SNL-WaterPower / WecOptTool-MATLAB

WEC Design Optimization Toolbox
GNU General Public License v3.0
12 stars 9 forks source link

NEMOH Mesh Sensitivity #173

Closed H0R5E closed 1 year ago

H0R5E commented 3 years ago

It has been noted in the RM3 study that particular meshes tend to cause instabilities in the NEMOH output (particularly for the coupled radiation damping coefficients). For the case in question (described by issue #4) the mesh and damping is as follows:

badmesh baddamping

Note, the unusual panelling of the top of the lower body, which contains many triangular elements. The panelling of the bottom of this body is similar to the panelling of the bottom of the top body, with more typical rectangular elements (.fig versions of the mesh figures are available below).

These triangular panels may not be the direct cause of the issue, however, as a successful run also contains such elements. From the RM3 optimization.m example:

goodmesh gooddamping

Trying a different meshing tool for the failing case may help determine if the mesh is the source of the issue. Otherwise, further investigation is required to determine what design parameters trigger these events, and a look at the panel by panel results from NEMOH would also be beneficial.

Mesh .fig files