Using Generalized Finite Difference Method and Parallel Computation to Build a Three-Dimensional Numerical Wave Flume

瀏覽統計 Email 通知 RSS Feed

簡歷

基本資料

Project title

Code/計畫編號

MOST106-2221-E019-036

Translated Name/計畫中文名

以廣義有限差分法與平行計算建立三維數值波浪水槽

Project Coordinator/計畫主持人

Chia-Ming Fan

Funding Organization/主管機關

National Science and Technology Council

Department/Unit

Department of Harbor and River Engineering

Website

https://www.grb.gov.tw/search/planDetail?id=12245336

Year

2017

Start date/計畫起

01-08-2017

Expected Completion/計畫迄

31-07-2018

Bugetid/研究經費

483千元

ResearchField/研究領域

土木水利工程

"本研究結合最新的無網格數值計算方法與繪圖卡平行計算，建立一個簡單、準確且高效率之三維數值波浪水槽，以研究三維非線性波浪傳遞問題，以及波浪與結構物互制等問題。本計畫內容為延伸前兩年計畫之研究成果，目的在建立真實三維平行計算數值波浪水槽，以提供給海洋防災工程規劃設計使用。廣義有限差分法是一種新的無網格法，可以節省網格建置與數值積分的工作，能大幅減少計算時間與人力，兼具傳統網格法與無網格法的優點，非常適合作為三維數值波浪水槽的空間離散方法；而本研究所採用的時間離散方法為二階Runge-Kutta 法，能兼具時間積分時之穩定性與高效率。除了廣義有限差分法與Runge-Kutta 法之外，本計畫採用繪圖卡平行計算方法將模式平行化，以增加電腦模擬效率，繪圖卡平行計算方法在近年來非常快速的發展與進步，可以在單一台電腦上完成數百顆計算核心同時模擬的效率，能大幅提高三維數值波浪水槽的電腦模擬效率。本計畫將以數個案例進行模式驗證與平行效率評估，例如：波浪與潛堤互制問題、波浪與洋流互制問題、海底滑坡導致海嘯產生與變形等問題。本研究所開發之電腦模擬模式與模擬結果可以提供給未來波浪問題與海洋工程相關研究參考使用。""In this project, a simple, accurate and efficient three-dimensional numerical wave flume will be constructed by using the generalized finite difference method (GFDM) and graphic processing unit (GPU)-based parallel computation. The three-dimensional numerical wave flume, developing in this project, will be very useful and essential to engineering applications and fundamental study in ocean engineering. The content of this project is an extension of our projects in previous two years, so the GFDM, second-order Runge-Kutta method and GPU parallel computation will be combined to form a highly-efficient three-dimensional numerical model for studying nonlinear waves propagation. The GFDM is a newly-developed meshless method, so the time-consuming tasks of mesh generation and numerical quadrature can be truly avoided. The GFDM remains the advantages from both of mesh-based numerical methods and meshless (meshfree) methods; therefore, the GFDM is adopted in this project for the spatial discretization of the numerical wave flume. Meanwhile, the second-order Runge-Kutta method, which is stable and accurate in time integration, is used for temporal discretization. Except for the GFDM and the second-order Runge-Kutta method, the GPU parallel computation is also included in this numerical wave flume to parallelize the computer codes and to enhance the efficiency of numerical simulation. The software and hardware of GPU parallel computation have been rapidly developed in the recent years and hundreds of computing cores can be implemented at the same time in a desktop. So, the computational efficiency of the three-dimensional numerical wave flume can be greatly improved by using GPU parallel computation. In this project, several three-dimensional numerical examples will be adopted to validate the accuracy, simplicity, stability and parallel efficiency of the developed model, such as the interactions between nonlinear waves and submerged breakwaters, the interactions between nonlinear waves and currents, the propagation of tsunami caused by underwater landslides, etc. The parallelized three-dimensional numerical wave flume and numerical results can be provided to future relevant study in ocean engineering."

Keyword(s)

三維數值波浪水槽
非線性波浪
廣義有限差分法
繪圖卡平行計算方法
three-dimensional numerical wave flume
nonlinear waves
generalized finite difference method
GPU parallel computation

DSpace CRIS

Using Generalized Finite Difference Method and Parallel Computation to Build a Three-Dimensional Numerical Wave Flume

基本資料

Description