Extrapolation of random wave field data via compressive sampling
Extrapolation of random wave field data via compressive sampling
复制标题
DOI：
10.1016/j.oceaneng.2018.03.044
复制
发表时间：
2018-06
影响因子：
5
通讯作者：
G. Malara;I. Kougioumtzoglou;F. Arena
中科院分区：
文献类型：
--
作者：
G. Malara;I. Kougioumtzoglou;F. Arena
估计海况的时空特性对于一些工程应用是至关重要的，例如涉及三维波浪与海洋结构物相互作用的应用。在这种情况下，开发一种允许仅使用相对较少的记录来推断关于波场的信息的技术是非常有影响力的，因为它允许最大限度地减少昂贵和复杂的测量技术的使用。提出了一种基于压缩抽样(CS)的自由面位移数据外推方法。该技术依赖于方向谱兼容的稀疏表示，并结合表达和求解L1范数优化问题。此外，通过使用自适应的基重加权过程，所开发的技术的精度显著提高。数值算例表明，该方法能够较好地再现自由面位移记录的时程，同时较好地捕捉了目标频谱和互相关函数的主要特征。
Estimating the space-time characteristics of a sea state is of crucial importance to a number of engineering applications, such as the ones involving three-dimensional waves interacting with marine structures. In this context, developing a technique that allows extrapolating information about the wave field utilizing only a relatively small number of records is highly impactful, as it allows minimizing the use of expensive and sophisticated measurement techniques. In this paper, a Compressive Sampling (CS) based technique is developed for extrapolating free surface displacement data. The technique relies on a directional spectrum compatible sparse representation in conjunction with formulating and solving an L1-norm optimization problem. Further, the accuracy of the developed technique is significantly enhanced via the use of an adaptive basis re-weighting procedure. Pertinent numerical examples demonstrate that the technique is capable of reconstructing the time history of a free surface displacement record successfully, while capturing the main features of the target frequency spectrum and of the cross-correlation function satisfactorily.