Ablaufautomatisierung zur virtuellen Optimierung von Blechhalterkraftverläufen

Transcription

1 Ablaufautomatisierung zur virtuellen Optimierung von Blechhalterkraftverläufen Procedure for Automated Virtual Optimization of Variable Blank Holder Force Distributions in Deep-Drawing Processes K. Wurster M. Liewald C. Blaich M. Mihm 8. Weimarer Optimierungs- und Stochastiktage WOST 8.0 Dorint Hotel, Weimar / Germany November 24th - 25th 2011

2 Content Deep-Drawing of Sheet Metal Car Body Components Deep-Drawing Die and Component Geometry Automated Detection of Sidewall Wrinkling in FEA Optimization of Blank Holder Force Distribution Summary and Outlook 2

3 Challenges in Deep-Drawing Processes Fender shaped test geometry Source: Mercedes-Benz Source: IFU, Blaich Car body: about 350 individual sheet metal components Multi-stage manufacturing process drawing dies require huge investments Virtual methods of tool-design for deep-drawing processes gain in importance today 3

4 Deep-Drawing Process die cavity blank holder punch slide plate pressure pistons bottom part of the tool Source: IFU, Schuler 4

5 Failure of Deep-Drawn Components Wrinkles of 1 st order in the flange area Wrinkles of 2 nd order in the sidewall area Surface defects Necking Cracks Spring-back Wrinkles and cracks restrict process window in deep-drawing Quality of deep-drawn components can be influenced by material flow Material flow can be influenced e.g. by blank holder force Source: IFU, Blaich, Beck 5

6 Content Deep-Drawing of sheet metal car body components Deep-Drawing Die and Component Geometry Automated Detection of Sidewall Wrinkling in FEA Optimization of Blank Holder Force Distribution Summary and Outlook 6

7 Deep-Drawing Die Set with Segment-Elastic Blank Holder matrice punch segment-elastic blank holder hydraulik counter pressure piston Source: IFU, Häussermann 7

8 Fender Shaped Component Geometry Sheet material: DC04 4 different corner radii 4 tapered walls 4 different wall lengths in circumferential direction Sheet thickness: 0.9 mm Draw depth: 100 mm Part shape is susceptible to reveal failure of sidewall wrinkling Source: IFU, Häussermann 8

9 Content Deep-Drawing of sheet metal car body components Deep-Drawing Die and Component Geometry Automated Detection of Sidewall Wrinkling in FEA Optimization of Blank Holder Force Distribution Summary and Outlook 9

10 Failure Detection in FEA Detection of cracks in FEA Challenges in sidewall wrinkling detection in FEA Thickness reduction Reliability, clearness Suitability for automated Distance to forming limit evaluation curve (FLC) in forming limit diagram (FLD) criterion for automated sidewall wrinkling detection is required 10

11 Wrinkling Criteria in FEA Identified wrinkling criteria base on Strain distribution in forming limit diagram Amount of local defomation energy Deviation of geometry, spatial inaccurateness Second principal stress 36 About 20 wrinkling criteria investigated at IFU

12 Wrinkling Criterion Based on Forming Limit Diagram - Procedure Testing for Conical Cup FE-calculations (LS-Dyna) Read pairs of variates of strain distribution in Postprocessing (LSPP) Calculateφ 1 / φ 2 ratios (MS Excel) Sortφ 1 / φ 2 ratios by value (MS Excel) wrinkle-free conical cup: conical cup with wrinkles: 12 major true strainφ log. Hauptformänderung 1 φ 1 sidewall wrinkling Falten sidewall 2. Art wrinkling log. Nebenformänderung φ2 minor true strainφ 2

13 Wrinkling Criterion Based on FLD Testing of Fender Shaped Geometry wrinkle-free area Sidewall wrinkling detection Reliability wrinkling limit curve gradient wrinkling area Clearness Gradient of wrinkling limit curve is definite value for optimization Qualifications for automated sidewall wrinkling detection fulfilled 13

14 Automated Detection of Sidewall Wrinkling FE-Model Batch Processing FE-Solver Command-File MS-Excel LSPP (Postprocessing) Simulation results + additionally generated results F optislang Wrinkling criterion sidewall wrinkling Problem specification file Objective function and constraints Sensitivity analysis Robustness evaluation Optimization (ARSM) 14

15 Content Deep-Drawing of sheet metal car body components Deep-Drawing Die and Component Geometry Automated Detection of Sidewall Wrinkling in FEA Optimization of Blank Holder Force Distribution Summary and Outlook 15

16 Optimization of Blank Holder Force Distribution - Overview rigid blank holder one single blank holder force blank holder force variable in space [kn] Blank holder force variable in space and time 16

17 Optimization of Blank Holder Force Variable in Space Initial Situation and Procedure Initial situation: Maximum thinning: 18.55% Blank holder force in each segment: 140 kn Sidewall wrinkling Optimization parameters: 10 blank holder forces (one force / segment) Target function: Minimization of thinning Constraints: Avoidance of sidewall wrinkles: φ1 / φ2 ratio < Procedure: Adaptive Response Surface Method (ARSM) 17

18 Optimization of Blank Holder Force Variable in Space - Results blank holder force variable in space [kn] Blank holder forces vary from 80.1 kn to kn No sidewall wrinkling Maximum thinning 14.97% Thinning reduced by 19% maximum thinning BADmax [%] Wrinkling limit curve 18

19 Optimization of Blank Holder Force Variable in Time Procedure Initial situation: 2100 kn Optimization result blank holder force variable in space 700 kn Optimization parameters: Tolerance margin method: 40 support points of blank holder forces in 4 significant blank holder segments (10 support points / segment) Procedure: Evolutionary Algorithm (E/A) Target function: Minimization of thickness reduction Constraints: Avoidance of sidewall wrinkles: φ1 / φ2 ratio <

20 Optimization of Blank Holder Force Variable in Time - Results maximum thinning BADmax [%] Blank holder forces vary from 80.1 kn to kn No sidewall wrinkling Maximum thinning 14.51% Thinning reduced by another 2% Wrinkling limit curve 20

21 Content Deep-Drawing of Sheet Metal Car Body Components Deep-Drawing Die and Component Geometry Automated Detection of Sidewall Wrinkling in FEA Optimization of Blank Holder Force Distribution Summary and Outlook 21

22 Summary Procedure for automated detection of sidewall wrinkling in FEA based on FLD developed and tested at IFU Gradient of wrinkling limit curve for mild steel DC04, 0.9 mm, h = 100 mm identified for conical cup (-1.22) and fender shaped geometry (-1.55) Implementation of tolerance margin method in FEA for optimization of blank holder force variable in space and time Optimization of blank holder force distribution of segment-elastic blank holder reduces maximum thinning by 21% Visual investigation of optimization results no longer necessary 22

23 Outlook Robustness analysis of optimization results Determination of safety margin for gradient of wrinkling limit curve within robustness analysis Integration of optimization of initial blank size Integration of optimization of heating zone for aluminum shock heat treatment method 23

24 2100 kn 700 kn Thank You Very Much For Your Attention! 24