篇名 | Ship Hull Optimization Based on New Neural Network |
---|---|
卷期 | 28:1 |
作者 | Yuan-Hang Hou 、 Xiao-Jing Jiang 、 Xiong-Hua Shi |
頁次 | 137-148 |
關鍵字 | approximate accuracy 、 FRBF neural network 、 hull form optimization 、 PSO algorithm 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 201702 |
DOI | 10.3966/199115592017022801011 |
Pointing at optimization design of hull form based on SBD (simulation based design) technology, a new neural network approximation technique is proposed. First, through using PSO (particle swarm optimization) algorithm training FRBF (flexible radial basis function) neural network weights, PSO-FRBF neural network algorithm is proposed. By comparison and analysis of the wave resistance coefficient of different methods, applicability and superiority of the new algorithm is proved. Then, Wigley hull is taken as example, with the principal dimensions and parameters as design variables, and variation of displacement as constraint condition, the total resistance optimization model is established through introducing PSO-FRBF wave resistance coefficient approximation model. Then the simulated annealing algorithm is used in the ship hull optimal design, and a reliable and reasonable optimized ship hull is obtained. The new neural network can provide fine technical support for related ship optimization design stage.