Cost and environmental considerations today often dictate to sail ships at lower speeds, the trend to apply “slow steaming” is ever increasing. For a number of ship types, e.g. container vessels, this leads to a growing operational envelope in which they need to perform optimal. Other ships like bulk carriers traditionally operate over a large envelope of different sailing conditions, ranging from ballast to partially laden to design draft conditions. Due to this broad range, hullform design poses special problems as there is no single point of design for which to optimise. The complexity of the design process is increased significantly with the number of conditions which need to be taken into account.
Another factor influencing ship hull performance is the surface condition which varies over time from a clean to heavy fouling and thus influences frictional resistance which is often the dominant part of the overall resistance. Modelling roughness effects on the basis of sand grain roughness in FreSCo+ allows to predict the detrimental effect of surface deterioration over time. The following figure 2 shows a comparison of the friction coefficient cF for a smooth surface (top) and for an assumed sand grain roughness of hR = 500 μm (bottom). The contour plots clearly indicate the different level of friction forces to be expected.
The full range of operational conditions including speed, draft, trim, surface condition and waves (added resistance) needs to be covered to make a life-cycle performance prediction. As the computational effort to run all these case in RANS predictions would be prohibitive, an artificial neural network – ANN - is applied to populate a complete multi-dimensional response surface for the resistance under different conditions. This ANN is fed with a tangible number of exact evaluations of total resistance from FreSCo+ predictions and evaluates the results for a large set of discrete points as indicated in the following figure.