報告人簡介
袁春鑫博士，本科畢業(yè)于中國海洋大學海洋科學專業(yè)，博士畢業(yè)于University College London數(shù)學系，目前就職于中國海洋大學數(shù)學科學學院、海洋數(shù)學技術(shù)聯(lián)合實驗室。主要從事非線性波動、海洋內(nèi)波、海洋渦旋與環(huán)流的研究，側(cè)重于數(shù)學理論與物理海洋學的交叉，其研究成果以第一作者身份發(fā)表在Journal of Fluid Mechanics、Journal of Physical Oceanography等國際權(quán)威期刊上。
內(nèi)容簡介
Over the last few decades,internal solitary waves in the ocean have garnered increased attention, owing to advances in understanding its great impact on oceanic energy cascade, climate system, marine engineering, underwater navigation and ocean ecological systems. Nevertheless, the primary control equation- the Navier-Stokes equation- is difficult to conduct analysis and to obtain the insight into their dynamics. Thus, we usually derive the reduced model based on physical laws and asymptotic method. In this presentation, we would like to present some work on the derivation and application of reduced model for internal solitary waves. AI method get more involved in the research of internal solitary waves and the fact is that pure data-driven Neural Network is usually difficult to obtain satisfactory results, the reason can be attributed to the lack of sufficient training data due to the difficultness of getting observational data of vertical structure and to the complexness of evolution of internal solitary waves which are in three spatial dimensions and one temporal dimension. Thus, one way to promote the performance of Neural Network is to get the aforementioned reduced models involved.
