Session: 06-05-03 Marine Hydrodynamics III

Paper Number: 79702

79702 - Wave Loads and Ship Motions Evaluated by Linear Strip and Panel Methods 

There is a growing interest in direct evaluation of wave loads in ship design. Despite a rapid development of the fully nonlinear Computational Fluid Dynamics (CFD), there are still room and demand for the classical linear methods based on the potential flow simplification. There are number of implementations based on these fundamental assumptions applied in the research and in the industry making it possible to have a rapid estimation of wave loads for a ship under design. In principle, these implementations can be divided into the strip-theory based methods and the panel methods. The strip methods consider only two-dimensional, transverse variation in the flow field around the ship hull; therefore, the hull is assumed to be slender. The panel methods analyze full three-dimensional variations in the flow field; thus, there are no restrictions on the shape of the analyzed structure. Additional problem in hydrodynamic analysis methods occurs from the modelling of ship forward speed.

This paper compares the results of a linear strip-theory based method and a linear panel method in terms of ship motions, radiation forces and wave loads. The results were validated against experimental data by performing a case study with a cargo vessel. The vessel had relatively simple hull geometry including long midship section similar to a barge and short fore and aft parts. Computer programs SEALOADS (strip method) and ANSYS AQWA (panel method) were used. SEALOADS applies Frank Close-fit method to solve the radiation problem in two dimensions (2D). ANSYS AQWA applies pulsating Green’s function source distributing method to solve the radiation in three dimensional (3D) flow domain. Therefore, the largest difference between the two applied methods in zero forward speed cases sums down to whether 2D or 3D effects in the flow field are considered. Both software codes apply artificial methods to consider forward speed of a vessel.

The results for added mass and damping were similar between the software codes, both with zero speed and forward speed. Only some coupling terms showed larger differences which was due to how the hull geometry was modelled in each code, especially at the ends of the vessel. Motion RAOs matched very well between the software codes and the experimental results. Forward speed increased differences only slightly in comparison to the zero speed cases. Global bending moments close to the midship matched between the software codes and the experimental results. Towards to the extremities of the hull ANSYS AQWA estimated different shear forces to SEALOADS and the experimental results. In the forward speed cases, analysis results for global loads were a bit more conservative against the experimental results than in the zero speed cases.

It is concluded that linear strip and panel methods estimate similar motions and global loads accurately when the ship hull is slender and simple. Different forward speed modelling methods increase the differences only a little when the basic idea behind the forward speed modelling is similar. Unfortunately, because of the simple hull geometry of the validating ship case, the drawn conclusions cannot be generalized. It is expected that the differences between the two potential flow methods increase with more complex hull geometries. Nonetheless, both linear potential flow methods can be applied for a fast initial studies of a vessel in design purposes.

Presenting Author: Aaro Karola Aalto University

Authors:

Aaro Karola Aalto University
Spyros Hirdaris Aalto University
Jerzy Matusiak Aalto University
Tommi Mikkola Aalto University