节点文献
应用人工神经网络理论的船舶运动包络预报方法
REAL TIME PREDICTION OF ENVELOPE OF SHIP MOTION USING NEURAL NETWORK
【摘要】 本文首先介绍了人工神经网络理论的基本内容以及前馈网络反向传播的训练方法,然后应用单隐含层网络作出船舶运动包络的实时预报。文中就一船舶的海上实测运动数据进行了实例计算,结果表明,本文方法的预报结果明显好于其它模型的自适应预报结果。
【Abstract】 The basic theory of neural network and the training method of back propagation for the feedforward net are described and the neuralnet with single hidden layer is used for real time prediction of the envelope of ship motions. As an example, the full scale duta of pitching motion of a ship are utilized for practical calculation. The results predicted by the neural network are obviously better than those by the other methods.
【关键词】 船舶运动;
人工神经网络;
包络;
实时预报;
【Key words】 Ship Motion; Neural network; Envelop; Real time prediction;
【Key words】 Ship Motion; Neural network; Envelop; Real time prediction;
- 【文献出处】 中国造船 ,Shipbuilding of China , 编辑部邮箱 ,1994年03期
- 【分类号】U675
- 【被引频次】8
- 【下载频次】89