Prediction of Forces and Moments in Multiphase Naval Simulations

Abstract

This thesis investigates the prediction of forces and moments in the static and dynamic planar motion mechanism (PMM) simulations of ship manoeuvring problems using the open-source Computational Fluid Dynamics (CFD) software package, OpenFOAM. Three well-established ships are used to investigate the static and dynamic PMM simulations: the Wigley hull (an analytical geometry), the DTC hull (a commercial ship), and the DTMB 5415 (a warship). The numerical simulations are performed using the InterFoam (for static cases) and InterDyMFoam (for dynamic cases) packages in OpenFOAM, which are incompressible multiphysics CFD solvers based on the volume of fluid (VOF) method to account for the multiple phases in a continuous regime (without needing to track the discrete air-water interface); the turbulence is modelled using Reynolds Averaged Navier-Stokes (RANS) approximations. The pressure and velocity are decoupled using the PIMPLE algorithm. The semi-implicit variant of multi-dimensional limiter for explicit solution (MULES) is used to solve the VOF Equation. Implicit schemes are used for time marching: local Euler scheme for steady cases; Euler scheme for unsteady cases. Two types of meshes are used in the current simulation: body-conforming mesh (generated by Pointwise) and castellated mesh (generated by snappyHexMesh tool). The castellated mesh is non-body conforming and generated via successive mesh refinement and adaptation using open-source software. A dynamic mesh technique is used in dynamic PMM simulations. Despite the lower accuracy of the castellated mesh generation method and the mesh deformation in dynamic PMM simulations, the results show an overall good agreement with experimental data and published numerical results. The relative errors remain small--for most of the static and dynamic cases under consideration--after a grid convergence study. There are three specific cases in which the error is significant: (1) at a high Froude number; (2) at a large drift angle; (3) for the specific combined yaw and drift case. In conclusion, the numerical method is capable of predicting force and moment coefficients for static and dynamic PMM simulations with good accuracy outside the three specific cases when the error is non-negligible.