Simulação CAE para previsão de cargas e análise de vida em fadiga Fabio Orefice Engenheiro de Aplicações CAE

Transcription

1 Simulação CAE para previsão de cargas e análise de vida em fadiga Fabio Orefice Engenheiro de Aplicações CAE Unrestricted Siemens AG 2016 Realize innovation.

2 Closed Loop System Driven Product Development The Strategic Role of Simulation and Test in Systems Engineering LMS Imagine.Lab LMS Test.Lab 1D Behavioral Simulation Virtual to Physical SIMCENTER 3D 3D CAE LMS Virtual.Lab LMS Samtech Suite Scalable 1D - 3D Simulation Incl. Real-Time Virtual to Physical Page 2 J

3 Page 3 APLICÁVEL À TODOS OS TIPOS DE INDÚSTRIA

4 Simulação CAE para determinação de cargas O que o meu cliente fará com o meu produto?

5 Which parameters influence fatigue? Three parameters Loads Load level Constant amplitude/variable amplitude Uni-axial/multi-axial σ, F Failure t, time Geometry Load configuration Notch severity Local stress state Material Basic properties Surface finish/manufacturing Residual stresses Temperature Page 5

6 Traditional Durability Design Process Physical prototype Design Load data collection Build Test Break: slow expensive quality problems Physical rig test Page 6

7 Introducing CAE (stress plots only) Physical prototype Design Load data collection Stress analysis: Stress range needed for fatigue! Dynamic load path? Where are the critical locations? Physical rig test Page 7

8 Numerical Durability Design Process Physical prototype Design Load data collection Virtual prototype Fatigue life prediction Load data analysis Loads FEM Page 8 Test schedule Numerical validation before testing: add time dimension to FEM less and better physical tests earlier important decisions Physical rig test

9 Simulation based Refinement Physical prototype Design Load data collection Virtual prototype Fatigue life prediction Loads Load data Prediction analysis Simulate component loads: required for time domain simulations of rig tests derive component tests Test Virtual test rig FEM Physical rig test Simulate driving on test track: schedule estimate loads before building prototypes Page 9

10 Durability Analysis Approaches Component loading to elastic stress history Component loading L 1 Elastic stress history t L 2 e x e xy e y e x e xy e y Quasi-static Superimpose elastic pseudo stress No rigid body motion and loads do not excite natural frequencies Inertia relief Accounts for rigid body motion Modal superposition Accounts for the excitation of natural frequencies Flexible bodies System level approach Automatically accounts for rigid body motion and the excitation of natural frequencies Transient analysis Large deflections Complex geometrical non-linearity (contact) Rotational Loading Vibration based (Random Fatigue) Stationary process Estimate local stress distribution t Page 10

11 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 11

12 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 12

13 Durability Road Loads Prediction: Current industry practice CAE data CAE data TEST data Full Vehicle Loads Hybrid CAE-TEST Full Vehicle Loads 100% CAE: Digital Test Track Suspension Loads Simulation set-up: Constrained Subsystem/ component MBS model Model excitations: Measured spindle forces (WFT, Test) Page 13 Most easy model set-up Directly apply measured loads Body loads not accurate Simulation set-up: (Predecessor) WFT measurements Unconstrained MBS model of predecessor & new vehicle Model excitations: Driving function (usually displacement) which is back-calculated from measurements LMS Virtual.Lab Motion-TWR Realistic simulation Accurate body loads Avoid tire and road modeling Still test data required Model complexity Simulation set-up: MBS vehicle model MBS durability tire model 3D road surface model Driver model Driveline model Model excitations 3D road profile True forward prediction. (No test or predecessor data required) All maneuvers are possible Tire parameter identification

14 Force (N) Attribute Name Unit C:\Programme\LMS\TW 3.3 QC\demo\general\nature.dmd,FORCE_long N C:\Programme\LMS\TW 3.3 QC\demo\general\nature_edt.dmd,FORCE_long N C:\Programme\LMS\TW 3.3 QC\demo\general\nature.dmd,FORCE_lat N C:\Programme\LMS\TW 3.3 QC\demo\general\nature_edt.dmd,FORCE_lat N C:\Programme\LMS\TW 3.3 QC\demo\general\nature.dmd,FORCE_vert N C:\Programme\LMS\TW 3.3 QC\demo\general\nature_edt.dmd,FORCE_vert N Zeit [s] 1 Stress (MPa) Component Fatigue Data Flow FORCE_long N 1 FORCE_lat N 2 FORCE_vert N 3 Loads Time (s) Groups (optional) Time (s) Load Function Set Load FE Assignment FALANCS (Solver) Post Process FE Model FE-Solver Mode Set Material & Parameters Page 14

15 Durability Road Loads Prediction: Current industry practice CAE data CAE data TEST data Full Vehicle Loads Hybrid CAE-TEST Full Vehicle Loads 100% CAE: Digital Test Track Suspension Loads Simulation set-up: Constrained Subsystem/ component MBS model Model excitations: Measured spindle forces (WFT, Test) Page 15 Most easy model set-up Directly apply measured loads Body loads not accurate Simulation set-up: (Predecessor) WFT measurements Unconstrained MBS model of predecessor & new vehicle Model excitations: Driving function (usually displacement) which is back-calculated from measurements LMS Virtual.Lab Motion-TWR Realistic simulation Accurate body loads Avoid tire and road modeling Still test data required Model complexity Simulation set-up: MBS vehicle model MBS durability tire model 3D road surface model Driver model Driveline model Model excitations 3D road profile True forward prediction. (No test or predecessor data required) All maneuvers are possible Tire parameter identification

16 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 16

17 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 17

18 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 18

19 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 19

20 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 20

21 How to predict loads? Multi-body simulation Durability loads prediction CAD Create or import Kinematic Joints or constraints Dynamic Forces (gravity, stiffness, damping) Flexible bodies Craig-Bampton or test deformation Solving Fast, robust, accurate Post-processing 2D/3D and root cause analysis Page 21

22 LMS Virtual.Lab Motion Capturing industry best modeling practices Mechatronics Modular assembly Verticals Automation CAD Create or Import Kinematics Joints Constraints Initial Conditions Dynamics Forces (Gravity, Friction, Damping, Stiffness, ) Flexible bodies CAE / Test modes superposition Nonlinear flex bodies Solving Fast Accurate Post-processing 2D / 3D Root cause analysis Durability NVH Acoustics Control system 1D model TEST data Parameterization / Optimization Page 22

23 LMS Virtual.Lab Motion TWR Time Waveform Replication Unique link with Test data For A road loads or durability simulation team TEST Measurements That needs to Apply measured signals to MBS/FE models Deliver component / body loads for durability analysis Benefits Realistic: same boundary conditions as in TEST Pragmatic approach: no tire model (F-Tire, CDTire, non-linear agriculture tire model) nor digitized road surface needed Robust: method deployed at BMW, Nissan, Honda, FAW, AGCO, CNH, Simulation results Technology Back-calculation of (arbitrary) set of measured sensor signals to driving functions in multi-body simulation Page 23 From any measurement to robust road loads prediction

24 LMS Virtual.Lab Motion TWR Hybrid Road: Test to CAE for Durability analysis Existing vehicle Vehicle A (test) MBS & Tire Model NEW vehicle or vehicle variant Body loads Enabled by Vehicle A (MBS) Back-calculate Use for prediction LMS Motion- TWR System identification Target simulation road profiles Equivalent road surface (invariant) Page 24 Validation accumulated Damage Fatigue Life

25 LMS Virtual.Lab Motion TWR Hybrid Road: Test to CAE for Durability analysis Explore different configurations in LMS Virtual.Lab Motion Variant A Compute and compare the component fatigue life in Virtual.Lab Durability and TecWare Variant B Equivalent road Variant C Page 25

26 Motion TWR workflow MBS model preparation Unconstrained vehicle with WFT MBS model setup Sensors Actuators Page 26

27 Durability Road Loads Prediction: Current industry practice CAE data CAE data TEST data Full Vehicle Loads Hybrid CAE-TEST Full Vehicle Loads 100% CAE: Digital Test Track Suspension Loads Simulation set-up: Constrained Subsystem/ component MBS model Model excitations: Measured spindle forces (WFT, Test) Page 27 Most easy model set-up Directly apply measured loads Body loads not accurate Simulation set-up: (Predecessor) WFT measurements Unconstrained MBS model of predecessor & new vehicle Model excitations: Driving function (usually displacement) which is back-calculated from measurements LMS Virtual.Lab Motion-TWR Realistic simulation Accurate body loads Avoid tire and road modeling Still test data required Model complexity Simulation set-up: MBS vehicle model MBS durability tire model 3D road surface model Driver model Driveline model Model excitations 3D road profile True forward prediction. (No test or predecessor data required) All maneuvers are possible Tire parameter identification

28 Full Vehicle Loads Cross Attribute Capabilities (LMS Virtual.Lab Motion & LMS Virtual.Lab Durability) Process scheme System Loads Test Rig Control Benefits No file transfer = no waste of time and no errors Modal participation factors always matches properly with FE modes Coordinate system orientation always correct System Model LMS Virtual. Lab Motion FE-solver Modal Participation Factors Structural FE Mesh Static and Normal Modes Fatigue setup Fatigue solver (LMS Virtual.Lab Durability) Post- process Page 28

29 Example of Full Vehicle CAE Analysis Durability evaluation at Front Truck Crossmember Page 29

30 Example of Full Vehicle CAE Analysis Durability evaluation at Full Vehicle Body Page 30

31 Example of Full Vehicle CAE Analysis Durability evaluation at Front Suspension Arm Page 31

32 Multibody Additional Examples Page 32

33 Multibody + Durability Additional Examples Page 33

34 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 34

35 Challenge to achieve good Road Loads Prediction Tire Representation Basic Handling Ride Comfort Durability Road noise LMS Standard Tire (till 10 Hz) (till 30 Hz) (till 50 Hz) (till 250 Hz) TNO Delft MF-Tire TNO Delft MF-Swift ITWM Comfort & Durability Tire FTire LMS Modal tire Frequency Page 35

36 Challenge to achieve good Road Loads Prediction Including Road Surface Model LMS Road Profile Interface Road element, referenced in each tyre Digitized Test Track definition for LMS CDTire and TNO MF-Tyre/MF-SWIFT Different road definitions Spline curve / spline surface ASCII model (RSM 1000) binary model (RSM 2000) for digitized test tracks 2D Road Profile *.rdf for TNO MF-Tyre/SWIFT Open CRG Road surface visualization Page 36

37 Challenge to achieve good Road Loads Prediction Including Mechatronic Systems Controls system integration using LMS Imagine.Lab 3 rd party interfaces (e.g. MATLAB/Simulink) FMI for Co-simulation and Model Exchange Real-Time simulations Page 37

38 Challenge to achieve good Road Loads Prediction Including Mechatronic Systems LMS Virtual.Lab Motion & LMS Imagine.Lab Amesim (Stryker Study w/ U.S. Army AMSAA) Model System Integration -hydraulic / pneumatic systems -powertrain matching -combustion, lubrication, etc. Page 38 AMSAA Coupled 3D Multi-body Dynamics & 1D Driveline System Simulation Vehicle Energy Management -Heat Exchangers -Cooling System -HVAC, Cabin Thermals Control System Validation -Electrical Networks -Hybrid Networks -HIL, SIL, MIL

39 Challenge to achieve good Road Loads Prediction Including Non-Linearity Driver feel Bump Bound/rebound Damper & Coil Spring prestress & hysteresis Tire Temp/Pressure effect Large Tire deformation Fuel economy Driver Comfort Braking Performance Judder & Squeal Driver feel Bushing non-linearity (hyper elasticity, Effect Bending Antiroll bar Geometrical nonlinearity of twistbeams Non-linearity leafsprings Friction forces, free play in steering rack composite gears Power steering Contact, friction, Steering feel temperature Page 39

40 Challenge to achieve good Road Loads Prediction Coupling with advanced soil model Page 40

41 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 41

42 Additional capabilities Hybrid analysis with experimental modes Experimental and Reduced FE Model (ERFEM) Import in Test formats Typically wireframe mesh Total component s mass properties at c.g. Reduced Modal Data Drastic data reduction Wireframe mesh for visualization Displacement data only at interfaces points Special Mass Matrix Case for mass, modal mass and stiffness Page 42

43 Additional capabilities Crack Propagation Prediction Page 43

44 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 44

45 LMS Driving Dynamics Automated Process for OEM Multiple Attributes & Performances to manage Chassis Subsystem Packaging Optimization Durability Loads Prediction Driving Dynamics Vehicle Dynamics K&C and Handling NVH and Road noise Vehicle Dynamics Ride Comfort Page 45

46 LMS Driving Dynamics Capturing industry best practices in 3D and 1D simulation Data Collection LMS Driving Dynamics Swift vertical Vehicle and Bushings Data LMS Virtual.Lab Composer Cost-effective customization LMS Virtual.Lab Motion LMS Imagine.Lab Amesim Reporting LMS Test.Lab LMS Motion Real-Time MS Excel Page 46 3d party and in-house tool support

47 Agenda Durability Road Loads Prediction Component Fatigue Data Flow TWR Time Waveform Replication Full Vehicle Loads Examples Challenge to achieve good Road Loads Prediction Tire Representation, Non Linearity, Road Surface Model, Mechatronic Systems, Coupling with advanced soil model Additional Capabilities Hybrid Approach and Crack Propagation Prediction LMS Driving Dynamics References Page 48

48 PSA PEUGEOT CITROËN selects LMS Virtual.Lab as integrated simulation platform for chassis engineering LMS wins multi-year contract for the deployment of +350 seats LMS Virtual.Lab to more than 500 users Covering all key vehicle performance attributes: NVH and Acoustics Motion 50 seats: suspension & chassis mechanisms (closures) vehicle body loads for durability Structures Pre-Post (incl. Crash) Structures Assembly (incl. Crash & Safety) Christophe Chanteur, CAE Full-Vehicle Department, Peugeot-Citroen PSA: The newly introduced modular approach for full-vehicle modeling matches our needs and is very useful to us. It enables us to create fullvehicle models starting from the models of their submechanisms, and also to define standard driving maneuvers much faster than before. Page 49

49 John Deere increases tractor endurance through early-process simulation using LMS Virtual.Lab John Deere Werke Mannheim, Germany, frontloads virtual simulation into the development process to further increase the quality and reliability of new tractor models LMS Virtual.Lab Motion and Durability provide valuable insight into how design variations influence tractor quality and reliability Virtual simulation helps John Deere compress tractor development cycles, while delivering better performance, flexibility, comfort and economy at comparable cost The introduction of Virtual.Lab Durability in the early design stages when there is still plenty of design freedom positively impacts tractor reliability and reduces the number of required prototype tests. Dr. Christian von Holst, Senior Engineer, John Deere Werke Mannheim Page 50

50 FAW (First Automobile Works) Customer request: Get accurate loads and fatigue life prediction early in the design process LMS Solution Technology transfer: Applying measured loads on vehicle model in most realistic manner ( Hybrid Road, combining Test and Simulation) LMS Virtual.Lab Motion LMS Motion-TWR LMS Virtual.Lab Durability LMS TecWare Comparison Results, Cobblestones 2, FX_RL 5 copyright LMS International Page 51

51 Case Study - Daimler Page 52

52 Dúvidas e Perguntas??? Page 53

53 Meus Contatos!!! Fabio Orefice Engenheiro de Aplicações CAE Rua Alegre, 443 São Caetano do Sul - SP Telefone: + 55(11) fabio.orefice@siemens.com siemens.com Page 54

54 Muito Obrigado!!!