【作者】 朱仁传； 杨云涛； 黄山； 查乐； 尚凡成；

【Author】 ZHU Ren-chuan;YANG Yun-tao;HUANG Shan;ZHA Le;SHANG Fan-chen;State Key Laboratory of Ocean Engineering,School of Naval Architecture,Ocean and Civil Engineering,Shanghai Jiao Tong University;China Ship Scientific Research Center;

【机构】 上海交通大学船舶海洋与建筑工程学院海洋工程国家重点实验室； 中国船舶科学研究中心；

【摘要】 频域势流理论一直是船舶运动计算和工程预报的重要基础。采用虚构的控制面将船体周围的流域分割为内域和外域,外域内域分别采用移动脉动源和兰金源格林函数构建边界积分方程,耦合联立边界条件建立了求解有航速船舶运动辐射绕射问题的频域势流方法。基于此计算了母型船及其一系列变形船体的水动力系数,通过机器学习建立了可供初步设计使用的水动力系数预报模型。计算研究表明区域分割法精度高,船型适用广;机器学习建立的预报模型,高效稳定,能有效支撑船舶初步设计。更多还原

【Abstract】 The fluid domain around ship is decomposed into inner and outer domain by an artificial control surface.In the inner domain Rankine source is adopted while in the outer domain translating and pulsating source is applied,which guarantees linear free surface condition and radiation condition automatically.On the control surface,the continuity conditions of velocity potential and its normal derivative are imposed to solve the boundary integral equations in a coupled way.Based on this method hydrodynamic coefficients of series data ships obtained by affine transforming from a mother type ship.Machine learning is used to model ship hydrodynamic and ship type by the data training.A predict model is obtained and verified by comparing with numerical calculation.It shows both the present hydrodynamic methods by numerical calculation and modeled by machine learning are stable with high precision.The latter and the former are fit for initial ship design and detailed design respectively.更多还原