ABSTRACT
Computational fluid dynamics based optimization is becoming increasing popular in hull form hydrodynamic designs; however, existing optimization studies commonly utilize gradient-free methods which can only handle a handful of design variables. Ship hull geometry is necessarily described by a large number of design variables since the hydrodynamic drag is strongly influenced by small changes in the hull shape. To break this limit, in this paper we utilize a gradient-based optimization approach along with the adjoint method for efficient gradient computation. To demonstrate its power, we optimize the stern region of the KVLCC2 tanker using more than 100 design variables subject to proper geometric constraints (volume and thickness) for practical shape. We also evaluate the impact of propellers on the optimization results. We neglect the free-surface since the Froude number is low and the design changes are in the stern region. Overall, we obtain 2.9% and 1.3% hydrodynamic drag reduction for the cases with and without the propeller, respectively.
