Multi-objective optimization of sealing structure of subsea pipeline connector based on developed FE model, sensitivity analysis, surrogate model and NSGA-II

• Published in: Journal of the Brazilian Society of Mechanical Sciences and Engineering Vol. 46; no. 2
• Main Authors: Jiao, Kefeng, Yun, Feihong, Hao, Xiaoquan, Wang, Gang, Yao, Shaoming, Jia, Peng, Wang, Xiangyu, Wang, Liquan
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2024; Springer Nature B.V
• Subjects: Artificial neural networks; Axial forces; Connectors; Contact pressure; Cylinders; Elastomers; Engineering; Finite element method; Genetic algorithms; Mathematical models; Mechanical Engineering; Multiple objective analysis; Optimization; Parameters; Performance indices; Pressure effects; Rubber; Sealing; Seals (stoppers); Sensitivity analysis; Sorting algorithms; Technical Paper; Two dimensional models; Underwater pipelines; Underwater structures; Shoulder protrusion variable; Effective variable recognition; Subsea pipeline connector; Rubber cylinder; Structural and material parameter optimization
• ISSN: 1678-5878; 1806-3691
• DOI: 10.1007/s40430-023-04643-7

A 2D axisymmetric finite element method with the penetration load is proposed in this paper to predict the sealing performance and the shoulder protrusion of the sealing structure of subsea pipeline connectors (SPC). The structural parameters and material parameters are optimized by combining the developed finite element model (FE model), sensitivity analysis, surrogate model and non-dominated sorting genetic algorithm (NSGA-II). Firstly, a 2D axisymmetric model considering the material parameters of the rubber cylinder, the condition of the shoulder protrusion, and the pressure effect and penetration effect of the internal fluid on the rubber cylinder is developed as the FE model. The effective design variables are selected through the sensitivity analysis, design variables are reduced from 10 to 6, and 4 artificial neural network models are trained as a surrogate model using the multiple-input single-output method. With the constraint of the shoulder protrusion variable proposed in this paper, the three performance indexes of the axial force and the maximum contact pressures of the inner and outer sides are optimized using the NSGA-II coupled the surrogate model. Pressure tests were carried out with the optimized rubber cylinder, which proved the effectiveness of the proposed method. This study provides a systematic and effective method for the multi-objective optimization of the sealing structure of SPC, especially for the structural parameter and material parameter optimization of elastomeric seals.