Data: Validation

OrcaWave performs a series of validation checks on your data. The results are grouped into information, errors and warnings:

If an error is reported, OrcaWave is unable to begin a calculation. You must fix all errors to begin a calculation.
If a warning is reported, OrcaWave is able to begin a calculation. Nevertheless we recommend you investigate any warnings to understand the impact they might have on your results.

Warning levels

The values of the warning levels are set on the calculation and output page, but they are used only for data validation. They are not used by OrcaWave during the calculation.

The minimum body volume is a warning level applied to the displaced volume of each body in your analysis.

The maximum panel aspect ratio is applied to each mesh panel. Long slender panels may give a poor quality discretisation of the integral equations and should generally be avoided. The length scale at which your mesh can resolve the problem is typically determined by the maximum dimension of the panels. Therefore, having many long slender panels can imply a relatively coarse resolution and a slow run time (because of an increased number of panels). In a worst case scenario, a large number of slender panels could cause badly conditioned rows in the matrix equations and lead to numerical errors.

Note:

Although slender panels should be avoided in general, they may be useful in the specific case of panels adjacent to the waterline in a second-order analysis. The accuracy of second-order loads can often benefit from such panels, e.g. by using vertical cosine spacing near the waterline.

The minimum panels per wavelength implies a maximum panel size. For instance, a value of 5 for this metric means a warning will be triggered if any panel is larger than 20% of the wavelength of any wave in the environment. Since the potential is assumed constant on each panel, smaller panels gives a better quality discretisation. Accurately resolved results typically need panels that are several times smaller than the wavelength.

If you select to perform validation of panel arrangement, then the following warning levels become active.

The maximum overlap of panels.

The maximum gap between panels.

The maximum angle between normals of adjacent panels.

Required memory estimate

The validation information includes an estimate of the memory that your model will require per thread during the OrcaWave calculation. This can inform your choice of thread count for parallel processing, which is especially important for models with large meshes. The most significant factors affecting the memory requirement are:

Model type – a restart analysis requires less memory, or none at all.
Solve type – including the source formulation and/or full QTFs requires more memory.
Symmetry planes – models with one or more symmetry plane in the global mesh require less memory.
Body mesh – larger meshes with more panels require more memory.
QTF calculation method – the indirect method usually requires more memory than the direct method.

Environment data validation

You must have at least one wave heading and wave headings must be unique.
You must have at least one wave frequency (or period) and they must be unique. Values of zero or infinity are not permitted in the list (but OrcaWave will automatically calculate the infinite-frequency limit of the added mass matrix).
When using a finite water depth, the Green's function for finite-depth water imposes restrictions on the range of wave frequencies.
The minimum QTF crossing angle must be less than or equal to the maximum. A warning will be given if the range excludes all possible pairs of your wave headings.
The minimum QTF period or frequency must be less than or equal to the maximum. A warning will be given if the range excludes all possible QTFs from one of the QTF frequency types included in your analysis.
In a full QTF calculation, the evaluation of the far-field (asymptotic zone) surface forcing on the second-order potential imposes a restriction on maximum wavelength.
In a model with Morison drag, a suitable wave spectrum must be defined in order to perform drag linearisation.
A warning will be given if a wave frequency exceeds $\omegahatest$, the estimated first irregular frequency, and you have not extended your mesh by including interior surface panels.

Note:

OrcaWave is able to compute a better estimate of $\omegahatest$ (i.e. closer to the true $\omegahat$) if you perform validation of panel arrangement. Therefore the warning message about irregular frequencies may change, or disappear, if you toggle the panel arrangement data item.

Body data validation

You must have at least one body included in the analysis.
An error will result if your body has zero mass or moments of inertia and the corresponding degrees of freedom are free to move: OrcaWave is unable to solve the body's equation of motion.
You must not specify circular chains of connections, in which one body is an indirect child of itself.

Field point validation

Field points must not be above the free surface, or below the seabed.

Mesh validation

The lists below explain the various tests that are performed. These are intended to catch the most common mistakes that can creep in. However, it is impossible to anticipate every possible situation, so you are recommended to always use the mesh view to check that the mesh looks visually correct.

Validation may raise error/warning messages that contain a list of panel indices indicating which panels have failed a particular test. You can visualise the panels involved by highlighting them in the mesh view. The mesh details page, which gives detailed information on every panel and every waterline, can be useful for understanding and fixing the error/warning.

Tip:	In some cases you may need to find the source data, in your mesh file, from which a particular panel originates. This can be done using the mesh file panel indices in the mesh view.

All panels are checked for the following criteria:

Panel centroids above the free surface or below the seabed.
Panel centroids on a symmetry plane (the exception to this is dipole panels, which are permitted to lie on a symmetry plane).
Panel edges with zero length or very small length (less than the length tolerance).
Panels with zero area or very small area (less than the square of the length tolerance).
Quadrilateral panels that are non-convex.
Quadrilateral panels constructed from vertices that were non-planar in the mesh file (i.e. projection length greater than the length tolerance).
Panels with an aspect ratio greater than the maximum panel aspect ratio.
Panels whose characteristic length, the maximum of the side lengths and diagonals, exceeds the size set by the minimum panels per wavelength.
Panels that should be on the free surface $\SF$ (e.g. panels in a damping lid or the free surface panelled zone of a QTF calculation) but which do not lie in the plane $Z=0$ and have a vertical normal.

Note:

If your mesh contains a symmetry plane, any error/warning messages will give a list of panel indices for panels in the first half (or quadrant) of the mesh that fail to meet the criterion.

Tip:

If OrcaWave warns about non-planar panels, consider using the divide non-planar panels option. You can use the mesh view to see the effect on the geometry. If your mesh has a lot of non-planar panels, using this option will increase the run time of a calculation so you might prefer to remesh instead.

Body meshes are checked for the following:

Body waterlines that are not closed. Incomplete waterlines can affect second-order load results and irregular frequency removal in the source formulation. In addition, waterlines must be closed for OrcaWave to generate free-surface panels, e.g. for the interior surface, a control surface or a free surface panelled zone.
Panels on the interior surface which do not have a vertical normal.
Panels from the body mesh file on the interior surface, if you have also specified to add interior surface panels.
Panels on the interior surface which are not contained within a waterline.
Volume $|\VInt|$ less than the minimum body volume, or water plane area $|\SI| \le 0$. These errors might indicate that the normals on the body surface are pointing in the wrong direction.
Dipole panels that are on the free surface (and therefore not wet on both sides).

If performing validation of panel arrangement, the following criteria are also checked:

Panels having an edge that contains the centroid of another panel or a field point.
Adjacent body panels that exceed the angle between normals warning level. This can indicate one or more incorrectly oriented panels. If the body has very sharp corners you may want to increase this warning level.
Gaps between body panels that exceed the gap warning level.
Panels that overlap by more than the overlap warning level.
Field points which are likely to generate meaningless results because they are inside a body and the option to detect and skip such points is not selected.

Parent model validation (for restart analyses)

The output option for intermediate results must be selected in the parent model file.
The parent model must not itself be a restart analysis.
The solve type of the child model (i.e. the current OrcaWave model) must be compatible with the parent.
The data file and results file of the parent model must be maintained in sync. OrcaWave checks some fundamental properties, such as the last modified time of the two files, but these checks are not comprehensive. You must take care to maintain the files in sync.
The calculation mesh (i.e. all body meshes and, if applicable, the damping lid mesh) must be identical to the calculation mesh when the parent model was run. OrcaWave will raise an error if it detects that the meshes do not match. For example, this might happen if the contents of a mesh file have changed, or if OrcaWave's meshing code has changed since the parent model was run.

Note:

There is no restriction on the meshes for control surfaces or the free surface panelled zone of a QTF calculation. They can be different from the parent model.