Body Structure & Chassis Optimization

Transcription

1 Dipl.-Ing. Elmar Teipen R&A Global Vehicle Dynamics Dipl.-Ing. Hessel van Dijk Ford of Europe Body Engineering Dr.-Ing. Michel Paas Ford of Europe Body Engineering Dr.-Ing. Axel Hänschke Ford of Europe Analytical Prototype Body Structure & Chassis Optimization For Improving Driving Maneuvers 2011 European Hyperworks Technology Conference Bonn

2 Agenda Why Body Structure Is Important To Vehicle Dynamics What Is The Body Doing During Driving? How To Investigate Structural Performance For VD? Customized Hyperstudy Approach MDO Application Examples Summary & Next Steps Page 2

3 Why Body Structure Is Important To Vehicle Dynamics Steering Straight Ahead Controllability (On-Center) Ride Compliance Response: Amount, Linearity, Delay Steering Wheel Torque Feedback Modulation (Torque vs. Response) Primary Ride Control / Comfort Secondary Ride Shake Rolling Feel / Plushness Impacts Solidity / Rigidity / Quality Impression Vehicle Dynamics Performance Influenced By Structural Stiffness & Weight Page 3

4 What Is The Body Doing During Driving? Front End Lateral Bending On-Center Compliance Steering Precision Response Delay, Non-Linearity Torsion Front-to-Rear Balance Compliance Body Shake, S&R, Rigidity Feel Vertical Bending Impact Harshness Front End Shake, (S&R) Trade-Off : Body / Chassis Lightweight Vehicle Dynamics Performance Page 4

5 How To Investigate Structural Performance For VD? Local EWS HPC Local EWS Hypermesh Preprocessing Competitor SW MNF-Creation ADAMS/Chassis Solving Hyperworks Postprocessing - ADAMS/Chassis: MBS Pre-Processing (With Vehicle Templates) & Solving - Embedded FE-Structures Of Body, Subframes, etc.» Manual Hypermesh Preparation (Superelements, Interface Nodes,...)» Modal Reduction With Competitor Software On HPC Supercomputer» Sequential MNF-Creation: Up To 3 h Per Run / Up To 16 GB mnf To Transfer» Implementation With Python-Templates (Manually Edited) - Hyperworks: VD Postprocessing (Based On Maneuver-Dependent Reports) - Body & Chassis Structural Modifications Based On Individual Results Time-Consuming Manual Interactions Limited Variations Not Efficient Page 5

6 Customized Hyperstudy Approach HPC TCL-Templates Preprocessing Radioss MNF-Creation ADAMS/Chassis Solving TCL-Templates Postprocessing Customized Hyperstudy Control & PP - Hyperstudy Customized Integration:» Selection Of Design Variables (In FE And MBS Templates) User Friendly» In-The-Loop Generation / Calculation Of Combined Models Automated Pre+Post» Fully Automated DOE Control From Local EWS No Manual Interaction» Data-Intensive Work Completely On HPC Supercomputer Reduced File Transfer» Semi-Parallel Runs Possible Reduced Run Time» Collection Of Final Results (Some kb Per Run) On EWS Reduced Data» Optimized License Application Reduced Cost Customized Hyperstudy = Key Enabler For Statistical FE + MBS Analysis on HPC Page 6

7 Application Examples: Stepsteer & Ride 4-Poster Vehicle Dynamics Metrics vs. Sheets - Steering (Sinesteer, Stepsteer, On-Center) - Ride (4-Poster, Road 10 Old Part, Blauwe Kei) - Handling - Braking BIP content within a Trimmed Body + Subframes Design Variables: Sheet Thicknesses + ADAMS Bushings Generic Responses» Wheel Loads (nominal, min, max)» Full Vehicle properties» Frt Axle Load, Total Weight Stepsteer Responses (Time Domain):» Peak Yaw Rate, Peak LatAcc» Steady-State LatAcc, Sideslip Angle,... 4-Poster Responses (Frequency Domain):» 50 PSDs (each 1 30 Hz) (Tpmnts, Seatrails, FrtBending, FulTors, Engine,...) Total Time To Complete: 2 Days (1000 Runs) Customized HS : Complete In-/Output Matrix Vehicle Dynamics Attribute Requirements In Body / Chassis Lightweight MDO Page 7

8 MDO Approach Meta Modelling HyperStudy DOE Data Mining Clustering CART Correlations Analysis Optimization Validation Page 8

9 MDO Approach Meta Modelling HyperStudy DOE Data Mining Optimization Validation Page 9

10 MDO Approach Meta Modelling HyperStudy DOE Data Mining Optimization Validation Page 10

11 MDO Approach Meta Modelling HyperStudy DOE Data Mining Optimization HyperStudy Validation Page 11

12 2011 European Hyperworks Technology Conference - Bonn Application Example: Body Full Vehicle Ride FV Frequency Responses: Ride Zoom: Full Torsion (1 30 Hz) Data-Mining Design Variables Page 12 Body DVs

13 MDO Toolkit Example: Frequency-Dependent Correlations Data-Mining FV Response Body DVs MDO Toolkit Available To Visualize Structural Effects Against Full Vehicle Metrics Page 13

14 MDO Subcase Extension Meta Modelling & Optimization 1. Total Mass 2. Stat. Stiffness 3. Dyn. Stiffness 4. Modes /Tors. Stiffn. 5. EUNCAP ODB 6. Rear Impact 7. LINCAP + + Meta Models incl. constraint eqs. 8. Vehicle Dynamics / Steering DVs RVs 9. Vehicle Dynamics / Ride Additional VD Loadcases To MDO Process = X-Attribute Trade-Off Studies Page 14

15 Summary Vehicle Dynamics Attribute Significantly Influenced By Structural Stiffness & Weight Lightweight Trend: Potentials & Risks For Vehicle Dynamics Standard Process Inefficient To Systematically Investigate Trade-Offs Customized Hyperstudy Enables X-Attribute MDO Including Vehicle Dynamics From Manual Iterations To Automatic Process From Local EWS To Supercomputer From Competitor Software To Radioss From Single Attribute Optimization To X-Attribute MDO Next Steps Process Refinements Implementation Of Structural Concepts Page 15

16 Contacts Elmar Teipen Dr. Michel Paas +49 (0) (0) Hessel van Dijk Dr. Axel Hänschke +49 (0) (0) Special Thanks Axel Flick, Altair Office Cologne Alexander Koch, Altair Office Cologne Thomas Kowalewski, Altair Office Cologne Arno Klein, P+Z Engineering Matthias Stenkamp, P+Z Engineering Page 16

17 Thank you for your attention!