Optimal design of variable gradient tube under axial dynamic crushing based on hybrid TSSA–GRNN method

• Published in: Structural and multidisciplinary optimization 2022, Vol.65 (1), Article 11
• Main Authors: Wang, Weiwei, Dai, Shijuan, Zhao, Wanzhong, Wang, Chunyan
• Format: Article
• Language: English
• Subjects: Adaptive search techniques; Computational Mathematics and Numerical Analysis; Crashworthiness; Cross-sections; Energy absorption; Engineering; Engineering Design; Genetic algorithms; Impact strength; Industrial Application Paper; Multiple objective analysis; Neural networks; Optimization; Search algorithms; Sorting algorithms; Statistical analysis; Stress concentration; Theoretical and Applied Mechanics; Thickness; Tubes
• ISSN: 1615-147X; 1615-1488
• DOI: 10.1007/s00158-021-03105-9

Cross-section shape and thickness distribution are essential to the dynamic crashing behaviors of thin-walled tubes. Therefore, this paper designs a novel functional gradient tube with three variable wall thicknesses along the axial direction. The best regular hexagon (RH) cross-section shape is predetermined by comparing with different cross-section tubes. On this basis, six circular fillets are added to each corner of the RH cross-section tube to reduce its stress concentration. Furthermore, to obtain the surrogate models more accurately and effectively, a hybrid TSSA–GRNN method is proposed by combing the adaptive t -distribution sparrow search algorithm (TSSA) and the generalized regression neural network (GRNN). Multi-objective optimization of the variable gradient tube is conducted by integrating the hybrid TSSA–GRNN method and the non-dominated sorting genetic algorithm (NSGA-II). The results show that the energy absorption, crashworthiness, and lightweight of the optimal variable gradient regular hexagon (VG-RH) tube are better than those obtained by the initial counterpart. The VG-RH tube can be recommended as a good absorber in engineering applications.