Machine Learning-Based Multi-Objective Prestress Optimization Framework of Suspend Dome Structure and Case Study

Mingliang Zhu,Jin Wang,Xiangchen Hu,Shilin Dong
DOI: https://doi.org/10.1016/j.engstruct.2024.118987
IF: 5.582
2025-01-01
Engineering Structures
Abstract:Intelligent optimization algorithms are widely employed for the prestress optimization of suspend domes to determine the optimal cable forces. This process involves numerous complex iterative calculations, with each iteration including a corresponding finite element analysis. Such a process is both cumbersome and time-consuming. In this study, a machine learning-based surrogate model integrated with an intelligent optimization algorithm was developed to address the prestress optimization problem in suspend domes. This optimization problem incorporates multiple objective constraints, including maximum vertical displacement, maximum lateral displacement of the support, uniformity of the cable force, and maximum cable force value. The efficiency and effectiveness of the proposed method were demonstrated through two case studies. Results show that the proposed method reduced computational time by 95 % and ensured convergence by effectively integrating the surrogate model with a multi-objective optimization strategy.
What problem does this paper attempt to address?