November 22, 2024
Persian Gulf University
فارسی
Rouhollah Fatehi
Academic Rank:
Associate professor
Address:
School of Engineering
Degree:
Ph.D in Mechanical Engineering
Phone:
07731222170
Faculty:
Faculty of Engineering
E-mail:
fatehi [at] pgu [dot] ac [dot] ir
Home
Education
Research ineterests
Research activities
Courses
Executive activities
Links
Image Gallery
Research
Title
Improving the Smoothed Particle Hydrodynamics Method to Simulate the Wave Maker Problem
Type
Article
Keywords
روش هيدروديناميك ذرات هموارشده؛ دقت؛ مرتبه همگرايي خطا؛ مشتق فضايي بهبوديافته؛ DualSPHysics
Journal
علوم و فنون دریایی
DOI
10.22113/jmst.2024.473084.2608
Researchers
H Ravanbakhsh (First researcher)
,
A. R. Faghih (Second researcher)
,
Rouhollah Fatehi (Third researcher)
Abstract
Improving the accuracy and order of convergence in the smoothed particle hydrodynamics method is still one of the challenges of the researchers. In this research, the first-order convergence scheme for the first spatial derivative in the continuity and conservation equations of linear momentum has been implemented in the DualSPHysics open-source code, in order to increase the accuracy and the order of convergence. By achieving more than three orders of improvement in the accuracy value, after implementing this method to simulate the Taylor-Green vortex benchmark problem in the authors' previous research, in this study, the simulation of the wavemaker benchmark problem by standard solvers and improved and compared with the analytical results through the calculation of the relative error of the velocity values at a certain position of the solution domain have been evaluated. The existence of stationary and moving solid boundary conditions as well as the free surface is one of the main challenges of the new problem. By using the improved solver and adjusting the effective parameters such as artificial viscosity value, density diffusion coefficient, neighborhood radius calculation coefficient and particle displacement coefficient, it is observed that the error reduction process is improved and its first-order convergence is achieved.