Validation of an SPH Sloshing Simulation by Experiments

Abstract

Sloshing is a violent fluid motion and is of current interest for many branches of the industry, among them gas shipping. Numerical methods are an important tool for analyzing sloshing. Among them, methods based on the smooth particle hydrodynamics (SPH) are particularly promising for analyzing violent fluid impacts. Previous work shows a good agreement in terms of free surface elevation between SPH simulation and experiments. An extensive comparison in terms of pressure in the tank is missing. This is due to the fact that availability of reliable and accurate pressure measurements is limited. Therefore sloshing experiments in a two-dimensional tank are performed. A regular one-degree-of-freedom motion with small amplitude is imposed for various frequencies around fluid natural frequency and three filling levels in range 17–40% of the tank length. By means of pressure sensors mounted on the vertical tank wall the pressure is measured for a non-impact type fluid motion. Free surface elevation is measured by wave probes and a high speed video recording is taken. An in-house SPH code is presented in detail. Standard SPH formulation is modified with the focus on implementation of the Verlet time scheme. The Verlet time integration scheme makes it possible to perform long time sloshing simulations due to its good momentum and energy conservation properties. A diffuse term coefficient is applied in the continuity equation. Investigated sloshing cases are without violent fluid impacts. Using artificial mass diffusion term in SPH simulations is expected not to significantly influence the pressure field. The paper shows that applying this technique with carefully chosen coefficient does not lead to any nonphysical phenomena in the SPH simulation for such a sensitive phenomenon as sloshing. By comparing the SPH simulations to the quasi-analytical multimodal method and experiments the code and diffuse term coefficient are validated.