| 摘要: |
| 内孤立波在海洋中的传播会携带能量和动量，不同振幅的内孤立波对海洋中的能量交换及海上工程等影响也不同，因此，研究内孤立波振幅与半波宽度、水深、分层条件、密度等水文特征参量之间的关系显得尤为重要。以往在研究中建立内孤立波振幅与它们之间的关系时，会受到不同理论有效适用范围的限制。本文借助实验室的水槽方法，设计了不同的水深、分层及密度条件下的内孤立波系列综合实验，发现内孤立波的振幅与半波宽度、水深、分层条件以及水体密度等参量之间并非简单线性关系。因此，利用机器学习的方法建立内孤立波振幅与上述参量之间的非线性关系，建立了支持向量机（SVM）和随机森林（RF）两种机器学习模型。将1266组实验数据建立样本库，其中包含训练集970组，测试集296组，对模型进行参数调优，最终通过测试集验证，SVM模型的平均相对误差为17.3％，RF模型的平均相对误差为15.5％。该方法适用于多种不同的水文条件，有效解决先前理论存在的适用性问题。 |
| 关键词: 内孤立波， 振幅， 水槽实验， 支持向量机， 随机森林 |
| DOI： |
| 分类号: |
| 基金项目:国家科技攻关计划,国家自然科学基金项目（面上项目，重点项目，重大项目） |
|
| Research on the Relationship between Wave Elements of Internal Solitary Waves Based on Machine Learning |
|
li zhi xin, wang jing
|
|
Ocean University of China
|
| Abstract: |
| The internal solitary waves propagate in the ocean carry huge energy and momentum. The influence of internal solitary waves with different amplitudes on the energy exchange and offshore engineering in the ocean is also different. Therefore, it is particularly important to study the relationship among the amplitude of the internal solitary waves, the half-wave width and the hydrological characteristic parameters such as depth, stratification, and density. In the past, the relationship between the amplitude of internal solitary wave and these parameters was limited by different theories. In this paper, with the help of flume in the laboratory, a series of comprehensive experiments under different depth, stratification and density were designed. It is found that the relationship among the amplitude of internal solitary waves, half wave width, depth, stratification and density is not a simple linear. Therefore, using machine learning method to establish the nonlinear relationship between the amplitude of internal solitary wave and the above parameters. We established two models including support vector machine (SVM) and random forest (RF), and built a sample database of 1266 sets including 970 training sets and 296 test sets. The parameters of the model have been optimized. Finally, the average relative error of the SVM model is 17.3 and that of the RF model is 15.5%, respectively. The results shows that the machine learning method is effective and feasible. This method can be applied to a variety of different hydrological conditions, which effectively solving applicability issues in the previous theory. |
| Key words: Internal solitary wave, amplitude, flume experiment, Support Vector Machine, Random Forest |