Optimal design of preform shape based on EFA-FEM-GA integrated methodology

• Published in: International journal of material forming 2021-09, Vol.14 (5), p.1043-1056
• Main Authors: Liu, Chengshang, Xu, Wujiao, Wang, Yu, Liu, Minyao
• Format: Article
• Language: English
• Subjects: CAE) and Design; Computational Intelligence; Computer simulation; Computer-Aided Engineering (CAD; Design optimization; Design parameters; Engineering; Finite element analysis; Finite element method; Forging; Forgings; Fourier analysis; Genetic algorithms; Machines; Manufacturing; Materials Science; Mechanical Engineering; Methodology; Modules; Original Research; Processes; Scaling factors; Shape factor; Shape optimization; Three dimensional analysis; Two dimensional analysis
• ISSN: 1960-6206; 1960-6214
• DOI: 10.1007/s12289-021-01620-0

Preform shape design for complex forging is an important and intractable aspect in the design of forging process. This paper presents an integrated methodology based on elliptic Fourier analysis (EFA), finite element method (FEM) and genetic algorithms (GA) to determine optimal preform shape. Firstly, an elliptic Fourier analysis, which has the advantages of wide versatility and short design cycle, is adopted as the transformation rule for the first time in this paper. The similarity between elliptic Fourier analysis and plastic forming process is demonstrated theoretically. Meanwhile, the main steps for using elliptic Fourier analysis module to design preform shape are introduced: two-dimensional slice, elliptic Fourier analysis and three-dimensional reconstruction. The preform shapes generated by elliptic Fourier analysis module can be simulated directly in finite element method module. Next, in order to control the deformation amount and material distribution of preform, two sets of design parameters, i.e., shape factor and triaxial scaling factors, are employed to control the preform shape before entering the finite element method module simulation. Then, these design parameters are optimized using the genetic algorithm module. Finally, taking a heavy forging with complex shapes as an example, its optimized design scheme is carried out in real forging production. The results show that the forging is produced without problems of crack, folds, underfilling and unreasonable flash distribution, which validates the effectiveness of the presented methodology. Furthermore, this integrated methodology could be extended to other complex forgings.