Previously experimental methods or empirical knowledge laid the foundation forship design. Performing a model test in a towing tank is time consuming andexpensive. In addition to this, studying detailed effects in a laboratory proves tobe difficult.
The exponential growth in computer power allows engineers to use computers andsoftware to solve Navier-Stokes equations using computational fluid dynamics(CFD). This method allows the engineers to test different models, in a time andcost-efficient manner.
In this master thesis, a study of CFD codes to perform a resistance prediction ontwo different hull design for the trawler, Roaldnes, have been made. The firstdesign being the current design from Seacon, Hull 1, the second being a designbased on findings from project report (Svoren, 2014), Hull 2. The differencebetween these hulls can be found at the aft shoulder where Season’s hull isdesigned with chines while the second design is without. Special emphasis wasmade on the inflow conditions for the propeller. This was done in order to compareand observe the effect of the changes made to the hull. Simcenter STAR CCM+ was chosen as CFD software, due to its user-friendly interface andpowerful built in post processing tools.
A detailed and thorough convergence study was performed to find the optimalsimulation setup. This study resulted in a mesh with approx. 2 mill cells, timestep of 0.01s, and a domain which stretched 4 lwl aft and forth, 5 lwl to the side,1.7 lwl up, and 3.5 lwl down. Due to symmetry, half of the model is simulated.A total of eight simulations on both hulls, at trawl and transit speed, and withand without nozzle and virtual disk, were performed. Hull 1 displayed a lowpressureridge over the sharp edge created by the chines. This lead to increasedvorticities in this area compared to Hull 2. These vorticities propagated onto thepropeller plane, where Hull 2 showed a slightly better wake field at trawl speed,and similar wake field at transit speed. A plot of the streamlines showed a lowvelocityswirl of water on the inside of each skeg at both velocities. These werehowever almost gone when introducing suction from the virtual disk simulatingthe propeller.Calculation of thrust reduction on Hull 2 showed an improvement of 3.75%compared to Hull 1. Full-scale calculations showed an improvement in total hullresistance of 6.7% at 4 knots, and 2.2% at 10 knots in favor of Hull 2.
Validation of results were made to known experimental data for a similar vessel,R/V Gunnerus. This investigation showed that the results obtained in this reportwas within an acceptable range from the reference vessel.
As CFD is a numerical tool to replicate reality, error sources, and uncertaintywill always be of presence. Nevertheless, it can be used to show trends. Resultsobtained in this report suggests that Hull 2 might produce a more uniform inflow,and a better wake field for the propellers, than what can be seen from Hull 1. Italso indicates that Hull 2 might produce lower thrust reduction, as well as anoverall lower hull resistance at both trawling and transit.
Norwegian University of Science and Technology
June 1, 2015