The thesis contributes to simulation-based shape optimisation by documenting the development of an efficient, adjoint RANS-based optimisation procedure. The tool facilitates the gradient-based CAD-free shape optimisation for two-phase flow problems. The first part of the thesis covers the development and validation of an unstructured, segregated volume-of-fluid based adjoint two-phase solver. The adjoint twophase solver is validated against results obtained from second-order accurate finitedifference studies. The adjoint solver approximates the gradient of an objective functional at the cost of a single adjoint solution, independent of the number of design parameters. Hence, the method suggests to exploit the discrete entities of the shape discretisation as design parameters. No a priori limitation of the dimensionality of the design space is required, although a sensitivity regularisation (filtering) step is usually necessary to yield a smooth shape update. The second part of the thesis is concerned with the development of an explicit sensitivity regularisation technique that also constitutes a local Lagrangian re-parameterisation of the discrete shape. The proposed strategy utilizes 0th- and 1st-order consistent kernel functions and considers both surface-normal and surface-tangential deformation components. Subsequently, a gradient-based CAD-free shape optimisation procedure is developed that embeds the adjoint two-phase solver, the sensitivity regularisation method and modules that provide ship technical constraints. The volume mesh is adapted to the regularised boundary deformation. The scripted procedure facilitates the automated optimisation of flow exposed geometries. The capabilities of the developed shape optimisation procedure are demonstrated during the application to different industry relevant optimisation problems.weiterlesen