Journal of Shanghai Jiao Tong University ›› 2021, Vol. 55 ›› Issue (1): 11-20.doi: 10.16183/j.cnki.jsjtu.2019.087

Special Issue: 《上海交通大学学报》2021年12期专题汇总专辑 《上海交通大学学报》2021年“海洋科学与工程”专题

Previous Articles     Next Articles

Multi-Objective Optimization of Three-Column Semi-Submersible Platforms Based on Surrogate Models

QIU Wenzhen, SONG Xingyu, ZHANG Xinshu()   

  1. State Key Laboratory of Ocean Engineering; Collaborative Innovation Center for Advanced Ship and Deep-Sea Exploration, Shanghai Jiao Tong University, Shanghai 200240, China
  • Received:2019-03-28 Online:2021-01-01 Published:2021-01-19
  • Contact: ZHANG Xinshu


In the initial design stage of a semi-submersible platform, the main particulars of the platform are the key factor affecting the hydrodynamic performance and construction cost. Therefore, multi-objective optimization of the main particulars of the semi-submersible platform is of great engineering significance. First, the design variables of each platform and sample database are determined by design of experiments. Then, the hydrodynamic performances of the semi-submersible platform are analyzed by using the panel method and Morison’s equation. The distribution of probes for estimating the wave elevations on the calm water surface is arranged, and the airgap can be computed. Based on the database obtained by numerical simulation, the surrogate models based on radial basis function (RBF) are established. Next, the formal parameters in RBF are obtained by using the leave-one-out cross validation method. The surrogate model can greatly improve the optimization efficiency. Finally, by using the multi-objective particle swarm optimization (MOPSO) method, taking safety and economy of offshore platforms as two optimization objectives, and taking platform stability, airgap and horizontal motion performance as constraints, the optimization program for the semi-submersible platform can be obtained. Through the detailed analyses of the optimization program for the semi-submersible platform, the most efficient design strategy for the three-column semi-submersible platform is proposed.

Key words: multi-objective particle swarm optimization (MOPSO), surrogate model, leave-one-out cross validation, radial basis function

CLC Number: