www.orcina.

com

L01 Default vessel

Introduction
In this example we use OrcaWave, the diffraction and radiation solver included with OrcaFlex since
version 11.0, to re-create the hydrodynamic data set used by the OrcaFlex default vessel. This is
the vessel type data that you see when you first add a new vessel to an OrcaFlex model. It was
generated for us in a 3rd party diffraction tool, many years ago. We have recreated it here to
demonstrate the use of OrcaWave as well as the import and subsequent use of those data in
OrcaFlex.

Note: OrcaWave has been extensively validated against WAMIT; for full details see the validation
study which is available for download from the Orcina website.

OrcaWave diffraction analysis

Mesh
The starting point for doing a diffraction analysis is to generate a suitable mesh of the vessel’s
hull form. OrcaWave does not currently include a mesh generation or editing tool, therefore a
suitable meshing tool is needed to generate a mesh file. OrcaWave accepts various mesh
formats including WAMIT .gdf, AQWA .dat and Hydrostar .hst formats (for the full list, see help).
In this example we have used Rhino to generate and edit meshes in WAMIT .gdf format, but there
are many other tools available.
In this case we have the mesh available from the original diffraction analysis and we use this as
the starting point. This mesh is in gdf format. The mesh makes use of half-symmetry along the
xz-plane to reduce the number of panels modelled in the diffraction analysis, and is truncated at
the free surface:

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 1 of 16

 www.orcina.com

The mesh in its original format contains 216 panels. The speed of the calculation and the
accuracy of the results are dependent on the number of panels contained in the mesh (similar to
the number of segments used to model an OrcaFlex line). Therefore mesh sensitivity studies are
usually performed in order to establish the optimum mesh density.
In this case we only intend to replicate the original diffraction analysis results. It should be noted
that our original vessel mesh has coarse resolution by modern standards, but this example is not
a meshing study, and we benefit from the fast calculation time that a coarse mesh provides. We
therefore only make a few minor improvements to the mesh as we work through the analysis.

Setting up the analysis

OrcaWave data files have the file extension .owd and the results files .owr. In this example we
will use several OrcaWave cases to compare the results generated by using different settings.
Open the file L01 Default vessel.owr.
The OrcaWave user interface is very simple compared to OrcaFlex and consists of several pages
into which the data to be used in the diffraction calculation are input. On the calculation & output
page, we choose the type of solve that we want to carry out.
Because we are re-creating the data used in the OrcaFlex default vessel, it’s worth listing the
diffraction results that it contains, and that we therefore want OrcaWave to generate:
• Displacement RAOs
• Load RAOs
• Wave drift QTFs (Newman’s approximation)
• Hydrostatic stiffness matrix
• Added mass and damping matrices
In addition to these data, OrcaWave is also capable of calculating full QTFs, sea state RAOs
(disturbance data) and multibody data, but these types of data are not covered in this example.
Other damping data is also generated and will be discussed later in this example.
The default OrcaFlex vessel also contains data for wind and current loading; note that this
information is not generated in a diffraction tool. See the OrcaFlex help for further details.
The diffraction solve type: box determines which sets of data OrcaWave calculates:

We require Newman QTFs to be calculated as well as the first order effects, therefore the
calculation type selected in this example is potential and source formulations.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 2 of 16

 www.orcina.com

OrcaWave then offers some further calculation choices:

Firstly we have a choice of load RAO calculation method, either Haskind or diffraction. The two
methods are mathematically equivalent and both methods can be calculated with no additional
computation cost, therefore we’ll use the default option of both. This means that two sets of load
RAOs will be available for import into OrcaFlex. The preferred load RAO calculation method choice
determines the default choice for which set is imported into OrcaFlex, but you can change your
mind at the time of import. However, note that the preferred choice also dictates which load
RAO set is used in the calculation of displacement RAOs.
Checking that load RAOs generated by the Haskind and the diffraction methods agree is a useful
quality check on the results.
There is a further choice of method used for the quadratic load calculation (QTFs) i.e. pressure
integration or control surface and again we will choose to use both, although this time there is an
additional computation cost and therefore the calculation time will be increased. The control
surface method requires an additional mesh, but it is often more accurate for vessels with sharp
corners / edges, and hence we have selected it as the preferred method so that it is the default
choice when we come to import the data into OrcaFlex.
The final setting that we discuss on the calculation & output page is the choice of linear solver
method. The direct solver is often faster for smaller meshes, the iterative solver is often faster for
larger meshes. A secondary factor that impacts performance is the number of headings. The
direct solver can solve for multiple headings without performing much extra work beyond what
is needed for a single heading. However, for iterative solvers the total work is proportional to the
number of headings. With this example the direct solver ran fractionally faster than the iterative
solver, but neither took longer than 1 minute to run on an 8-processor laptop.
It is possible that memory consumption, instead of calculation time, might be a limiting factor in
the production of OrcaWave output. Diffraction calculations require the solution of matrices
whose dimensions are directly related to the number of panels in the calculation mesh. As a
result, a refined mesh can demand a large quantity of your computer memory during
calculation. You should ensure that the OrcaWave thread count (which can be set from the Tools
menu) is low enough that the analysis will fit your available memory. OrcaWave presents the
estimated memory consumption per thread on the validation page, to help inform your choice.
The full details of all the calculation choices, and of the remaining data items on this page, can be
found in the OrcaWave help file.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 3 of 16

 www.orcina.com

On the environment page we define the water depth and the wave periods and headings that we
want to include. The original default vessel data was calculated using 400m water depth,
therefore we have opted to use the same here. However for interest, we also ran the analysis
with 100m water depth (the default OrcaFlex water depth). The influence that this parameter has
on the results is shown later in this document, in the Comparison of results – 1st order section.
The 24 wave periods and 9 wave headings listed on the environment page are the periods and
headings that were considered in the original analysis.

Jumping to the inertia page, the vessel’s centre of gravity, mass and moment of inertia tensor are
specified. We are specifying the inertia by a matrix (for a general body). The centre of gravity, mass
and moment of inertia tensor have been copied directly from the default OrcaFlex vessel. In
OrcaFlex, the inertia matrix is defined about the centre of mass, therefore the corresponding
inertia origin has been selected in OrcaWave.
The constraints page provides the means to apply external stiffness and damping matrices. The
stiffness matrix would, for example, be used to apply the mooring stiffness matrix if you were
generating displacement RAOs for the vessel in its moored condition. The damping matrix can
be used to apply additional sources of damping to the body. In this case you will see that we
have applied additional roll damping. This is discussed later in this document.
On the bodies page we can define the body mesh file. OrcaWave can perform multibody analysis,
so there could be more than one body specified here, but in this example we are considering
only a single body. The length for OrcaFlex import is set to 103m, the length of the default vessel.
This property is not used in OrcaWave; it can be set here for convenience so that the OrcaFlex
data item is automatically set upon import.
The body mesh file specifies the name of the file that contains the vessel mesh. This mesh file is
our final version of the vessel mesh. We have made use of half symmetry in the mesh file, so we
need to set the symmetry data item accordingly. The control surface mesh file specifies the name
of the file that contains the control surface mesh, which will be used in the quadratic load
calculation. We will discuss this in more detail later in this example.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 4 of 16

 www.orcina.com

To demonstrate the mesh validation tools we are going to first study the original mesh. To do
this, reset the model (F12), and then on the calculation & output page untick the divide non-planar
panels box. Also, on the bodies page untick the add interior surface panels… box.
Then click on the mesh view tab. You will see an OrcaFlex style view of the mesh or meshes. The
view can be zoomed, panned and rotated in the same way as with an OrcaFlex view. In the list of
view options down the right-hand side, make sure that the body mesh and waterlines boxes are
ticked, but the control surface mesh and dry body panels are not ticked. You will now see the
vessel’s mesh with the waterline highlighted in blue:

Next, click on the validation tab. OrcaWave performs a validation of the mesh or meshes, and in
this case two warnings are reported.
Specifically, these warnings refer to the following issues:
a) Several panels in the mesh are non-planar.
b) Irregular frequency effects are expected from Body1’s mesh.
These are common warnings, so we will explain each one here. Warning a) is a result of the mesh
being quite coarse for the hull curvature that is being modelled. In OrcaWave all panels must be
planar i.e. each panel must have all of its vertices lying on a single plane. Triangular panels, with
only 3 vertices, are naturally planar, however quadrilateral panels may not be. If the vertices that
make up a panel in the mesh file are non-planar then OrcaWave will project them onto a
common plane. In other words, it will move vertices slightly to make the panels planar. Warning
a) is telling you that this has happened, and a list of the affected panels is given on the validation
page. To see which panels the numbers listed refer to, click the show in mesh view option. The
affected panels are then highlighted:

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 5 of 16

 www.orcina.com

If you zoom in on these panels, you will see the effect that making them planar has had:

It is possible that the vertices / panels being moved in this way will have a negligible effect on the
results. However the easiest way to improve the mesh so that non-planar panels don’t exist is to
triangulate them i.e. split each non-planar panel into two triangular panels. This obviously
increases the number of panels in the mesh, and therefore the calculation time. So it might be
worthwhile running sensitivity checks to see what effect non-planar panels versus more panels
has on the accuracy of the results and the run time.
In this example the run time is very quick, so we have triangulated the non-planar panels to
create a modified mesh, which will get rid of warning a). This can either be done in the mesher
(e.g. Rhino) or you can get OrcaWave to do this for you by checking the divide non-planar panels
box on the calculation & output page. Tick this box and look at the mesh view again to see the

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 6 of 16

 www.orcina.com

effect on the panels (note you will have to click on clear in the highlight panels section to remove
the previously highlighted panels).
Warning b) tells us that there is a risk of irregular frequency effects occurring. Details of how
these effects arise are given in the help file theory section on this topic. The key thing to note
here is that they can cause erroneous results if not accounted for correctly. The recommended
method for removing irregular frequency effects is enabled in OrcaWave by supplying mesh
panels that cover the interior free surface. These panels can be included in your mesh file or you
can ask OrcaWave to add these panels to your mesh by ticking the add interior surface panels…
box on the bodies page.
Asking OrcaWave to mesh the free surface for you may lead to mesh quality issues (e.g. panels
with high aspect ratio if the body waterline has very short segments). So adding the free surface
mesh to your body mesh file gives you more control over the quality of this mesh. However in
this case, OrcaWave meshes the free surface well, using the triangulation method.
Tick the relevant box (add interior surface panels…) on the bodies page, and then go back to the
mesh view page, check the interior lid panels box is ticked and you should see the interior surface
panels showing in blue. On the validation page OrcaWave now reports no errors.

If the dry body panels box is also ticked on the mesh view page, you will see that we have also
added some panels above the water line. These have been added for visualisation purposes
only; when imported into OrcaFlex the mesh file can be used to add graphics to the vessel, so
showing the extents of the hull above the surface is useful. Note however that the panels above
the free surface will not be used in the OrcaWave calculation, and OrcaWave automatically
detects that they are above the free surface, so inclusion of these panels in the mesh file will not
slow down the OrcaWave run.

The changes made to remove non-planar panels and irregular frequency effects have increased
the mesh from 216 to 940 panels, but the run time is in the order of ~1 minute so the penalty for
having this improved mesh is not too onerous.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 7 of 16

 www.orcina.com

Note that a more detailed mesh validation check is possible, by ticking the perform validation of
panel arrangement box on the calculation & output page. This performs some further checks (e.g.
overlaps and gaps) when looking at the validation page but takes longer to do, so this is often
used when first importing a mesh, and turned off for later work once you are happy with the
quality of your mesh.

Finally, we need to add a further mesh to represent the control surface. This is needed in this
case because we have opted to calculate the quadratic load using the control surface method as
well as the default pressure integral method. The latter does not need an additional mesh, but
here we are going to compare results from the two methods. The control surface mesh has
again been created in Rhino. This mesh needs to surround the body, have the same symmetry as
the body mesh, and (for a partially submerged body) must include the free surface between the
control surface and the body waterline.
This mesh has already been added to this model, so on the mesh view page, ticking the control
surface mesh box will show the control surface in yellow:

You can examine the different parts to this mesh by ticking / unticking the relevant boxes and
manipulating the view.
Note that one further check that it is always worth making is to check the direction of the panel
normals. Ticking this box in the list displays the direction of the interior surface of each panel. If
any panel normals are pointing in the wrong direction then you must return to the mesher to
correct this.

This is the conclusion of the set up for our analysis. We have specified:
• required outputs,
• the applied environment values of water depth, wave heading and wave period,
• the analysed body via its inertia and mesh description, and
• additional meshes of enclosed free surface and control surface for use in the calculation

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 8 of 16

 www.orcina.com

We have also reviewed the OrcaWave validation of these inputs. Now we can generate and view
the OrcaWave outputs.

Viewing the results

Re-open the L01 Default vessel.owr file, which contains the results of the OrcaWave analysis. The
results are presented across two pages: graphs and tables.
The graphs should be recognisable to anyone familiar with the OrcaFlex ‘check’ vessel data plots.
The tables provide the full set of results, spread across a number of pages, in tabular form.
These results can be exported in spreadsheet format, however as we will be using them in
OrcaFlex, then there is no need to do this, as OrcaFlex recognises the owr file.

OrcaFlex dynamic analysis

Import into OrcaFlex
There are two ways to get the vessel diffraction data into OrcaFlex. If you are starting a brand-
new model, you can simply open the OrcaWave .owr file in OrcaFlex, either by dragging and
dropping it into an OrcaFlex window, or via the file | open menu command. This method will
create a new model with a vessel in it and with the vessel type data populated with the
OrcaWave diffraction results. If the water depth used in the OrcaWave analysis was anything
other than infinity, then that will have been imported also, and the water depth in the OrcaFlex

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 9 of 16

 www.orcina.com

model will have been set to this value. There are a number of other parameters that are also
conveniently set when using this method: see help for details.
The second method of importing vessel data is via the import… button on the vessel type data
form, as you would do to import hydrodynamic data that had been generated by any other
diffraction tool. We will demonstrate this method here.
In OrcaFlex, open the model L01 Default vessel.sim. In this model we are comparing the original
OrcaFlex default vessel with the ‘new’ vessel data generated by OrcaWave. We have therefore
added two vessels to the model; the ‘original vessel’, which is currently hidden (ctrl + h on the
object in the model browser to show it) is represented with a green outline, and the ‘OrcaWave
vessel’, as a red meshed object. Note that the OrcaWave vessel has made use of the panel mesh
file that was used in the OrcaWave analysis! Hence we have a much better visual outline of the
vessel’s shape.
To see the import process, first reset the model (press F12). Then open the OrcaWave default
vessel data form from the model browser. This is in the list of vessel types:

The vessel type data form contains the data that has been imported from OrcaWave already,
however you can repeat the process by clicking the Import… button at the bottom of the data
form and selecting the OrcaWave results file L01 Default vessel.owr. The import vessel data form
then appears:

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 10 of 16

 www.orcina.com

Note that the centre of gravity is reported here. OrcaFlex uses the position of the centre of
gravity to deduce how the origins used by the diffraction tool relate to the origins used by
OrcaFlex. Therefore it is important that the centre of gravity is set before the diffraction data is
imported. This data is set on the vessel type data form, on the structure page. In this case, it has
already been correctly set.
The clear existing data check box allows you to delete all existing data from the vessel’s data form
before importing the new data. Ticking this box ensures that you don’t accidentally use any
default vessel data in your analysis.
The requested data page gives you control over which data sets contained in the diffraction
output file are imported into OrcaFlex. In this case we are choosing to import everything that is
available (we did not calculate Full QTFs or Sea state RAOs, hence these are greyed out). The
calculation methods for load RAOs and quadratic loads are as per the ‘preferred’ settings we
selected in the OrcaWave analysis, but these could be changed here if we wished.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 11 of 16

 www.orcina.com

To import the data, press the import button and check any warning messages that appear (in this
case there are none).

Comparison of results – 1st order

We ran several OrcaWave cases to check out the impact of the various parameter settings.
To compare the 1st order results from OrcaWave with the data in the default OrcaFlex vessel we
used the report vessel response option (see OrcaFlex help for details). This allows us to report the
displacement RAOs for specified points on the vessel, and for different wave directions. By
checking these results we will be comparing both the displacement RAO calculation and, because
the displacement RAOs are derived from the load RAOs, it will check the load RAO calculation
also.
Firstly, we checked the impact of the water depth in the diffraction calculation. OrcaFlex has a
default water depth of 100m, so we looked at the impact of setting the water depth in OrcaWave
to 100m, versus setting it to 400m. Below, we have plotted the surge RAOs for the 0 degree
direction, at the vessel origin (for brevity we have chosen to only show the plots that display
results of interest).

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 12 of 16

 www.orcina.com

It is clear from the above left-hand plot that setting the water depth in OrcaWave to 100m has a
significant impact on the surge results at longer wave periods when compared to the original
default vessel data. Setting the water depth to 400m in the OrcaWave calculation produces
excellent agreement, as shown in the right-hand plot. This demonstrates the importance of using
the appropriate water depth in the diffraction calculation, particularly if you are considering a
relatively shallow water application.
Next, we compare the plots for sway and roll, for the 45-degree direction, at the vessel origin. In
the first two plots the OrcaWave calculation included no external damping so that we can show
what a body motion resonance often looks like:

Here, it is clear that we have significant disagreement between the OrcaWave data and the
existing default vessel data, in the region around 8.5s period. Closer examination of the data
showed that the amplitude graphs show smooth variation in the added mass and damping, and
also in the load RAOs, through the 8.5s region. The spikes only occur in the displacement RAOs.
This is classic behaviour for a body motion resonance, in this case a roll resonance. Potential
theory typically overestimates this motion because it does not take into account viscous
damping. It is common practice to apply an external damping matrix to the body in the
diffraction analysis to account for the effect of viscous damping.
In this case, we can see that the original vessel data has some linear roll damping applied (as
other damping), so the OrcaWave analysis was repeated with an external damping matrix applied
(see the constraints page of the OrcaWave file to see where this is applied. The same value also
appears in OrcaFlex on the other damping page of the vessel type data form). Note that
OrcaWave is also able to calculate the roll damping as a percentage of critical, if the critical
damping coefficient is known.

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 13 of 16

 www.orcina.com

So far all of the OrcaWave results shown have been calculated using the default method of load
RAO calculation i.e. the Haskind method. In the plots below we also include the results from the
diffraction method:

Applying the external damping has clearly had the desired effect, and both sets of OrcaWave
displacement RAOs plotted show good agreement with the original results, although it is noted
that the diffraction method appears to give slightly better agreement than the Haskind method,
probably suggesting that diffraction was the method that was used in the original analysis.
At this point we would usually do a mesh sensitivity check to see if more panels created better
agreement between the two load RAO calculation methods, however because the original vessel
analysis was done with a particular mesh density, we are limited to using that in this comparison.
The subject of mesh convergence is covered in example L02 OC6 Semi-sub.

The above graphs show the amplitude of the generated RAOs, but we also need to check the
phase. This can also be done by plotting the data; however you could instead run a simulation in
a regular wave so that you can visually compare the response.
If you have re-set the model, then re-open the sim file L01 Default vessel.sim.
Open the workspace file L01 Default vessel phase check.wrk. Run the replay to see how the two
vessels respond to the default wave loading: an H = 7m, T = 8s, Dean stream wave (remember to
‘show’ the original vessel). You should be able to see that the pitching motion of the two vessels
is slightly out of phase. We investigated this issue by also running a diffraction analysis in WAMIT,
which showed that the WAMIT results agreed with the OrcaWave results. It is possible that either
the RAO origin or the RAO phase origin used by OrcaFlex for the original vessel data is not
specified correctly.
We are satisfied that the results generated by OrcaWave are correct.

Comparison of results – 2nd order

We also generated wave drift loads in the form of Newman QTFs in the OrcaWave analysis. This
was done using both the pressure integration and the control surface methods for comparison.
The control surface method requires an additional mesh, as described previously, and is
expected to be the more accurate method for bodies of this shape. We also compared the
results with the original vessel QTFs, however it is noted that the data are reported about
different origins (the original vessel QTF data is reported about the centre of gravity, the

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 14 of 16

 www.orcina.com

OrcaWave results are reported about the vessel origin). Force QTF data can therefore be directly
compared, but moments cannot. However it is noted that the results can be reported relative to
an origin of your choice if you do so via the API, hence allowing direct comparison of the
different sets of results. This is beyond the scope of this example, please refer to the API help file
for further details.
The plots below show the surge and heave results for the 45-degree wave direction, comparing
the two OrcaWave methods with the original vessel results:

It is known that quadratic loads are often the slowest to converge when performing a mesh
convergence study and in this case it is likely that the mesh should be refined further to get better
agreement between the methods. The effects of mesh refinement will be looked at more closely
in example L02; in this case we are restricted to using the original mesh which, while appropriate
for RAO generation, is possibly too coarse for the quadratic load calculation.
Setting the remaining data
For completeness, we also set the current and wind load data for the OrcaWave vessel, by using
copy and paste to transfer those data sets from the original vessel.
OrcaFlex also requires the RAO and QTF data to span the range of frequencies involved in the
simulation. The data imported from OrcaWave includes wave periods between 4 and 22
seconds. If the wave train modelled in OrcaFlex includes components outside of this range then
OrcaFlex will make some assumptions about how the RAO and QTF data is applied to these
components and will issue a warning, e.g.:

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 15 of 16

 www.orcina.com

To avoid this warning, you can define data for zero and infinity periods in your RAO and QTF
tables, following the guidance in the help file, which is what we have done in this model.
Finally, both the wire frame and the shaded image in OrcaFlex can also utilise the panel mesh.
The mesh can be imported as panels at the same time as the diffraction data is imported. This
panel data can be found on the drawing page of the vessel type data form.

Switch to shaded graphics mode (ctrl + g) to see the panels shaded:

Diffraction: L01 Default vessel 07/01/2022 16:33 Page 16 of 16