검색 상세

Numerical Study of Sheet Metal Deep Drawing Process by Artificial Intelligence Multi-Objective Optimization Strategy

인공지능 다목적 최적화 전략에 의한 판재 딥드로잉의 수치적인 연구

  • Feng Guo (곽풍, 서강대학교 일반대학원)

초록 (요약문)

In sheet metal forming, persistent challenges such as failure, wrinkling, and springback necessitate innovative solutions. This study elucidates a novel methodology tailored to mitigate these prevalent defects. Our primary aim was to bolster formability while concurrently diminishing defects throughout the forming cycle. To achieve this goal, we amalgamated deep neural network, genetic algorithm, and Monte Carlo simulation techniques, collectively denoted as DNN-GA-MCS. In our quest to optimize process parameters, encompassing the S-VBHF, friction coefficient, and drawbead restraining force (DBRF). Especially, we deployed a segmented and variable blank holder force (S-VBHF) trajectory, facilitating precise adjustments to the blank holder force (BHF). In addition, we also try to optimize the process variables and sheet geometry variables at the same time. The Forming Limit Diagram (FLD) predicted by machine learning algorithms based on GISSMO damage model provided a comprehensive assessment for evaluating sheet failure dynamics during the forming process. Our proposed methodology underwent stringent validation via numerical simulations, with the several FEA models providing the benchmark. Empirical outcomes underscored substantial enhancements in formed sheet quality: marked reductions of 8.33% in failure, 10.81% in wrinkling, and 5.88% in springback. In summation, our advanced methodology manifests potential in refining sheet metal forming processes, emphasizing its efficacy in substantially curtailing defects.

more

목차

Chapter 1 Introduction 1
1.1 Background 1
1.2 Literature review 6
1.3 Research motivations 12
Chapter 2 Artificial intelligence modelling training sets construction 15
2.1 Material properties 15
2.1.1 Experimental implementation 16
2.1.2 Yld2004-18p yield criterion 17
2.2 Finite element analysis models and design variables 19
2.2.1 Cylindrical cup FEA model 19
2.2.2 Oil pan FEA model 21
2.3 Training set sampling and validation 24
2.3.1 Determining training sampling points by LHD method 24
2.3.2 Training set k-fold cross-validation 26
2.4 Machine Learning Algorithm for Prediction of FLD 27
2.4.1 GISSMO damage model 28
2.4.2 FLD predictions related fracture experiment 30
2.4.3 Fracture characteristic of triaxiality curve 31
2.4.4 DNN model for predicting the FLD 33
Chapter 3 Optimal processing parameter design methodology 36
3.1 Design optimization problem description 36
3.2 Objective functions definition 37
3.3 Training data sets pre-processing 39
Chapter 4 Artificial intelligence-driven multi-objective optimization 41
4.1 Artificial intelligence modelling 43
4.1.1 Deep neural network (DNN) model 43
4.1.2 Artificial intelligence modelling validation 44
4.2 Multi-objective Optimization Algorithm based on NSGA-II 45
4.3 Stochastic analysis Technology based on MCS 46
Chapter 5 Results 48
5.1 Artificial intelligence (AI) modelling result 50
5.1.1 AI modelling validation result 50
5.1.2 AI modelling approximation result 52
5.1.3 RSM modelling sensitivity 56
5.2 Multi-objective Optimization Algorithm result 59
5.3 Stochastic analysis technology result 60
5.3.1 Process parameter importance analysis result 61
5.3.2 Process parameter correlation analysis result 63
5.4 Numerical simulation result 66
5.4.1 Cylindrical cup model result 66
5.4.2 Oil pan model result 69
5.5 Material flow variation analysis results 73
5.6 Comparison results of advanced technologies 80
Chapter 6 Application and Discussion 82
6.1 Application of different FEA model 82
6.2 NSM Stamp-param OPT variables optimization package 87
6.3 Discussion 89
Chapter 7 Conclusion 92
7.1 Summary 92
7.2 Innovation 94
7.3 Future work 95
APPENDEX A. Cylindrical cup FEA model Design of experiment 97
References 98

more
반출 Meta View 목록