PENERAPAN METODE CORRECTED SMOOTHED PARTICLE HYDRODYNAMICS PADA ALIRAN DI ANTARA DUA PELAT DENGAN GANGGUAN BALOK DAN HEAT SOURCE

• Main Author: Maharani Septiani, Livia
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39566
• Institution: Institut Teknologi Bandung

SPH is a Lagrangian method that represents a system using a set of particles which move in a particular domain. Each particle has its own material properties such as density, velocity, and temperature, which evolve according to density, momentum, and energy equations. The equations are approximated using an integral function representation called kernel approximation. The kernel approximation is then discretized using particles. Discretization is done by replacing representation integrals and their derivatives by the sum of the values of the neighboring particles property in the local domain of a particle. This local domain is called the support domain. Support domains that are located on the domain boundary will be truncated so that inconsistencies occur which cause approximation inaccuracy.The acceleration of each particle is calculated. Then, it is used to calculate velocity and position of the particles on the next time step. In addition to acceleration, density and temperature are also calculated at each time step. The SPH method is used to simulate incompressible ow, which are Poiseuille ow, Couette ow, and ow over square cylindrical obstale which acts as a heat source. The ow cases are simulated using Weakly-Compressible SPH formulation. From the simulation results, it can be seen that CSPH can improve kernel functions and derivatives. Thus, particle properties at the domain boundary can be approximated more precisely. The use of articial viscosity can reduce the numerical oscillation that occurs in the simulation results. In the stagnation and wake area, there are still signicant dierences between the SPH solution and FVM. These dierences are caused by particle clustering in the SPH simulation that can be overcome by using particle shifting method. In addition, particle number needs to be increased so that the turbulence phenomenon can be modeled better.