesh max internal energy 6448Nmm max displacement 279mm I - I - 51
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 Analysis: comparison of Displacement 2.5mm meshing 3,50E+02 3,00E+02 2,50E+02 Displacement [mm] 2,00E+02 1,50E+02 1,00E+02 5,00E+01 0,00E+00 0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00 55,00 Time [ms] 2.5_1 max 297,53 2.5_1n max 290,50 2.5_2n max 287,30 2.5_3 max 278,79 2.5_5 max 288,75 2.5_6 max 293,74 2.5_6n max 297,34 2.5_7 max 282,87 2.5_7n max 311,40 60,00 65,00 70,00 75,00 80,00 85,00 90,00 95,00 100,00 V2-5_4n: 2.5 mm mesh, 25 deg, more spotweld max internal energy 6477Nmm max displacement 280mm I - I - 52
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung v2.5_6: 2.5mm mesh, 0deg mapped stamping data MpCCI max internal energy 63412Nmm max displacement 293mm v2.5_7: 2.5mm mesh, 25deg mapped stamping data MpCCI max internal energy 6416Nmm max displacement 285mm I - I - 53
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 v2.5_6n: 2.5mm mesh, 0deg DYNAIN max internal energy 6360Nmm max displacement 297mm v2.5_7n: 2.5mm mesh, 25deg DYNAIN max internal energy 6392Nmm max displacement 311mm I - I - 54
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Analysis of Results/ Conclusion Analysis: Comparison max. displacement 400 Compression is nearly independent from the element length in a range from 10mm to 2.5 mm for the same element orientation max displacement [mm] 375 350 325 300 275 250 225 200 175 150 125 c 100 75 50 25 0 0 1 2 3 4 5 6 7 8 9 10 11 12 average element length [mm] Analysis of Results/ Conclusion Analysis: Comparison max. displacement The max displacement difference for 0 mesh and 25 mesh is small for finer meshes and big for coarser meshes 20 18 16 14 12 10 8 c 6 4 2 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 average element length [mm] I - I - 55
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 Analysis of Results/ Conclusion Analysis: Influence on mesh translation and 350,00 reumbering 300,00 For 10 mm the variation is about 8mm where as the deviation for smaller elements sizes is about 3mm for the same element orientation 250,00 200,00 350,00 300,00 250,00 200,00 150 00 300,00 2_1 max 290,09 2_1_ren max 290,09 2_1_dis max 293,09 3.5_1 max 289,05 3.5_1_ren max 289,05 3.5_1_dis max 286,29 3.5_1_ren max 288,70 250,00 200,00 150,00 10_1 max 279,79 10_1_ren max 271,64 10_1_dis max 278,53 Analysis of Results/ Conclusion Analysis: Influence on results of very small elements. C o m p ariso n o f F u ll In teg ratio n 3,50E+02 3,00E+02 For 1,25 mm the BT (typ2) element seems to be to weak. For smaller elements sizes, full integration seems to have better results. Displacement [mm] 2,50E+02 2,00E+02 1,50E+02 1,00E+02 5,00E+01 0,00E+00 0,00 3,50E+02 6,00 12,00 18,00 24,00 30,00 36,00 42,00 48,00 54,00 60,00 Time [ms] C om parison of BT 66,00 72,00 2.5_1n m ax 290,50 2_1n m ax 282,73 1.25_1n m ax 296,78 78,00 84,00 90,00 96,00 3,00E+02 Displacement [mm] 2,50E+02 2,00E+02 1,50E+02 1,00E+02 5,00E+01 0,00E+00 0,00 5,00 10,00 15,00 20,00 25,00 30,00 35,00 40,00 45,00 50,00 55,00 60,00 65,00 70,00 75,00 80,00 85,00 90,00 95,00 100,00 2.5_1 max 297,53 2_1 max 290,09 1.25_1 max 303,92 Time [ms] I - I - 56
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Analysis of Results/ Conclusion Analysis: comparison max displacement variation variation for different mesh sizes displ. (mm) 310 290 270 250 230 210 190 170 150 1 2 3 4 5 6 7 8 9 run 10mm 5 mm 2.5 mm Analysis of Results/ Conclusion Analysis: Comparison of integration method Calculation Time CPU Time [s] 1,80E+04 1,60E+04 Belytschko-Tsay Fully Integration 1,40E+04 1,20E+04 1,00E+04 8,00E+03 6,00E+03 4,00E+03 2,00E+03 0,00E+00 10_1/10_1n 10_2/10_2n 10_4/10_4n 5_1/5_1n 5_2/5_2n 5_4/5_4n 2.5_1/2.5_1n 2.5_2/2.5_2n 2.5_4/2.5_4n I - I - 57
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 Analysis of Results/ Conclusion Results Results for the displacement and the internal energy seem to be smooth and stable in a range from 15mm to 2,5 mm for orthogonal element orientation Different element orientation give different results for coarser meshes Finer mesh is not so sensitive for different element orientation, integration method, number of spotwelds, mapping, small changes in the input (renumbering, moving the model in space) Mapping tools are easy to use for Crash coupling. The influence of the mapping was getting smaller for smaller element sizes for the influenced zone was getting smaller and the crash mode was very stable in our example. Analysis of Results/ Conclusion Conclusion If you know the collapse mode of a part you can use a coarse mesh which should be orthogonal in the collapse direction (so you can achieve superconvergence ) If you doesn't know the collapse mode of a part; Please use finer meshes No one knows the exact collapse mode of all the parts in a vehicle! Meshing rules for orthogonal /Mapping/Integration schemes meshes are important for coarser meshes but not important for finer meshes. Creation of finer meshes can be automated by TEC ODM! I - I - 58
4. LS-DYNA Anwenderforum, Bamberg 2005 Modellierung Analysis of Results/ Outlook Outlook The crashbox sample will be applied to a complete vehicle to find out about the time saving potential and the influence on the results. Thank you for your attention! Please ask some questions I - I - 59
Modellierung 4. LS-DYNA Anwenderforum, Bamberg 2005 I - I - 60