Multi-objective optimization of semi-submersible platforms using particle swam optimization algorithm based on surrogate model. (15th April 2019)

• Record Type: Journal Article
• Authors: Qiu, Wenzhen; Song, Xingyu; Shi, Kaiyuan; Zhang, Xinshu; Yuan, Zhiming; You, Yunxiang
• Is Part Of: Ocean engineering. Volume 178(2019)
• Page Start: 388
• Page End: 409
• Publication Date: 2019-04-15
• Subjects: Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2019.02.039
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: An Innovative Semi-submersible platform Optimization Program (ISOP) has been developed to solve the multi-objective optimization problem for semi-submersible platforms (SEMI). Three types of SEMIs, including semi-submersible floating production unit (SEMI FPU), heave and vortex induced motion (VIM) suppressed semi-submersible (HVS) and semi-submersible floating drilling unit (SEMI FDU) are selected for case studies. The hydrodynamic performances of three types of semi-submersible platforms are analyzed by using panel method and Morison's equation. In order to improve the computing efficiency, the hydrodynamic performances for different hull forms during optimization process are estimated by the surrogate models, which are built by artificial neural network prediction method and Inverse Multi-Quadric (IMQ) radial basis function (RBF). The accuracy of surrogate models is ensured by performing leave-one-out cross validation (LOOCV). The most probable maximum (MPM) heave motion and total weight, representing the safety and economy, respectively, are chosen as the two objectives for optimization. The transverse metacentric height, the MPM surge motion, and the most probable minimum (MPMin) airgap are selected as constraints. Based on surrogate models, multi-objective particle swarm optimization (MOPSO) is employed to search for the Pareto-optimal solutions. A Computational Fluid Dynamics (CFD) tool is adopted to validate the proposed model for the prediction of the motion responses. By comparing the obtained Pareto-optimal solutions with the initial design using simple panel method plus Morison's equation, it is confirmed that the MPM heave motions for SEMI FPU, HVS and SEMI FDU can be suppressed by up to 12.68%, 11.92%, and 14.96%, respectively, and the total weights can be reduced by up to 12.16%, 13.00%, and 24.91%, respectively. Through the detailed analyses of optimization results, the most efficient design strategies for semi-submersible platforms are discussed and proposed. Highlights: Surrogate models are adopted to evaluate the stability, motion responses and airgap for semi-submersible platforms. The multi-objective particle swarm optimization is used for the multi-objective optimization of SEMIs. The most efficient design strategies for three types of semi-submersible platforms are proposed and discussed.