论文标题
部分可观测时空混沌系统的无模型预测
Development of a Hybrid Simulation and Experiment Test Platform for Dynamic Positioning Vessels
论文作者
论文摘要
储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。
The harsh ocean environment and complex operating condition require high dynamic positioning (DP) capability of offshore vessel. The design, development and performance evaluation of DP system are generally carried out by numerical simulations or scale model experiments. Compared with the time-consuming and laborious experiment, the simulation is convenient and low cost, but its results lack practical reference due to oversimplification of the model. Therefore, this paper presents a hybrid simulation and experiment test platform for DP vessels. Its characteristics are: the realistic calculation of environmental loads and motion response, the consistency of algorithms and parameters for simulation and experiment greatly shortening the time of experiment adjusting, switchable and online renewable controller facilitating algorithm testing. The test platform can test the performance of DP system and determine the operational time window. In the hydrodynamic simulation, the six degree-of-freedom model is used to describe the dynamic response of the DP vessel, considering the fluid memory effect and frequency-dependent hydrodynamic parameters. In the experiment, the similarity theory based on the same Froude number is used to ensure the consistency of control parameters with simulation. Finally, a case study of DP shuttle tanker is used to verify the credibility of the test platform.