基于相空间重构技术的舰船摇荡极短期预报  

Extreme short term prediction of ship motions based on phase-space reconstruction

作　　者：蔡烽[1] 万林[1] 石爱国[1] 

机构地区：[1]海军大连舰艇学院

出　　处：《水动力学研究与进展（A辑）》2005年第6期780-784,共5页Chinese Journal of Hydrodynamics

摘　　要：舰船运动受各种非线性因素影响严重,且舰船自身6个自由度的运动间存在相互耦合,在很多情况下表现为混沌。本文提出了基于相空间重构技术的舰船摇荡极短期预报模型,包括加权一阶局域法多步预报模型和RBF神经网络局域法多步预报模型。仿真表明,两种模型均能有效预报舰船摇荡极短期运动。Seriously affected by nonlinear factors and coupling among motions in 6 freedoms, ship motion at sea even appears as chaotic phenomena in many occasions. Based on the chaos analysis of ship motion, two kinds of time series multi-steps forecast model, based on phase-space reconstruction, are proposed including add-weighted one-rank local-region multi-steps forecast model and RBFNN local-region multi-steps forecast model. Simulations show these two methods are effective for chaotic time series prediction, including extreme short-term prediction of ship motions.

关 键 词：相空间重构 混沌 神经网络 极短期预报 舰船摇荡 

分 类 号：U661.321[交通运输工程—船舶及航道工程] TP391.9[交通运输工程—船舶与海洋工程]