Shipyards these days see an increase in customers that specify combined speed and seakeeping ability design requirements. This requires the shipyard to make a prediction of the additional installed power required to maintain a certain speed when waves are encountered. The additio
...
Shipyards these days see an increase in customers that specify combined speed and seakeeping ability design requirements. This requires the shipyard to make a prediction of the additional installed power required to maintain a certain speed when waves are encountered. The additional required installed power is directly related to the average extra resistance that the vessel is subjected to when it’s sailing in waves. This extra resistance is known as the time-averaged added resistance in waves. In the maritime industry Computational Fluid Dynamics (CFD) is increasingly used for resistance predictions as it promises cheaper and faster predictions than model testing. The result does come without the comforting ’truth’ of the towing tank. In this study, the applicability of CFD for the estimation of the time-averaged added resistance in regular head waves is researched by assessing the error and uncertainty of the solution. For fast sailing vessels, no standard procedure for the estimation of the timeaveraged added resistance in waves using CFD has yet been developed. Therefore the secondary research objective is to establish such a procedure. For this research the resistance predictions are done for the Fast Displacement Ship (FDS) hull form. Extensive research was conducted on this hull form by the Cooperative Research Ships (CRS) organisation. Their model tests results are used for the validation of the solution. The discretisation error is determined through a procedure developed by L.Eça and M.Hoekstra [25] which is based on a grid refinement study. The time-averaged added resistance is estimated by simulating the vessel both in calm water and waves. The time-averaged calm water and total resistance in waves are determined from these simulations. The time-averaged added resistance estimate is then calculated by subtracting the calm water resistance from the total resistance. First a grid topology is optimised to simulate the incoming waves as well as the vessels response to them accurately and efficiently. Grid sensitivity studies of the simulation of incoming waves as well as simulations of the static vessel in waves and the vessel subjected to forced motion are used to determine an efficient topology. To determine if the vessel’s response is accurate, it is compared to the solution from potential flow code solver PRECAL. The comparison proved that accurately propagating waves and accurate vessel response to the waves and motions are achieved on grids with less than 3 M cells in total. Verification estimated an uncertainty that varies between 0.5 % and 1.3 % for the time-averaged total resistance in waves and between 15.1 and 36.2% for the time-averaged calm water resistance on grids with a total number of cells ranging between 1.3 and 6.6 M. Comparison with the results from the model test revealed that an error of 1.4 % was present in the timeaveraged added resistance estimate. This error is smaller than the uncertainty margin of the model test result. Using the proposed method, the time-averaged calm water resistance estimate didn’t converge well, resulting in a large discretisation uncertainty. As the added resistance prediction is dependent on the calm water resistance prediction, it is also affected by this uncertainty. Therefore it’s concluded that the proposed method for the estimation of the time-averaged wave added resistance using CFD is not yet applicable in its proposed form. However, by using the proposed method, it is possible to estimate the time-averaged total resistance in waves accurately and efficiently. Therefore it’s concluded that further research is required to improve the uncertainty present in the time-averaged added resistance due to the uncertainty seen in the calm water resistance for the used grids.