基于人工神经网络的船舶操纵性能预报方法  

PREDICTION METHOD FOR SHIP MANEUVERABILITY BASED ON ARTIFICIAL NEURAL NETWORKS

作　　者：范佘明[1] 祝伟敏[1] 李志春 

机构地区：[1]中国船舶及海洋工程设计研究院

出　　处：《中国造船》1999年第4期7-12,共6页Shipbuilding of China

摘　　要：本文以单桨肥大型船回转性能为例，探讨应用改进的ＢＰ神经网络建立船舶操纵性预报 数学模型的方法。研究结果表明，只要选择适当的学习样本、网络结构和学习算法，该数学模 型的预报精度在工程上可以接受。In this paper, the prediction of ship maneuverability based on an improved artificial neural network is studied, and the turning performance of single screw full form ships are taken for an example. It is verified that the prediction accuracy of this method is acceptable in engineering practice, if the learning samples, network's structure and learning algorithm are reasonable.

关 键 词：船舶 操纵性 回转性 人工神经网络 预报 

分 类 号：U661.33[交通运输工程—船舶及航道工程] TP18[交通运输工程—船舶与海洋工程]