【摘要】 船舶波浪载荷预测对船舶结构安全与设计优化至关重要，准确获取波浪载荷面临诸多挑战。本文构建多维度影响因素特征数据集，融合实船航行、水池试验与数值仿真多源数据，经清洗预处理后，明确波高、波浪周期等关键参数构建多维特征向量，并合理划分数据集；对随机森林模型进行优化与水动力约束融合，采用网格搜索、交叉验证结合遗传算法优化算法参数，并评估参数优化效果。针对3类典型船型开展数值仿真与模型验证，研究表明，优化后的随机森林算法预测精度高、计算高效，能为船舶波浪载荷预测提供可靠方案。更多还原

【Abstract】 The prediction of ship wave loads is of vital importance to the structural safety and design optimization of ships. There are many challenges in accurately obtaining wave loads. This paper constructs a multi-dimensional characteristic data set of influencing factors, integrates multi-source data such as real ship navigation, pool tests and numerical simulation. After cleaning and preprocessing, key parameters such as wave height and wave period are identified to construct multidimensional feature vectors, and the data set is reasonably divided. The random forest model is optimized and integrated with hydrodynamic constraints. Grid search, cross-validation combined with genetic algorithm is adopted to optimize the algorithm parameters, and the optimization effect of the parameters is evaluated. Numerical simulation and model verification were carried out for three typical ship types. The research shows that the optimized random forest algorithm has high prediction accuracy and efficient calculation, and can provide a reliable solution for the prediction of ship wave loads.更多还原