Identification modeling of ship maneuvering motion based on physics-informed neural network

This dataset contains the simulation data used in the study: “Identification modeling of ship maneuvering motion based on physics-informed neural network (PINN)”. The dataset includes simulated maneuvering motion data of the KVLCC2 ship used for training and validating the PINN model. The simulation data were generated using the Abkowitz-type ship maneuvering model with hydrodynamic derivatives obtained from literature. The dataset contains the state variables and control inputs used for system identification, including surge velocity (u), sway velocity (v), yaw rate (r), rudder angle (δ), and the corresponding accelerations. The simulation data include three maneuvering conditions: • 35°turning circle maneuver • 10°/10°zigzag maneuver • 20°/20°zigzag maneuver Each maneuver was simulated for 1000 s with a sampling interval of 1 s, resulting in a total of 3000 samples used for training. The experimental validation data used in the paper are obtained from the SIMMAN 2008 workshop database (http://www.simman2008.dk/) and are not included in this repository. Interested readers should obtain these data directly from the official SIMMAN database. These data are provided to support the reproducibility of the numerical experiments reported in the paper.本数据集用于支撑题为「基于物理知情神经网络（Physics-Informed Neural Network, PINN）」的船舶操纵运动辨识建模研究，包含相关仿真数据。 数据集包含用于训练与验证PINN模型的KVLCC2型船舶的仿真操纵运动数据。该仿真数据基于阿博科茨（Abkowitz）型船舶操纵模型生成，其水动力导数取自公开学术文献。 数据集涵盖用于系统辨识的状态变量与控制输入数据，包括纵荡速度（u）、横荡速度（v）、转首角速度（r）、舵角（δ）以及对应的加速度项。 仿真数据包含三类操纵工况： • 35°回转圈操纵试验 • 10°/10°之字操纵试验 • 20°/20°之字操纵试验 每类操纵场景均进行了1000秒的仿真，采样间隔为1秒，总计生成3000组用于训练的样本数据。 论文中使用的实验验证数据取自SIMMAN 2008研讨会数据库（http://www.simman2008.dk/），未包含于本数据集仓库中。感兴趣的读者可直接从SIMMAN官方数据库获取该类数据。 本数据集的发布旨在支撑论文中数值试验的可复现性研究。