Paulina Swenne, HAW Kiel
Mast design and the resulting exhaust gas propagation are important considerations in the design process for special-purpose ships and yachts. Exhaust gas in certain deck areas can cause technical disadvantage or passenger discomfort. Thus, exhaust gas propagation must be assessed carefully for example via wind tunnel tests or numerical analysis. However, such numerical analysis can be difficult to integrate into existing shipyard infrastructures. The aim of this master thesis was to determine how exhaust gases disperse and spread around the vessel under realistic operating and environmental conditions. Therefore, a computational fluid dynamics (CFD) workflow was developed and integrated into an existing shipyard infrastructure.
Requirements for hardware and software, as well as the interfaces and their connections, have been defined. A dedicated Linux-based computational environment was set up to subsequently establish a workflow in an open-source environment using OpenFOAM. A complete CFD workflow was implemented, including preprocessing like geometry preparation and mesh generation as well as simulation execution and postprocessing. Since appropriate meshing is a key factor for a successful numerical simulation, both the geometric preparation and the meshing were carefully developed and verified through a mesh study. A time-step study was then conducted using the resulting mesh to ensure numerical reliability. The exhaust gas propagation was simulated using two different approaches: Reynolds-Averaged Navier–Stokes Equations (RANSE) and Detached Eddy Simulation (DES).
For the purpose of this thesis, the flow is simplified and assumed to be unsteady and incompressible, since the flow velocity and the associated Mach number are both well below the speed of sound. To visualize the exhaust gas, a passive scalar was used, which allows the exhaust gas to be transported by the surrounding air without affecting the actual air flow. Since the focus was placed on process integration, only the upper part of the superstructure was modelled to minimize the discretization and computational effort. Both turbulence modelling approaches were systematically compared in terms of accuracy, computational cost, and suitability for industrial use in a shipyard context in order to demonstrate which method is the most suitable for representing the propagation of exhaust gases. Finally, the numerical results were validated against existing wind tunnel test data to optimize the exhaust system configuration.
The thesis showed that it is possible to set up a numerical analysis workflow for industrial use with an open-source framework. According to the simulation of exhaust gas propagation the DES approach provides reliable results. Validation against wind tunnel test data supported the numerical results. Further geometric modifications can be simulated with low effort due to the created workflow.
The results provide detailed insights into exhaust gas propagation and highlight the differences between the two investigated methods. The DES approach is more suitable for industrial investigation of exhaust gas propagation in a shipyard context. Overall, this thesis generated an integrated CFD workflow which enables reliable predictions while meeting the practical requirements of shipyard operations.