A fast GA-ANN model and application in multi-objective optimization of the sealing ring for the subsea pipeline connector with regard of the penetration load

• Published in: Journal of mechanical science and technology Vol. 38; no. 1; pp. 309 - 322
• Main Authors: Jiao, Kefeng, Yun, Feihong, Hao, Xiaoquan, Wang, Gang, Yao, Shaoming, Jia, Peng, Wang, Xiangyu, Wang, Liquan
• Format: Journal Article
• Language: English
• Published: Seoul Korean Society of Mechanical Engineers 2024; Springer Nature B.V; 대한기계학회
• Subjects: Contact pressure; Control; Dynamical Systems; Engineering; Finite element method; Hydrostatic pressure; Industrial and Production Engineering; Mechanical Engineering; Multiple objective analysis; Optimization; Original Article; Response surface methodology; Sealing; Underwater pipelines; Underwater structures; Vibration; 기계공학; Sealing performance; NSGA-II; Artificial neural network; Genetic algorithm; RSM; Structure optimization
• ISSN: 1738-494X; 1976-3824
• DOI: 10.1007/s12206-023-1225-8

In this paper, a finite element method is proposed to predict the contact pressure of the metal seal in the subsea pipeline connector with regard of the penetration load on the critical sealing surface. Further, a fast GA-ANN model is proposed on the basis of GA-ANN to optimize the sealing structure of the connector, which uses only 3–6 % of the calculation time compared to the traditional GA-ANN model. The fast GA-ANN is further coupled with NSGA-II in the multiple objective optimization of the sealing structure and the results are compared with that of the response surface methodology (RSM). In terms of the number of valid candidate points and the optimal candidate point, NSGA-II coupled fast GA-ANN model performs much better than RSM. The hydrostatic pressure tests were carried out with the optimal sealing structure by the fast GA-ANN and the results meet the design requirements very well.