Numerical modelling of aerated-water wave impacts

Numerical modelling of aerated-water wave impacts

van der Eijk, Martin (TU Delft Mechanical, Maritime and Materials Engineering; TU Delft Marine and Transport Technology; TU Delft Ship Hydromechanics and Structures)

Wellens, Peter (mentor)

Delft University of Technology

Marine Technology

2018-11-15

A substantial part of structural damage for conventional vessels is caused by complex free surface events like slamming, breaking waves and green water. These events lead to the interaction of air and water where air can be entrained in water. The resulting air bubbles can considerably affect the evolution of the pressure caused by hydrodynamic impact loading. Besides a {cushioning effect on the impact pressure and a possible increase of the compressibility of the mixture higher than air, earlier documentation conclude that due to aeration the acting forces can increase due to long lasting time of the pressure and the resonance between oscillating air pockets & generated pressure waves.Nowadays, methods used to predict these forces, assume that the fluid is incompressible. This can lead to an {underestimation of the forces. The aim of the graduation project is to evaluate the effect of a homogeneous mixture of water and air on the wave impact, specifically for a green water event. Based on the documentation of the numerical method ComFLOW, for the evaluation an improved extended numerical solver in Fortran and MATLAB is built.ComFLOW is described as a robust finite volume method which can model these impact loadings by solving the governing laws -continuity equation and conservation of momentum- for a two-phase seperated-flow using the first-order fractional step method and the second-order Adams-Bashforth timestepping scheme. Even with a coarse grid, a clear distinction between water and air can be maintained by combining the Volume-of-Fluid approach with the local height function and a constant line reconstructed free surface (SLIC-iVOF). However, the method is not capable to model the compressibility for a dispersed flow by assuming that inertial forces on the aeration are dominant: a homogeneous mixture of air and water.In this report an extension is developed based on two volume fractions: the entrained air and air above the free surface, and Kapila's five-equation model. Considering the implementation and improvement of ComFLOW itself, the representation of capillary forces is improved by a new developed Continuum Surface Force (CSF) model for SLIC. Further, taking into account the transpose of the velocity gradient in the diffusive term, not done by ComFLOW, lead to no significant underestimation of the viscosity around the free surface for Re=O(35) and an error reduction of 2.7% for the 2D rising bubble. Besides the neglected term in the diffusive term, for the interpolation of the density around steep interfaces for SLIC, like wave impacts, the cell-weighted averaging method leads to less spurious velocities around the surface than the gravity-consistent averaging method used by ComFLOW. The last revised implementation is a {modified local height function, documented by ComFLOW as strictly mass conserving, which is made more strictly mass conserving due not overlapping of the heights.The modelling of compression waves and bubble interaction is verified with the 1D shock tube and 2D helium shock bubble. For a simplified green water event, the results of the dam break tested for aeration levels 0.1, 1 and 5% showed that the effect of the aeration after the impact is most relevant and can achieve impact forces more than two times higher than the initial impact due to the combination of a plunging wave, mixing of water and air, air pocket oscillation and the longitudinal acoustic mode. The effect of aeration becomes significant for Ma=O(0.10). These observations are based on the dam break results with obstacle.

CFD
Volume of fluid
wave impacts
aeration
Compressible Flow
green water

http://resolver.tudelft.nl/uuid:9778a824-6d72-4960-a812-4615a53c2c41

2019-02-02

Student theses

master thesis

© 2018 Martin van der Eijk