Naval architects draw inspiration from previous designs, literature reviews, statistical regression models and systematicseries. In this paper, a complementary approach, using simulation-driven design, is presented: exploration of the multidimensionaldesign-space using first-principles methods. Here the vessel is modelled parametrically with the freevariablesthat define the design-space. The design-space is then populated by systematic variation of these variablesusing “design of experiments” approach to explore the design-space in an efficient manner. A multi-dimensional Krigingtechnique is used to “fit” a meta-model to the simulation-calculated performance measures of each of the designvariants. The subsequent meta-model can be used for three key functions: visualisation of the design-space byinterpolation of iso-parametric surfaces; answering “what if … ?” style questions (e.g. “what if the length is increased by0.5m?”); providing performance data, in an extremely efficient manner, for simulation-driven formal optimisation. Thekey benefit of the proposed method is that it allows the design team to quickly explore the design-space and build up aknowledge base ahead of an anticipated project. This then allows quick interrogation of the numerical model series tosubstantiate design decisions during the bidding and tendering process as well as provide the data necessary for formaldesign optimisation.The proposed method is elaborated using an example of a fast luxury motor-yacht of approximately 70m in length. Thevessel’s calm water hydrodynamics are studied for speeds up to 20kn by means of SHIPFLOW (FLOWTECH); seakeepingperformance is determined using Seakeeper (FormSys); intact stability is assessed with Hydromax (FormSys).The parametric model is realized within the FRIENDSHIP-Framework (FRIENDSHIP SYSTEMS) which also serves asthe integration platform for the analysis software.
展开▼
