基于IPSO-BP神经网络的最小阻力船形优化设计  

Minimum Resistance Hull Form Optimization Design Based on IPSO-BP Neural Network

作　　者：侯远杭[1] 刘飞[2] 梁霄[1] 

机构地区：[1]大连海事大学交通运输装备与海洋工程学院,辽宁大连116026 [2]北京四方继保自动化股份有限公司,北京100085

出　　处：《上海交通大学学报》2016年第8期1193-1199,共7页Journal of Shanghai Jiaotong University

基　　金：中央高校基本科研业务费专项资金(3132016066;3132016215;3132016339;3132016358);国家自然科学基金(51579022)资助

摘　　要：针对船形优化设计问题,为了克服主尺度与型线分开设计的不足,将主尺度与表征船体形状的参数联合组成设计空间来进行优化设计,同时以不同航速的船体阻力加权总和作为优化目标,引入IPSO-BP神经网络建立兴波阻力系数的近似预报模型,以主尺度和船形修改系数作为设计变量,以排水体积的变化量作为约束条件,分别利用遗传算法与模拟退火算法对Wigley船形进行了优化计算.计算结果证明了IPSO-BP方法建立兴波阻力系数近似模型的优良性能,得出的光顺合理的优化船形证明了该方法的可行性与合理性.Aimed at hull form optimization design issue,in order to overcome the deficiency of respectively design of principal dimensions and hull lines,the principal dimensions and hull shape parameters were combined to form the design space for optimization design.In the optimization process,the weighted sum of total resistance with different speed,in which the wave-making resistance coefficient was approximately simulated by the IPSO-BP neural network,was proposed as the objective function,with the principal dimensions and parameters of the front half hull modification function as design variables,and variation of displacement as constraint condition.The GA and SA algorithms were both used for the optimization of Wigley hull.The results show that the excellent performance of the IPSO-BP method for the approximate model of the wave resistance coefficient.The optimal hull obtained prove the feasibility and rationality of the method,which can open up a new concept and provide technical support for ship project demonstration and preliminary design.

关 键 词：船形优化 IPSO-BP神经网络 近似精度 优化算法 

分 类 号：U661.1[交通运输工程—船舶及航道工程]