Combined Orcaflex Example

W Orcina
Production Risers: A01 Catenary and Wave Systems Page 1 of 14

A01 Catenary and Wave Systems
Introduction
In these examples, four types of riser system are shown. These are a simple catenary, a lazy
wave, a steep wave and a pliant wave.
On opening each simulation file, the default Workspace will present wireframe and shaded
graphics views of the system.
The examples also show:
Variation of drag coefficient with Reynolds Number.
Bend stiffener modelling.
Modelling of buoyancy modules.
Extending the simulation.
Modelling tethers and their clamps.
All the examples in this set have regular waves applied. These are repeated waves with the same
height and period. The Dean Stream non-linear wave theory has been applied because it is
accurate over a wide range of water depths.
The build up (Stage 0), where wave heights are ramped up from zero to the required value, is
given a duration of one wave cycle. Less than this and the rapid application of the wave could
generate unwanted transient loads.
The main analysis (Stage 1) is given a duration of five wave cycles. This is sufficient for most
systems of this type to achieve a settled response.
W Orcina

1. Catenary Riser
The example has a single riser in a simple catenary extending from a vessel to the seabed.
When you open the simulation file you will see side elevation views of the catenary in both
wireframe and shaded format. Beside them are two graphs. The upper graph is of top connection
End Force against End Force Ez-Angle. The lower is a range graph of normal drag coefficient
along the catenary length.
1.1. Building the model
Open the model browser via the Model
dropdown menu (or use the shortcut F6)
then double click on Catenary Hose to
open its Line Data Form.
The Connection table shows End A is
attached to the vessel FPSO and End B is
Anchored.
The Anchored option gives access to an
additional data item called Height above
seabed which is important to get the correct
seabed interaction at that connection.
Seabed contact takes account of the line
contact diameter. This diameter is found in the Contact page of the Line Types Data Form (note
that a line contact diameter of ~ means that the line outer diameter is used as the contact
diameter). If the connection z coordinate was set at zero (centreline directly on the seabed) then
the line bottom edge would be placed one radius beneath the seabed.
If you set Height above seabed to zero then OrcaFlex offsets the end connection z coordinate by
the line contact radius so the bottom edge of the line sits flush to the seabed. More details are in
System Modelling: Data and Results | Lines | Line Data | Connections in OrcaFlex Help.
The End Orientations have also been set in this model. These specify the directions of the end
fittings. If the riser is aligned with that direction at the end fitting then no moment is produced. It
is therefore called the No-moment direction. See an explanation of this in System Modelling:
Data and Results | Lines | Line Ends | No-moment Direction of OrcaFlex Help.
You can view the local axes, including the line end orientations, by selecting Axes | Local Axes
in the View dropdown menu (or use the shortcut Ctrl + Y). The key point is that OrcaFlex adopts
an End A to End B convention. This means that the end direction (i.e. the local z-axis direction
on which the no-moment direction is based) is into the line at End A and out of the line at End B.
In this example, the riser end connections are pinned (zero connection stiffness). This means the
connection is free to rotate and no end moments will be produced so changing the end
orientations will not change the behaviour of the system.
However it is good practice always to set the line End Orientations. The model is then set up
ready in case you change to a built-in connection (non-zero connection stiffness) at a later stage.
Also Bend Stiffener design loads need End Force and End Force Ez-Angle, its direction relative
to that no-moment direction. To obtain the correct design loads therefore the no-moment direction
W Orcina

needs to be set. A more detailed explanation is in System Modelling: Data and Results | Lines |
Line Ends | End Force and End Force Ez-Angle in OrcaFlex Help.
Identifying the line segmentation required is a balancing act. You want sufficient refinement that
the model can capture the response accurately but not so much that runs are slow with no
significant improvement in results. It is best to put detailed segmentation only where it is needed.
Where the detail is required will depend on what you are modelling.
In this example the area of interest is the touchdown point where curvature is expected to be high.
Double click 'Catenary Hose' on the model browser to open its Line Data Form. Look at the
segmentation for the line on the Structure page.
The riser is made up of three sections with the same properties but different segmentation. There
is detailed segmentation in the middle section, the touchdown region, where it has a target
segment length of 0.3m. There is coarser segmentation either side where the riser remains fairly
straight.
From the Model Browser, select Variable Data | Drag Coefficient | Generic Drag. This contains
a table relating Reynolds Number to Drag Coefficient the data is intended for illustrative
purposes only. Select the Profile button to see it as a graph.
Go to the Drag & Lift page of the Line Type Data Form (select 10 flexible variable Cd in the
Model Browser). Generic Drag is selected for the Normal Drag Coefficient. Note that the drag
coefficient in the y direction is specified to be ~ which is interpreted as the same as the x
direction in other words the drag is isotropic. At each time step OrcaFlex will calculate
Reynolds Number at each line node and apply the appropriate Cd value.
The Reynolds Number Calculation data item on the Sea page of the Environment data form is
used to determine how OrcaFlex interprets the data specifying Cd vs Re. This data item should be
set to match the conventions used by your Cd vs Re data source. For more information see
System Modelling: Data and Results | Environment | Sea Data in OrcaFlex Help.
1.2. Results
Edit Replay Parameters on the Replay dropdown menu will give you access to the replay
settings. Choose Latest Wave to see the final cycle of the run. You will see the line nodes turn
white as they contact the seabed.
Look at the range graph of the line x-Drag Coefficient for the Latest Wave. This shows how the
normal drag coefficient varies along the line length. It gives the maximum, mean and minimum
values experienced through the final wave cycle. The graph shows a variation between 0.4 and
1.2. There is significant variation along the whole length.
The graph of Catenary Hose End Force against End Ez-Angle at the FPSO (End A) shows the
Bend Stiffener design loads discussed previously. A moving cross on the curve shows where on
the curve the animation has reached. It is not a closed loop, indicating that maybe a slightly
longer run is required to let it settle.
The simulation duration is five wave cycles. You can extend the run without having to start at the
beginning again. Select the Calculation | Extend Dynamic Simulation dropdown menu option
and enter how much more simulation time is required. The model state will change from
Completed to Paused and you can then set it running again for this additional time.
W Orcina

The graphs discussed above were all set up in advance using the workspace facility. To create a
range graph of effective tension, go to Select Results in the Results dropdown menu. This takes
you to the OrcaFlex results form. In the Result Type area at the left of the form are the types of
plots available. Greyed out options are those that are not valid for the type of analysis carried out
in this example.
Make sure Range Graph is selected. This result is only valid for lines so the only option in the
Object area is Catenary Hose. OrcaFlex also needs to know what time period of the run you
want results for. In this example, select Latest Wave in the Period area.
Finally you need to select the variable you want plotted. For lines there are many results so they
have been sorted into groups. By selecting the group in the Show list, you can hide and show
different variables. Select the Forces group, if it isnt already, then select Effective Tension in
the Variable area.
Press the Show button to produce the plot. The range graph will now be at the top left. It shows
the expected tension distribution for a simple catenary.
W Orcina

2. Lazy Wave Riser
A lazy wave configuration is similar to catenary but has support provided at about midwater by
distributed buoyancy modules. 'Lazy' means that the riser centreline is near parallel with the
seabed on contact while 'Wave' describes the line shape as a result of the buoyancy modules.
This example has two risers positioned exactly on top of each other to allow comparison of two
modelling methods. These are:
10 Lazy Wave Distributed: Buoyancy is applied using a line section that has the
equivalent smeared properties of the line and buoyancy modules combined. This
approach is more efficient and convenient for the design stage and load analysis. It
allows easy variation in buoyancy application and does not need as much refinement of
line segmentation.
10 Lazy Wave Discrete: Buoyancy is applied as discrete modules attached to the line.
This approach is more useful in the final stages of an analysis when the as-built module
details are known. It allows a check that module size and pitch do not allow excessive
sag of the riser between the modules; the curtain-railing effect
A wireframe plan view and a shaded side elevation of the risers will be presented when you open
the simulation file. You will also see range graphs of curvature for both options.
The plan view shows that this is a cross-weather case; weather heading is shown in the top right
of the view. The current is in the riser plane while the waves are normal to it. The vessel has also
been rotated so it remains heading into the waves to represent a weather-vaning system.
Both Risers have the same end connection
settings and the same number of sections.
As with the simple catenary model, refined
segmentation has been applied where
needed.
For this example it is the sag, hog and
touchdown regions that require the detail.
The suspended length below the hang-off,
and the length lying on the seabed, do not
experience significant curvature and so
allow coarser segmentation.
We will first consider the riser with smeared
properties, 10 Lazy Wave Distributed. Open the Line Data Form for this line. You can do this
by double clicking the line name on the Model Browser. Now look at the Structure page on the
form.
The riser is made up of five sections for the suspended length, sag, hog, touchdown and length on
the seabed. Segmentation differs between the sections as mentioned above, however the hog
region (Section 3) also has a different line type. It uses 10 +Floats instead of 10 flexible.
This +Floats type contains the smeared properties of riser and float.
W Orcina

To see the details of these properties select the Line Types button at the bottom left of the form.
It will take you to the Line Types Data Form. You can also access it directly via the Model
Browser.
Select View Mode | All so you can compare the properties of the two line types. If you move
through the pages on the form you will see that 10 +Floats has increased OD, mass,
hydrodynamic coefficients and contact diameter. The drag and stress diameters are set to match
the values of 10 flexible. There is a variable drag coefficient called 10+Floats drag which is
based on the Generic Drag specified for 10 flexible and adjusted to account for the extra
normal drag from the buoyancy modules.
There is no need to calculate any of these changes, just run the OrcaFlex Line Type Wizard to
create them.
The Line Type Wizard is only fully accessible when the program is in the Reset state. So exit the
form and select 'Reset' from the Calculation drop down menu (or press F12). Now open the Line
Types data form again.
If you are going to use the Line Type Wizard then you first need the select the line type where
you want it to put the data. So select the '10+Floats' and then select the Wizard button at the
top right of the form. You are now in the Line Type Wizard facility.
Because we have already set up an example, it has remembered what we have done and will be
showing you the Line with Floats special category option. Select the Next button at the bottom
right to move to the page where you input data.
Note the hydrodynamic properties for the buoyancy module in this example are for illustration
only and do not reflect actual values.
This facility allows you to easily experiment with module properties and pitch (spacing) to get the
configuration you wish. You can find more information on how the properties are calculated in
System Modelling: Data and Results | Lines | Line with Floats in OrcaFlex Help
Now we will consider the riser with discrete modules attached, 10 Lazy Wave Discrete. Open
its Line Data Form via the Model Browser. If we look at the Structure page the only difference
is that Section 3 (the hog region) now has 10 flexible line type and its segmentation is more
refined.
Move to the Attachments page. There are 20 attachments from an arc length of 81.25m to
128.75m. The spacing is 2.5m which is the same pitch as used for the Distributed method. The
finer segmentation of 0.625m is to make sure there is a node at each location for the module to
connect to and that there are four segments between each module to show if the riser sags.
Each attachment is called Module and represents an individual buoyancy module. To see the
properties, select Attachment Types at the bottom left of the form. This will take you to the
Attachment Types Data Form. Go to the Clump Types page and select View Mode |
Individual to see all the module data at once.
The properties are not quite the same as those used by the Line Types Wizard.
Line Type Wizard: Total Drag = Drag of Modules + Drag of Exposed Riser Length
Line and Attachments: Total Drag = Drag of Modules + Drag of Total Riser Length
Therefore the attachments should only contain the additional properties the modules would add to
the bare riser or you would be applying loads twice.
W Orcina

In addition, notice the Align with option is set to Line Axes. The module z-axis will align
with the line node z-axis and move with it. This is appropriate for a module that wraps around a
line. The alternative, Global Axes, is more appropriate for clumps that are attached by a swivel
or tether so do not rotate with the line.
The detailed profiling within the module is not included in this example. You could add a curved
plate shape to each module to model this refinement. If so then the clumps will need to be
replaced by 6 degree of freedom buoys and the single line would need to be broken into a series
of shorter lines connected between these buoys. This is a far more complex model which should
only be built if detailed analysis of the curtain-railing effect is essential.
2.2. Results
If you have reset the simulation then you must reload the simulation file in order to access results.
Look at the animation through the latest wave. The blue section indicates distributed buoyancy.
The pink clumps indicate the individual modules. The two lines are moving in unison. This is
even though the plan view shows waves being produced at the hangoff and moving down the line
due to vessel surge motions.
Look at the curvature range graphs for both lines through the latest wave. The lower plot is the
Distributed method and the upper plot is the Discrete method. Note that these curvature plots
show the magnitude of the curvature vector and so are rectified, i.e. always positive. Therefore
the curve for the trough region has been reflected to become a peak.
There are sudden changes in curvature at about 80m and 130m: these are points of inflection
where the buoyed length begins and ends. The other sudden drop is at about 165m, the
touchdown point. The variation of curvature in the discrete buoyed region is the curtain-railing
effect. Each little peak is where a buoyancy module is located.
There is also a slight curvature increase near the seabed termination. This is because the anchored
end (End B) is being held exactly on the seabed. In fact the riser would sink into the seabed
slightly and the node in front of End B is able to do so. Hence the slight increase of curvature.
This slight rise is not significant to the system response.
The two range graph plots show there is little difference in extreme curvature between the two
modelling methods.
W Orcina

3. Steep Wave Riser
A steep wave configuration is similar to the lazy wave but has a near vertical connection at the
seabed. Steep means that the riser centreline is near vertical at the lowest end while 'Wave'
describes the line shape as a result of the buoyancy modules.
When you open the file you will see four views of the system. The riser extends from the side of
the vessel down to a PLEM. The plan view shows weather is near head seas for the vessel but
normal to the riser plane a transverse case. This is typical for West African systems.
The two detailed shaded views at top and bottom right show Bend Stiffeners (BSRs) located at
the hang-off and PLEM connections respectively. There are two methods available for modelling
BSRs in OrcaFlex; both are demonstrated here for comparison.
Two duplicate risers are included in this model, one for each style of BSR used. When you first
open up the model, one of them is hidden. Make sure the Model Browser is in View by Groups
mode; this is set by right-mouse clicking in the Model Browser and checking that the View by
Groups option is ticked. There are two groups shown at the bottom of the Model Browser;
Separate Line BSRs and Attachment BSRs. Select the Separate Line BSRs group and Ctrl + H
to show the objects in this group. The two systems are modelled directly on top of each other, so
very little changes in the model views except for red coloured BSRs now being visible on top of
the green coloured ones in the two right-hand views. Press Ctrl + H again to hide this group while
we first consider the Attachment BSRs group in detail.
In each group, the riser is built in a similar
manner as for a lazy wave (see previously),
however the lower end is vertical rather
than horizontal when it reaches the seabed,
hence the name steep.
The buoyancy has been applied by
distributed properties but it could also be
modelled as discrete modules if required;
see the Lazy Wave example for details.
The PLEM is represented by two shapes set
as Drawing types. They have no contact
stiffness so will not interact with the riser; they are present for visualisation only.
Go to the 10 Steep Wave Line Data Form. The z coordinate for End B says 5m but the height
above seabed says 4.822m. This is because OrcaFlex assumes lines have hemispherical ends and
so for clearance purposes the height above seabed still takes account of line contact radius.
However in this example the z coordinate is what we are interested in because it defines the pivot
point for the line.
Both line ends in this example have built-in connections (Infinite connection stiffness). Local
bend protection is provided by Bend Stiffeners (BSRs), included in the model.

W Orcina

3.1.1. Modelling Bend Stiffeners as Attachments
The simplest way to model a BSR is as an attachment to a line. This method allows you to model
the polymeric section of the BSR, but not the rigid steel mount section, therefore if fitting loads
are required they must be obtained through a separate calculation outside of OrcaFlex. Go to the
Attachments page of the 10 Steep Wave Line Data Form. Two BSRs are attached, one at End
A (hangoff) and one at End B (PLEM). Each has been given a unique name (the Name column of
the table) so they are easily identified when extracting results.
Now look at the Structure page. The riser is made up of six sections. These represent the regions
BSR, Suspended, Sag, Hog, Suspended, BSR. Select the Profile Graph button at the bottom left
of the form. This will show you how the diameters of the riser vary. You can see the red line
showing the riser. The buoyed section is very clear with its much larger OD. The blue lines at
right and left are the BSRs. You can see how they taper and wrap around the riser.
It is very important to remember when creating a BSR in this way that the refinement of the BSR
depends on the segmentation of the section it wraps around. If the section has 0.5m segmentation
then that is also the segmentation of your BSR, even if you input BSR profile data at 0.1m
intervals. Always check this profile plot to check if the resulting profile is appropriate. It is also a
good way to check that your BSRs are the right way round, i.e. with the root at the end of the line.

The BSRs can be clearly seen in the Shaded Views. Views can be switched between Wireframe
and Shaded by selecting the Shaded Graphics Mode in the View dropdown menu, or by using
the keyboard shortcut, Ctrl + G.
The views below show the hang-off and the PLEM BSRs (both in green). You can see how the
bend stiffener tapers as it extends away from the line end.

To see the attachment properties select the Attachment Types button at the bottom of the Line
Data Form and go to the Stiffener Types page. Each attachment has properties determined by a
line type and a length of 3m, as determined by the BSR profile see below.
To see the physical properties of the BSRs, close the Attachments Form and select the Line
Types button instead. Take a look at the line properties for the lines Topside BSR Properties
W Orcina

and Seabed BSR Properties. They have been given the Homogenous Pipe category with
typical BSR polymer material properties so OrcaFlex can calculate mass, stiffness, etc.
However BSRs have an Outer Diameter that changes along the arc length. Looking at the
Geometry & Mass page you will see that the Outer Diameter is a variable data item rather than a
constant value. To see the details right click on the form and select Edit Variable Data from the
dropdown menu. You will now be in the Variable Data Form, which can also be opened directly
from the Model Browser.
Within the Data Source Type Structure is Line Type Outer Diameter. This contains two
Variable Data Sources, one for each BSR. They have identical profiles except Seabed BSR OD
is the reverse of Topside BSR OD. This is because Topside BSR has its root closest to End A
while the Seabed BSR has its tip closest to End A. The Reverse button at the bottom of the page
allows OrcaFlex to invert the table automatically for you.
Select Profile at the bottom of the table to take a look at the distribution. Note also that each OD
table gives values up to an arc length of 3m. This maximum arc length is what determines the
attachment length so you do need to input the data for the whole polymeric BSR length.
Back on the Structure page of the Line Data Form, note that the first and last sections of line
have fine segmentation. These sections are 5m long, the first 3m of which are protected by the
BSRs. The fine segmentation has been deliberately continued for 2m beyond the tip of the BSRs.
It is good practice to check the detailed curvature at the point where the riser exits the BSR,
because over bending can occur just beyond the BSR tip if it is loaded beyond its design
capability.
More details on BSRs, and Bend Limiter modelling can be found in System Modelling: Data and
Results | Lines | Modelling Bend Restrictors in OrcaFlex Help.

3.1.2. Modelling Bend Stiffeners as Separate Lines
If you require more detail in your bend stiffener model, for example if you need to include the
steel root section, or model the rattle space or friction effects between the BSR and the protected
line, then you need to model the BSR as a separate line rather than as an attachment.
In this method, we utilize the Line Contact model to define the relationships between the
stiffeners and the protected line. This method allows the BSR to be modelled as separate sections
to represent both the steel mount and the profiled cone section.
In the model, hide the Attachment BSRs group and show the Separate Line BSRs group. The
BSRs are now shown in red, with a short grey section (the steel root) at the end where they
connect to the vessel or the PLEM. The two BSR lines are called TopEnd BSR and Seabed End
BSR.
Open the Line Data Form for the line TopEnd BSR, via the Model Browser. If we look at the
Structure page, there are two sections in the line, one of line type Steel Root and one of type
BSR Cone. These line types have been defined on the Line Type Data Form, and in fact the
BSR Cone line is identical to the Topside BSR Properties line used in the attachments method,
with the exception that it has been made a different colour, for clarity.
Now look at the Line Data form for either BSR line. On the Structure page, note that the BSR
Cone section of line has been given relatively short segments, 0.25m. This is so that the changing
properties of the BSR as it reduces in diameter are captured accurately. Both BSRs have End A
W Orcina

built-in to their connection points (they have Infinite connection stiffness) and End B is Free.
The End Orientation of End A has been set appropriately, as discussed in the Catenary Riser
example. The BSR lines also have torsion included, as indicated by the box at the top of the Line
Data Form. Torsion is required by the Line Contact model, which is discussed below. The end
positions of both BSRs take into account the 0.5m steel root section.
Look at the Contents page; note that the BSR lines have Free Flooding contents.
Close the data form and reset the model (press F12). The two right hand views showing the BSRs
close-up now show that the BSRs and riser lines are separate lines and are hanging independently.
Run statics (press F9) and youll see the riser quickly gets pulled inside the BSRs. This happens
because a Line Contact relationship exists between them.
Double click on the Line Contact Data item, which is located towards the top of the Model
Browser. A relationship is defined on the Relationships page for each of the BSRs; these specify
that the 10 Steep Wave1 line is positioned Inside each of the BSRs. Containment is enabled
for both, indicating that the inner line (riser) is shielded by the outer line (BSR). Stiffness and
friction values are also assigned here. The Line Contact model is described in more detail in the
Help file section System Modelling: Data and Results | Lines | Line Contact | Modelling, and in
Example D04: J-Tube Pull-In.

3.2. Results
Re-open the simulation file and load A01 Steep Wave Riser Attachment BSR.wrk from the
Workspace dropdown window. This will present the curvature and bend moment range graphs for
the attachment-type BSR at the hang-off, and the portion of riser that is inside it. The right-hand
two plots are for the riser, the left-hand two are for the BSR. The curvatures for the two lines are
very similar (top two plots), as youd expect, however the bend moments for the two lines are
very different (bottom two plots), showing that the BSR experiences significantly more of the
load than the riser does, as it is designed to do.
Open the A01 Steep Wave Riser Line BSR.wrk workspace to see the same results for the
separate line type BSR. This time the curvature plots are quite different. The steel root of the BSR
is clearly visible, with zero curvature (remember the steel root is not modelled with the
attachment method), and also the riser is free to move in a different manner to the BSR as there is
a small clearance between the two. The separate line model allows the two lines to move
independently, limited only by the contact between the outer diameter of one and the bore of the
other, whereas in the attachment model the BSR and riser act as one, with no relative movement
between them.

W Orcina

4. Pliant Wave Riser
A pliant wave configuration is a lazy wave with the addition of a tether restraining the touchdown
point. This example shows two methods of modelling the tether arrangement.
Simple Tether: Port Side. The riser is a continuous line. A link restrains the riser and a
clump attachment gives the clamp properties. This is sufficient for most cases. The link
component has no mass or hydrodynamic properties but if the link is small compared to
the riser length this is not expected to make significant changes to the response.
Detailed Tether: Starboard Side. The riser is split into two lines connected together by a
6D (6 degrees of freedom) buoy that represents the clamp. This allows you to extract
Bend Protection requirements either side of the clamp and look at more complex
tethering arrangements. It also allows the link to be replaced by a line, as in this example.
When you open the file, you will see a series of views of the system. To the left are a shaded side
elevation and a wireframe plan view of the two risers. To the right are wireframe elevation views
of the tether attachment methods. The upper one is Simple, the lower is Detailed.
Make sure the Model Browser is in View by Groups mode; this is set by right-mouse clicking in
the Model Browser and checking that the View by Groups option is ticked.
As with the preceding examples, this is
not a coupled analysis. Risers in shallow
water have negligible effect on the
motions of an FPSO so a fully coupled
analysis is not required.
Instead the vessel motions are applied
through displacement RAOs and moorings
need only be included if interference is of
interest.
The plan view in the OrcaFlex model
shows the vessel offset out of the riser
plane. This represents the mean and 2
nd

order offsets due to the environmental loads, which are applied as the vessel initial position.
The 1
st
order motions are then applied in the dynamic simulation using vessel displacement
RAOs. These oscillations are applied about the initial position.
The risers extend from the vessel to the seabed in a wave formation and have an additional tether
restraint near the touchdown. The risers extend to port and starboard of the vessel, and so will
experience the same loading in the head sea conditions applied here. This will allow easy
comparison.
The riser hog bend has distributed buoyancy properties in this example but discrete modules
could be used if preferred. See the Lazy Wave example for details of this.
W Orcina

X Y
Z
0.9 m
OrcaFlex 9.4a60: Pliant WaveRiser.sim(modified13:17on28/06/2010by OrcaFlex 9.4a60) (azimuth=270; elevation=1)
Time: 35.0000s
Open the 10 Pliant Simple Line Data Form via the Model Browser. In this example there are
two smeared buoyancy sections, 10P+SmallBuoys which is shown in red in the view and
10P+BigBuoys which is shown in green.
It is a common practice to vary the size and pitch of modules along the riser length. Smaller
modules closer to the hang-off can reduce the risk of compression ahead of the modules where
the increased drag and inertia of the buoyed length means it cannot fall through the water as
rapidly as the bare length. Larger modules closer to the seabed termination or tether (as in this
example) increase the mean tension and reduce the risk of compression at the termination or the
tether going slack.
Go to the Line Types Data Form via the button at the bottom left of the Line Data Form. Select
either of the buoyed line types and select the Wizard button at the top right of the form.
You will not see the full Line Type Wizard settings unless you reset the simulation but the
summary form you will see is sufficient to show what the buoyancy module settings were. The
two types have the same pitch but the small modules have diameter and length of 1.2m and the
big modules have diameter and length of 1.45m.
Close the Wizard and look at the Drawing page of the Line Types Data Form. Each line type has
a different colour and pen thickness so they can be identified easily in the view.
The components of the simple model are contained
in the Simple Tether group. This set extends to
port. The attachment is shown in the figure to the
right.
The riser is a continuous line of 232.5m length and
the clamp is a clump attached at a node 176m along
it. See the Attachments page of the 10 Pliant
Simple Line Data Form to see this location. Section
4 of the riser stops at this arc length to ensure there is
a node at that precise location.
Open the Attachment Types Data Form via the Attachment Types button at the bottom left of
the Line Data Form to see the clamp settings. Physical properties are for those of the steelwork
alone. As for the buoyancy modules discussed in the Lazy Wave example, the clump attachment
should only contain loads that are additional to the risers loads.
Hydrodynamic properties for Clamp Type have not been set in this example because they rarely
have a significant effect on the response of the system.
The tether is modelled as a 9.9m long link and attached at
the same node as the clamp. Select Tether Pliant Simple
from the Model Browser to open the Link Data Form. The
link type is Tether so it will apply tension when stretched,
but will go slack rather than enter compression.
This arrangement applies clamp and tether loads at a single
point, the node they are attached to. Curvatures very local to
this location will have lower accuracy.
If you are interested in the distribution of the local load and
identification of any bend protection requirements then a
W Orcina

more detailed model might be required. An example of this is the Detailed Tether group. The
attachment is shown in the figure to the left.
The riser is split into two separate lines, above and below the clamp. They are connected via a 6D
Buoy. The end connections are built-in so that moments are transferred from one riser length to
the other via the buoy.
This buoy represents the clamp and so has the properties of the clamp, the riser within it and also
any riser contents. Because the clamp is 1m long, each riser length has been shortened by 0.5m so
the total length is still 232.5m. Hydrodynamic properties assigned to the buoy exclude any
contribution from the clamp itself, but include contributions from the riser to allow comparison
with the simpler Clump method above.
You could model the clamp as a line type but the addition of a 6D buoy allows convenient
extraction of bend protection loads either side and allows the link to be replaced by a line.
In the detailed model the tether is now a line Tether Pliant Detailed. Its connection is offset
from the centreline of the riser by 0.2m so its length has been shortened by the same amount. The
tether properties were also produced by the Line Type Wizard but this time the Special
Category was Rope/Wire.
Look at the Limits page of the Line Types Data Form. You will notice that Compression is
limited has been ticked for the line type Tether. This means that the line will go slack rather
than enter compression, just like the link. The tether axial stiffness also matches the stiffness of
the link used in the Simple version.
The Detailed version therefore allows assessment of the resulting loads on the clamp and the
effectiveness of any intended bend protection. Bend stiffeners can be modelled using either of the
methods shown in the Steep Riser example.
4.2. Results
The views compare the two modelling methods. The detailed views to the right are of the simple
method (top) and the detailed method (below). The resulting configurations are very similar.
An efficient model of a system will contain details critical to the system response but simplify or
exclude details that have negligible effect. In this example, the simple clamp had no
hydrodynamic properties while the detailed clamp included them. The difference in results is
negligible, so clamp hydrodynamics could be excluded.
Now load A01 Pliant Wave Riser Curvature.wrk from the Workspace dropdown menu. It will
present four curvature range plots. On the left are the results for the length above the clamp. The
upper plot is for the simple configuration, the lower plot for the detailed configuration.
To make the axes match double click on the upper plot to access the Graph Properties Form. Set
the Y axis maximum value to 0.12 on the Axes page then select OK. This shows there is slightly
more curvature in the detailed model at the clamp but other differences are small.
The right hand plots are for the length below the clamp. There is a slight difference in curvature
in the touchdown region, 10m to 30m on the detailed plot, but again this is small.

W Orcina

Production Risers: A02 Midwater Arch Systems Page 1 of 16
A02 Midwater Arch Systems
Introduction
In this example, three types of riser system are shown: a lazy S (both simple and detailed), a
steep S and a pliant S. An S configuration is similar to a catenary but has support provided at
about midwater by an arch structure.
On opening each simulation file, the default Workspace will present views in wireframe and
shaded graphics of the system.
The examples also show:
Modelling of midwater arches.
Contact between Lines and with Shapes, including friction.
Modelling bridles and tethers.
Using Winches to assist statics.
Setting the model north pointer.
W Orcina

1. Lazy S Simple
A lazy S configuration is similar to a catenary but has additional support provided at about
midwater by an arch structure. 'Lazy' means that the riser centreline is near parallel with the
seabed on contact while 'S' describes the line shape as a result of the arch.
When you open the simulation file you will see four shaded views. These show the midwater
arch in plan view, side and elevation views as well as a side elevation of the whole system.

This example has three risers and an umbilical extending from a vessel to the seabed. All four
lines pass over the same midwater arch.
There are a range of arch designs used in the industry. This example represents a structure with
two buoyancy floats with four gutters passing over the top. The example shows the basic
principles so the analyst can apply them to alternative arch structures as well.

Make sure the Model Browser is in View by Groups, set via the Model Browser View
dropdown menu.

W Orcina

X
Y
20 m
N
W
S
E
Or caFlex 9.4a63: Lazy SSimple.sim (modified 16:39 on 01/07/2010 by Or caFlex 9.4a63) ( azimuth=270; elevation=90)
Time: 60.0000s
The arch plan view shows a compass.
The settings for this can be found on
the Drawing page of the General Data
Form. Because there are a significant
number of similar objects in this
model they are identified relative to
these compass directions.
The plan view also shows the weather
directions. Current is out of the riser
plane and waves are 45 off current.
The vessel is head to waves with an
offset collinear to waves. This offset
is applied by moving the vessel initial
position.
The flexibles have each been split into two lines.
Upper catenary from hangoff to arch clamp.
Lower catenary from arch clamp to seabed.
Neighbouring risers have also been given different upper catenary lengths to reduce the risk of
contact when loading is out of the riser plane.
Correct modelling of the line end restraints at the arch is very important. The arch clamps will
transfer moments between the flexibles and the arch. If a pinned connection is set then this
transfer cannot occur directly and the resulting system response may be incorrect. Therefore a
built-in connection is recommended. For details on built-in connections and no-moment
directions see the Simple Catenary discussion in A01 Catenary and Wave Systems.

The arch itself is built up of several objects.
Arch. A 6D lumped buoy that provides the physical and hydrodynamic properties.
Vertical Support. A group of shapes that provide vertical restraint to the flexibles.
Horizontal Support. A group of shapes that provide horizontal restraint to the flexibles.
Arch Drawings. A group of shapes that provide the image of the arch structure.

Arch is a 6D (6 degrees of freedom) lumped buoy that contains the physical and hydrodynamic
properties of the whole arch structure. With this option OrcaFlex requires you to determine these
overall properties.
Open the Arch 6D Buoy Data Form and look at the properties page. The buoy origin is at the
top of the arch (clamp height) and on the centreline. This is a convenient location when attaching
flexibles because their location is typically given relative to this point. The Centre of Mass and
Centre of Volume have therefore been specified relative to this location too.
W Orcina


Figure 1-1: Arch Local Axes
The buoy rotational hydrodynamic terms have not been set. For a large, well restrained, structure
such as an arch the rotational terms do not have a significant effect on the system response.
Therefore in this example they have not been included.
Note that the actual mass and volume of the arch have been given. It is important to apply the
correct values because dynamic analysis uses them for physical and hydrodynamic inertia as well
as net buoyancy.
If the structure will trap water as it moves then this trapped water should be included in the total
mass and volume of the structure so that the physical and hydrodynamic inertia are correct.
Also note that arches typically have complex shapes and overall hydrodynamic properties are
therefore uncertain. It is recommended that sensitivity studies are carried out on the
hydrodynamic terms used. This should include the rotational terms if it is suspected they are
significant for a particular system.

The Arch and most of the Support shapes are hidden in the views. What you are seeing is a
group of shapes in the Arch Drawings group that give the appearance of the actual structure.
These Elastic Solid type shapes have zero stiffness so are for visualisation only. This is a useful
feature when presenting to clients who may not be analysts themselves. It allows them to see the
model in a way they are familiar with.
To see any hidden object right click on it in the Model Browser then select Show from the
dropdown menu. The same process but selecting Hide will make it invisible again. Note however
that Hide/Show only affects the views. The object is still present in the mathematical model and
will still affect the system.
The views below show the hidden support structure as it is built up.
The arch gutters that the flexibles pass through are modelling in this example by shapes. Shapes
interact with the nodes of lines and so line segmentation is refined in the region of shape contact.
The Model Browser group Vertical Support contains shapes representing the bottom of the
gutter. These are Elastic Solid shapes with stiffness so will produce a reaction force.
Gutter W (the left hand view) provides the curved lower catenary support. Gutter E (the right
hand view) provides the curved upper catenary support. Clamp provides the flat part in
between.
x
x
z
y
W Orcina


Figure 1-2: Arch Vertical Support Components
Because the gutter bottoms have the same radius this group of shapes has been extended beneath
all the gutters instead of a group being required for each one.

Figure 1-3: Arch Vertical Support
Friction is applied between the flexibles and these shapes. Friction between lines and shapes is
not set on the Line Types Data Form but instead on the Solid Friction Coefficients Data Form that
can be accessed via the Model Browser. Note the shape-line friction is only active in dynamics.

A very simple arch model might contain just these vertical supports. The risers are able to roll on
and off the arch but the only lateral restraints while on the arch would be friction and the bend
stiffness of the riser. Riser curvature at the clamp location would therefore be highly
conservative.
Also all lateral moments would be transferred at the clamp connection only rather than distributed
along the gutter wall and so produce errors in the yaw response of the arch. Therefore it is
recommended that the gutter walls are represented in some form.
The Model Browser group Horizontal Support contains shapes to represent the walls. The walls
in this example are very simple but should be sufficient for a design iteration stage. Each wall has
its own Model Browser group and contains shapes for the upper catenary, lower catenary and
clamp regions.
Note the curved walls are made from Curved Plate shapes with a sweep of about 130. Having a
greater radius than the vertical supports, if they swept a full 360 to make a cylinder then they
might interact with the tethers. This could provide additional, unrealistic restraint to the arch.
W Orcina


Figure 1-4: Arch Horizontal Support
The support structures combine to produce the following.

Figure 1-5: Arch Horizontal and Vertical Supports
The shapes in the support group generate effective boundaries for the movement of the lines.
They are not intended to represent specific pieces of kit. For example the bottoms of the gutters
might be made of pipework or plate steel. These would be quite thin structures and yet the shapes
here have significant thickness.
Although the actual structure may have thin walls it is not advisable to model them as such. If a
shape is say 2mm thick then a node entering slightly more than 1mm will be pushed to the other
side of the shape. It is better to make the shape as thick as you can without interfering with other
lines or buoys. The nodes would then have to enter further into the shape to end up on the wrong
side.
The result is a more robust system for loading applied in a range of directions. To see an
alternative method that does use thinner shapes see the Lazy S Detailed example later in this
document.
The arch is restrained by bridles and tethers. These are contained in the Tethers and Bridles
group. The bridles are built as lines attached to the arch at their top connections (End A). The
tethers, also lines, are anchored onto the seabed at their bottom connections (End B).
The tethers and bridles are connected to each other via 3D buoys. A 3D buoy has 3 degrees of
freedom, translation but no rotation. It can be used to connect lines that have no bend stiffness
because rotational moments do not need to be transferred from one side to the other. The 3D
buoys in this example have been given negligible properties because their function is just to
connect three lines together.
W Orcina

If you need to connect lines that do have bend stiffness see the Pliant Wave discussion in A01
Catenary and Wave Systems.
Shapes with zero stiffness represent the gravity base and the tether posts for visualisation only.

The bridles and tethers in this example have been assumed chain. Actual chain data should be
used but if this is yet to be decided then Generic chain properties can been produced using the
Line Type Wizard routine. If you reset the simulation file you can then see the Wizard in action.
Select '2" chain' on the Line Types Data Form then click on Wizard at the top right of the
form. You are now in the Line Type Wizard routine. Step through it to see how the '2" Chain'
properties were built.
Note that, unlike the risers, the chain properties have 'Compression is limited' on the Limits page
of the Line Types Form turned on. This is because chains go slack rather than resisting
compression.
1.2. Results
Load the Lazy S Results.wrk workspace via the Workspace dropdown menu. This will present
four time histories.
The top left shows the Arch Surge (X) motion. This has settled into a cyclic response. However
the Arch Yaw (Rotation 3) shows a more complex response with a double oscillation per wave
instead of the single oscillation seen in the Arch Surge.
This is not unusual because the Arch is responding to transverse loads from the flexible upper and
lower catenaries. These have different configurations in the water column so will have different
load magnitudes and phases.
This complex response is seen in the Arch Pitch (Rotation 2) at the top right and the tether tension
at anchor at bottom right. These both show a slight rise in response at the time the Arch sees its
second yaw oscillation.
Now take a look at the Detailed example in the following section.
W Orcina

2. Lazy S Detailed
The more detailed version of the Lazy S system has the following additional complexity.
Arch properties from a composite structure.
Refined gutter wall modelling.

Discussion of the overall system construction is contained in the Lazy S Simple section earlier in
this document. Only the changes to the system will be discussed here.
Open the Simple model in a separate copy of OrcaFlex on the same machine. This will allow
comparison of the two modelling methods.
Make sure your Model Browser is in Group View. It can be set via the Model Browser View
dropdown menu.

The physical and hydrodynamic properties of the arch can be produced by a single lumped buoy
with the overall properties calculated and specified by the user. Alternatively components of the
arch can be produced by other objects attached to a common buoy. The user specifies the
properties for these components and OrcaFlex calculates the overall properties.
W Orcina

Adding components to the arch increases the size, complexity and runtime of the model. It is
therefore not recommended for standard load and fatigue analysis. However it can be helpful for
investigation of specific problems.
For example having hydrodynamic loads calculated for each major object means if one end of the
arch experiences very different flow from the other then the resulting moments on the arch will be
captured.
Also it can be convenient to have a buoyancy tank modelled as a line if you want to consider
some cases with the tank flooded. This saves having to replace the overall buoy properties for
empty with those for flooded. All you have to do is change the line contents density.
In this example only the two buoyancy tanks have been separated out. Each has been modelled as
a series of single segment lines attached at both ends to the 6D buoy. Being single segment lines
the stiffness properties are not important. The lines are in the Tanks group of Arch Structure.
The advantage of using a series of single segment lines is that loads are calculated for each line
and applied at its end connections, distributing the loads along the tank length. This will capture
any resulting moments from change of flow across the length.

Figure 2-1: Arch as 6D Buoy and Two Lines
On the LineTypes Data Form you can see the properties of the tanks in Buoyancy Tank type. It
is important to be careful setting the line Cd values. Lines use the surface area in the axial drag
formula, not the end area. The line axial Cd therefore needs to be calculated so it gives the correct
resultant drag.
In this example the two tanks are also side by side so one will shield the other when flow is along
the Arch local x axis. However OrcaFlex does not automatically calculate shielding so both tanks
will experience undisturbed drag when in reality the downstream tank will see reduced flow. The
options for obtaining the correct horizontal normal drag are therefore:
Divide the horizontal drag between the two tanks by halving the horizontal flow Cd, so
non-isotropic normal Cd. Resulting pitch moments will not be exactly right because both
tanks see drag but the differences will be typically small and the arch is unlikely to be
sensitive to them.
Turn wake effects on so the downstream tank sees reduced flow. However wake
reduction will not include the effects of neighbouring structures such as supports and
gutters and again there are more calculations so runs are slower.
The first of these has been applied in this model.
W Orcina

In the vertical direction (line local x) both tanks see full flow so Normal Cd is 0.7 the full amount.
In the horizontal direction (line local y) the normal Cd has been halved to 0.35 so each tank will
produce half the drag.
Open the Buoyancy Tank East 1 Line Data Form via the Model Browser. Note that Torsion has
been turned on and the ends connections are built in (stiffness of infinity). This is not necessary
for a single segment line because there can be no rotation along its length. However it does
remove the warning message that is produced when non-isotropic properties are applied and
torsion is turned off.
The remainder of the arch properties are then calculated and applied to the 6D buoy so the
resultant physical and hydrodynamic properties match those used in the simple model.
Compare the data for the Arch buoy in the two models. Taking away the tank contributions has
changed the volume, mass, moments of inertia, centre of mass and volume and the drag areas.
Note the line objects have been hidden from view and shapes with zero stiffness have been used
for visualisation in this example. For more information on hiding and showing objects see the
Lazy S Simple model discussion earlier in this document.

The detailed model views show the gutter walls modelled in two different methods.
The 10 S Gutter group within Horizontal Support has walls created by short, single segment
lines representing the struts of the gutter. Line on line contact is then turned on for these struts
and the riser so they can interact with each other.

Figure 2-2: Gutter Walls with Line on Line Contact
Look at the Structures page of the 10 Up S Line Data Form. Line on line contact is
computationally intensive so Clash Check has only been turned on for sections that are
expected to interact with the struts.
The line properties are Wall Dummy Type and can be found in the Line Type Data Form. Note
the type has negligible properties so the struts have little effect on the arch response but there is a
significant contact diameter (Contact page) for clashing. If you want to give actual properties for
these struts then remember to remove their contribution from the Arch properties.
Contact stiffness is set high enough to stop one line pulling through the other but not so high to
cause a near rigid stop of the line that would produce rattle. OrcaFlex is a global analysis package
so the contact is to get the correct response, not to model local deformation from contact. See
Theory | Line Theory | Clashing in OrcaFlex help for more information on this.
W Orcina

This wall modelling method allows detailed control of the gutter wall shape and you can extend
this so the gutter base and edges. However this method does increase the complexity of the model
resulting in larger files and slower runs. It is best suited for checking as-built structures to see if
risers bend around gutter edges or if fatigue issues arise from point loading from the gutter struts.
For the remaining three gutters the walls are produced by shapes. Each wall has two curved
plates and a rectangular block. Because these closely resemble the appearance of the actual
structures they have not been hidden so the view shows the shapes

Figure 2-3: Gutter Walls with Line on Shape Contact
If the gutter walls have a constant radius in plan view then this is a convenient way of modelling
them. There is improved curvature and moment modelling without a penalty on file size and
runtime when compared to the simple model.
However the shapes need to be thin or they will protrude through the neighbouring gutter walls
and interact with the wrong lines. As discussed with the simple model earlier, this increases the
risk of the line nodes being pushed to the outer face of the wall instead of the inner one. A higher
shape stiffness is typically needed to reduce the amount the node will sink into the shape.
The greatest risk of nodes passing to the wrong side is during the statics convergence search when
the software is seeking line configurations that provide a load balance. During the search nodes
can move inside the shapes and so be pushed to the wrong face.
Line on line contact also has a problem in that it is not present in statics so the static search can
find a solution outside the gutter.
Both problems have been solved here by the same method. Temporary winches have been
attached to the risers and umbilical to hold them inside the gutters during the static search. You
can find them inside the group for each flexible.
A link could be used instead but it is important to make sure in either case that the tensions are
zero before releasing them. Otherwise there will be a step change in load. This could cause
transients in the system response and slow down runs as the implicit timestep algorithm works to
follow the rapid change. It is easier to produce a zero tension at release with a winch because it
can pay out line.
In this example the winches pay out during the build-up period (Stage 0) and then released at the
start of Stage 1. Open the Winch Data Form for 10 up N Hold. This shows Release at Start of
Stage is set to 1 for Stage 1. Also the winch pays out 1m of wire over Stage 0 so it becomes
slack as the riser comes to rest against the gutter wall.
W Orcina

2.2. Results
Load the Lazy S Results.wrk workspace via the Workspace dropdown menu. Make sure this is
also loaded for the Lazy S Simple model so the results can be compared.
Some of the plots below have matching axes. To do this in OrcaFlex, double click on each graph
to access and modify the plot parameters. Note all Simple model plots are on the left and Detailed
model plots are on the right.

The plots below show variation of tension at the bottom of a tether with time. The plot axes have
the same scales. Tension variation is reduced in the more detailed model. The minimum tension is
now 10te instead of 6te.

Figure 2-4: Tether Tension: Simple (Left) and Detailed (Right)

The plots of Arch Yaw (Rotation 3) time history have the same range but the simple model has a
negative range, the Detailed model a positive range. The simple model has a yaw variation of
between 179 and 181 (-181 and -179). The detailed model has a yaw variation between 177 and
183. This is an increase from 1 to 3. The increased yaw is due to the more accurate
modelling of the moments on the arch from the riser contact with the wall.
Figure 2-5: Arch Yaw: Simple (Left) and Detailed (Right)

OrcaFlex9.4a66: LazyS Simple2.sim(modified 15:58 on 05/07/2010 byOrcaFlex9.4a66)
Time History: ArchTie_S Effective Tension at End B
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
T
i e
_
S
E
f f e
c
t i v
e
T
e
n
s
i o
n
(
k
N
)
a
t E
n
d
B
300
250
200
150
100
50
OrcaFlex9.4a66: LazyS Detailed2.sim(modified 15:58 on 05/07/2010 byOrcaFlex9.4a66)
Time History: ArchTie_S Effective Tension at End B
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
T
i e
_
S
E
f f e
c
t i v
e
T
e
n
s
i o
n
(
k
N
)
a
t E
n
d
B
300
250
200
150
100
50
Time History: Arch Rotation 3
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
R
o
t a
t i o
n
3
(
d
e
g
)
-176
-177
-178
-179
-180
-181
-182
-183
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
R
o
t a
t i o
n
3
(
d
e
g
)
183
182
181
180
179
178
177
176
W Orcina

Arch systems have complex motions so a single change can vary more than one response. An
example of this is shown in the plots below. The detailed model reduces the Arch Pitch variation
(Rotation 2) and shifts the mean Arch Surge (X).
Figure 2-6: Arch Pitch: Simple and Detailed

Figure 2-7: Arch Surge: Simple (Left) and Detailed (Right)
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
R
o
t a
t i o
n
2
(
d
e
g
)
2
0
-2
-4
-6
-8
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
R
o
t a
t i o
n
2
(
d
e
g
)
2
0
-2
-4
-6
-8
Time History: Arch X
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
X
(
m
)
-100
-101
-102
-103
-104
-105
-106
Time History: Arch X
Time (s)
60 50 40 30 20 10 0 -10
A
r
c
h
X
(
m
)
-100
-101
-102
-103
-104
-105
-106
W Orcina

3. Steep S Riser
The Steep S system here has a similar
arrangement to that seen in the Lazy S
Simple example earlier.
However, instead of the lower catenary
descending and lying on the seabed it
terminates at a near vertical
connection.
Steep means that the riser centreline
is near vertical at the lowest end while
'S' describes the line shape as a result
of the Arch.

This example has two risers and an umbilical descending from an FPSO to the seabed. The
central umbilical has a longer upper catenary to keep it out of the plane of the risers and reduce
the risk of contact.
When you open the simulation file you will see two views of the system and two plots. The
upper view is a wireframe plan of the system. Note the North pointer in this case has been chosen
to be in the riser plane and not aligned with a Global axis. The lower view is a shaded elevation of
the system.
dropdown menu.
Details of Arch construction have been discussed in the earlier Lazy S systems. Therefore only
the seabed termination will be considered here. The lower catenary is vertical rather than
horizontal when it reaches the Arch Gravity Base. The Gravity Base is included in the model
as a shape with zero stiffness and is for visualisation only.
End B of each flexible lower catenary is anchored 5m above the seabed so on top of the arch
gravity base. Double click on any of the three Lower lines in the Model Browser to see the
settings.
The line End B declination in this example is 170, indicating the line is heading down into an
end connection that is 10 (180-170) from vertical. The azimuth is 10 because anchored
connection orientations are relative to the Global axes directions. The riser system is on a heading
10 anti-clockwise from the Global X direction.
The connection stiffness is infinity so it is built in. For more information on end connection
settings see the A01 Catenary and Wave Systems | Simple Catenary example.
The lower catenary lengths have to be optimised carefully. If the flexibles are too slack then they
could enter compression as the arch moves. If too taut then they could become the tethers for the
arch system. Typically they are tauter than seen in the Lazy systems.
W Orcina

A taut line will want to descend vertically during the static search. It will resist the shape restraint
more than in the Lazy system. This means more care is usually needed to get an acceptable static
convergence. It is easier for the static search to find a solution where the line is on the wrong side
of the shape, wrong being the one you dont want.
Temporary restraints can assist the convergence. Take a look at the Lazy S Detailed discussion
earlier. Alternatively a spline could be used to approximately define the required line shape. This
is selected in the Statics table on the Line Data Form. Its settings are specified on the Spline
Starting Shape page of the same form.
The spline does not need to be very precise because the static analysis will refine the line
position. It just needs to be enough to direct the line to the required side of the shape at the start
of the static analysis.
Reset the file by selecting Reset in the Calculation dropdown window. The splines will appear as
grey lines. The control points will appear as crosses.
Control points can be dragged by holding down the left mouse button. The control points will
move in the plane of the view and the spline will adjust accordingly. Remember to check the
result in plan as well as elevation view.
If the lines leap around during the static convergence then the search may need a bit more
damping. Look at the Full Statics Convergence page of the Line Data Form. Min Damping
has been increased from 1.0 to 5.0 while Max Damping has been increased from 10 to 50. This
will reduce overshoot but can make the convergence slower.
For more assistance with Statics convergence issues contact us for the Static Convergence
Guide knowledgebase article.

3.2. Results
If the model was Reset then reload the simulation file. Two instantaneous range graphs are
presented to the right of the views. The upper plot is for 10 Upper East riser and the lower is
for 10 Lower East riser. They show the variation of effective tension along the length at any
instant in time.
Run the replay (Replay dropdown menu) and the tension distributions will also vary, matching
what is happening in the views. The tension varies steadily in this case.
These instantaneous range graphs are also helpful in identifying where a line response originated.
They will also show axial waves in the line that might not be observed with other plots.
W Orcina

4. Pliant S Riser
A pliant S configuration is a lazy S with the addition of a tether restraining the touchdown point.
This example is the same as the Steep S but the lower termination has been changed to Lazy with
restraining tethers.
When you open this file you will see a shaded elevation view of the arch arrangement.
dropdown menu.

The line is built in the same manner as for the
previous S models. Therefore only the
termination change will be discussed
The line is made pliant by the addition of a tether
close to touchdown. In this example the tether
has been modelled as an OrcaFlex link attached
to the riser. No clamp details have been included.
This is the simplest way of modelling the tether
and is sufficient for most cases. For discussion of
the alternatives see the documentation for A01
Catenary and Wave Systems | Pliant Wave
example.
Again splines have been used to assist in static
convergence. See the Steep S discussion earlier
for details.

4.2. Results
Two plots are shown. The top right is a tension range graph for the Umbilical lower catenary,
Umb Lower. This shows the effective tension variation along the length through the final wave
cycle. The tension step change at about 50m arc length is where the tether is attached.
The lower right is a time history of arch motion in the Global Y direction. It shows a settled
cyclic response has been achieved.

W Orcina

Production Risers: A03 Jumper to High Tower Page 1 of 4
A03 Production Risers: Jumper to High
Tower
1. Introduction

Flexible jumpers connect a spread-moored FPSO to a high riser tower in deep water, i.e. the
tower top is close to the surface. The lines descend to the seabed as rigid structures clamped
around the tower. An installation of this type can be found in the Girassol field.
As this is a deep water system, the tower top will flex under the weather and jumper loads. The
tower has therefore been modelled as a very stiff line with one end fixed to the seabed, i.e. a
vertical cantilever. Tower motion depends on loads from all the jumpers as well as the waves and
current, so all the jumpers need to be included in the model.
If the tower motions are pre-defined, it could be modelled as a vessel with the motions specified
as harmonic motions or time histories. This would mean that individual jumpers could be
analysed, if required, as they would have no effect on the modelled motion of the tower.
W Orcina

The default workspace shows a shaded view of the tower top and the jumpers.
2. Building the model
Eight jumpers extend from the vessel side to the
top of the tower. They leave the vessel through
bellmouths inclined at 30 to vertical. The
bellmouths are modelled explicitly using Curved
Plate shape objects. Contact between the jumper
and the bellmouth takes account of the jumper
diameter.
Bellmouth thickness is artificially increased to
reduce the risk of the jumper cutting through the
solid and escaping. Nodes push to the nearest face
of the shape so you want the outer surface to be
well out of the way so that line nodes dont get
pushed to the wrong surface. Some
experimentation may be needed to find acceptable
values of thickness and stiffness (very high
stiffness is effective in constraining the jumper,
but may require a reduced integration timestep for
stability).
Jumper segmentation is chosen so that up to 10
nodes can make contact with the bellmouth.
Again, some experimentation is advisable here to ensure that the modelling is adequate.
Additional bellmouth shapes are superimposed on the hollow ones. These extra shapes fill in the
hollow and have no stiffness. To fill the hollow of the bellmouth simply copy the original shape
in the Model Browser and changed the curved plate setting from hollow to solid. Then set the
shape type to trapped water so it shields the jumper length within the bellmouth from direct
wave and current loading. See Help section | Theory | Shape Theory | for more details.
The riser tower is a slender vertical cantilever supported by a float at the upper end. We model it
as a Line with a 6D Buoy at the top to represent the float (lines cannot be directly connected to
each other so an intermediate buoy is required. In this case we use a 6D Buoy to allow the riser
top end to rotate under load).
The line representing the tower core is built-in (encastr) at both ends, as are the four production
risers. The risers are clamped to the outside of the tower using the Line Contact model. Open the
Line Contact Data form from the Model Browser. Two sets of relationships have been defined;
firstly, each riser is clamped to the tower core by means of an Around relationship, which
spaces clamps along the towers length at locations defined on the Penetrator Locations page.
This page also defines how far offset from the tower centreline the risers are held (1m in this
case), as well as the contact diameter and area of each clamp. The second set of line contact
relationships specifies that each risers outer diameter can (if the environmental conditions are
severe enough) contact the outer diameter of the tower core i.e. Outside relationships. For more
details of the available line contact relationships, see Help | System Modelling: Data and Results |
Lines | Line Contact | Modelling.
W Orcina

Close this data form and open the Workspace A03 Jumper to High Tower Clamps.wrk to see
one set of clamps in the shaded graphics mode.
One of the Production Jumpers is modelled with slugs of gas flowing along the line. Open the
Line Data Form for the Product S Out line and click on the Contents page. Here, ten slugs of
gas, each 80m long, have been set up to flow along the line from the tower towards the vessel.
The flow velocity has been set up using a variable data item called Start Up. Right-mouse click
anywhere on the form to select Edit Variable Data to open up the data form for this. Start-up
conditions are modelled with the flow velocity being 0m/s initially, increasing up to a constant
10m/s during the simulation (this is deliberately set to a high value so that the effects of the slugs
moving through the line can actually be seen in the model). Note that the negative velocity
indicates that the flow is moving from End B towards End A of the line; a positive velocity would
indicate the flow is moving in the opposite direction.
In this example, the current profile (velocity variation with depth) is applied as a 7
th
power law
rather than interpolating between specified depths. Take a look on the | Environment | Current |
data page. The model contains both a Swell and a Wind Wave, therefore the model has been run
for a minimum duration of Swell Period * Wind Wave Period (16s*9s=144s).
All the lines have the normal drag coefficient specified as a relationship with Reynolds Number.
Note this relationship is a generic form and you should check what relationship is appropriate for
your lines.
The vessel is using the default data for a 103m vessel. This has then been Froude scaled for a
130m long vessel. The original vessel length is specified in the Structure tab of the Vessel Type
page. The shaded drawing is also scaled by setting it to 103m draw size on the Shaded Drawing
Tab. The required length of 130m is specified on the Vessel page.
Beware: Use Froude scaling with caution. The method has many limitations and is intended for
use in research studies or pre-bid analysis. It should not be used on an actual project. See the Help
| Theory | Vessels | Vessel Data for details.

3. Results
Press Ctrl +T to return to the default view. Look at the animation through Stage 1. Dynamic
movements of the tower are small. The main dynamic loading occurs at the FPSO end of the
jumpers and is caused mainly by roll motion of the FPSO.
Load the A03 Jumper to High Tower Bellmouth.wrk workspace. This shows the balcony region
in shaded graphics and an instantaneous range graph of y-curvature for one of the jumpers. It
shows the section closest to the bellmouth (arc length 0m to 10m). Restart the replay and observe
the reverse curvature through the wave cycle.
The bellmouth covers an arc length of about 3m and is designed to restrict curvature to about 0.2
rad/m (5m bend radius). Note in this model, the jumpers are pinned at the inner end of the
bellmouth (Arc Length = 0) so the curvature here is zero.
Load the A03 Jumper to High Tower Slug Flow.wrk workspace. The graph in the upper right of
the screen shows the instantaneous range graph of the contents density for the line that has been
modelled with slug flow. As the animation plays you can see the gas slugs (lower density)
moving along the jumper. You can also see the effect of the slug movement in the top left hand
W Orcina

view. The blue line is the one that contains the slug flow; you can clearly see it jump in response
to the slugs of gas moving along the line.
W Orcina

Production Risers: A04 Disconnecting Turret System Page 1 of 6
A04 Disconnecting Turret System
Introduction

This model represents an FPSO system with a releasable turret. Disconnecting systems are found
in regions where typhoons or ice flows are common. It allows the vessel to be removed until
conditions improve.
The turret can be lowered from the vessel with winches. However, downtime is expensive so
operators prefer to keep producing until the last moment then drop the turret and drive off. The
turret drops below the severe weather until the FPSO returns to reconnect.
The turret system needs to be designed so there are acceptable responses in all three phases.
Connected. The turret is attached to the vessel and moves with it.
Disconnecting. The risers are disconnected from topside and the turret is released.
Disconnected. The turret is settled deeper in the water.
This example shows the disconnecting phase. The turret is released then the FPSO moves off
station. When you open this file you will see a shaded view of the system.
Make sure the Model Browser is in View by Groups, set via the Model Browser View
dropdown menu.
W Orcina

Building the model
Open the Environment Data Form via the Model Browser and move to the Current Page. This
current profile has been set to change direction through the water column. The Rotations (deg)
column in the Profile table shows the deviations from the specified Surface Direction of 0.
View Vertical Profile 3D View shows this rotation. You can control this view in the same way
as you can the standard view windows. Click on the arrows at the top left to rotate the view
around and see the different headings.
The Model Browser shows the system in two groups, Vessel System and Disconnecting
System. Select Disconnecting System in the Model Browser. This contains the structures
making up the Turret as well as the moorings, risers and umbilicals.
Open Turret from the Model Browser. It is a 6D buoy with a Spar Buoy type that gives the
physical and hydrodynamic properties of the real turret buoy. Note that these values need to be as
accurate as possible to effectively model this system when the buoy is released from the vessel.
The model has been set up with the turret origin at its base. Right click anywhere on the 6D Buoy
Data Form and select Properties from the dropdown list. This shows the Centre of Mass is 6m
below the Centre of Volume. It is a common practice in this kind of system to keep the Centre of
Mass low down on the Turret to improve its pitch stability when disconnected.
This example has been set up so clearance information can be obtained between the Turret and its
Protective Cage during the disconnection. Any reaction force from contact between the two will
also be captured.
Clearance information is available between lines, so a single line, Shell of Turret, has been
attached to the buoy at both ends. The buoy is 24m long. However the line representing its shape
is specified as 14m long. This is because lines are treated as having hemispherical ends for the
purposes of contact and clearance calculations. Adding the contact radius hemispheres at both
ends gives a length that matches the turret for contact purposes.

Figure 1: Turret (Left) and Shell of Turret (Right)
The line is required to supply contact information but not physical or hydrodynamic properties.
To see how this has been done open the line type Turret Contact from the Model Browser. The
Geometry and Mass page on the Line Type Data Form shows the outer diameter and mass are
negligible. The Drag, Lift & Added Mass page shows the hydrodynamic coefficients have also
been set to zero.
W Orcina

Now look at the Contact page. The contact diameter has been set to match the outer diameter of
the buoy. See the Lazy S Detailed Example in A02 Midwater Arch Systems for more
discussion of Line on Line contact.
The turret will be held up against a bearing plate. This is modelled by the buoy vertices
contacting a rectangular solid, Turret Bearing Plate, in the Vessel System group. The buoy
vertices therefore need to be given the correct contact area.

Figure 2: Turret and Turret Bearing Plate (Left) and Turret Lines and Vertices (Right)
Each cylinder is considered as a rectangular box for contact purposes. Go to the Drawing page of
the Turret 6D Buoy Data Form. Select Draw square cylinders to see these boxes in the view.
To see them clearly select the wireframe view by un-checking the Shaded Graphics Mode option
in the View dropdown menu. Then hide all objects except Turret, by right clicking on the
Model Browser and selecting Hide. With three cylinders in the stack this means there are 16
vertices with 4 on the top face. The buoy contact area is distributed evenly between all of these
vertices.
Go to the Contact page of the Turret 6D Buoy Data Form. The Total Contact Area has been set as
316m
2
. The four vertices at the top therefore give a top area of 79m
2
, the area of the top face.
Solid Friction Coefficients on the Model Browser specifies the friction between the shape and
the buoy. This area can be used for setting shape friction with any combination of lines or buoys.
Note that shape-line friction is only active in dynamics and does not apply to line on line contact.

This example shows one way of modelling the contact and clearances which is appropriate for a
cylindrical turret. However there are alternative disconnecting turret geometries and alternative
ways of modelling the contact with baseplates and guides. A conical turret, for example, could
have a Curved Plate shape attached to the buoy to produce the cone, or rings of lines could be
used.
It is important that all lines attached to the Turret are included in the model or the actual system
response will not be produced. In this example three mooring lines and 6 Lazy Wave Risers are
connected to the bottom of the 6D Buoy. For details of these configurations see the Example Sets
A01 Simple Catenary and Lazy Wave Systems and C03 Turret Moored FPSO.

X
Y
Z
3 m
OrcaFlex 9.4a70: A04 Disconnecting Turret System.dat (modified 14:25 on 13/07/2010 by OrcaFlex 9.3c) (azimuth=264; elevation=15)
Reset
W Orcina

Select the Vessel System group in the Model Browser. This consists of the FPSO, the
moonpool, the protective cage and the restraints for the Turret.
Moonpool is a cylinder shape representing the FPSO moonpool. Because it is a Trapped
Water shape the wave and current effects are suppressed inside it. So the Turret buoy is shielded
while inside the moonpool.
The Moonpool Cage group is the protective
structure around the Turret. It is made of single
segment lines attached to the FPSO. The lines
have clashing turned on so they can produce
contact reaction forces as well as clearance
information for Shell of Turret.
Note the slight gap between the vertical and
sloped lines. This gap has been left so that they
do not contact each other. Therefore any contact
results are only between the Shell of Turret and
its protective cage.
The Moonpool Restraints group contains the objects that hold the Turret inside the Moonpool.
These are the Turret Bearing Plate mentioned earlier and the temporary links that hold the turret
within the moonpool.
The links restrain all degrees of freedom. Open VertLatch in the Model Browser to see the
settings. High stiffnesses are required to restrain the large Turret. To prevent any rattle a high
damping has also been specified.
Release at Start of Stage is set to 2. So at the start of Stage 2 the links will release the Turret,
and then have no more effect on the model. The release has been timed to occur at the wave crest.
It is advisable to also check the release at trough, mid descent and mid rise to find the most
conservative condition.
You can split your analysis into as many stages as you like and these are specified on the
Dynamics page of the General Data Form. For more details see Theory | Dynamic Analysis in
the OrcaFlex help file.
After release of the buoy, the FPSO will move forward. Open FPSO from the Model Browser
and go to the Calculation page. The Primary Motion has been set to Prescribed. Now move to
the Prescribed Motion page. This shows that the vessel is stationary in Stage 0 and Stage 1,
then moves forward with a velocity ramping up from 0m/s to 2m/s during Stage 2 and remains
constant at 2m/s during Stage 3.
The stages can therefore be summarised as shown below.
Stage 0 Build up of Waves -17.5s to 0.0s
Stage 1 Full Waves applied 0.0s to 17.5s
Stage 2 Links released at start and
vessel moves forward with
increasing velocity
17.5s to 22.5s
Stage 3 Vessel moves forward with no
change in velocity
22.5s to 70.0s
W Orcina

At the top right of the FPSO Vessel Data Form the Length has been set as 206m. If this was set as
~ then OrcaFlex would assume the required vessel was the same length as defined on the
Vessel Types Data Form. However in this data results in OrcaFlex scaling the vessel type data to
match the specified vessel length of 206m.
Now select Vessel Types. On the Structures page of the Vessel Type Data Form the Vessel
Type length has been set at 103m, half the length of the required vessel. OrcaFlex will therefore
apply Froude scaling on the vessel data, increasing the response based on the ratio of required
length to vessel type length.
However, this scaling method does not take account of all the response changes with increased
vessel size. It is intended for concept and pre-bid analysis only where the vessel has not yet been
selected. For more details see System Modelling: Data and Results | Vessel Data in the
OrcaFlex help file.
Where the vessel is known then the actual data for that vessel should be used or the response and
results will be wrong.
The turret release is a discrete event occurring during the simulation. This rapid change requires a
small time step to capture the change without the run becoming unstable. However this is an
isolated event and retaining this small time step for the whole simulation would result in a
significantly slower run.
Instead this example uses the Implicit Variable Time Step Integration scheme. This will reduce
the time step when the turret release causes sudden changes and will increase the time step again
when the change is past.
However, the variable time step algorithm should be used with care. Every time a timestep is
changed a small amount of energy is added to the system. This is the nature of variable time
steps. Constant time step should always be the first choice (whether implicit or explicit) and
variable time step should only be used to assist with isolated discrete events where runtime with a
small step is impractical.
Another option is to use the explicit solver, if it yields reasonable runtimes. Explicit solvers have
the advantage that stable simulations are almost always accurate this is not the case for implicit
solvers. If in doubt, it is often wise to perform spot-checks of your implicit solver results against
results produced by the explicit solver.
Results
The view shows an elevation of the system and three time history plots. The upper right plot is
Turret Z position. The lower right plot is the FPSO X motion while the lower left plot is the clash
force from the turret contacting the protective cage.
Run the replay for the whole simulation (Replay dropdown menu). A vertical line on each time
history will show where the replay has got to.
The Turret will drop and then the FPSO will move away to be clear of it. Note the turret in this
system does not bounce back enough to risk hitting the FPSO hull but this is something that
should always be checked for.
Now modify the view so you can see the touchdown region of the left hand mooring line. Run the
replay again and watch how the chain bounces around as it is laid rapidly onto the seabed.
W Orcina

The clash force plot reports that another line contacts the Shell of Turret at about 19 seconds. If
you hide the FPSO and Moonpool (right click on the Model Browser) you will be able to see the
protective cage. Zoom in on the cage and run the replay. The cage lines will change from pink to
white when the Turret contacts them.

W Orcina

Production Risers: A05 Steel Catenary Riser Systems Page 1 of 10
A05 Steel Catenary Riser Systems
Introduction

This example contains three examples of steel catenary risers (SCRs). These are:
Catenary with Spar
Catenary with SemiSub
Lazy Wave with FPSO
The example also illustrates:
User specified units
Homogeneous Pipe properties
Adding Coatings and Linings
Modelling strakes
Tapered stress joints (TSJs)
Flex joints
API RP 2RD Stress Code Check
Vessel drawing techniques

W Orcina

1. Catenary with Spar
A steel riser descends from a Spar to the seabed in a simple catenary configuration transitioning
to a flowline after 7800ft. The riser is hung off the Spar at the soft tank with a Flexjoint and the
upper 4000ft of riser is straked to suppress Vortex Induced Vibration (VIV). Note VIV analysis is
not included in this example.

When you open the simulation file you will see a wireframe elevation view of the system and
three plots. Make sure your Model Browser is in Groups View which can be selected from the
Model Browser View dropdown menu.
Details on constructing a simple catenary configuration are contained in the Catenary Risers
example of the A01 Catenary and Wave Systems set. This discussion will only consider
differences from that example.
The example has been set up with User defined units. Open General via the Model Browser.
OrcaFlex allows you to use SI, US or User defined (customised) units.
If the units are changed, then OrcaFlex converts all the data in the model into the new units. This
is useful if project data is supplied in a range of units. Data can be input in the units supplied then
the model can be converted back to the preferred units.
More information is in System Modelling: Data and Results | General Data of OrcaFlex help.

The model browser shows a Spar Group and a Riser.
The Spar Group contains a vessel object called Spar and a number of shapes. The shapes are
all Elastic Solid type with zero stiffness. They are for visualisation only generating the Spar
appearance that is seen in the wireframe and shaded views. The Spar vessel object is hidden in
all the views. Hide and Show options are accessed by right clicking on the object in the model
browser and selecting from the Dropdown menu.
Now double click on Spar in the Model Browser to access the Vessel Data Form. The vessel
has been offset in the direction of the waves and current. There is also a Trim of 6 with the
weather. The offset and trim represent the mean and 2
nd
order offset contributions from
environmental effects not explicitly calculated in the analysis. For example the wind, wave drift,
current.
X
Z
1000 ft
Or caFlex 9.4a70: Catenar y with Spar.sim (modified 09:38 on 14/07/2010 by OrcaFlex 9.4a70) (azimuth=270; elevation=0)
Replay Time: 45.60s
W Orcina

The Calculation page shows that Static Analysis and Primary Motion are turned off. The vessel
1
st
order dynamic motions are included through displacement RAOs specified by the
Superimposed Motion. These 1
st
order dynamic motions are applied relative to the Spar offset
position.
Select Vessel Types at the top right of the form and the Vessel Type Data Form will open. The
Displacement RAOs page shows data has been supplied for wave incident angles of 0, 45 and
90 only.
Now look at the Conventions page. At the bottom of the page Symmetry has been specified as in
the XZ and YZ planes. These planes are both vertical because the OrcaFlex vessel z axis points
up. OrcaFlex will determine the RAOs for the required heading based on these rules of symmetry.
Now open Riser from the Model Browser. The Riser Line Data Form shows the top connection
(End A) attached to the Spar and the bottom connection (End B) on the seabed.
This example has a Flex
Joint at the End A which is
modelled in two parts.
The Connection Stiffness
table states that End A has
variable bend stiffness
Flexjoint Stiffness. Right
click anywhere on the form
and select Edit Variable
Data from the dropdown
menu.
The Profile button at the
bottom right of the Bend
Connection Stiffness page
(Structure folder) shows the
specified non-linear relationship between Angle Change and Bend Moment. End A has a no-
moment direction 8 from vertically down and pointing towards End B. (Azimuth 0 and
Declination 172). Angle changes are calculated relative to this no-moment direction.
It is important to input the correct no-moment direction and bend connection stiffness relationship
because these will affect the stresses generated in the top of the riser.

Close the Variable Data Form to return to the Line Data Form. The remaining properties of the
Flex joint, both physical and hydrodynamic, are produced by Line Type Flexjoint which is
applied to the top 9ft of the riser.
Move to the Contents page. Contents density and internal pressure have been specified. Both
must be set correctly to obtain the correct wall tension along the riser length.
Move back to the Structure page. The rest of the line is made of Straked, Riser Pipe and
Flowline line types. There are several sections to each line type. These allow the segmentation
to be optimised. More detail is in regions where significant curvature is expected or there is a
change in properties, less detail in the regions where the line has less curvature and is not subject
to significant variation in loading during dynamics.
W Orcina

Select Line Types at the top right of the Line Data Form. Set the Line Type Data Form View
Mode to All then select the Category page. All four line types have been set as Homogeneous
Pipe. This means OrcaFlex will determine the overall properties based on the dimensions and
material properties supplied.
The construction of a Homogeneous Pipe is a load-bearing pipe with non-load-bearing coatings
and linings added. Details can be found in System Modelling: Data and Results | Lines |
Homogeneous Pipe of OrcaFlex help.
The Geometry & Mass page shows the outer and inner diameter of the load bearing pipe and its
material density. The Structure page of the Line Type Form is where Youngs Modulus is
specified so pipe stiffness can be calculated.
The Coatings and Linings page shows the additional layers applied. These will change the
outer and inner diameters and increase the mass. In this example no linings are applied.
The Flexjoint has no coatings while the Riser Pipe and Flowline have about 2.5 of abrasion
protection. However the Straked type has two coatings, the abrasion protection and the Strake.
Therefore it has a variable coating thickness called Coating and Strake applied.
Right click anywhere on the form to access the Variable Data Form. The Coatings or Linings
page shows the properties of these layers. The layers are listed in order of application, so the first
layer is the abrasion protection and the second is the Strake. The Strake thickness has been set to
give the correct displacement. The actual diameter of the structure would be greater than this.
Close the Variable Data Form.
The Drag, Lift & Added Mass page shows a high normal drag coefficient for the Strake, as
would be expected. This drag coefficients use the diameter of the coated pipe as the reference.
This diameter is shown in the Drag/Lift Diameter columns. A value of ~ indicates that no
coatings have been applied and the bare steel diameter on the Geometry & Mass page will be
used. You may need to adjust your drag coefficient if its source uses a different reference
diameter. See System Modelling: Data and Results | Lines | Line Types | Drag, Lift & Added
Mass Data in OrcaFlex Help.
Now select the Code Checks View Mode at the top left of the form. The API RP 2RD stress
code check has been implemented within OrcaFlex. In this case the yield stress (SMYS) has been
set for steel and a factor (Cf) of 1.2 applied, appropriate for extreme conditions.
Return to the Model Browser and open the Environment Data Form. The data on the Waves page
shows that this example is a regular wave analysis. The default Dean Stream Theory has been
selected because this is robust for a wide range of water depths. The Current page shows the
current heading is collinear with waves in this example, they have the same heading. Note that
OrcaFlex specifies Wave and Current headings Towards a direction.
On the Seabed page the vertical and shear stiffness have been increased from default to represent
a stiffer seabed. Seabed stiffness has a significant effect on SCR fatigue so these settings must be
set carefully.
In this example a Linear Seabed has been applied. OrcaFlex has the option for a more complex
Non-Linear Seabed but it is recommended this is used on a case by case basis because it requires
OrcaFlex to carry out significantly more computation. More details on the Seabed options can be
found in Theory | Environmental Theory | Seabed Theory of OrcaFlex Help.
W Orcina

Now move to the General Data Form via the Model Browser. The Dynamics Page shows a build
up period (Stage 0) of one wave period then a run for five wave periods (Stage 1). This usually
sufficient to obtain a regular wave settled response for most systems.
The Integration & Time Steps page shows that a Fixed Step Implicit Time Integration scheme
has been applied. This is the default setting and will provide fast, good quality results for SCR
systems.

1.2. Results
Three plots are shown in OrcaFlex. The right hand side plots are instantaneous range graphs of
Utilisation along the riser length. Utilisation is a result from the API RP 2RD code check and a
value greater than 1.0 indicates the limit has been exceeded. The lower plot shows the variation
along the full riser length from the hangoff at 0ft to the seabed termination at 9810ft. The upper
plot shows the details for the first 50ft of riser.
The lower left hand plot is of utilisation against End Force Ez-Angle at the Flexjoint. The End
Force Ez-Angle is the rotational deflection from the no-moment direction. As would be expected
an increased angle gives an increased utilisation. However the non-linearity of the flexjoint
stiffness means at greater angles the increase in utilisation is reduced.
Stress results are also calculated and plots can be produced from the Results Form (Results
Dropdown menu).
Run the replay via the Replay dropdown menu. The instantaneous range graphs are synchronised
with the view to the left and show how the utilisation distribution changes with time. Note the
step change in utilisation at the transition of Flex Joint to Riser and Riser to Flowline.
Variation with time is most significant at the hangoff and touchdown regions as is typical with an
SCR configuration
The plot of utilisation against Ez-angle is for the latest wave. A moving cross hair will appear on
the plot when replay reaches that latest wave. This cross denoting where on the plot the animation
has reached.
W Orcina

2. Catenary with SemiSub
Two steel risers descend from a SemiSub pontoon in a simple catenary configuration,
transitioning to a flowline after 7800ft. The risers are hung off the outer edge of the forward
pontoon. One SCR has a Flexjoint connection at the hangoff, the other has a Tapered Stress Joint
(TSJ). Otherwise the risers are identical.

When you open the simulation file you will see two wireframe views of the system. One is a side
elevation of the configuration and the other is a detailed view of the SemiSub and connections.

The SCR configuration is similar to the one attached to the Spar in the earlier example. Only
differences will be discussed in this section.
Both risers hang off the SemiSub with an angle of 6 from vertical. With the SemiSub offset on a
Far heading by 10% of water depth the riser deflection from vertical will increase. To prevent
excessive stress at the hangoff from this increased curvature local bend protection is used.
The Riser with Flexjoint has local stress relief provided by a non-linear end connection
stiffness. This method is discussed in more detail in the Spar example earlier.
The Riser with TSJ provides local reinforcement to the riser wall with a tapered stress joint.
This is a line section with a tapered outer diameter. At the hangoff the root of the TSJ has a larger
outer diameter and it then decreases linearly to the tip where the outer diameter matches the rest
of the riser. The section therefore needs to be assigned a line type with a variable outer diameter.
Open the Line Types Data Form via the Model Browser. Put View Mode on All so you can
compare the properties of line type TSJ with the others and move to the Geometry & Mass
page. Instead of a constant outer diameter TSJ has TSJ Profile.
X
Z
1000 ft
OrcaFl ex 9.4a70: Catenary wi th Semi Sub.dat (modi fi ed 16:15 on 14/07/2010 by OrcaFl ex 9.4a70) (azi muth=270; el evati on=0)
Reset
X
Y
Z
80 ft
OrcaFl ex 9.4a70: Catenary wi th Semi Sub.dat (modi fi ed 16:15 on 14/07/2010 by OrcaFl ex 9.4a70) (azi muth=323; el evati on=9)
Reset
W Orcina

OrcaFlex9.4a70: Catenarywith SemiSub.sim(modified 13:27 on 16/07/2010 byOrcaFlex9.4a70)
Riser with TSJ profile
Inner surface Outer surface
Arc Length (ft)
50 40 30 20 10 0
R
a
d
i
u
s

(
f
t
)
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
Right click on the form and select Edit Variable Data to move to the Variable Data Form. The
Line Type Outer Diameter page is found in the Structure folder. A table shows the TSJ outer
diameter. The diameter order is defined following the usual line conventions (End A to End B).
So in this example the taper is Root to Tip. If you wanted to reverse this (because the TSJ was at
End B for instance) then the Reverse button at the bottom of the table will do this for you.
Close the Variable Data Form and move to the Structure page. Note that the bending, axial and
torsional stiffness for the TSJ are not specified because these will vary along the line.
Set View Mode to Code Checks. The Design case factor, Cf, has been set at 1.0 for
operational conditions. Note that no minimum wall thickness can be specified for the TSJ because
it varies with length. It is calculated based on the specified Outer and Inner Diameters of the
Geometry & Mass page.
Open Riser with TSJ from the Model Browser. On the Structure page the first section has the
line type TSJ. The section length is greyed out because this is specified on the Line Type Outer
Diameter page of the Variable Data Form.
Select the Profile Graph button at the bottom of
the Line Data Form to see resulting riser diameter
variation. Double click on the graph to set the X
axis from 0ft to 50ft to see the TSJ region more
clearly.
The green line is the inner diameter while the red
line is the outer diameter. The taper is modelled
by a series of segments with reducing outer
diameters. The refinement is controlled by the
number of segments in that section.
The increase of outer diameter further along the
line is from Coatings and Strakes on the riser.

The SemiSub is modelled as a vessel. Its visual representation could be generated by attaching
shapes that have no stiffness. Instead this example illustrates using the Vessel Type wireframe
drawing.
Open the Semi Displacement RAOs Vessel Type Data Form from the Model Browser and
move to the Drawing page. The SemiSub image has been produced by vertices connected by
edges.
In the shaded graphics mode OrcaFlex will draw a solid shape based on the wireframe vertices
and edges. When making use of this facility you need specify the wireframe with some care as
described in the System Modelling: Data and Results | Vessels | Vessel Types Data: Shaded
Drawing help topic.
In essence OrcaFlex partitions the vertices into sets of connected vertices. Two vertices are
deemed to be connected if there exists a path of edges between the two vertices. Then, for each
set of connected vertices, the smallest convex shape enclosing the set is drawn. If you have
concave shapes then you need to subdivide your vertices into groups with no sharing of vertices.

W Orcina


Figure 1: Shaded View of SemiSub: Vertices Shared (Left) and Not Shared (Right)
Each box of the SemiSub shape has therefore been given its own unique set of 8 vertices and 12
edges to produce the required shape in both wireframe and shaded view. The figures above show
the shaded shape results with and without sharing.
2.2. Results
Two plots are generated by the default workspace when the simulation file is opened. These are
API RP 2RD utilisation range graphs. They show the variation through the latest wave of
utilisation along the first 100ft of riser extending from the Hangoff.
The upper plot is for the riser with Flexjoint ,the lower for the riser with TSJ. They show different
utilisation distributions because of the way the devices work but neither exceed the allowable
limit of 1.0.

W Orcina

X
Z
600 ft
OrcaFlex 9.4a70: Lazy Wave with FPSO .sim(modified 11:51 on 20/07/2010 by OrcaFlex 9.4a70) (azimuth=270; elevation=0)
Statics Complete
3. Lazy Wave with FPSO
A steel riser descends from the bottom of an
FPSO external turret in a Lazy Wave
Configuration. Buoyancy is applied to the riser
length from 3515ft to 4705ft and a Flexjoint
provides local stress relief at the Turret
connection.
When you open the simulation file you will see a
wireframe elevation view of the configuration
and three plots.
The red section of riser shows the buoyancy
location. The white section is seabed contact.

Discussion on SCR properties and Flexjoint modelling can be found in the Catenary with Spar
example earlier in this document. A detailed discussion on building a Lazy Wave configuration
can be found in the Lazy Wave Riser example of A01 Catenary and Wave Systems.
The FPSO is modelled using Orcaflex default vessel properties. These are for a vessel of 338ft
long while the vessel in this example 984ft long. The vessel response has therefore been increased
using Froude scaling. This facility is discussed in the A04 Disconnecting Turret System. In an
actual project these properties should be replaced by the actual properties of the required FPSO or
the system responses will be incorrect.
This example does not include a detailed turret that will rotate in relation to the vessel. It is
simply modelled for visualisation as a shape with zero stiffness.
In reality the turret will rotate about its vertical axis so that its orientation will remain reasonably
constant when the vessel weathervanes. Therefore hangoff positions and orientations will need to
be modified whenever the vessel changes heading.
Turret hangoff locations are generally supplied in polar coordinates (radius and heading) relative
to the turret centreline. It is therefore more convenient if OrcaFlex allows connections to be
specified in Polar Coordinates too.
The vessel origin has therefore been moved to the turret centre at keel level. All FPSO data and
connections have been adjusted to also be relative to this new origin.
Polar coordinates are accessed in Reset Mode only. Therefore reset the model via the Calculation
dropdown menu then open the All Objects Data Form via the Model Browser. Make sure it is
set up to Show Connections and Included Connections has Lines selected. Now move to the
Polar Coordinates page. The connection position can now be specified as a radius (R), heading
(Theta) and height (Z) from the vessel origin.
The resulting Cartesian Coordinate positions are shown on the Positions page.
W Orcina

3.2. Results
If you reset the simulation then reload the simulation file. The default workspace for this file will
load the wireframe view and two range graph plots for the latest wave responses.
The plots show the distribution of API RP 2RD utilisation. The distribution over the range from
hangoff to seabed is shown, together with a more detailed plot covering just the first 500ft from
hangoff. The wave shape gives additional compliance to the system, distributing the utilisation
over more of the length and so relieving the hangoff location.

W Orcina

Drilling Risers: B01 Drilling Riser Page 1 of 5
B01 Drilling Riser

A tensioned drilling riser descends from a semi-submersible drilling vessel to a BOP (blow out
preventer) on the seabed. A drill string is modelled running inside the riser down to the BOP and
carries on into the casing below the seabed.
The complete system is around 2000m long from top to bottom; therefore the model is easier to
understand if viewed in sections, using Workspaces. The file opens up with the default workspace
loaded, which shows two shaded views of the semi-sub.
First load up the Workspace B01 Drilling Riser top views.wrk. This shows the topside part of
the system, which has a drilling riser descending through the moonpool of a semi-sub, down to
the seabed.
The semi-sub motion is represented by a vessel. The semi-sub shape could be defined by
importing a .x file or by adding shapes to the vessel but in this case it is represented by a
wireframe. The main semi-sub shape (red part) has been specified on the vessel type drawing tab.
The drill tower (green part) is defined on the drawing tab for the vessel named Semisub.
This allows a general representation of the vessel to be stored with the vessel type (main shape)
and then customised individually on the vessel page each time it is used (drill tower). It also
allows the vessel and vessel type parts to be different colours, as in this case. For an explanation
of how the solid shape is filled in, see the example A05 Steel Catenary Riser Systems or the Help
file section System Modelling: Data and Results | Vessels | Vessel Types: Shaded Drawing.
W Orcina

The moonpool edges are modelled as dummy lines 'MP FWD', 'MP PORT', 'MP AFT' and 'MP
STBD' attached to the platform. You can see them in the current view as yellow lines around the
moonpool edge. The dummy lines are there so that clearance information can be obtained
between the riser and moonpool. The lines need only have one segment, as line on line contact
occurs between segments rather than between nodes. Because the lines are only single segments,
with both ends connected to the vessel, their behaviour does not require any stiffness or other
information, and they have negligible properties.
Clearance results are always reported between lines, so there is no need to turn clashing contact
on for these lines if all you are interested in is whether they come into contact or not. Clash
checking is only required if you want to show how the lines react as a result of the contact.
All of the shapes used to model the deck structures are Drawing type Shapes, which means that
they have zero stiffness and are there for visualisation only.
The components of interest in this system are as follows: -
At the very top of the screen the bottom of a winch, coloured blue, can be seen. This winch
supports the drill string, which is the grey line seen passing through the upper deck before
disappearing inside the riser. Open the Drill string support winch data form: the winch is set to
keep the tension constant during the dynamic simulation by using the Specified Tension Change
option and setting its value to zero.
The rotary table is also modelled using four links to provide the lateral restraint. They are
modelled as very long links (500m) so that they dont prevent axial movement of the drill string.
These links are currently hidden in the model; make sure that the View by Groups option is
selected in the Model Browser, and then find the group named Rotary Table (within the
Semisub Group). Click on the Rotary Table group and then press Ctrl + H to show the entire
group. Repeat to hide the group again.
The riser has a slip-joint at the tensioners; this is required to allow the line to move axially but not
laterally. In this instance, it is modelled with a single segment line called Riser Slip Joint
(coloured pink in the model). The slip joint line has a very low axial stiffness to allow axial
motion of the nodes at each end but high bend stiffness to stop lateral motions. The bending
stiffness is set similar to that of the riser itself. It could be argued that it needs to be stiffer.
The four tensioners, in the usual cruciform arrangement, are modelled using four Links of
spring/damper type. In this case their stiffnesses are non-linear but their damping is linear with
velocity (a non-linear option is available). The non-linear stiffness is shown in the plot below.
The characteristics have been set to give a lesser axial stiffness from 4m to 12m link length, a
stroke of 4m. Beyond this range, the stiffness should increase, and we have chosen to double
the stiffness.

W Orcina

OrcaFlex9.6c: B01 Drilling Riser.sim(modified 14:37 on 01/07/2013 byOrcaFlex9.6c)
Tens_AFT Stiffness Profile
Length (m)
15 10 5 0
T
e
n
s
i
o
n

(
k
N
)
3600
3400
3200
3000
2800
2600
2400
2200

The tensioner ring is modelled as a 6D Buoy; the tensioners and the slip joint all connect to it. It
is visible in this view as a red cube. The buoy has been given negligible properties, as its purpose
here is simply to act as a connection point.
The riser (in yellow) and the choke and kill lines (purple) all connect to the tensioner ring as well.
The choke and kill lines each have a line contact relationship defined with the riser, which
enables the lines to be clamped to each other at specified locations. The first two clamps are
visible in the current view, as thin orange rings around the choke and kill lines (open workspace
B01 Drilling Riser tensioner detail.wrk for a closer view).
These three lines extend all the way down to the seabed, where they terminate to the BOP. Open
the Workspace B01 Drilling Riser bottom views.wrk. The blue cylindrical object is the BOP.
The riser, choke and kill lines are all connected to another 6D Buoy (red cube) with negligible
properties, called Flex Joint. This ensures that the BOP sees appropriate total moments rather
than individual ones.
The flex joint itself is modelled by connecting this buoy to a length of line (the short thick red line
visible in the top right-hand view) which has its connection to the buoy encastr, but has finite,
non-zero connection stiffness at its lower end. Note that this is one method of creating a flex joint,
which is appropriate when the joint coincides with a line end. In OrcaFlex you can also create
flex joints mid-line by using the Flex Joint type of Attachment: see the Help file section System
Modelling:Data and Results | Lines | Attachments | Flex Joints for further details.
In the simulation, the semi-sub has been set up with a prescribed motion, making it move with a
constant speed in the Global X direction. Run the replay and you can see the effect of the flex
joint, preventing the moment from being transmitted from the riser to the casing. The casing line
is connected to the Flex Joint and extends down through the BOP and into the seabed. Its lower
end can be seen in the left-hand model view; it is the lowest point in the model.
W Orcina

The drill string is modelled inside the riser and casing. Click on the Riser Group in the model
browser and select Ctrl + H to hide the contents of this group. The grey drill string is now visible.
Ideally the drill string should be modelled as a continuous line, extending from the winch on the
semi-sub right down to the drill head, however in this model we have split it into two sections,
joined by a further 6D Buoy (visible as a pale blue cube). The reason for this will be explained
below.
Finally at the bottom end of the drill string we have added a 3D Buoy (visible as a blue dot) to
add some mass to represent the drill head assembly.
Make the Riser Group visible again (Ctrl + H in the Model Browser) and open the Line Contact
Data form, located towards the top of the Model Browser. The line contact relationships for the
choke and kill lines (as discussed earlier) are defined here, on the Relationships tab. The riser
restraint goes around the choke and kill lines, like a collar (see sketch below). The Penetrator
Locations tab shows where those collars are, and their offset from the riser centreline.

We have also defined the relationships for the drill string lines; the upper and lower sections of
drill string are located Inside the riser and casing respectively. Containment is enabled for these
relationships this means that the inner line is shielded by the outer line so it doesnt see
environment forces. The inner line sees fluid loads from the outer lines contents.
The line contact model in OrcaFlex only allows you to have one line inside two other lines if they
both have contents that are free flooding. In this case, the riser and the casing line contents are
both mud (i.e. not free flooding), and this is the reason why the drill string has been split into two
sections. Torsion needs to be included for any lines identified as the splined line in a line contact
relationship, and for both lines if the relationship type is Around.
Contact stiffnesses for the line contact relationships have been set so that they are high enough to
restrain correctly, but not so high that they cause rattle. Note that line contact is different to line
W Orcina

clashing, therefore we do not need to turn on Clash Checking for these lines. For more
information about the different contact models in OrcaFlex, contact Orcina for a copy of the
knowledge base document Modelling Contact.
The casing line has a p-y model, which means that there is no normal resistance from the seabed,
but there is lateral resistance. Open the Casing line data form (in the riser group) to see that the P-
y Model1 is specified at the top of the form. Click on the P-y Models button at the bottom of
the form to see the details of the P-y Model defined.
The lower drill string (which is inside the casing) does not have a p-y model, so it would usually
see a huge normal reaction force from the seabed, which would push it upwards. In this model
however, the drill string is shielded from contact with the seabed by means of the Containment
Enabled part of the line contact model. Containment shields the inner line from environmental
fluid forces as well as from contact with elastic solids or the seabed.

2. Results
Open up the B01 Drilling Riser top views.wrk workspace again and look at the motion of the
slip joint and tensioners when you replay the animation.
Open the results workspace B01 Drilling Riser tensioner.wrk. This bring up results graphs
showing the tension time histories of the four links and the effective tension in the riser.
Open the results workspace B01 Drilling Riser clearance.wrk. Look at the Line Clearance range
graphs for the riser. The graphs show the overall clearance to each of the dummy lines
representing the moonpool edge. These allow for the line diameters. Minimum clearance is just
less than 4 m to the aft edge.
W Orcina

Drilling Risers: B02 Workover Riser Page 1 of 2
B02 Workover Riser
X
Z
30 m
X
Z
OrcaFlex 9.2a25: A2.2 Workover riser.sim(modified 09:51 on 27/03/2008 by OrcaFlex 9.2a25) (azimuth=270; elevation=0)
Time: 50.0000s

A workover riser is deployed from a semi-submersible drilling vessel. The lubricator assembly is
suspended in the drilling derrick and the riser hangs below it, with much of its weight supported
by a pair of tensioners hung off the drill floor. Clearance is tight where the riser passes through
the drill floor, and this area is an important point of interest in the analysis.
Building The Model
Open the Model Browser. The data are collected into two Groups - the Semisub Group and the
Riser Group (from the Browser View menu, select View by Groups).
The Semisub Group comprises the Semisub Vessel itself, plus a number of Shapes and Lines
which are used to supplement the drawing and provide facilities for modelling contact with the
riser at the drill floor.
A detailed model of the drilling vessel has been assembled using the wire frame drawing facilities
supplemented by the Shapes. This is not strictly necessary for the analysis, but is helpful in
understanding the system geometry. Look through the modelling to see how we have done it.
Example A02.01 Drilling Riser talks through some of these issues in more detail.
The pontoons and columns are specified on the Vessel Types form. The derrick and parts of the
deck are specified on the Vessel form. Change the pen colour on the Vessel form to see which
part is which. The drill floor in the derrick and parts of the main deck are modelled as Shapes.
Switch to the shaded graphics workspace for the full effect.
The Riser Group consists of (from the top down):
Top hang-off - a winch applying a constant tension
Lubricator sub-group comprising two linked items modelled as 6D Buoys connected by links
Two tensioners
Tensioned riser
BOP stack at the seabed, modelled as a Shape
The riser has been modelled from the bottom up: default with OrcaFlex is top down, but OrcaFlex
does allow you to reverse this. Note that the Riser Top End data item (on the line data form) is
set to End B so OrcaFlex knows it has been reversed. The order of the line sections on the data
form is still specified from End A towards End B.
W Orcina

Drilling Risers: B02 Workover Riser Page 2 of 2
The edges of the hole in the drilling platform are modelled using single segment lines, which
matches the approach used in example A01.01. However, differently to A01.01, clashing has
been enabled between the drill string and the moonpool edges. Rather than only reporting
clearance results, these moonpool edges will resist penetration by the drill string, and report the
force that arises during a contact event.
Line clashing is enabled by checking the clash check box on the line data form for all lines
involved in contact. This can be applied section by section, and should only be used where it will
be needed, as it is computationally demanding. Once lines are able to clash, their interaction is
governed by the contact stiffness on the line types data form.
Contact stiffness is typically set lower than the real-world contact stiffness, beginning to
experiment with values close to 1000 kN/m. Contact stiffnesses are evaluated by judging whether
the penetration between clashing lines is acceptably small, and this is typically achieved without
using very high values. For low contact velocities, line clashing results should not be sensitive to
the values entered for contact stiffness, but a sensitivity study is recommended to test this
assumption. For more advice on clashing and clash results, check the OrcaFlex help at Help |
Theory | Line Theory | Clashing.
Results
Zoom in closer to the drill floor. Replay the last wave cycle and see how the riser makes sliding
contact with the hole in the middle of the drill floor. This is indicated by the line segments
lighting up white where contact occurs.
For a detailed report on the contact, load workspace B02 Workover Riser results.wrk. It includes
a Line Clashing Report for the riser for the latest wave. The report gives details of when each
contact event started and ended, total impulse applied, peak clash force and energy, and
maximum penetration between the clashing lines.
Typically, clashing issues are analysed with OrcaFlex using a number of simulations. Models are
run without line clashing, and clearance results are inspected to look for cases where line clashing
does occur. This reveals any incidence of clashing behaviour without allowing the line clashing
algorithm to increase simulation run time.
The cases where clashing is predicted by the line clearance results are run again, with line
clashing activated for the regions in the system where clashing may occur. Two approaches can
be taken at this point; the OrcaFlex results for clash force and energy can be used, and the
sensitivity of the results to line contact stiffness and segmentation should be tested. Alternatively,
the lines contact velocity can be used as input to a detailed element model of the line cross
section.
OrcaFlex is a global analysis tool, and therefore does not include local deformation of a line due
to impact. The local stresses that may arise in a line as the result of a contact event will not be
evaluated by the OrcaFlex line model, and so this detailed local analysis is often performed in
alternative modelling packages.
CL
12 January 2009
W Orcina

Drilling Risers: B03 SPAR Page 1 of 2
B03 Spar
X
Z
200 m
X
Z
OrcaFlex 9.2a25: A2.3 SPAR.sim(modified 11:03 on 27/03/2008 by OrcaFlex 9.2a25) (azimuth=270; elevation=0)
Time: 40.0000s

A riser descends from a Spar to the seabed via a guide down a central moonpool. A constant
tension device supports the top of the riser.
Building the model
The Spar is modelled as a vessel with its response to wave loads given as RAOs. Vessel drawing
facilities are not used: instead we use two cylindrical Shapes to represent the hull and deck.
The riser is suspended at the lower edge of the deck and hangs down inside a moonpool. The
water inside the moonpool is isolated from wave and current influences and will move with the
Spar. The effect is modelled by the 'Moonpool Water' shape which is defined as Trapped Water.
Trapped water removes wave and current effects from within the shape. Any effects from the line
itself moving in the water are retained.Trapped water does not offer additional inertia to the
flooded object, but is only used to shield a region of the model from environmental loads.
Fixed guides are fitted in the moonpool. These restrain the riser in the buoy x and y directions but
allow free motion in the z direction (we neglect friction between the riser and the guides). Three
pairs of links spaced vertically, i.e. along the riser centreline, model the guides. These links are
hidden in the view window using the OrcaFlex 'hide' option. Open the model browser and click
on Guide 1X with the mouse right hand button. If you are using View by Groups, there will be a
group called Hidden Links. You will see an option to 'show' the link. If this is selected, the link
will appear in the view window. You can also show the whole hidden links group by right-
clicking on the group name. Hiding is carried out the same way. For visualisation purposes only,
short cylinders with zero stiffness are placed at the guide locations.
The links are specified as Spring/Dampers, not Tethers, so that the spring can take compression.
Damping is also applied. Note that the ends of the links are attached to and move with the Spar.
Note also that each link is made very long so that vertical movement of the node to which it is
attached makes very little difference to the force in the link, and the link remains near-normal to
the riser.
Look at 'Guide 1X'. The stiffness profile has resistance to compression when the link length is
0 m to 304.5 m and resistance to tension when the length is 305.1 m to 609.6 m. From 304.5 m to
305.1 m, a gap of 600 mm, the link has no action (zero tension) as the riser is within the guide -
W Orcina

Drilling Risers: B03 SPAR Page 2 of 2
i.e. the guide has a "rattle space" of 600 mm. The link does also have some damping, which will
act even when the line is within the rattle space, depending on its velocity.
It is not generally necessary to give the link the true contact stiffness between a riser and a steel
guide tube. The important criterion is that the riser should not penetrate the guide so much as to
invalidate the model; in practice this usually means that a few millimetres penetration is quite
acceptable. If you have some idea what contact force you are expecting, this will enable you to set
an acceptable stiffness. If not, then a preliminary trial run will give you some idea of what you
need. In general, it is not a good idea to set the stiffness too high as this may lead to noise in the
results and possibly to integration instability and the need to re-run with a shorter time step.
Because we are modelling short duration contact in this simulation, and allowing the riser to
rattle, the explicit integration scheme is used. Short time steps are required to accurately capture
the riser motion in the guides, and therefore the run time benefits available from using long time
steps with the implicit integration scheme will compromise the accuracy of our results in this
case. The explicit run time is not particularly long for this case.
Results
Open the workspace B03 Spar.wrk and look at the animation through the latest wave. The overall
motion appears gentle. Viewing the windows that show the Spar, there is substantial riser motion
inside the moonpool.
For this model the shaded graphics view (CTRL+G) is unhelpful as the interesting part of the
model - the riser inside the Spar is obscured.
Look at the time history of tension in Guide 3X (at the base of the Spar). The contact force
reverses direction (changes sign) and disappears briefly once per wave cycle. The force is
irregular and spiky as we should expect for a rattling contact.
Another workspace is provided with this example, B03 riser waves.wrk, which presents results
that show waves being transmitted down the length of the riser due to the spar motion. The
instantaneous range graph of riser X-motion shows that as the top end of the riser swings from
side to side in the waves, the riser response moves down the line.
As a final part of the investigation into the riser motion, a modal analysis of the riser can be
performed. Once statics has been completed, Modal Analysis is available on the Results menu.
For the lines in the model, OrcaFlex will report their natural modes of oscillation. For this riser,
there is a transverse mode that corresponds to the shape of the waves we are observing, and this
mode has a period very close to the wave period in the simulation. It is quite possible that this
natural mode of the riser motion is being excited by the waves in the model. As a further check on
this possibility, the right-click menu for a time history offers a Spectral Density. Take a spectral
density of riser curvature or tension from the workspaces, and note the 8 s period content. If this
was to give unacceptable results, it may be appropriate to run the model with an irregular seastate,
as this is less likely to excite a resonance in the system.
CEL
30
th
June 2008

W Orcina

Drilling Risers: B04 Clashing Risers Page 1 of 2
B04 Clashing Risers
X Y
Z
80 m
X Y
Z OrcaFlex 9.2a25: A2.4 Clashing Risers.sim(modified 15:28 on 28/03/2008 by OrcaFlex 9.2a25) (azimuth=280; elevation=0)
Time: 90.0003s

Two drill strings are hung from a MODU. Pairs of strings are used if one string has to be
disconnected and drilling is required to continue. In this model one string is connected to the
wellhead while the other has been released and is being raised to the drilling vessel.
This simulation looks at a snap-shot of the raising operation, assuming that the total free riser
length remains constant, and applies a regular design wave. The risers are close enough together
that hydrodynamic effects due to the wake shed by the free riser lead to clashing between the two
drill strings. The clashing is analysed, and the effect of the wake behind the free riser is included
in the model.
Building the model
The MODU is modelled as a vessel with its response to wave loads given as RAOs. Vessel
drawing facilities have been used to show the pontoons and bracing as blocks of colour, while
some of the vessel types drawing lines have a thickness so that the deck is indicated, but we can
see past it.
The riser on the wellhead has a slipjoint as its first section, modelled as a single line segment. The
slip joint line type has the bending stiffness of the riser, but a low axial stiffness, to represent the
stroke of the slipjoint. Tension in this riser is provided by links, attached at the second node on
the line. See example A02.01 Drilling Riser for details of this type of arrangement.
Both ends of the active drilling riser are modelled as flexjoints having some finite connection
stiffness.
The free riser does not have a flex joint, but simply hangs straight down from the vessel. Statics is
performed in still water, without current applied. The current is ramped in dynamics. This is
because line-on-line contact does not operate during statics, and so we will not see any of the
expected clashing. The worst outcome would be that the lines are coincident after statics (with
current) and there would be a sudden large reaction force when dynamics begins.
Because we are ramping the current after statics, the main simulation is run for 10 wave periods.
This is necessary to let the free riser drift into a stable position. Results should be taken from the
latest wave of the simulation.
W Orcina

Drilling Risers: B04 Clashing Risers Page 2 of 2
Contact between the risers is included using line on line clashing. The edges of the moonpool are
also included in the model, but clashing is not enabled for these lines. Clearance results are still
available, without requiring that line clashing be activated.
As we are evaluating clashing between two lines, the explicit integration scheme is used.
Clashing typically requires a short time step in order to accurately resolve short-duration contacts
in time. When the integration time step is forced to a small value for the sake of accuracy, the
implicit or explicit scheme would be using time steps of a similar size. When this situation arises,
the explicit integration scheme offers faster run times, as there is no iteration performed at each
time step. A final consideration when evaluating short-duration events is the logging interval used
to capture results from the model. This has been reduced to 0.05 s from the default value of 0.1 s.
However, an even shorter value may be required if close inspection of the results shows poor
resolution. It is important to note that the size of simulation files is directly affected by logging
interval, so the largest interval which generates well-resolved results is usually preferred.
To model the wake shed behind the free riser (which begins the simulation upstream of the
drilling riser) and also behind the connected riser, which is upstream once the free riser has
drifted, we have enabled wake shedding on the risers. The wake model used was created using the
Wake models button on the Drag & Wake tab of the free riser data form. Two models from
industry literature are available, and we have chosen the Blevins model; an empirical formulation
that includes both drag reduction on downstream lines and lift effects which act towards the wake
centre. For more detail on this feature, select part of the data, and press the F1 key to call up the
relevant part of the OrcaFlex help.
Wake effects are treated in a similar manner to line clashing, we must set lines to shed a wake,
and also enable other lines to react to it. These effects can be activated section-by-section on
lines, because the wake effects require further computation, it is preferable to apply them only
where they will be required. The two wakes currently included in the model are independent, and
will not interact.
Results
The supplied workspace B04 Clashing Risers results.wrk includes quite a few graphs. A plot of
the X position of the free risers End B is presented, so that we can judge whether the free riser
has drifted to a stable position in the ramped current. The riser X position is beginning to settle
near to the end of the simulation.
A time history of clash force between the two risers is shown, taking results from 178 m arc
length on the connected riser. The arc length was chosen by referring to the clash force range
graph, which is shown at the bottom right.
Finally, a clashing report for the connected riser gives further detailed information on the pipes
interaction, and a time history of tension at 178 m arc length shows that the contact is not severe
enough to change the drilling risers tensile behaviour.
Perhaps the most useful result here is to open the file default workspace, and watch a reply of the
simulation. In the right-hand window we have zoomed in on the region over which clashing
occurs, and it is drawn white in the shaded view. It can be seen that the clashing is a sliding
contact, with the free riser brushing against the connected drill string as it drifts.
W Orcina

Drilling Risers: B06 Running BOP Page 1 of 3
B06 Running BOP

A blow out preventer (BOP) stack is lowered to the seabed for connection to the wellhead. Guide
wires are used to restrain the BOP while it is lowered.
The drilling vessel uses displacement RAOs while the lowering is simulated in a low seastate.
The BOP is modelled using a 6D spar buoy (shown in red in the model). A wave packet including
the maximum expected wave has been applied to the system.
The BOP hydrodynamic properties are set on a cylinder-by-cylinder basis. They are intended to
give a reasonable impression of the environmental loads on the buoy although it will be
restrained quite strongly by the guide wires and winch. Slam loads are also applied by means of a
slam coefficient applied to lower-most cylinder in the stack. Open the Data Form for the BOP 6D
Buoy, in the BOP Group, to see how these properties have been assigned. In this example, the
slam force exit coefficient is left as zero, as we are only interested in the lowering operation.
We are using a Winch for lowering the BOP through the water column. The Winch object has no
mass, displacement, drag or added mass, therefore we will not capture all of the loading on the
lowering gear. The BOP should be sufficiently restrained by the guide wires so that this
approximation will not cause too much inaccuracy.
If the winch wire properties are significant then it can be changed to a Winch-Line hybrid. Part of
the wire would be modelled by an OrcaFlex Line object. A Winch would then be attached to the
free end of the Line to provide the length that pays out.
The guide wires are long OrcaFlex lines, which run from the semi-sub down to a template on the
seabed. At the seabed end of each line, a section is modelled with a variable outer diameter
profile to represent the guide spike on the template. The BOP has four guides (coloured yellow)
W Orcina

that slide down the guide wires and ultimately locate on the template spikes. Line Contact is used
to model this, with the Inside relationship being used in this instance. The guide lines are
inside the guides, which means that the penetrator balls are located on the guide lines and the
guides themselves are fitted with smooth spline surfaces (see the OrcaFlex Help section System
Modelling: Data and Results | Lines | Line Contact | Data for further explanation of these terms).
In this model the Around relationship could also have been used which would have switched
which line of each pair was splined and which had penetrators. Each relationship has its pros and
cons and which one is best to choose depends on a number of factors, however often either
relationship is suitable.
In this particular model, the guide wires have high axial stiffness but very low bend stiffness,
which is typical for a wire; however this combination can result in some strange spline behaviour!
If the Around relationship was used then the spline would be fitted between the nodes of the
guide wires. If the segmentation is quite coarse and a force is applied to the spline by the
penetrator pulling on it (as happens in this model if the BOP moves sideways) then the nodes of
the splined line can turn (due to the low bend stiffness) which forces the spline itself to deflect
away from the line that carries it. The spline is forced to fit through each node, aligned with the
node directions. This phenomenon is undesirable and is clearly visible in the screen shot below: -

The solution to this is to shorten the segment length, however on a line that is 250m long this
could mean a lot more segments, which would subsequently increase the run time. An alternative
is to do what we have done in this model, which is to use the Inside Line Contact relationship
instead. The segments on the guide wires still need to be fairly short and the penetrators (located
at the nodes) need to be positioned close enough to ensure that there is always at least one
penetrator in contact with the spline surfaces of the guide tubes, but they can be longer than they
would need to be with the Around option. See Example B01 Drilling Riser for further
explanation of this type of line contact relationship.
Open the Line Contact Data Form from the Model Browser to see the Line Contact relationships.
Note that Containment is Enabled is ticked for all the relationships. The guide tubes are free
flooding so OrcaFlex will calculate which parts of the guide wires are shielded by the guides, and
will calculate the fluid forces acting on them accordingly.
W Orcina

The Line Contact relationships also sort out the line positions for us in Statics; Line Contact is
active during Statics, therefore the penetrator lines are forced inside their respective splined lines,
as required.
Elastic solid shapes are used to model the guide cones located on the bottom of the guide tubes;
usually we recommend that thin-walled shapes are made thicker to reduce the risk of nodes
accidentally picking up the outer surface of the shape rather than the inner. However in this case
the contact between the wires and guide tubes sufficiently restrains the motion of the BOP and
prevents this from happening, therefore the cones can be modelled with their real dimensions.
During dynamics, the Winch payout is ramped up to a maximum speed, and ramped down again
when the BOP is near the seabed. You can view a graph of the payout rate by looking at the
Variable Data item BOP Lower Rate in the model browser. Click the Profile button to see the
graph. The purpose of this ramping is to avoid sudden shock loads in the Winch, and to give a
gentle set-down for the BOP on the template.

2. Results
Open workspace B06 Running BOP lowering.wrk and watch the replay of the whole
simulation. Note the vertical line on the timehistory of BOP Z. It shows where the animations
have reached. The impact of the waves passing the Buoy as it enters the water can be seen in the
top right-hand view.
Now open workspace B06 Running BOP results.wrk and note how the BOP X and Y positions
vary with time. While the BOP is lowering on the guide wires, it is loosely aligned with its final
position. Once the guides come into contact with the alignment spikes, the BOP is finely aligned
before it finally comes to rest on the template.

W Orcina

Moorings: C05 Single Point Mooring Page 1 of 2
C05 Single Point Mooring
X
Z
100 m
X
Z
OrcaFlex 9.2a24: B5 Single Point Mooring.sim(modified 09:41 on 10/03/2008 by OrcaFlex 9.1c) (azimuth=270; elevation=0)
Time: 3600.0000s

An OrcaFlex default vessel is moored using three legs and a short chafe chain from the
connecting node. This system is often used as a temporary mooring. In OrcaFlex we aim to
inspect the load on the moorings and also quantify the mooring offset for this vessel in this sea
state. The simulation uses irregular waves acting for a duration of 1 hour.
Building the Model
The main mooring lines are composed of chain sections, which are drawn yellow, with a
polyester rope mid-section, drawn blue. The top end connection point is known as a node, and has
been modelled in OrcaFlex using a 3D buoy. The buoy is net heavy in the water, and has no
hydrodynamic properties. We are assuming that the mooring system motion will be dominated by
the hydrodynamics of the lines this demonstrates that it is not always necessary to enter data for
every object attribute.
As the chains have no bending stiffness there is no need for the node to be able to rotate or
transmit bend moments so a 3D buoy is appropriate for the node.
The vessel motion is calculated in all six degrees of freedom, so that the mooring offset in these
environmental conditions can be evaluated taking into account the restraint from the moorings.
The chains have a high stiffness, but quite a coarse segmentation, as we are not concerned about
their curvature. If the simulation is repeated with refined segmentation in the moorings, especially
over the touchdown region, the vessel offset and mooring tensions are only slightly affected, but
the curvatures do change.
The vessel is not included in statics, in order to demonstrate that the drift motion is calculated.
The mooring lines end orientations have been set with Use Static Line End Orientations from
the OrcaFlex Model menu.
Results
Tensions for the chafe chain and also for the most loaded mooring line are plotted for the period
of simulation when the largest loads occur (open Workspace C05 Single Point Mooring to see
these). We have taken statistics for the top line tension to find the time of greatest tension before
the time histories were plotted.
W Orcina

Moorings: C05 Single Point Mooring Page 2 of 2
The vessel excursion envelope is plotted on an x-y graph. X-y graphs can be used to plot any two
values from a simulation. They are often used for envelopes of position, or for other coupled data,
such as tension-angle plots for bend stiffener design. The initial slow drift (due to the vessels
exclusion from statics) can be clearly seen on the x-y graph.
W Orcina

Moorings: C06 CALM Buoy Page 1 of 4
C06 CALM Buoy
Introduction

A CALM buoy is moored by six equally spaced mooring lines. A shuttle tanker is moored to the
buoy by a hawser, with fluid transfer through a floating hose. This is a coupled analysis example.
The example also discusses the Spar Buoy Short Wave Issue.

Building the model
The CALM buoy is modelled as a 6D spar
buoy in which the object is represented as a
stack of co-axial cylinders. Take a look at
|System Modelling Data and Results|6D
Buoys|Spar Buoy and Towed Fish Properties|
in the helpfile. In this case we have used a
total of 6 cylinders of which Cylinders 1 and
2 are representations of the upperworks, and
Cylinder 5 represents the projecting skirt.
It is important to ensure that the physical and
hydrodynamic properties of the buoy are as
close as possible to the real CALM buoy
properties. Take a look at |System Modelling
Data and Results|6D Buoys|Modelling a
Surface-Piercing Buoy| in the helpfile for
further discussion of data preparation for
CALM buoys and the like.
The tanker is modelled as a free body subject
to wind, wave and current loading. Look at
the Vessel page. The primary motion has
6DofF motions calculated. These include the
hydrodynamic and wind drags, the wave drift
loading and the wave loads (1
st
order), which are specified in the Included Effects box. Because
the 1
st
order motions are included the displacement RAOs are not required. Therefore
superimposed motion is set to none to avoid applying that motion component twice.
Note that the wind load on the buoy is represented by a global applied load because 6DofF buoys
only experience wind load directly via wings.

W Orcina

Beware
This type of buoy analysis is only valid if the wave length is greater than 3x diameter of the buoy.
The Hs 2m by Tz 6s storm applied to this buoy infringed this requirement so component waves
have been filtered. A more detailed discussion of this is contained in the knowledgebase article
Spar Buoy Short Wave Issue which we have attached to the end of this document.
Read the OrcaFlex Help section |Theory|Line Theory|Interaction with the Sea Surface| for details
on modelling floating hoses.

Results
The view shows a shaded view of buoy and tanker. Look at the animation through Stage 1. Note
how the hose and tanker drift due to the environmental loading. Note that the drift of the tanker is
exaggerated because it is not included in the static calculation. We therefore see the current and
wind drift in the dynamics stage.
Instantaneous range graphs of the hose effective tension and curvature change through the
animation. The graphs show results from the tanker to the buoy (an arc length of 0m is the tanker
connection while an arc length of 120m is the buoy connection). You can see waves travelling up
and down the hose length.

Spar Buoys Short Wave Issue
This note documents an issue that can arise when modelling near-surface spar buoys.
A random sea is modelled in OrcaFlex using a large number of regular component waves, whose
periods and amplitudes are chosen to match the specified spectrum. You can see the wave
components chosen and their wavelengths (In OrcaFlex 8.7 onwards) by clicking the 'View Wave
Components' button on the Waves page of the Environment data form.
To cover the high frequency tail of the spectrum, the components include some short period
waves. In some cases they can go down to a period whose corresponding wavelength is shorter
than the diameter of a spar buoy in the model, and this raises a modelling issue. Note that the
issue is only really relevant for near-surface buoys, since short waves do not penetrate far down in
the water column.
A wave whose length is short compared with the buoy diameter will have effects that vary
significantly across the diameter of the buoy. So to analyse a short wave's effects accurately
OrcaFlex would have to calculate the fluid load contributions (buoyancy, drag etc.) at a number
of points across the diameter of the buoy. But in OrcaFlex spar buoys cannot be subdivided in
that radial direction - they can only be it subdivided in the buoy's axial direction, by dividing the
buoy up into a stack of cylinders.
OrcaFlex calculates the fluid kinematics (velocity, acceleration, surface elevation, surface slope
etc.) based on the sea surface and fluid kinematics at just one point on each cylinder of the buoy,
and assumes that the values at that point apply right across the cylinder's diameter. This is fine
for wave components a lot longer than the buoy diameter. But for waves shorter than about 3
W Orcina

times the buoy diameter the fluid kinematics, and hence the loads, average out since different
parts of the buoy diameter are seeing different phases of the wave at any one instant.
The result of this is that the effects of the wave components shorter than about 3 diameters are
exaggerated in OrcaFlex, since it cannot allow for the fact the effect of such components are
averaged across the buoy diameter.
The figure below illustrates the point for the calculation of the buoyancy force and moment
(similar problems arise for drag, added mass force etc.), for three cases where the wavelength is
less than 3 times diameter.
2d
2d/3
d/3

OrcaFlex determines the water surface slope at the centre of the buoy and then assumes this slope
is constant across the buoy. Of course in a random sea the surface slope is a combination of the
slopes of the various individual components that represent the random sea (and OrcaFlex allows
for this), but longer waves have smaller slopes so the short waves tend to dominate the surface
slope. So for simplicity the figure only shows the situation where there is just a single short wave
component present, and it illustrates how its wavelength affects the buoyancy calculation.
OrcaFlex calculates the buoyancy force and the centre of buoyancy (i.e. the point where it acts)
by calculating the wetted volume and centre of wetted volume of each cylinder of the buoy. And
each cylinders wetted volume and wetted centroid is calculated by assuming that the surface
through that cylinder is the tangent plane to the wave surface directly above or below the volume
centre of that cylinder.
The shaded section of the buoy therefore indicates what OrcaFlex will assume for the immersed
volume and hence the magnitude of the buoyancy force and the centre of buoyancy, and so the
size of the buoyancy moment applied.
As the wavelength reduces, so the discrepancy increases between the actual and the modelled
righting moment. For example in the third case (shortest wave), the actual righting moment due
to buoyancy will be small, since the wave crests and troughs are distributed fairly evenly across
the buoy, but the calculated buoyancy moment will be large since the shaded area is almost all on
W Orcina

one side of the buoy, so OrcaFlex will calculate that the centre of buoyancy is offset by nearly
half a diameter from the buoy centre.
You can work around this problem by removing (or scaling down) the short wave components.
There are two ways to do this in OrcaFlex, as follows.
Copy the wave components chosen by OrcaFlex, by clicking the 'View Wave Components'
button and then copying the frequency, period, amplitude and phase lag columns. Then
change the wave type to 'User Specified Components' and paste the wave components back
into OrcaFlex. Finally, delete (or scale down) those components that are shorter than about 3
diameters.
Alternatively you can copy the spectral density values, by clicking the 'View Spectrum'
button, then right clicking and selecting Values, and then copying the table of spectral
frequencies and density values. Then change the wave type to 'User Specified Spectrum' and
paste the spectral frequency and density values back into OrcaFlex. Finally, delete (or scale
down) the last few entries to truncate (or reduce) the spectrum for wavelengths less than
about 3 diameters.
Note that these two methods will not give exactly the same results, since in the latter OrcaFlex
will rediscretise the truncated spectrum. But they should give statistically equivalent results.

W Orcina

Moorings: C07 Metocean Buoy in Deep Water Page 1 of 3
C07 Metocean Buoy in Deep Water
Introduction
The example shows a metocean buoy moored in deep water. It is held in place by a long mooring
line anchored at the seabed. A regular wave train has been applied.

Building the model
The metocean buoy is modelled as a 6D Spar
Buoy so as to capture the interaction with the
sea surface. See the help topic | System
Modelling Data and Results | 6D Buoys |
Modelling a Surface-Piercing Buoy. A
smaller buoy at mid-depth is represented by a
Clump attached to the line. The mooring line
itself is made up of several line types with
different properties.

The metocean buoy geometry is defined in
terms of a stack of cylinders of given
diameter and length. The stack base and the
centre of mass are positioned relative to a
user-defined origin. In this case the origin lies
on the buoy axis 2m above the stack base.
The centre of mass is placed at the origin.
The program calculates the position of the
centre of buoyancy as described below.
The cylinder geometry is used to determine the buoyancy forces on the buoy, taking account of
the instantaneous position of the water surface and its slope, relative to the buoy. Therefore the
centre of buoyancy position is a product of the cylinder geometry and the buoy position.
In this example we have considered the tail (bottom three cylinders) as free-flooding. This is
sometimes used to increase buoy stability. Therefore the bottom cylinders have been given an
inner diameter. The water inside is considered trapped for horizontal motions (the walls will
prevent flow) but free for vertical motions (open ended). This trapped water is automatically
included in the buoy properties by the program. In cases where the vertical inertia of the trapped
water is important, it can be represented by a dummy line with appropriate properties attached to
the buoy instead.
The Bulk Modulus can be assumed to be infinite (this data is only required for submerged
systems where the buoy compressibility has a significant effect on the buoyancy).
For complex shapes the geometry is best represented by something like a strip model. For
example, we subdivide the tail tube into a number of cylindrical sections, and may use a number
X
Z
80 m
OrcaFlex 9.3a13: B07 Metocean Buoy in Deep Water.sim(modified 14:01 on 13/01/2009 by OrcaFlex 9.3a13) (azimuth=270; elevation=0)
Replay Time: 95.90s
W Orcina

of thin slices to represent the tapered sections of the top and bottom of the hull. When we
consider the forces acting on these slices we must use values which are appropriate for the whole
shape, not the individual slice.
This strip method is important when modelling surface piercing objects. For example compare a
cylinder and a cone of identical volume and height. Both are immersed to half their height. Yet
the displacement for the cone is significantly greater than for the cylinder. The spar buoy
discretises the cone into a series of cylinders as shown by the dotted image below.

OrcaFlex calculates the hydrodynamic forces on each component cylinder independently.
However it does not calculate shielding of one cylinder by another. This must be accounted for
when setting drag and added mass values.
Drag areas in the normal direction are simply area x length for each cylinder. However remember
the diameter must be set to give the correct displacement. With more complex structures this may
be different from the actual diameter. Use the actual diameter when calculating drag area.
Drag in the axial direction needs to be considered carefully because only the part of the cross-
sectional area exposed to the fluid is relevant. This may differ on the upper and lower parts of a
cylinder, in which case some judgement is required. In this case the exposed areas have been
considered from the base up, however OrcaFlex does not differentiate between exposed areas up
and down so will apply for both directions. You need to consider both options..
When determining the coefficients to apply, split the cylinders into groups. In this instant there is
a hull group (cylinders 4 and 5) and three short cylinders, the counterweight (cylinder 7) and the
tail (cylinders 6 and 8). Apply the resulting coefficient to all the cylinders in the group.
Apply the similar methodology for added mass shielding and coefficients.
OrcaFlex does not include radiation damping (wave-making at the water surface), although this is
an effective way of extracting energy from a surface-piercing buoy, particularly for small
amplitude motions where the quadratic drag terms become extremely small. Radiation damping
forces vary linearly with velocity and are represented in OrcaFlex by the unit damping force and
unit damping moment terms. It often proves impossible to obtain theoretical values for these and
they are normally set by adjusting the values until the OrcaFlex motions match the results of still
water heave and rock tests.
W Orcina

Results
Look at the close up view of the buoy while running the latest wave animation. The buoy follows
the local water surface variations well in terms of its vertical motion (heave) but less well in terms
of its inclination (pitch) That means that buoy heave is a good measure of wave height but buoy
pitch is not a good measure of wave slope.
Look at upper graph of buoy X and Z motions. The X motion has a mean drift: if the simulation
had been run for a longer duration this would have settled.
Now look at the lower graph of buoy Z against Sea Surface Z at the buoy. The graph very
quickly settles to a straight line at 45, with Z amplitude equal to wave amplitude. This confirms
the visual observation that the buoy heave motion is a good measure of wave height.
Also be aware fo the short wave issue where the wavelength is less than 3x the buoy diameter.
This is detailed in Example C06.
W Orcina

Moorings: C08 Fish Farm Page 1 of 2
C08 Fish Farm

Introduction
Two fish farm cages are modelled. They are connected to each other via moorings to give a 2 cell
structure. Be sure to have Model Browser on View by Groups.
Building the model
The important result for this model is the behaviour of the net as a complete entity. This allows
the net model to be simplified in two ways:
The net mesh needs to be modelled in sufficient refinement to show the distribution of
loading. This means an equivalent mesh can be generated that has the same resultant loads
but does not need to show each individual knot and line. This is basically the same as
defining the mesh refinement on a surface for an FE model.
Detailed motions at each knot and the length of rope between are not required. Therefore the
lines can be single segment and the knots can be 3DofF buoys. Pinned connections between
the two can be used because bend stiffness is negligible so moment transfer would be too.
These nets are suspended below floating rings. Again single segment lines are used. However the
plastic ring does have bend stiffness so the bending moments need to be transferred. Therefore
6DofF buoys and built in connections (end connection stiffness of Infinity) are used.
W Orcina

Moorings: C08 Fish Farm Page 2 of 2
The whole structure is then moored using more lines.
When building this type of structure use the Group facility to build a basic unit then copy, paste
and move within the model browser. Remember you can right click on a model browser group to
translate or rotate all objects within it as a whole set. Also, you can copy and paste objects by
using the standard Ctrl+C and Ctrl+V shortcut keys.
Results
Replay the animation to see the cells respond to the applied loading.
W Orcina

Riser Installation: D02 Pull In Analysis Page 1 of 3
D02 Pull In Analysis
X
Y
50 m
OrcaFlex 9.2a25: C1.2 Pull In Analysis.dat (modified 16:10 on 31/03/2008 by OrcaFlex 9.2a25) (azimuth=315; elevation=90)
Time: 91.0002s

Introduction
A line has been laid on the seabed in a specified route. It is to be pulled in to a wellhead funnel by
a winch. A gateway ahead of the funnel controls the pipe entry so it approaches at an angle that
will not cause the pullhead to jam.
On opening, the default view shows a wireframe representation of the lay route. The grey line is
the position that the line begins in. The wellhead is at the left.

Building the model
The line is laid in a prescribed path, represented by the grey line in the view.
Look at the Prescribed Starting Shape on the Pipe page. The track is made up in a similar manner
to a childs railway track. Lengths of straight and curved pieces are specified, each beginning
where the last ended. Note the radius of each curved piece is constant.
You can input these pieces by specifying the length and radius or the location that each point
ends. OrcaFlex will determine the other set from what you input. The initial direction of the first
piece also needs to be specified, though it is usual to have it match the Line end Azimuth angle at
End A.
During statics OrcaFlex will lay the line out along the track. Each nodal position is specified from
the distance along the track. Note that OrcaFlex also allows for any axial strain from the As Laid
Tension.
In this model we know the as-laid position of the line and so we do not want OrcaFlex to shift far
to find a static equilibrium position. This is accomplished by using strongly restrained Step 2
statics. The statics method has been set to Mag of Std. Error / Change. The Mag of Std.
Change data item at the end of the table has been reduced from 0.2 to 0.01. The effect of this
parameter is to restrict the line to making only very small motions during a single static iteration.
Finally some extra damping and more iterations are added to assist in settling the system.
W Orcina

This method of settling statics, where a line is forced to remain close to a known equilibrium
position is quite widely applicable. It could be used for lines passing over chutes or through
guides, where the line must stay close to a spline fit starting position.
The gateway is modelled as two lines of the required diameter. They and the pipe are given
contact stiffness values and the clashing option found on the Structure page of the line data forms
is turned on (see below for more on this). Accurate contact stiffness values are rarely known for
pipes: in this case, all we require is that the gateway effectively constrains the pipe, so an
arbitrary value can be applied. Take a look at | Theory | Line Theory | Clashing | in the helpfile for
interpretation of the clash forces.
To turn the clashing routine on, say 'yes' to Clash check on the 'Pipe', 'Gateway 1' and 'Gateway 2'
pages. Note that activating the Clash check will result in slower runs so it should only be applied
where required. For 'Pipe' it has only been selected for the first 30 m of the pipe; the length
expected to contact the gateway during the pull-in.
Remember also that clashing between lines is not active during the static stage. This is not an
issue in this example because contact occurs during the dynamic stage only.
To help visualisation in the wire frame view, two Shapes, 'Gate 1 Shape' and Gate 2 Shape', have
been positioned to coincide with the Gateway lines. Zero stiffness has been specified so that no
contact forces are applied to the lines by the shapes; contact occurs purely from the line-on-line
contact. Since contact between a line and a Shape takes place at the nodes of the line only, it
would not be appropriate to use the Shapes as guides. This would demand very fine segmentation
of the Pipe line and so a bigger matrix to solve.
The analysis is carried out with the following stages:
Note that winch pull-in is denoted by negative payout the standard OrcaFlex convention.

Stage Time Winch Payout Pipe End A
'To Gate' 'To Hub'
Statics Zero Tension Zero Tension Anchored
0 -1 s to 0 s Hold Relax Settle
1 0 s to 46 s -87.84 m Relax To Gateway
2 46 s to 86 s Release -15 m Into Funnel
3 86 s to 91 s - Hold Settle
Note that 'Relax' means the winch length varies to keep the tension at 0 kN. Hold means that the
winch length does not change during that stage.

It is important that winch lengths in OrcaFlex never reach zero (or close to zero) as this causes
convergence problems. However, it is convenient to pull-in to a connection point on the winch as
this exactly defines the final pulled-in position, and makes the calculation of the pull-in length
required more straightforward. In this example, each winch has three connection points. The first
and second connection points are both anchored to the seabed with a (vertical) distance of 2 m
W Orcina

between them. If we pull in to the second connection point, we can easily calculate the pull-in
required (total winch length at the start of the pull-in minus 2 m) and avoid the problems
associated with zero length winches. For the To Gate winch this starting length is simply the
static length of the winch (with zero tension applied).
The winch 'To Gate' is attached to End A of the pipe, on the seabed in the middle of the gateway
then 2 m vertically above this point. In Stage 1, this winch pulls the pipe to the gateway; it is then
released. The winch 'To Hub' is attached to End A of the pipe, the far end of the funnel on the
seabed and 2 m vertically above this point. In Stage 2 this winch pulls the pipe from the gateway
up into the funnel.
The pull-in model requires two winches because there is no interaction between winch wires and
other structures. If only one winch was used, it would pull the line straight into the funnel,
missing the gateway.
At the end of the pull-in, the system is allowed to settle for a few seconds.
The prescribed shape is intended for lines that have all their length in contact with the seabed. If
the line has End A off the seabed then OrcaFlex will specify the nodal X and Y coordinates first,
and then add Z. The line height above seabed varies linearly from the raised End A down to
touchdown. This does not represent a realistic situation, as the line ought to hang in a catenary
shape to touchdown.
Also the line segmentation needs to be refined enough that it can follow any curves in the track.
For example a segmentation length of 5 m will not be able to follow a bend radius of 1 m.

Results
Go to the default view, CTRL+T, and look at the animation through the whole simulation. Try
both wire frame and shaded graphics views (toggle between then using CTRL+G). Note that if
you simply press Ctrl+G from the the default wireframe, you must zoom in to drop below the sea
surface. Watch the line being pulled first to the gateway.
Then load the results workspace. This shows close up views of the funnel and tension time
histories for the winches. Run the replay again and watch the line curve around one of the
gateway lines.
Both graphs show winch tension overshoots and oscillations. These are a consequence of
speeding up the operation to save processing time - the actual pull-in operation would probably
take an hour or more compared to just over a minute here.
Slowing the rate of winch pull-in will reduce the overshoots and oscillation amplitudes, but at the
cost of longer run time.

W Orcina

Riser Installation: D03 Lay on Tower Page 1 of 4
D03 Lay on Tower
Introduction
This is an example of installing a riser
over a subsea tower in 300m water. The
riser is being paid out from the moonpool
of the installation vessel, and is supported
from an auxiliary winch at the bow of the
ship. There is a clamp about 10m above
the attachment point of the winch, which
has to be placed in the centre of the
support tower.
This example shows:
Guiding a line into position using
shapes
Friction between lines and
shapes.
Clearance checks with moonpool
edges.
Getting physical and
hydrodynamic properties for a
winch wire
Shielding the length inside the
moonpool from direct wave and
current loading.

The installation will take place very slowly and only in calm weather, so a series of static
snapshots would be sufficient. However it is simple and convenient to run the analysis
dynamically and the replay gives a valuable view of the operation.
You need to view the model browser in the View by Groups mode.
W Orcina

Building the model
Look at the Vessel Structure group.
The vessel has a shape attached
representing the inside of the
moonpool, Moonpool Shielding.
It is blue in the view here.
The shape type is trapped water so
any nodes within the shape will
have the wave and current load
algorithms turned off. This
represents shielding from direct
wave and current loading.
When the nodes leave the shape
these forces are turned back on
again.
Note this is not the same as setting drag and added mass to zero. The line will be moving in the
moonpool water and so relative flow will not be zero. This means drag and added mass on the
line will not be zero.
The boundaries of the moonpool edge are marked by single segment lines in Moonpool Top
and Moonpool Bottom. These lines have negligible physical and hydrodynamic properties but
have a contact radius of 1m (contact diameter 2m). Each is positioned with centreline 1m from
the moonpool edge so the line outer edge is at the boundary. Clearance information can then be
obtained between the flowline and the moonpool. The boundaries are pale blue in the view above.
Now look at the Winches and Flowline group.
The flowline is split into two lines connected either side of
the clamp, see Flowline. They have encastre
connections to the clamp (infinite connection stiffness).
More refinement of segment length is applied in the
region of the tower so the lines can better follow the curve
of the arch.

The clamp is modelled as a 6DofF Spar Buoy, clamp buoy, with small diameter at the ends and
a larger one in the middle. Rotation about the local x and y axis are restrained by the flowline
connections and reaction from the arch.
For restraint about the local z axis the CofG is offset slightly below the centreline of the buoy.
This is to provide a righting moment to prevent free spin. The alternative would be to turn on
torsional stiffness but this means a larger matrix to solve so can give slower runs and convergence
is harder.
The buoy contact area is 3.6m
2
. Contact is with shape vertices. You need to change the buoy
drawing settings to Draw square cylinders and the view to wireframe if you want to see them.
There are 24 vertices in total so each represents an area of 0.15m
2
.
W Orcina

When the clamp is being slid, only four vertices will be in contact, those for the larger diameter.
This means a contact area of 0.6m
2
. When dropped in place there will be eight vertices in contact
(for the smaller diameter cylinders) so 1.2m
2
. The exact values are not critical in this application
but it shows the principle.
Both flowline lengths and the clamp buoy interact with the arch shapes. This reaction includes
friction. Take a look at the Solid Friction Coefficients page in the model browser for settings.
The installation operation here is performed using two winches, a main and an auxiliary.
The main winch represents the main riser handling winch in the moonpool of the lay vessel; It is
attached to the flowline End A (the top end) and paid out to lower the flowline slightly. Because
the winch wire is short and remains in air the stiffness and length of the winch object are
sufficient properties.
The auxiliary winch is at the stern and to lift the riser into position. It is much longer with most of
the length submerged. See the structure in the Clamp Control Winch group.
The mass, displacement and hydrodynamic loading on this
wire could be significant for the overall system behaviour.
Therefore most of the winch wire length has been modelled
as a line, Aux Winch Wire. The winch object itself, Aux
Winch, has only sufficient length to allow pay in without
ending up with zero length.
The Aux Winch Wire is attached to the Lower Flowline
via a short link, Winch Wire to Flowline. This avoids the
need to break the flowline into two and use a buoy. Because
no moments would be transferred a short link is sufficient.
The tower arrangement is shown in the Tower Structure
group. It uses a dummy vessel, Tower, as a common axis
system and all calculations are set as None. The vessel
also provides the graphics for the tower legs. The arch and
hole for the clamp are produced by shapes with contact
stiffness so they generate a reaction force.

The winches are active in different stages to enable the required order. The stages are as listed
below and finish with an 8sec settle:
Stage Main Winch Auxilliary Winch
Payout Rate Payout Rate
-8s to 0s 0m ~ -5m -0.625m/s
0s to 16s 5m 0.625m/s 0m ~
16s to 32s 5m 0.625m/s 0m ~
32s to 48s 0m ~ -4m -0.5m/s
48s to 64s 0m ~ 0m ~
Stage 0, the wave build-up stage, can be used here as no waves are applied.
W Orcina

Results
The default workspace shows a number of system views. Look at the animation for the whole
simulation. Note how the clamp drops into its slot.
Open the workspace D03 Curvature and Contact.wrk. The right hand plots show curvature
about the local y axis (curving in the vertical plane). The left hand graph shows clearance of the
flowline from the moonpool edges.
Run the replay and note how the curvature changes close to the arch. Also how the minimum
clearance (clear water) varies along the arc length as the line is paid out and moved.

W Orcina

Riser Installation: D04 J-Tube Pull In Page 1 of 4
D04 J-Tube Pull-In
1. Introduction
This example shows how to model a typical J-tube pull-in operation. It includes paying a line (a
riser) out from a vessel over a chute whilst simultaneously pulling in the other end up a J-tube. It
demonstrates the Line Contact model and introduces some tips for controlling the line position in
the static solution. Winch objects are used to control both the pay-out and the pay-in of the riser,
and friction is included in the contact between the riser and the J-tube wall. Note that objects have
been modelled not to scale, for clarity.
The areas of interest in this type of analysis would most likely be the seabed clearance, possible
jamming of the pull-head in the J-tube and pull in loads on the winch. This model also captures
flexing of the J-tube in response to the pull-in loads and the forces placed on the clamps holding it
to the main structure.
When you open the simulation file the default workspace is opened up which shows several views
of the J-tube and vessel/chute arrangement. Replay the simulation and you will see the riser and
pull head being pulled up into the J-tube. Notice that the riser and J-tube highlight white when
contact occurs. The bottom right hand view shows the line sliding along the deck surface.

W Orcina

Ensure that View by Groups is selected (right-mouse click in the Model Browser to find this
option).
Select the Workspace D04 J-Tube Pull In vessel.wrk from the Menu bar. This shows a shaded
view of the vessel and the top end of the riser passing over a chute. The end of the line is
connected to a winch object called Payout. At present, it is not possible to add extra line length
during a simulation, therefore the full length of line that you want to include in your analysis
needs to be in the model from the start. This means that you may have to start with an overly long
frictionless deck shape to accommodate this. An Elastic Solid Shape is used to model the deck
surface; this is named Deck and is currently
hidden in the model. In the Model Browser, select
this object, right-mouse click and select show to
make it visible (or use the shortcut Ctrl+ H).
The riser passes through a moonpool and will be
shielded from the environmental loads in this
region. This is modelled by adding a Trapped Water
type shape to the vessel (named Moonpool
Shielding in the model), which will modify the
fluid motion within the shapes boundary to remove
the environmental contribution.
The chute is also modelled using Elastic Solid
Shapes. Notice that the chute has been split into two
parts, a left hand and a right hand section. This is
the preferred way to model a tight chute; if the chute is modelled with just one shape then
OrcaFlex can have problems when trying to find a static solution because the contact force for the
shape keeps switching directions as the line nodes bounce between the side walls. With the chute
split into two shapes then the contact force is constant for both shapes, as they have a wall each.
The J-tube itself is modelled as a line with pre bend. Open the J-tube Line Data Form and look at
the Pre-bend page to see how this is done. Note that bending is given relative to the nodal x and y
directions. Try turning nodal axes on when setting these so you can remind yourself if you rotate
about x or y. Positive direction follows the RHS
rule. Both ends of the line are built-in (they have
infinite connection stiffness) and therefore have
their End Orientation angles set appropriately (see
System Modelling: Data and Results | Lines | Line
Ends | No-moment Direction of OrcaFlex Help for
more details).
Note however that End B of the J-tube is
connected to a 6D buoy with negligible properties
called Connection. This is there simply to enable
the bellmouth shape and the J-tube line to be
connected to each other, while leaving the bottom
end of the J-tube free to deflect in response to
loads arising from the pull in operation.
The Line Contact model is used to specify that the
W Orcina

riser line should pass through the bore of the J-tube. Because the J-tube is modelled as a line, it is
able to flex under loading as the riser is pulled through it. A second Line Contact relationship is
used to model the clamps holding the J-tube to the structure.
Open the Line Contact Data Form from the Model Browser. On the Relationships page there are
two relationships defined, one between the Riser and J-tube lines and another between the Tower
and J-tube lines. In the first relationship, the riser is the Penetrating Line and the J-tube is the
Splined Line. The relationship is of type Inside; this means that the J-tube is fitted with a
smooth spline surface and the riser is fitted with penetrators that will contact with it. The
penetrators are positioned at the nodes of the riser line.
Containment has been enabled; this means that the inner line (the riser) will be shielded from the
environmental fluid forces by the outer line (the J-tube) wherever it is positioned inside it. The
contents of the J-tube line are set to Free Flooding therefore the external pressure and buoyancy
force acting on the inner line are as for a normal (non-contained) line.
The second relationship is of type Around. This means that penetrators on the Tower line are
placed around the J tube line at the positions and offset as defined on the Penetrator Locations
page.
A friction coefficient has been assigned to apply axial friction between the riser and the J-tube.
Making the J-tube the splined line means that a spline tube is created which has the inner
diameter of the line type as the bore diameter, and that has infinite wall thickness. This means that
the static solution is found easily, as the section of the riser that is detected to be between the ends
of the J-tube is pushed into the centre of the spline tube.
Select D04 J-Tube Pull In j-tube.wrk from the Menu bar. Reset the model (F12) and then run
statics again (F9) and watch how the lines behave. Note that during Full Statics the lines do not
interact with each other; the Line Contact model only comes into play during Whole System
Statics and this can be seen clearly when the riser is suddenly pulled inside the J-tube.
An Elastic Solid Shape has been positioned at the bottom end of the J-tube to model the bell
mouth. Because the real bell mouth will have a relatively thin wall thickness, a Drawing type
Shape has been used to show the bell mouth visually. A Drawing type Shape has no stiffness and
therefore does not interact with other objects. A second bell mouth shape has been created with a
much thicker wall to act as the structural boundary. This shape is an Elastic Solid type, therefore
is has a stiffness and will interact with the line nodes that contact it. This item is hidden in the
model (called Bell mouth guide). Making this boundary shape thicker than the physical
structure removes the risk of line nodes that penetrate into the shape being pushed to the outside
surface instead of the inside one. This happens if the nodes penetrate past the mid-point of the
shapes thickness, which can happen if the shape is thin-walled; nodes are pushed to the nearest
surface hence in this situation the outer surface becomes the nearer one.

Open the Payout winch data form; the dynamic simulation is split into three stages to allow the
winch to be controlled by stage. Stage 0 is the usual build-up stage, where the waves are allowed
to gradually build up, and the winch is stationary during this period. Stage 1 is included to allow
the winch to change smoothly from stationary to a non-zero payout rate by using the Specified
Payout Rate Change option (a sudden change in winch speed can cause instability or run
slowdown in the dynamic simulation, unless very short time steps are used). The winch then pays
out 40m of line at a constant rate during Stage 2 dynamics.
W Orcina

At the other end of the riser, a second winch object Up Tube is used to pull the riser into and up
the J-tube. Because the winch wire has to pass through the bore of the J-tube, and Winch objects
dont interact with Shapes, part of the winch wire is modelled with a line. OrcaFlex allows you to
model different sections of a line with different properties; hence a single line is used in this
model to represent a section of the winch wire, the pull-in head and the riser. Double click on the
Riser line in the model browser to see the various sections. The Winch object is used to pull the
line through the J-tube and it needs to be long enough so that its length does not reach zero at the
end of the pull in.
The riser line needs to find a solution in statics where it is on top of the vessel deck and over the
chute, as well as being inside the J-tube. The Line Contact model ensures that the line is forced
inside the J-tube bore, however some assistance is required to help find the correct position on the
vessel deck. To do this the statics method used is Spline, which gives the line a good starting
point. If you reset the model again (F12) you will see the spline visible as a white line. Also now
visible is an additional winch object ChuteWinch connecting the riser to a point on the vessel
close to the chute (press Ctrl + T to see the entire model view). This winch is temporary; it pulls
the riser into place in statics, pays out a little in the build-up stage of dynamics to release any
tension, and then releases completely at the start of Stage 1 dynamics so that it doesnt contribute
to the simulation.
OrcaFlex needs to know some information about the sense of the riser line, because some of the
calculations performed need to know which end of the line (End A or End B) is at the top and
which is at the bottom. For example, if a line contacts the seabed, the touchdown results point is
determined by starting at the top end and moving towards the bottom end until the first node in
contact with the seabed is found. In this example, both ends of the line are lifted up, so there
would be two touchdown points if the line contacted the seabed. To indicate that we want the
touchdown point to be reported as the point closest to End B that touches the seabed, we need to
set the Top End data item on the line data form to End B (rather than the default setting of
End A). Full details of the calculations that use this data can be found in the Help file section
System Modelling: Data and Results | Lines | Line Data.
The model also needs some damping applied during Whole System Statics to assist with finding
the static solution. This is applied on the General Data Form, on the Statics page.

3. Results
Re-open the simulation file and select D04 J-Tube Pull In results.wrk. Range graphs for the
tension and curvature in the Riser line are displayed, along with a graph showing the force
exerted on the clamps holding the J-tube to the structure (bottom left).
A time history plot of the Up Tube winch tension is also shown. In some cases, using the around
line contact relationship instead of the inside relationship will give a smoother pull-in (and
therefore a smoother winch tension result plot), due to the fact that with the around relationship
you are pulling the smooth spline surface over the penetrator ring at the mouth of the J-tube,
rather than the other way round. In this instance however, both relationships give similar results,
so the inside option has been used here as it gives an easier static solution.
W Orcina

Pipelay and Recovery: E01 Simple Rigid Stinger Page 1 of 5
E01 Simple Rigid Stinger
1. Introduction
In these examples, two different methods of modelling a simple rigid stinger are presented. The
first uses Links to model the reactions between the lay pipe and the rollers; the second uses the
Line Contact model to do the same thing. The two models have been created using an identical
lay pipe and vessel, and the four rollers have been positioned in the same places. The results are
therefore directly comparable. Which method you choose to use will depend on which you find
easiest to build, and also on whether you need to model axial movement of the pipe at some stage;
the line contact model can handle axial movement, the link model cannot.

2. OrcaLay Style Model with Additional Lateral Restraint
X
Z
10 m
OrcaFlex 9.3a48: C02.01 OrcaLay Model plus lateral restraint.sim(modified 15:27 on 08/07/2009 by OrcaFlex 9.3a47) (azimuth=270; elevation=0)
Statics Complete

This model (E01 OrcaLay plus lateral restraint.sim) is based on the simple method used in
OrcaLay for multiple static analyses of lay configurations. The reactions between the lay pipe
and the stinger rollers are generated by Tether type Link objects, which have strength in tension
only. This has the advantage of speed and simplicity over more detailed treatments, but does not
allow for movement at the tensioner. It therefore represents the common situation where the pipe
is held fixed during welding operations.
The OrcaLay model is two-dimensional and consequently has tethers in the vertical plane only.
This model extends the idea by adding tethers to provide lateral restraint as well as vertical, and
can therefore deal with three-dimensional motions due to out-of-plane environmental conditions.
This type of model is not suitable for modelling the situation where the tensioner allows axial
movement of the pipe, because the roller locations (link connections) would move with the pipe,
which is unrealistic.
W Orcina

The method allows for lift off the rollers, but the method used in setting up the geometry
implicitly assumes the pipe lies on the stinger over its entire length. Any significant departure
from this condition will lead to some inaccuracy in the geometry. Nonetheless, it does provide a
simple and robust technique for modelling many pipelay operations.
2.1. Building the Model
The active parts of the model consist of a lay vessel, the pipe being laid and a set of roller boxes.
These are best seen with the model browser set to View by Groups.
Each roller box is represented by three tethers and a three roller shapes. The latter have no
stiffness (because they are drawing type shapes) and serve purely as a visual indicator of the
roller positions. The physical restraint is provided by the three links. These are positioned so that
they become taut when the pipe comes into contact with a roller. Because tethers have no
strength in compression, they limit movement in only one sense of direction. The tethers are
extremely long so that the tether tension has a negligibly small component along the axis of the
pipe. In principle the two lateral tethers could be replaced by a single spring/damper with
appropriate non-linear properties, but the two-tether approach is conceptually simpler and, as a
result of the symmetry of the system, easy to build. The tethers can be hidden in the model if you
do not wish to see them.
The lay pipe is built-in at its connection to the vessel with infinite stiffness and declination 90.
The target segment length is set to 2 m over the upper section of the line (the roller region) and in
the touchdown region. Longer segments are used elsewhere. Note that the lengths of the sections
that lie over the stinger have been set to match the roller spacing allowing for the stinger radius.
The length of the first line section is set to the distance between the connection point on the vessel
and the first roller. This means a node is at each roller position because links can only attach to
nodes.
The vessel is the OrcaFlex default Vessel, which is treated as fixed in position during the static
calculation. The lay pipe and its supports are placed on the vessels axis of symmetry but can
easily be repositioned by using the Move facility in the model browser. Right click on an object
or group and you will see the link to the Move facility.
The model also contains a 6D Buoy placed at the centre of the roller radius and used to position
the solids and links representing the roller boxes. It basically provides a ramp local axis system.
The procedure to add an extra roller to the stinger is as follows:
1. Duplicate one of the groups representing the roller boxes,
2. Change the connection for the 6D Buoy from Vessel to Free, so that objects can be
attached to it.
3. Change all the Vessel connections in the new group to 6D Buoy, so their positions are
now specified in the ramp (buoy) local axis system.
4. Change Rotation2 for the Buoy to position the new roller box along the stinger radius.
This rotates the group with the buoy into the required position on the ramp radius.
5. Reconnect the new Links, Shape and finally the 6D Buoy to the vessel so everything is
now attached to the vessel and will move with it.
6. Adjust the line length to achieve the desired top tension, as described below.

W Orcina

2.2. Initial Set Up
In general, a pipelay configuration will be set up to achieve a specific level of top tension. In
OrcaFlex this is done via the Line Setup Wizard which is accessed through the Calculation
menu on the OrcaFlex main window. This allows either the lines length or its anchor position to
be altered to achieve a target condition. In this model the geometry of the upper sections of the
line must not be changed, as this will affect the link connection locations, so the Wizard was used
to alter the anchor position to achieve a top tension of 150 kN in still water.
The environmental conditions were then set as follows:
Slab Current 0.5 m/s 20 off the bow of the vessel.
Jonswap wave, Hs 1m, Tz 4s, 45 off the bow of the vessel. The wave preview facility was used
to identify the largest event in the first three hours of the simulation and the Simulation Time
Origin on the Wave page was set accordingly so that a 150 s simulation includes the largest
wave. The wave history applied can be seen by clicking on the View Profile button on the
Waves Preview page of the environment data form (note that the position used in the waves
preview is the point where the pipe connects to the vessel, which is the position where we want to
detect the largest wave).

2.3. Results
Loading the simulation file also loads the default workspace so results summaries are
automatically generated. The tension range graph (top left) shows that the maximum tension at
the tensioner is about 200kN, which is close to 1.5 times the nominal tensioner load. It also
shows that the entire pipe remains in tension throughout the simulation. The time history of
Effective Tension at End A (top right) shows how the tensioner load varies. The range graph of
Maximum von Mises stress (bottom left) shows that peak stresses are below 400 MPa at all times.
Loading the workspace file E01 Roller4 Loads.wrk shows what is happening at the fourth roller
box. The tension in the upper link (bottom right), representing contact between the lay pipe and
the lower roller, remains positive throughout showing that the pipe never lifts off the roller.
Contact with the port roller (identified as tension in the starboard link) and the pipe is much more
chaotic, with the pipe moving from side to side throughout the simulation as a result of the out-of-
plane environmental conditions. The time histories show that there is frequent intermittent
contact with the port roller, but no contact with the starboard roller (i.e. no tension in the port
link). This motion is clearly visible in the shaded graphics view if you run the replay.
Finally, note that the analysis uses a constant normal drag coefficient of 1.2 for the lay pipe. This
simple approach is likely to be conservative and the use of a variable drag coefficient based on
the local instantaneous Reynolds number would probably reduce stress levels.

W Orcina

3. Line Contact Stinger
3.1. Building the Model
In this model (E01 Line Contact Stinger.sim), we use the same vessel and lay pipe as used in the
previous example, but this time we use Line Contact to model the roller boxes. This method gives
a model that comprises of only three components (the vessel, the pipe and the stinger), however
we have also added some extra shapes for visual purposes.
The stinger is modelled using a Line. Open the workspace file E01 Line Contact Stinger
Views.wrk. In this model, drawing-type Shapes have been used to represent the rollers
themselves, in exactly the same way as the previous model. A 6D Buoy was again utilised to
make the placement of the roller shapes easier.
The line used to model the contact points of the rollers is currently hidden. In the model browser,
click on the Stinger line and select Ctrl+H to show it. An unusual looking line will appear with
large circular plates attached to it. Open the Line Contact Data Form from the Model Browser.
There are five contact relationships set up; the first has the stinger Around the pipe, which has
the purpose of ensuring that the pipe doesnt find solutions below the stinger. A further four
contact relationships have the stinger Outside the pipe to define the sides and bottom surfaces of
the roller boxes.
Look at the Penetrator Locations page; the rollers are defined by using large diameter penetrators,
which are offset from the stinger lines axis so that they form appropriate boundaries. Back on the
Relationships page, notice that the Around relationship has a much greater contact stiffness than
the other relationships. This ensures that this relationship dominates, and the pipe always finds a
position that is inside the guides. If you reset the model and then run statics, you will see the
tussle between the guides and the lower rollers taking place, with the guides winning due to their
greater stiffness.
The stinger line itself has pre-bend assigned to it, so that it forms the curve of the stinger. Open
the Stinger Line Data Form and look at the Pre-bend page to see this. It is set at 0.00625rad/m
so 160m radius (Radius = 1/Curvature)
3.2. Results
Re-open the simulation file. The file will open with the same workspace as the previous example;
the results are almost identical. Load the workspace file E01 Line Contact Stinger Roller4
Loads.wrk to see the loads on the last roller box. This time, the line contact results are displayed
rather than the link tensions as in the previous example. As a comparison, the results for the line
contact force at End B of the Stinger line (i.e. the last roller box) and the vertical link tension for
the same roller set in the previous model are plotted side by side below, using the same axis
scales. The left-hand graph is the Line Contact model, the right-hand graph is the Links model:
W Orcina

OrcaFlex9.6b: E01 Line Contact Stinger.sim(modified 09:26 on 02/01/2013 byOrcaFlex9.6b)
Time History: Stinger Line Contact Force at End B
Time (s)
150 100 50 0
S
t i n
g
e
r
L
i n
e
C
o
n
t a
c
t F
o
r
c
e
(
k
N
/ m
)
a
t E
n
d
B
50
40
30
20

OrcaFlex9.6b: E01 OrcaLayplus lateral restraint.sim(modified 12:28 on 21/12/2012 byOrcaFlex9.6b5)
Time History: UpperLink4 Tension
Time (s)
150 100 50 0
U
p
p
e
r
L
i n
k
4
T
e
n
s
i o
n
(
k
N
)
50
40
30
20

W Orcina

Pipelay and Recovery: E02 Rigid Hinged Stinger Page 1 of 5
E02 Rigid Hinged Stinger with Piggyback
Line

This model represents a rigid stinger hinged off the back of a lay vessel with the rollers not
represented as individual supports but as a continuous support with a constant radius of curvature.
This support is provided in the model by a set of elastic solid shapes. An additional line is
included to model a cable piggybacked to the pipe.
Because the interaction between the lay pipe and the stinger can tolerate axial movement of the
pipe, this model allows the pipe to move in and out of the two pipe tensioners. The model
therefore represents the situation where welding operations have been suspended and the pipe is
moving relative to the vessel so as to limit stresses. Modelling both tensioners is not generally
necessary, but it has been done here to demonstrate how easily it can be done. A third tensioner is
included on the piggyback line.
The method allows for lift-off the stinger and axial movement of the pipe, however the individual
loads applied to the pipe are approximated by contact with a continuous curved surface, so local
loads due to rollers are neglected. Nonetheless, the model provides a quick method of assessing
how a stinger set-up will behave, before embarking on a more detailed analysis.

1. The Model
The active parts of the model consist of a Lay Vessel, the pipe being laid, a piggyback cable and a
stinger. Open the Workspace file E02 Rigid Hinged Stinger Chute Exit.wrk to see a view
looking up at the point where the pipe exits the stinger. The stinger is modelled by two sets of
lines connected to a 6D Lumped Buoy with negligible properties. The hydrodynamic behaviour
of the buoy is determined by the lines attached to it, it has no significant properties of its own. A
W Orcina

further consequence of using a 6D Buoy as a frame of reference is that the stinger is rigid and
therefore this model provides no information on internal loads within the stinger.
The way in which the model was constructed is best seen with the model browser set to View by
Groups. There are two main groups called LayBarge and Pipe and Stinger plus a further
group called Static Guides which contains additional link and winch objects that are present for
modelling purposes and do not represent permanent parts of the system.
The first group contains the lay vessel, the pipe being laid and two constant tension winches used
to tension the lay pipe. There is also the piggyback cable and a third constant tension winch to
tension it.
The second group contains a 6D Buoy with negligible properties, a set of elastic solids used for
the boundaries that support the lay pipe, two stiff single-segment lines used to hinge the stinger to
the vessel, and two further groups containing the sets of single segment lines representing the
sides of the stinger. Each of these side groups contains thirteen such lines, each one 4m in length.
The main reason for modelling these as individual lines rather than individual segments in a
single line is that they can be independently ballasted more easily this way simply by changing
the contents density on a line-by-line basis. An additional advantage is that, because the lines are
connected at both ends to the 6D Buoy and have no intermediate nodes, they have no influence on
the stability of the numerical integration. These lines are all within a hidden elastic shape and
consequently show contact throughout the simulation. This doesnt affect the position of the
stinger though; all line ends and the shape are connected to the buoy, so they all move as one unit.
Setting up sets of lines like these is made straightforward by a combination of the duplication
facilities in the model browser and the All Objects Data form. By placing the set of lines in a
group, the user has the option of duplicating and moving groups using standard model browser
facilities and then re-naming the lines, or starting each set from a new line and repeating the
duplication and positioning process. Judicious choice of names at the start can make this process
run much more smoothly.
There are four elastic solids representing the interaction between the lay pipe and the stinger. In
fact only three of these are active because the one called Dummy Stinger base is a Drawing type
Shape which is used purely to indicate the curve over which the pipe is laid. The actual Stinger
Base is hidden and can be made visible by right-clicking on its icon in the model browser and
selecting Show. This reveals it to be a sector of a cylinder with radius equal to the stinger
radius. The reason for using such a bulky restraint is to ensure the pipe remains on top of the
stinger at the end of the statics calculation. The pipe assist and cable assist winches (in the
Static Guides group) are present for the same purpose; they help guide the lines into the correct
positions in statics, pay out to release any tension during Stage 0 dynamics, and then release and
disappear as soon as Stage 1 dynamics starts. The stinger assist link serves a similar purpose; it
controls the stinger buoy during statics, has no tension in the converged static state and releases at
the start of the dynamic simulation.
Friction is modelled between the Lay Pipe and the Stinger Base by assigning a friction coefficient
between them. Open the Solid Friction Coefficients form from the Model Browser to see where
this is set.
The lay pipe is built in at its connection to the vessel with infinite stiffness and declination 94.
The target segment length is set to 2.5m over the upper section of the line, increasing to 10m in
the suspended length. The seabed portion has 2.5m segments in the touch down area increasing
to 20m towards the anchor. Apart from the first section, the pipe is made of line type Lay Pipe
W Orcina

whose properties were derived with the Homogeneous Pipe line type category to be those of a
300 mm OD, 20mm WT steel pipe.
The first section of the pipe uses line type Lay Pipe Slider this is identical to the line type used
for the rest of the pipe except that its axial stiffness is set extremely low. This allows the upper
part of the line to move in response to the tension applied by the two tensioner winches. The first
of these is attached to the vessel at the same point as the pipe and to the pipe at arclength 20m, the
end of Section 1. The second tensioner is attached to the vessel at the same point and to the pipe
at arclength 30m. Both tensioners apply a force of 250 kN. It is not strictly necessary to model
two independent tensioners in this way, as one winch applying twice the force would produce
very similar results. However, users often prefer their models to resemble the physical
arrangement more closely, hence the approach taken here. Note that the first section of line
contains a single segment. This is necessary to ensure static convergence of the line, which for
more segments would prove extremely difficult because of the low axial stiffness. Note also that
the Homogeneous Pipe category does not allow the line type axial stiffness to be altered, so the
category for Lay Pipe Slider is set to General.
The piggyback cable is set up in the same way as the lay pipe, only this time using a single winch
object to model the tensioner.
The vessel is an OrcaFlex default vessel, which is treated as fixed during the static calculation.
The lay pipe and its supports are placed on the vessels axis of symmetry but can easily be
repositioned by using the Move facility in the model browser.
2. Initial Set Up
In general, a pipelay configuration will be set up to achieve a specific level of top tension. In this
case, the tension is applied directly by the two tensioner winches each applying a tension of
250kN in still water to give a total tension of 500 kN in the static calculation. In this model, we
wish to model the situation where installation is paused and the tensioners are allowing the pipe
to move relative to the vessel to minimise the dynamic loads.
Line Contact was used to piggyback the cable onto the lay pipe. Open the Line Contact Data
Form from the Model Browser; there are two contact relationships set up, one of which defines
the clamp positions that hold the cable to the pipe, and the other defines an Outside relationship,
which allows the outer diameters of the two lines to contact each other. For more details on the
line contact relationships, see the Help file System Modelling: Data and Results | Lines | Line
Contact | Data.
Line Contact only becomes active during Whole System Statics, in other words it is not
considered during Line Statics. In this model, this means that at the end of Line Statics the pipe
and the piggyback cable could be laying in positions that overlap each other (because at this stage
they dont interact with each other). When Whole System Statics starts (and Line Contact comes
into play), if the lines are penetrating into each other quite deeply it results in the initial line
contact force being unrealistically large, and statics therefore struggles to converge.
To prevent this, the Cable assist winches have been positioned so that they hold the cable away
from the pipe during line statics. The stiffness of these winches has been set to a value that is low
enough that the stiffness of the line contact clamps overrides them once Whole System Statics
starts. The images below show the two stages of the static calculation.
W Orcina


Pipe and Cable Positions after Line Statics (note the thin grey ring around the purple pipe this and
the thin orange ring above it shows the clamp i.e. it is there but not yet active).

Pipe and Cable Positions after Whole System Statics (Statics Complete cable has been pulled inside
the orange clamp ring)
The Whole System Statics Convergence Parameters were also adjusted (on the General Data
Form) to increase the Min and Max Damping applied, helping the system to converge.
Reset the model (F12) and run statics (F9) to see these winches, and the Line Contact model in
action.

The environmental conditions were set as follows:
Slab Current 1.0 m/s 20 off the stern of the vessel.
Jonswap wave, Hs 1m, Tz 8s, 45 off the stern of the vessel.
The simulation was run for 100s post build-up without searching for a particularly onerous wave
packet.
W Orcina

3. Results
Re-load the simulation file to load the default workspace, so results summaries are automatically
generated.
The range graph of von Mises stress for the whole simulation shows the variation over the entire
length of the line. Dynamic variations are small and the highest stresses occur in the stinger
region, as expected. The peak stress of just over 300,000 kPa (300 MPa) would be acceptable for
good quality steel.
The time history of tension for Pipe Tensioner1 shows the tension is held constant at 250kN while
the corresponding plot for length (top left) shows that there is a small amount of pay out and haul
in, but that the length remains close to 21 m throughout the simulation. Since this tensioner acts
over the first 20 m of the lay pipe, this shows that the initial set-up of the line is acceptable and
that the sliding section of line is not being unreasonably stretched or squashed. It is inadvisable to
let the link length get too small or instability can occur. This is because F=k*dL/Lo and you are
making Lo smaller so small change in dL gives a large change in load.
Loading workspace file E02 Rigid Hinged Stinger Chute Results.wrk opens time histories of in-
plane rotation for the stinger buoy and total contact force for the base of the stinger. These
summarise the motion of the stinger and the loads applied to it by the pipe. The workspace also
opens a range graph of solid contact force for the pipe. This shows the force per unit length
imposed on the pipe by the stinger support. Given knowledge of the roller spacing on a real
stinger, this data can be used to estimate the loads applied to individual supports. (Note that the
solid gives total force (kN) while the line gives contact force per unit length (kN/m). This is why
the magnitudes plotted differ).
OrcaFlex simulation files can sometimes become very large due to the amount of results data that
they contain. One option available to help reduce the file size is to increase the logging interval
(i.e. the frequency with which the results are recorded), which is a setting on the Dynamics tab of
the General data form. In this case, we have increased the logging interval to 0.5s, which is
sufficient for the rate of change of loading in this model because there are no high frequency
results to log.

W Orcina

Pipelay and Recovery: E03 Rigid Hinged Stinger with Rollers Page 1 of 5
E03 Rigid Hinged Stinger with Rollers

This model represents a rigid stinger hinged off the back of a lay vessel with the rollers
represented as individual supports modelled as pairs of single segment lines with clashing
enabled. There are rollers on both stinger and the lay barge. The lay pipe is restrained laterally
by the V-shaped configuration of the roller lines. An alternative method of modelling rollers on a
stinger is shown in Example E04 Articulated Stinger.
Line clashing is not applied in statics, so measures are required to control the position of the lay
pipe during this part of the calculation. The dynamic simulation is carried out in two stages:
firstly the line is lowered onto the rollers by fictitious winches without wave environment. This
simulation is run for long enough for the system to settle and the final state is preserved to
provide the initial conditions for subsequent cases including wave motion.
Because the clashing interaction between the lay pipe and the stinger can tolerate axial movement
of the pipe, this model allows the pipe to move in and out of the tensioner. The model therefore
represents the situation where welding operations have been suspended and the pipe is moving
relative to the vessel so as to limit stresses. Note the clashing interaction between lay pipe and
rollers is frictionless.
The method allows for lift off the stinger and axial movement of the line. The individual rollers
are modelled explicitly and represented by lines that will generate a reaction force when the lay
pipe contacts them. It therefore provides a detailed model of the pipe lay operation with all
relevant features accurately represented. The model is complex, but as described below its
W Orcina

construction and modification can be helped enormously by taking full advantage of the facilities
offered by the OrcaFlex model browser.
1. The Model
The active parts of the model consist of a Lay Vessel, the pipe being laid and a stinger. The
stinger is modelled by two sets of lines connected to a 6D Lumped buoy with negligible
properties. The hydrodynamic behaviour of the buoy is determined by the lines attached to it, it
has no significant properties of its own. A further consequence of using a 6D Buoy as a frame of
reference is that the stinger is rigid and therefore this model provides no information on internal
loads within the stinger.
Groups. There are two main groups called LayBarge and Pipe and Stinger plus a further
group called Temporary Elements that contains additional objects that are present purely for
modelling purposes and play no part in the ultimate dynamic simulation.
The first group contains the lay vessel, the pipe being laid, a constant tension winch used to
tension the lay pipe and a nested group called Vessel Box1 containing two single-segment lines
used to represent a pair of rollers connected to the vessel. In this model, there is only one set of
rollers on the vessel, although in practice there would usually be more. Creation of extra rollers
can easily be done by duplicating the group and then moving it using standard facilities of the
model browser.
The second group Stinger contains a 6D Buoy with negligible properties, two stiff single
segment lines used to connect the buoy to the lay vessel, a group called Runners and a second
group called Rollers. Each of these groups contains nested groups which contain sets of lines.
Runners contains two further groups containing the sets of single segment lines representing the
sides of the stinger called Port Side and Starboard Side. Each of these groups contains thirteen
single segment lines 4m in length. The main reason for modelling these as individual lines rather
than individual segments in a single line is that they can be independently ballasted more easily
this way simply by changing the contents density on a line by line basis. An additional advantage
is that, because the lines are connected at both ends to the 6D Buoy and have no intermediate
nodes, they have no influence on the stability of the numerical integration.
Rollers contains eight groups named Box1, 2,,8 each containing two single-segment lines
representing a pair of rollers. Placing the lines in groups this way makes it easier to reposition
them when optimising the roller layout. Note that the final group, Box8, contains two extra
vertical lines to limit the lateral motion of the lay pipe at the end of the stinger.
Setting up sets of lines like these is made straightforward by a combination of the duplication
facilities in the model browser and the All Objects Data form. By placing the set of lines in a
group, the user has the option of duplicating and moving groups using standard model browser
facilities and then re-naming the lines, or starting each set from a new line and repeating the
duplication and positioning process. Careful choice of names at the start can make this process
run much more smoothly.
The fictitious winches Winch1, 2, 3 are present to lift the pipe above the rollers at the start of the
dynamic simulation. They pay out during the build-up stage of the simulation and release at the
start of Stage 1, disappearing from the model. The payout for each winch is set so that the tension
W Orcina

falls to zero before the release occurs. Also the length of the build-up is set to a relatively long
20s to soften the impact between pipe and rollers and minimise transients in the system.
The target segment length is set to 2.5 m over the upper section of the line, increasing initially to
10m in the suspended length, followed by a reduction to 2.5m at touchdown. The seabed portion
has 10m segments throughout. Apart from the first section, the pipe is made of line type Lay
Pipe whose properties were derived with the Homogeneous Pipe line type category to be those
of a 200 mm OD, 10mm WT steel pipe.
part of the line to move in response to the tension applied by the tensioner winch. This is
attached to the pipe at its point of attachment to the vessel and at arclength 20m, the end of
Section 1. Attaching the tensioner to the line at the vessel end means the tension in the pipe at
this point includes the tensioner load. The tensioner applies a force of 200 kN. Note that the first
section of line contains a single segment. This is necessary to ensure static convergence of the
line, which for more segments would prove extremely difficult because of the low axial stiffness.
Note also that the line type category has been set to General to allow the axial stiffness to be
modified.
There are additional line types used for the pivots, runners and rollers.
2. Initial Set Up
In general, a pipelay configuration will be set up to achieve a specific level of top tension. In this
case, the tension is applied directly by the tensioner winch applying a tension of 200 kN in still
water. In this model, we wish to model the situation where installation is paused and the
tensioner allows the pipe to move relative to the vessel to minimise the dynamic loads.
Step 1
The first step in the process is to run the Stage 1 model to the end of the simulation, as shown in
file E03 Rigid Hinged Stinger with Rollers - Stage 1.sim. Running the replay for this model
shows how the pipe is lowered onto the stinger during the build-up stage by the winches, which
then release and disappear during Stage1. Loading the simulation file also loads the default
workspace so results summaries are automatically generated. These are time histories for the pipe
and the tensioner, which demonstrate that the system, although not totally settled, is behaving
calmly by the end of the simulation. There is no wave environment in this simulation, so any
motion present results entirely from lowering the pipe onto the stinger.
Step 2
We save the final configuration at the end of Step 1 back into a new data file by first going to the
Model menu and selecting Use specified Starting Shape for Lines (Time=50s). This writes the
position of the pipe and all other lines in the model into their data forms as a User Specified
Step 1 starting shape. It also sets the stinger buoy position in the same way. We then do a Save
As to save this as a new data file, in this case called E03 Rigid Hinged Stinger with Rollers -
Stage 2.dat
W Orcina

We want to prevent the static calculation from altering this line shape, which we do as follows:
On the statics page of the General data form we select Separate Buoy and Line statics and
exclude all buoy degrees of freedom from statics. On the data form for line Pipe we set the Step
2 statics method to None.
Since the lay pipe and stinger are close to equilibrium, we no longer need the group called
Temporary elements and could delete it. Alternatively, we could set the winch tensions and the
link stiffness to zero which would leave them in the model, but doing nothing, which is what has
been done here, although the entire group has been hidden from view.
In reality there is no such thing as a perfect constant tension winch and pay out and haul in will
only occur when the tension in the winch strays outside some deadband region. In this model, the
Tensioner winch has been set up with an assumed range of 5% for the deadband region. This
is achieved by setting the winch type to Detailed and the Deadband (+/-) on the Detailed
Properties page to 10 kN (5% of 200kN) . We also set the damping terms for Haul in and Pay
out to 0.1 kN/(m/s).
All that remains to be done before carrying out a real case simulation is to impose some
environmental loads. In this case a spreading stern sea based on the Jonswap spectrum has been
applied. The spectrum has been split into 5 wave directions with a spreading exponent of 24. This
is combined with a 0.5 m/s slab current onto the port bow (quartering seas). Note that imposing a
current at this stage, rather than in Step 1 may mean the system takes some time to settle down,
but this is preferable to including current in Step 1 which would restrict the subsequent analysis to
a limited range of environments. Of course, if the environment is close to unidirectional, then the
current could be included in Step 1. In this case the current has been imposed immediately, but in
some cases it may be necessary to ramp it during the build-up stage to help if static convergence
proves difficult.

The simulation was run for 50s post build-up with a Simulation Time Origin of 250s which
imposes a particularly steep wave rise on the system. The time step was reduced to 0.05s in this
model, to help keep the iteration count below the recommended 10 iterations per time step. The
simulation file is saved as E03 Rigid Hinged Stinger with Rollers - Stage 2.sim.
3. Results
Loading the simulation file E03 Rigid Hinged Stinger with Rollers - Stage 2.sim also loads the
default workspace. This shows two Range Graphs for the first 200m of lay pipe and Time
Histories for the tensioner.
The range graph of von Mises stress for the whole simulation shows the variation over the upper
part of the line; the stresses are low elsewhere. Dynamic variations are small and the highest
stresses occur in the stinger region, as expected. The peak stress of just over 273,000 kPa
(273 MPa) would be acceptable for good quality steel.
The range graph of Line Clash Force gives an indication of the interaction between the lay pipe
and the rollers. In this model, there is continuous contact between the pipe and the rollers, so
these forces will be reliable. Where there is intermittent contact, the clash forces will be sensitive
to the value of contact stiffness and time step. In these cases clash energy is a better indicator of
the severity of the contact.
W Orcina

The time histories of tension for the Tensioner show how the load varies between the limits
defined by the nominal tension and its associated deadband, 200 10 kN. The corresponding
plot for length shows that there is a small amount of pay out and haul in, but that the length
remains close to 13.6 m throughout the simulation. Since this tensioner acts over the first 20 m of
the lay pipe, this shows that the initial set-up of the line is not perfect and that the sliding section
of line is being squashed by about a third its unloaded length. This is not significant, but this part
of the model needs to be monitored to ensure the sliding length is not deformed excessively. Too
much stretch and the bending results will be affected, too much squashing and the sliding
segment will produce compressive loads that affect the tension applied to the pipe.
The simulation was run with the implicit integration scheme, and with a constant timestep of
0.02s; this was required to produce a good quality tension results plot for the tensioner. If a longer
timestep is used, the tension results for this model are much noisier.

W Orcina

Pipelay and Recovery: E04 Articulated Stinger Page 1 of 4
E04 Articulated Stinger
1. Introduction

This model represents a three-section articulated stinger hinged off the back of a lay vessel. Each
section of the stinger can hinge relative to the adjoining sections. Rollers are represented as
individual supports, which restrain the lay pipes motion both laterally and vertically.
This example demonstrates the use of the Line Contact model to create the rollers on the stinger,
however the method shown in Example E03 Rigid Hinged Stinger with Rollers could
alternatively be used in this example if preferred. The advantage of the method used here is that,
because it uses the Line Contact model, it is included in Statics and (if required) friction can be
included. Example E03 uses the Line Clashing model, which is not included in statics, and
therefore a two-step approach must be taken. In addition, the Line Clashing model does not
include friction effects.
Because the interaction between the lay pipe and the stinger can tolerate axial movement of the
pipe, this model allows the pipe to move in and out of the tensioner. The model therefore
represents the situation where welding operations have been suspended and the pipe is moving
relative to the vessel so as to limit stresses.
The method allows for lift off the stinger and axial movement of the line. The individual rollers
are modelled explicitly and represented by contact areas that will generate a reaction force when
the lay pipe contacts them. It therefore provides a detailed model of the pipe lay operation with
W Orcina

all relevant features accurately represented. The model is quite complex, but as described below
its construction and modification can be helped enormously by taking full advantage of the
facilities offered by the OrcaFlex model browser.

2. Building the Model
The active parts of the model consist of a lay vessel, the pipe being laid and a stinger. The stinger
is made up of three identical sections.
Groups. There are three main groups: LayBarge and Pipe, Stinger and Visual Objects.
Visual Objects are inside the Stinger group in the model.
The first group contains the lay vessel, the pipe being laid and a constant tension winch used to
tension the lay pipe. In this model there are no rollers on the vessel, although in practice there
would usually be one or more.
The Stinger group is split into three sub-groups, one for each stinger section. Each section
consists of a stinger line that represents the framework and buoyancy of the section, two short
singlesegment pivot lines used to connect the sections to each other or the vessel, and two links
used to limit the relative rotations of the sections. The unstretched length and stiffness of these
links determines the stinger radius and can be calculated with simple geometry.
All the above objects are connected to 6D Buoys with negligible properties, one buoy for each
section. The hydrodynamic behaviour of the buoys is determined by the lines attached to them,
they have no significant properties of their own. A further consequence of using 6D Buoys as a
frame of reference is that each stinger section is rigid and therefore this model provides no
information on internal loads within the stinger, apart from the loads in the links used to limit
rotation.
In addition to the active parts, there are also a number of drawing-type Shape objects in the
model. The method used to create the rollers makes the stinger simple to set up, but it does mean
that its appearance is a little unusual. Shapes are therefore used to create a more visually correct
representation of the stinger. These are found in the third group Visual Objects.
To see the active parts of the stinger, select the Stinger group in the Model Browser, right-
mouse click and select Show (or Ctrl + H). The large orange disks used to define the rollers will
now be visible.
Hide the Visual Objects group and Press Ctrl + T to switch to the default view. The active
objects that make up the three sections of the stinger will now be visible; they are coloured red,
yellow and green so that you can see which objects belong to which stinger section. Open the
Line Contact Data Form from the Model Browser to see the Line Contact relationships. In this
model there are fifteen relationships set up, which might seem complicated, however once the
relationships for the first stinger section are created, they can be easily copied and pasted to
quickly create the rest. For more details about using the Line Contact model in this way, see
Example E01 Line Contact Stinger.
Setting up sets of objects like these is made straightforward by a combination of the duplication
facilities in the model browser and the All Objects Data form. By placing the set of objects in a
Group, the user has the option of duplicating and moving each group using standard model
W Orcina

browser facilities and then re-naming as required. Careful choice of names at the start can make
this process run much more smoothly.
The target segment length is set to 2.0 m over upper section of the line increasing to 10m in the
suspended length before reducing to 2m again at touchdown. The seabed portion has 10m
segments throughout. Apart from the first section, the pipe is made of line type Lay Pipe whose
properties were derived with the Homogeneous Pipe line type category to be those of a 250 mm
OD, 15mm WT steel pipe.
part of the line to move in response to the tension applied by the tensioner winch. This is
attached to the vessel at the same point as the pipe and to the pipe at arclength 20m, the end of
Section 1. The tensioner applies a force of 400 kN. The tensioner winch is attached to the
laypipe at both ends to ensure that the Effective Tension reported at End A is equal to the
tensioner load. Note that the first section of line contains a single segment. This is necessary to
ensure static convergence of the line, which for more segments would prove extremely difficult
because of the low axial stiffness. Note also that the line type category was set to General to
allow the axial stiffness to be modified.
There are additional line types used for the pivots and stinger sections.
All that remains to be done before carrying out a real case simulation is to impose some
environmental loads. In this case a 2m Hs, 8s Tz stern sea based on the Jonswap spectrum has
been applied combined with a 1.0 m/s surface current onto the port beam.
A small amount of damping was applied in Whole System Statics (on the General Data Form) to
help find a static solution. The simulation was run for 100s post build-up with a Simulation Time
Origin of 9400s, which imposes a particularly steep wave rise on the system.

3. Results
Load the workspace E04 Articulated Stinger Results.wrk to view a set of plots that summarise
the behaviour of the system.
The range graph of maximum von Mises stress for Stage 1 of the simulation (top right-hand
graph) shows the variation over the upper part of the line; the stresses are low elsewhere.
Dynamic variations are large and the highest stresses occur in the stinger region, as expected.
The most severe stresses occur at the first and last rollers of the stinger and are fairly high at
around 500 MPa. This may be acceptable, depending on the material used, and given the
relatively severe environment.
The bottom right-hand graph shows the range graph for Line Contact along the section of the lay
pipe that is on contact with the stinger. The six spikes in contact force correspond to the six roller
positions. The bottom left-hand graph shows the instantaneous result; run the replay to see this
vary with time. Notice how the fifth roller has intermittent contact with the pipe, but all others
have continuous contact.
W Orcina

The time history of length for the Tensioner shows that there is a small amount of pay out and
haul in, but that the length remains close to 19 m throughout the simulation. Since this tensioner
acts over the first 20 m of the lay pipe, this shows that the initial set-up of the line is sufficiently
accurate. In this case, the model is behaving satisfactorily, but it is important to monitor this part
of the model to ensure the sliding length is not deformed excessively. Too much stretch and the
bending results will be affected, too much squashing and the sliding segment will produce
compressive loads that affect the tension applied to the pipe.
Finally, note that all these results are based on a constant normal drag coefficient of 1.2. Some
reduction in loads might be achieved by using Reynolds number-dependent drag.

W Orcina

Pipelay and Recovery: E05 Pipe Davit Lift Page 1 of 2
E05 Pipe Davit Lift
X
Z
30 m
X
Z
OrcaFlex 9.2a24: C4.1 Pipe Davit Lift.sim(modified 09:48 on 10/03/2008 by OrcaFlex 9.1c) (azimuth=270; elevation=0)
Time: 900.0000s

The free end of a steel pipe is lifted from the seabed by davits to bring it alongside a workboat.
The lift is modelled dynamically in small waves and no current.
Building the model
The line is initially laid out straight on the seabed. We use Catenary Step 1 statics followed by
Full Statics for this. Note that the initial position of the free end (End A) is deliberately chosen to
over-stretch the pipe, and the As Laid Tension is set to a low but non-zero value. These both help
statics convergence. When statics is converged, End A has moved to its equilibrium position.
Three winches are equally spaced on the vessel side and attached to the pipe near End A. As the
winches haul in, lifting the pipe, the pipe free end also moves horizontally towards the anchor. To
accommodate this, the workboat has to move forward.
The analysis is carried out in five stages as follows:
Stage 0 System settles - no haul in -2s to 0s
Stage 1
Winches pull-in about 45% of total;
workboat accelerates
0s to 300s
Stage 2
workboat moves at constant speed
300s to 600s
Stage 3
Winches pull-in last 10% of total;
workboat decelerates to zero speed
600s to 800s
Stage 4 System settles 800s to 900s
Note that the winches all pull in slightly different amounts to achieve the lift.
The winches are set to Specified Tension = 0 in statics. This means the pipe is laid out with no
load from the winches, and the starting length for each winch wire is set to give zero tension with
no slack. In subsequent stages, Specified Payout is negative, meaning that the winches haul in.
W Orcina

Workboat movement is defined on the Prescribed Motion page on the vessel data form. At each
stage we define the change in velocity. OrcaFlex arranges that this occurs smoothly by applying a
constant acceleration through the stage.
It should be mentioned that the default RAO data in OrcaFlex is in use for this workboat model.
The default RAO data are based on a slender tanker hull form, and so are not appropriate for a
workboat-shaped vessel. More appropriate data should be available for real project work.
A brief settling time is provided at the start and a longer one at the end of the simulation.
End B of the line is anchored to the seabed. The model would need to include a very long length
of line in order for there to be no tension variation at the seabed anchor due to the lift. In this
example, there is a tension rise at End B, because we have included less length of line in the
model.
Results
A results workspace is provided open the file E05 Pipe Davit Lift.wrk.
The tension time history through the simulation for all three winches is shown. Wave action
generates variations in winch tension acting at 4.5 s period (the wave period) throughout the
simulation. The simulation is run slowly to minimise other dynamic effects due to shock loading
on the winches, but even so, there are visible jolts at several points:
t = 0 when the winches start to lift the pipe,
t = 600s where the rate of lift is suddenly reduced,
t = 800s where the lift stops.
These are artificial effects due to abrupt changes during the simulation and can be discounted for
practical purposes. Zoom in (Hold Alt and drag to draw a box on part of a graph with the mouse)
to see the wave period variations as smooth curves, and the higher frequency variation that occurs
at the times listed above.
The jolts can be minimised or avoided completely by using time history control of the winch
payout. Time history control is used when the Whole Simulation winch control option is
chosen. The time history is generated outside OrcaFlex, and can include ramping at changes to
payout rate.
The instantaneous value range graph of pipe curvature is plotted. As the simulation replay
progresses, you can see the curvature changing. Maximum bending occurs at the attachment of
Winch C, with a further lower peak at touchdown.
W Orcina

E05 Pipe Davit Lift
X
Z
30 m
X
Z
OrcaFlex 9.2a24: C4.1 Pipe Davit Lift.sim(modified 09:48 on 10/03/2008 by OrcaFlex 9.1c) (azimuth=270; elevation=0)
Time: 900.0000s

The free end of a steel pipe is lifted from the seabed by davits to bring it alongside a workboat.
The lift is modelled dynamically in small waves and no current.
Building the model
The line is initially laid out straight on the seabed. We use Catenary Step 1 statics followed by
Full Statics for this. Note that the initial position of the free end (End A) is deliberately chosen to
over-stretch the pipe, and the As Laid Tension is set to a low but non-zero value. These both help
statics convergence. When statics is converged, End A has moved to its equilibrium position.
Three winches are equally spaced on the vessel side and attached to the pipe near End A. As the
winches haul in, lifting the pipe, the pipe free end also moves horizontally towards the anchor. To
accommodate this, the workboat has to move forward.
The analysis is carried out in five stages as follows:
Stage 0 System settles - no haul in -2s to 0s
Stage 1
workboat accelerates
0s to 300s
Stage 2
workboat moves at constant speed
300s to 600s
Stage 3
Winches pull-in last 10% of total;
workboat decelerates to zero speed
600s to 800s
Note that the winches all pull in slightly different amounts to achieve the lift.
The winches are set to Specified Tension = 0 in statics. This means the pipe is laid out with no
load from the winches, and the starting length for each winch wire is set to give zero tension with
no slack. In subsequent stages, Specified Payout is negative, meaning that the winches haul in.
W Orcina

Workboat movement is defined on the Prescribed Motion page on the vessel data form. At each
stage we define the change in velocity. OrcaFlex arranges that this occurs smoothly by applying a
constant acceleration through the stage.
It should be mentioned that the default RAO data in OrcaFlex is in use for this workboat model.
The default RAO data are based on a slender tanker hull form, and so are not appropriate for a
workboat-shaped vessel. More appropriate data should be available for real project work.
A brief settling time is provided at the start and a longer one at the end of the simulation.
End B of the line is anchored to the seabed. The model would need to include a very long length
of line in order for there to be no tension variation at the seabed anchor due to the lift. In this
example, there is a tension rise at End B, because we have included less length of line in the
model.
Results
A results workspace is provided open the file E05 Pipe Davit Lift.wrk.
The tension time history through the simulation for all three winches is shown. Wave action
generates variations in winch tension acting at 4.5 s period (the wave period) throughout the
simulation. The simulation is run slowly to minimise other dynamic effects due to shock loading
on the winches, but even so, there are visible jolts at several points:
t = 0 when the winches start to lift the pipe,
t = 600s where the rate of lift is suddenly reduced,
t = 800s where the lift stops.
These are artificial effects due to abrupt changes during the simulation and can be discounted for
practical purposes. Zoom in (Hold Alt and drag to draw a box on part of a graph with the mouse)
to see the wave period variations as smooth curves, and the higher frequency variation that occurs
at the times listed above.
The jolts can be minimised or avoided completely by using time history control of the winch
payout. Time history control is used when the Whole Simulation winch control option is
chosen. The time history is generated outside OrcaFlex, and can include ramping at changes to
payout rate.
The instantaneous value range graph of pipe curvature is plotted. As the simulation replay
progresses, you can see the curvature changing. Maximum bending occurs at the attachment of
Winch C, with a further lower peak at touchdown.
W Orcina

Pipelay and Recovery: E06 Midline Pull-up Page 1 of 2
E06 Midline Pull-up
X
Z
6 m
OrcaFlex 9.2a24: C4.2 Midline pull-up.sim(modified 10:12 on 10/03/2008 by OrcaFlex 9.1c) (azimuth=270; elevation=0)
Time: 15.0000s

A long umbilical cable is laid on the seabed. The mid point is then lifted a short distance, e.g. to
position the umbilical in a plough prior to burial. The lift is carried out in still water.
Building the model
The umbilical is positioned on the seabed using a prescribed shape. This is a simple way of laying
a line out straight on the seabed with both ends free. Since both ends are free, and the umbilical is
assumed to be lying in an as-laid position, line step 2 statics has been turned off. This prevents
OrcaFlex statics from moving the line away from where it is known to be to a different
equilibrium position. This requires us to use separate buoy and line statics on the General data
form.
Segmentation is fine near the lift point to give good definition of bending, coarser elsewhere
where detailed results are less important, to reduce computation time.
The model represents lifting at a point remote from both ends. Both ends are left free, and should
be sufficiently far away from the lifting point that the end condition has no effect on the lift. We
check this later.
The analysis is carried out with the following stages:
Stage 0 System settles -0.1s to 0s
Stage 1 Winch pulls in 4 m 0s to 10s
In statics, the winch is set to zero tension. This sets the initial wire length such that tension is zero
with no slack. In Stage 1 the winch raises the middle of the line 4 m, i.e. payout is -4 m.
The winch tension is taken as a result below. It should be noted that the winch wire has no mass
or hydrodynamic properties. If this model was to be extended to deeper water, the missing winch
wire contributions could begin to affect the results.
W Orcina

Pipelay and Recovery: E06 Midline Pull-up Page 2 of 2
Results
During the replay, watch the changes in the instantaneous value range graph of curvature for the
central part of the umbilical (1500 to 1600 m). This can be found in the workspace E06 Midline
pull-up.wrk.
Other results of interest include winch tension and the tension distribution in the whole length of
the umbilical. Note that tension effects make themselves felt right to the free ends.
As mentioned earlier, the line ends are not supposed to be moved as a result of the lift, and this is
true for our case. If we were to run this simulation with a shorter umbilical, then the line ends
would move, and this would affect the winch tension and line curvature. It is generally
recommended to inspect truncated line end conditions, looking for confirmation that the model
was not truncated too soon.
W Orcina

Payload Handling: F02 Passive Compensation Page 1 of 4
F02 Passive Compensation

Introduction
During the lowering process the heave of the vessel can result in severe motions of the payload
and/or high loads on the crane.
To reduce these effects, heave compensation devices are often used. These can be active systems
that adjust winch payout to keep the payload at a constant depth or passive systems where the
payload is allowed to move but there is additional compliance with damping to reduce the loads
on the crane when it heaves upwards.
This example considers the passive system, using a CRANEMASTER device as an example. The
principles can also be applied to other devices. You need to view the model browser in the View
by Groups mode. There are three systems present, no device, linear and non-linear
compensation.
W Orcina

Building the model
A typical system consists of a sliding rod that has motion controlled and limited by combinations
of hydraulic, pneumatic and spring mechanisms. OrcaFlex does not need to model each
component in detail, just the net effect on stiffness and damping.
Note the payload is very simple in this case to better show the action of the compensator.
The mechanism is modelled using the combination of a single segment line section and a
spring/damper link. The single segment line helps the statics to solve more robustly.
The line section represents the structure of the compensator (in this case 1m long) and is
positioned in the cranewire length at the location of the device. Look at Crane Wire Linear or
Crane Wire Non Linear.
It is a single segment so that it will behave as a stiff stick and we dont have to worry about
bending. We then give this segment negligible mass and displacement, default bend stiffness
(because it is not used) and a low axial stiffness (for the slide).
The actual mass and stiffness of the Compensator Rod are added as clump weights at the top and
bottom of the segment. In this example 80% of the weight is at the top so the cranewire will take
that load directly. The remaining 20% is at the bottom so will be affected by the compensator
mechanism.
The result is a segment that is correct for mass and displacement and will slide freely axially
while not bending. See the drilling riser examples for more applications of slide in slip joints.
The sliding now needs to be controlled. This is done with the link. We position the link parallel to
the segment, each link end attached to each end of the segment.
The first step is to identify the static load that the device is required to support and what stroke
position you want it to be in when experiencing that load. In this instance the payload produces
195.7kN tension at the bottom of the Compensator Rod. See the End B static tension from Crane
Wire. This is in the No Heave Compensation group.
This mechanism has the stroke just offset from the compressive limit when supporting this load.
This example has the stroke at 0.05m for this load. The devices actually settle at 0.0865m with a
tension of 196.1kN indicating the load distribution is a bit more complex but this is a small
difference so the method is a reasonable estimate.
Because the mechanism being modelled is a CRANEMASTER device, the compressive stroke
limit is a stiff spring rather than hard stops. Activation of the spring is modelled as a rapid linear
increase in stiffness if the stroke reduces from 0.05m.
For the extension a limit has not been build because it should not be reached if the device is
applied correctly. Instead the stiffness rises in a nonlinear manner to provide increased resistance
to motion. This change is chosen to match the characteristics of the CRANEMASTER device.
The figure below shows the stiffness characteristics for this example. The slope is a steep linear
from 0m to 0.05m, representing the compression limiter spring, then a gentle rise to represent the
net hydraulics/pneumatics. The table can be seen in the Passive Heave Linear and Passive
Heave Non Linear links.
W Orcina

Tension vs Extension
-600
-400
-200
0
200
400
600
0 0.2 0.4 0.6 0.8 1
Stroke (m)
T
e
n
s
i
o
n

(
k
N
)

The other part of the control mechanism is the damping of the motion. The CRANEMASTER
mechanism provides more damping on the return than on the extension. This has been reproduced
by applying non-linear damping.
The figure below shows the damping characteristics for this example. Both extension (+ve
velocity) and compression (-ve velocity) show a nonlinear curve. However the compressive one is
much steeper. The data can also be seen in the Passive Heave Non Linear link.
Tension vs Velocity
-500
-400
-300
-200
-100
0
100
-1.0 -0.5 0.0 0.5 1.0
Velocity (m/s)
T
e
n
s
i
o
n

(
k
N
)

The example also includes a linear damping value of 250kN/(m/s) for comparison. This is in the
Passive Heave Linear link.

W Orcina

Results
The three graphs presented are crane wire top tension with no compensation, with linear and non
linear damping. Set the Y axes to a common scale of 140kN to 280kN to aid comparison (double-
click on the plot to open up the Graph Properties data form).
The top right hand plot shows the crane loading with no compensation. The loads are large and
there is a lot of noise. The top left shows the response with non linear damping. Amplitudes are
reduced and the noise has also reduced. The bottom left shows the response with linear damping.
In this example the damping relationship can be simplified with little change in response.

W Orcina

Payload Handling: F03 PID Controlled Active Winch Page 1 of 3
F03 PID Controlled Active Heave
Compensation
X
Z
20 m
X
Z
x
z
x
z
x
z
x
z
OrcaFlex 9.2a24: C3.3 PID Controlled Active Winch.sim(modified 11:23 on 24/03/2008 by OrcaFlex 9.2a24) (azimuth=270; elevation=0)
Time: 100.0004s

This package-lowering example uses a winch to model an active heave compensation device.
1. Introduction
An external function written in Python is used to control the length of a winch, which is holding
an object at midwater. An OrcaFlex Line object is used to model the majority of the work wire;
using a line means that the wire physical and hydrodynamic properties are accounted for. The
winch between the vessel and work wire is controlled with a PID controller via the OrcaFlex
programming interface, OrcFxAPI.
Python is used to write the external code in this example; however MATLAB, C++ or Delphi
could have been used instead.
2. Building the Model
The package being lowered is a simple 6D lumped buoy, with links used as rigging. An OrcaFlex
line is used between the links and the lowering winch. Links and winches have no mass or
hydrodynamic forces, so to use a winch for the entire work wire would mean that we miss the
forces on the wire in the water.
The vessel uses displacement RAOs, and so moves in response to the wave loading imposed on
the model. Our aim is to use the winch to compensate for the vessel motion and maintain the
package at a constant depth, i.e. active heave compensation.
To control the winch, a PID controller (proportional integral derivative controller) was
chosen. This is a generic control-loop feedback mechanism, which is widely used in control
systems. The PID control algorithm has been coded in Python, a dynamic scripting language, and
interfaces with the model via the OrcFxAPI. The Python code, contained in the file
PIDController.py, can be viewed with any text editor, however we recommend Notepad++ as it
presents files in an easily readable format.
W Orcina

Full details of the Python Interface are not included here; instead, the user is directed to the help
file for the OrcaFlex API, which can be found on the Orcina website
http://www.orcina.com/SoftwareProducts/OrcaFlex/Documentation/OrcFxAPIHelp/
In the model, OrcaFlex is made aware of the Python code on the Variable Data form, under the
External Functions heading. Click on PIDWinchControl in the Model Browser to see this. Once
the PIDController.py file has been identified as the File Name, the name of the function to be
used to control the winch is then selected from a drop-down list in the Function Name box
(OrcaFlex queries the contents of the Python file automatically, and lists the available functions).
We also need to define an Initial Value (i.e. the value passed to the external function the first time
it is called), and how regularly the external function should be called. In this case, we are using
the Implicit integration scheme, so the function will be called on each iteration of each timestep
regardless of what is entered in the TimeStep box.
Once the external function is defined, we can use it to control the winch. This is done on the
winch data form. The name of the external function, PIDWinchControl, is used as a variable
data item for the winch payout rate throughout the simulation. A new tab appears on the winch
data form, headed External Functions. This tab can be used to pass information to the external
function.
In this example, we use the tab to tell the function the following information:
the name of the object to be controlled (Template)
the name of the object result to be controlled (Z)
the control start time, which is set at 20s so that the motion of the package with and
without heave compensation can be compared
the target depth at which the package should be held (85m), so Z = -85
a min and max payout rate for the winch (optional)
PID control parameters, k0, kP, kI and kD

3. Results
Opening the simulation file opens up the default workspace file, which displays some important
results. The variation in winch tension is shown (top right) to illustrate the loads experienced
during the motion. The template weight in water is 19.75 te, and so a mean tension of approx. 200
kN is reasonable. There is a shock load on the winch when it begins to control the package depth
at 20s. This is because of a step change in acceleration.
Time histories of Winch length and Template Z position are shown below the tension plot; the
latter clearly showing the effect of the PID controller holding the template close to the required
depth. Run the replay to see this in action in the wireframe model view. Vessel motion near the
stern is also plotted (top left), and has a range of up to 4 m. The corresponding range in template
depth, once the heave compensation is operating, is less than 0.5 m. The effectiveness of the
compensation depends on the PID controller parameters; the maximum capability of the real
equipment should be considered when the model is built i.e. the parameters must be tuned to
reflect the performance of the real system. If this is not done then it is possible to achieve better
or even worse performance in a simulation than may be realistic.
W Orcina

Note: Further External Function examples in both Python and C++ and available for download on
the Orcina website http://www.orcina.com/Support/index.php
W Orcina

Deployment: G04 Anchor Last Deployment Page 1 of 2
G04 Anchor Last Deployment
X
Z
100 m
X
Z
Or caFlex 9.3a17: C04.06 Anc hor - last Deployment.s im ( modified 12:03 on 02/03/2009 by Or c aFlex 9.3a17) ( az imuth=270; elevation=0)
Replay Time: - 10.00s
X
Z
100 m
X
Z
Or caFlex 9.3a17: C04.06 Anc hor - last Deployment.s im ( modified 12:03 on 02/03/2009 by Or c aFlex 9.3a17) ( azimuth=270; elevation=0)
Replay Time: 60.47s
X
Z
100 m
X
Z
OrcaFlex 9.1c: C4.6 Anchor-last Deployment.sim(modified 10:30 on 10/03/2008 by OrcaFlex 9.1c) (azimuth=270; elevation=0)
Time: 250.0010s

This model shows a method commonly adopted for the deployment of a deep water
oceanographic mooring. The mooring is streamed out behind a moving vessel, with the top end
buoy as the first object in the water, furthest from the vessel. When the whole length of the
mooring has been deployed, the anchor is tied off at the vessel stern, and the vessel proceeds to
the correct mooring position. Without stopping the vessel the anchor is cut free. The whole
mooring then free-falls until the anchor hits the seabed.
The simulation is used to determine the final position of the anchor relative to the vessel's
position when the anchor ties are cut. It is also used to determine the tensions in the line.
Note that the mooring line may go slack when the anchor strikes the seabed. If the line is not
perfectly torque balanced it is possible for loops and kinks to form in the line. This may lead to
early failure of the mooring. The model assumes that the line is perfectly torque-balanced and that
the installation is carried out in still water.
Building the model
The mooring is represented by a single Mooring Line attached between the Buoy and the
Anchor, both modelled for simplicity as 3D buoys. A more detailed model would use 6D
buoys, to allow us to study the pitch and roll motions of these objects and model the surface
piercing more accurately if required.
The Anchor is attached to the Lay Vessel using a Release Link.
Open the Release Link Link Data Form via the Model Browser. The link is set to release at
Start of Stage 1. This is 10 s into the simulation. The release represents the action of cutting the
anchor free.
Rather than starting the analysis with the vessel stationary and waiting for it to reach steady state
this example includes the steady forward velocity in statics. Open the General Data Form from
the Model Browser and look a the Statics page. The Starting Velocity has been specified. The
Starting Velocity models the constant initial forward speed as a current applied in statics.
The Lay Vessel motion during the dynamic simulation is specified on the Prescribed Motion page
of the Vessel Data Form. OrcaFlex warns if the starting velocity for the model is not the same as
W Orcina

Deployment: G04 Anchor Last Deployment Page 2 of 2
the velocity of the ship at the start of the dynamic simulation because a step change can generate
unwanted transients or model instability.
Note that the velocity direction on the Statics page of the General data Form is given relative to
the Global X axis. The direction on the Prescribed Motion page of the Vessel Data Form can also
be specified relative to Global X but in this Example it is relative to the Vessel Heading instead.
OrcaFlex converts the initial dynamic heading so it is also relative to Global X and compares it to
the static heading. If they differ OrcaFlex will also give a warning.
The Lay vessel uses OrcaFlex default data, which represents a tanker. There are no waves, and so
the motion is unaffected, but if waves were added then actual RAO data for the vessel would be
required.
The Anchor 3D Buoy Data Form shows a contact area and seabed friction coefficient have been
specified. The Mooring does not contact the seabed and the friction coefficient of its Line Type
Wire is set to 0.0, see the Friction page of the Line Type Data Form.
Results
Three time history plots are presented.
Two plots are of the mooring effective tension at the Buoy (Top Right) and the Anchor (Bottom
Right). The changes in effective tension caused by the buoy submerging and by the anchor hitting
the seabed are clearly identifiable in the time histories.
The time history of the Buoy Z position is shown at the bottom left. The buoy maximum depth is
about 47 m, compared with the settled depth of about 42 m. This temporary additional depth is
important when choosing the depth rating of the buoyancy unit.

W Orcina

Offloading Systems: H01 Chinese Lantern Page 1 of 1
H01 Chinese Lantern
Introduction
Risers descend from the underside of a moored CALM buoy to its base on the seabed in a
Chinese Lantern configuration.
Building the model
The CALM buoy is modelled using the 6D Buoy Spar
option see example C06 CALM Buoy for a
discussion of some of the issues involved.
The risers are forced to bend outwards by angling the
end fittings and adding buoyancy modules. Press
CTRL+Y to turn the local axes view on. You will need
to switch to wireframe view to see this. CTRL G will
toggle between the wireframe and shaded views. Note
the heading of the Z-axis at each end of the risers.
Open the Line data form for Riser 1. Hose make-up is
quite detailed, with typically 0.2m long segments.
Fittings 1 and 2 represent the steel end fittings of
individual hoses and reinforced sections are used at
top and bottom ends. The Attachments page shows
where buoyancy is attached.
Torsion can be important if the CALM buoy offset is
large or environmental loading is severe. It is worth checking critical cases with torsion on if you
are going to analyse torsion free.
In this model there are two systems, with and without torsion. They have been positioned so they
see identical loading.
Results
Look at the animation through the latest wave. You can see that the upstream hoses kink
inwards for half of the wave cycle, i.e. they bend inwards rather than out in response to the wave
loading. This is seen in cases with and without torsion.
Note these graphs have the arc length on the Y axis rather than the usual X axis. The numbers are
also in descending order instead of increasing. This is achieved on the Select Results page by
choosing Arc length axis: Vertical and Arc length axis inverted. These are to be found below
the objects table.
Compare the curvature range plots for identical hoses with and without torsion. The upper one is
with torsion. Make sure the axes of the plots have the same range. There is a change in curvature
distribution and so the extreme value increases when torsion is present.
W Orcina

Offloading Systems: H02 Jacket To Semisub Page 1 of 2
H02 Jacket to Semisub
X
Z
20 m
X
Z
OrcaFlex 9.3a45: D02 Jacket to Semisub.sim(modified 15:12 on 26/06/2009 by OrcaFlex 9.3a45) (azimuth=270; elevation=0)
Replay Time: -10.00s

A cable catenary is suspended between a semi-submersible rig and fixed jacket. It is analysed in a
random sea.
Building the model
Both the jacket and the semi-submersible are modelled as vessels. Although the jacket does not
move, representing it as a vessel allows the vertex and edge or shaded drawing facilities to be
used and objects to be attached to it. To ensure it does not move, the jacket primary and
superimposed motions are set to 'None'.
From the Model Browser, select 'Cable'. Note that the Semisub connection has the End
Orientation Gamma angle set at 90. When torsion is not applied, gamma can be used to define
the orientation of End x and y about z. In this example, setting gamma to 90 makes the x axis of
the end connection normal to the semisub side, and y parallel to it. This matches the x and y
orientations at the Jacket End. Select the view and press CTRL+Y to show the line end axes, then
CTRL+ALT+Y to show the line axes. Press CTRL+Y and CTRL+ALT+Y again to turn them off.
The main concern in this example is clashing between the cable and the jacket structure. This is
best assessed visually. A jacket strut has been made from a cylinder shape and has then been
attached to the jacket vessel. We will consider whether shape contact or line on line clashing is
most appropriate for this example.
Line on line clashing considers contact between line segments, and allows smooth modelling of
sliding contacts. Using segments for clashing interaction also means that fine line segmentation is
not automatically necessary to model clashing accurately. However, line clashing does not
include friction between two objects in contact, and is not available during a static analysis.
Lines contact shapes at the line nodes. This means that fine segmentation is often necessary for
modelling contact between lines and shapes. In this case, the cable is likely to wrap around the
jacket member during a long contact, and fine segmentation will be needed to allow the cable to
follow the curvature of the jacket members surface. Friction has been included in this model
between the cable and the jacket member shape.
W Orcina

Offloading Systems: H02 Jacket To Semisub Page 2 of 2
The units of line clashing stiffness are kN/m, which is intended as a change in reaction force with
increasing depth of penetration between two line segments. No dependence on the length of line
in contact is included, because it is assumed that the line segment length is longer than the line
diameter. This implies that line on line clashing is expected to only occur between individual
segments along each line in contact, and not multiple segments of one line simultaneously.
Because this simulation involves one line wrapping around the other, using shape contact is more
appropriate for this situation. The cable line segmentation is already required to be quite fine in
order that the cable can follow the member surface accurately, so enough segments are present for
well-modelled shape contact.
Finally, some contact between the cable and the edge of the semisubmersible pontoon was noted,
and so a shape was placed at this location also. The line segmentation is refined to capture
curvature caused by contact at this sharp edge.
Results
Time histories and range graphs of the clash events are available in the provided workspace.
However, for this type of analysis, it is often sufficient to watch a replay. Opening the shaded
view workspace will show one view of the whole model, and one close up of the clashing regions.
You can observe the clashing while the simulation replay is running although it may be
necessary to zoom in further on the view.
OrcaFlex colours clashing objects in white, to identify whether contact has occurred as easily as
possible. The clashing with the large jacket line is easy to see, whereas the clashing between the
cable and pontoon shape is harder to spot.
W Orcina

Offloading Systems: H03 Floating and Stowed Lines Page 1 of 2
H03 Floating and Stowed Lines

Introduction
A floating hose, attached to the stern of an FSU (Floating Storage Unit), is stowed by curling it
around and attaching the free end to the vessel side. Interaction with the vessel sides needs to be
considered. The hose is released, and during the analysis, the protective chain at the FSU breaks.
Building the model
The vessel hull is modelled as a solid so that clashing can be considered. However note that the
complex loads due to turbulence and water trapped between the vessel side and the approaching
hose are not considered and there is no shielding from the vessel hull included in the model. The
hull is modelled as a simple rectangular block attached to the vessel.
The hose is attached to the stern of the vessel and on the port side, near the bow. It is modelled as
a single line with the chain clamp and the Mid-line Connector (MBC) set up as line properties.
The hoses are modelled using line sections to represent the different parts of the hose. These are
steel flanges with and without floats, clamps and composite hose lengths. The segmentation is
fine enough to capture curvature in the main hose but could be further refined local to the hose
end connections if required.
The clamp and MBC could be modelled as 6D spar buoys and the hose length split into three to
insert them. However this does mean modifications to prevent the buoys spinning on their local z
W Orcina

Offloading Systems: H03 Floating and Stowed Lines Page 2 of 2
axes, i.e. including torsion on the hose line. Making them line properties avoids this issue and,
with the implicit time integration scheme, does not incur a runtime penalty.
The end of the hose dropped into the water. As the hose swings away from the vessel under the
influence of the waves and current, the load on the retaining chain at the stern increases. We
model the chain breaking using a nonlinear link between chain and clamp.
The link has a linear increase in tension
from 0kN at 0m to the required breaking
tension of 500kN at 1m. Beyond 1m the
tension drops to zero and so the link
cannot resist further extension.
This means there is no slack state before
the link breaks this is not expected to
make significant difference to the results.
The rest of the chain will go slack instead.
During the static search it is possible for
the link to extend beyond the break
point. A parallel link temp parallel is
added to protect it. This link takes the load
during statics and then releases at Stage 0.
During the simulation there are discrete events occurring through the run, dropped hose ends,
breaking lines, hose impacting the hull and the free ends lashing around. With an implicit time
step integration scheme it is hard to be sure the time step is sufficiently small to follow these
discrete responses accurately. It may run with a high number of iterations per step as it struggles
to follow the response. Or the step may be too large to capture higher frequency responses.
To be sure of obtaining accurate results for such an analysis it is simpler to use the explicit time
step integration scheme, as shown in this example.
More information on the two time step integration schemes can be found in the OrcaFlex Help
topic System Modelling: Data and Results | General Data | Integration & Time Steps.
Results
The range graph of solid contact force shows that the hose contacts the hull of the vessel. Looking
at the time history of contact force at a point in the region where contact occurs shows more detail
of the contact loads.
Load Workspace H03 Floating and Stowed Lines - tension.wrk from the Workspace dropdown
menu. Look at the time history of tension in the chain (Top Right) and the Hose (Bottom Right) at
the clamp. The chain protects the hose by taking peaks of tension but the chain itself suffers
snatch loading. After 31.5 s the chain has broken and the hose takes the entire load itself.
Other results of interest include the hose curvature; which is quite high because it is moving on
the surface of the waves. The precise behaviour of the hose curvature will vary dependent on the
local conditions at the time of the release. This analysis will therefore yield results to characterise
the release behaviour, but they should not be assumed to represent the maximum loads possible
on the hose or the chain.
OrcaFl ex 9.3a48: D03 Fl oati ng and Stowed Li nes.si m (modi fi ed 14:25 on 14/07/2009 by OrcaFl ex 9.3a48)
snappi ng l i nk Sti ffness Profi l e
Length (m)
2 1.5 1 0.5 0
T
e
n
s
i o
n

(
k
N
)
60
50
40
30
20
10
0
W Orcina

Towed Systems: I01 Streamer Array Page 1 of 2
I01 Streamer Array
X
Y
40 m
X
Y
OrcaFlex 9.2a29: E01 Streamer Array.sim(modified 14:08 on 22/04/2008 by OrcaFlex 9.2a29) (azimuth=270; elevation=90)
Time: 30.0008s

This is an example of a seismic streamer model. The model represents the Port half of the system
only i.e. two hydrophone streamers plus one air gun towed from a ship.
Building the model
The diverter is modelled as three elements: a 6D Buoy (Towed Fish option) to represent the
buoyancy tank; a heavy line at the base to give stability; and a wing to represent the lifting
surfaces. The wing generates lift and drag forces as the diverter is towed through the water. See
'Prandtl199' on the Wing Types data form. Pressing the Graph button will show you the lift and
drag coefficients versus incidence angle. Select any item on the Wing Types data form and press
F1 to get more details from the on-line help.
Lines cannot be connected directly to lines in OrcaFlex. Branching is therefore achieved using 3D
buoys. These are acceptable, as transfer of bending moment between lines is not important for
this example. (Since a 3D buoy has no rotational degrees of freedom, it does not respond to
applied moments.)
Rather than modelling the two streamers in full, most of each length is represented by a "sea
anchor", a 3D buoy with area and drag coefficient set to generate a drag force equivalent to that
of the streamer it represents. Look at 'StrmTail A' data.
The buoy properties are only set for the global X direction. Because a 3D buoy cannot rotate, the
drag force will therefore always act along the global X axis. The drag specified is intended to
represent only the axial drag of the truncated line. No normal drag or added mass is applied to the
drag anchor, as the single clump will not correctly represent the distribution of these properties
along the real streamer away from the tow vessel.
The motion of the whole system through the water can be modelled within OrcaFlex in two ways.
The vessel can be given a constant forward velocity. However this means that the system origins
are constantly moving in relation to the global origin. Also the system will keep moving out of
the view window.
Alternatively, as the system is moving in a straight line, we can achieve the same effect by
holding the system stationary and applying a current. This is the method applied here. It should be
noted that the OrcaFlex default RAO data has been used, which is not appropriate for a tow
W Orcina

Towed Systems: I01 Streamer Array Page 2 of 2
vessel. In production models, it would be expected that the real vessel RAO data would be
supplied and used for analysis.
Results
Go to the default view, CTRL+T, and look at the animation through the latest wave. Look in
elevation view, CTRL+E, as well as plan view, CTRL+P. Zoom in to look closely at the motion
of the diverter.
The workspace file plots time histories of the lift, drag and incidence angle for the diverter wing.
Note how the lift is far greater than the drag and that the incidence angle only varies by 2.5.

W Orcina

Defence: J01 Deployment with Sub Page 1 of 2
J01 Deployment with Sub
X
Z
30 m
OrcaFlex 9.3a47: F01 Deployment with Sub.sim(modified 14:50 on 29/04/2008 by OrcaFlex 9.2a29) (azimuth=270; elevation=0)
Replay Time: 38.30s

Introduction
A submarine tows a sensor array. It releases components in turn to position them on the seabed.
Building the model
The system is to be installed using a submarine. With the model browser in group view, open the
group Sub. This contains a vessel called 'Submarine'. The vessel has shapes attached to give a
Submarine shape. Notice a key symbol by the shape Hull. This means it cannot be moved. You
can lock and unlock an object via the right hand mouse button.
A line extends behind the vessel, Subline. The far end is attached to the release mechanism that
consists of a deployer (a 6D spar buoy) and two links. These links release the anchors.
Each sensor group contains an anchor, a float and a sensor line. A link holds the float close to the
anchor while towing then releases on deployment.
A line (Baseline) also connects the two anchors together so that a required maximum separation is
not exceeded.
The floats are modelled as 3D buoys, i.e. translations only, because the detailed motions of the
floats are not of interest.
The two anchors are modelled as lumped 6D buoys to better model interaction with the seabed,
for example if one edge will contact first. Note that seabed contact occurs via the buoy vertices. A
seabed friction coefficient of 0.2 has been applied and contact area is set at 0.05m
2
overall. This is
applied equally between all the vertices. Only the bottom vertices will contact so the actual
contact area will be 0.025m
2
.
If we had left the area as ~ then the overall contact area would have been determined from the
buoys volume divided by its height, i.e. 0.005m
3
/1m=0.005m
2
so actual contact area of
0.0025m
2
.
W Orcina

Defence: J01 Deployment with Sub Page 2 of 2
The analysis is carried out in the following stages:
Stage Vessel Links Time
0 Proceeds at 1.5m/s

-10sec to 0sec
1 Proceeds at 1.5m/s 1st anchor released so anchor and
float sink towards seabed
0sec to 5sec
2 Proceeds at 1.5m/s 1st float released so rises up 5sec to 35sec
3 Proceeds at 1.5m/s 2nd anchor released so anchor and
float sink towards seabed
35sec to 40sec
4 Proceeds at 1.5m/s 2nd float released so rises up 40sec to 100sec
A forward speed of 1.5m/s is specified in statics, together with a small cross current.
Corresponding hydrodynamic drag forces are then applied to the model when determining the
initial static configuration. The transition to dynamics is then smooth, with no acceleration
required.
Results
Look at the animation through the whole simulation to see the deployment procedure.
W Orcina

Miscellaneous: Z01 Line on Line Impact Page 1 of 2
Z01 Line on Line Impact

Introduction
OrcaFlex allows you to prevent lines passing through each other in the dynamic analysis. When
the option is selected on both lines it generates a resistive force when one line contacts the other.
This routine also takes account of line outer diameters when assessing contact.
On opening, the default Workspace for the folder sets up a shaded graphics view of the system
and selected results. The view shows a yellow line that is pinned at the top and free at the bottom,
the HAMMER. The red line, the ANVIL, is fixed at both ends and suspended across the
Hammers path.
Building the model
Two lines are generated called 'Anvil, the line that is hit, and 'Hammer', the one that swings to hit
the first. 'Hammer' is made to swing simply by releasing End A at the start of Stage 1, i.e. after
the build-up period.
To prevent one line passing through the other, both require contact stiffness to be set and clash
check to be enabled.
Contact stiffness and damping are specified in the Line Types data form on the 'Contact tab.
Accurate values are rarely known, but this may not be important. In many cases, all that is
required is that one line should not pass through the other, in which case it is sufficient simply to
W Orcina

Miscellaneous: Z01 Line on Line Impact Page 2 of 2
set the contact stiffness to a sufficiently high value. For this example, contact stiffness is set to
5MN/m and contact damping is set to zero.
Matters become more difficult if accurate loads are needed for design purposes. For further
details see |Theory|Line Theory|Clashing|.
The clashing calculation is enabled on the line data page. Go to Anvil and look at the Structure
tab. The Clash Check is selected. You will find the same for Hammer.
Note that the Clash check will result in slower runs so should only be applied where required.
Run the case without clashing first and look at the clearance results. From these you can identify
which lines and where on the lines there is a risk of contact. Then turn on the clash routine for
those line sections.
This model is run with the explicit time integration. Contact will also work with implicit time
integration schemes but in this example the lines are in air so have little damping. The resulting
sudden changes in loading are better captured with explicit.
Danger
Remember that the contact point may move due to lines sliding along each other. Note no friction
is included in this slide. If they then slide onto a section that has contact turned off then the lines
will pass through each other.
See the examples B04 Clashing Risers and Z02 Line on Line Slide for other applications.
Results
Look at the animation through Stage 1, i.e. from the moment 'Hammer' is released. It swings
across to 'Anvil' which it pulls with it before it swings back and away.
Look at the instantaneous range graph of Line Clash Force for 'Hammer' through the simulation.
The contact moves between 50m and 60m from the pivot end (End A).
Now look at the Anvil time history at an arc length of 51m. The clash force is highest on initial
impact, at about t = 4.2s. There is intermittent contact - i.e. the Hammer strikes the Anvil and
rebounds repeatedly.
Now go to the Results form and select Line Clashing Report for the Anvil for the period from 4 to
5s. This gives details of the magnitude and duration of each contact event.
W Orcina

Miscellaneous: Z02 Line on Line Slide Page 1 of 2
Z02 Line on Line Slide

Introduction
The line clashing routine can be used to model one line sliding over another. However remember
that friction is not included for line-on-line contact and that line on line contact will only occur in
dynamics. A fun example of this is given below.
On opening, the default Workspace for the folder sets up a shaded graphics view of the system. A
boy scout slides down an aerial ropeway by holding onto the ends of a short rope wrapped over it.
He then flies off the end to hit the ground.
Disclaimer: No real boy scouts were injured in the making of this example.
Building the model
The system is submerged but the seawater density has been set at 1.28kg/m, the density of air. A
solid fixed in space makes the platform that the scout stands on at the start of the analysis.
The scout is modelled as a 6D buoy, which is fixed in statics. Properties for a typical 10 year old
scout have been applied but drag and added mass are ignored as he is moving in air, so generated
loads will be negligible. The arms and legs are then drawn using the buoy drawing facilities.
Solids are attached to the buoy to make the head and body shapes.
The buoy representing the scout has also had seabed friction enabled. Look at the Contact tab on
Boy Scout. A high friction coefficient has been applied and a total contact area. Contact only
W Orcina

Miscellaneous: Z02 Line on Line Slide Page 2 of 2
occurs at the buoy vertices. This one has 16 (Drawing tab). The total area has therefore been set
as 0.16m
2
so each vertex represents 0.01m
2

The ropeway is a simple line inclined and held at both ends. The hanger is also a simple rope.
Because there is no friction with line on line contact modelling the ropeway has a large normal
drag coefficient applied to the middle section (Rope with extra drag for friction line type).
Both the ropeway and the hanger have a contact stiffness and clashing check running so they
cannot pass through each other. See Z01 Line on Line Impact for details.
To prevent the Hanger passing through the Ropeway, both lines must have a contact stiffness
assigned and 'Clash Check' selected. Contact stiffness is specified in the Line Types data form. It
should be sufficiently high that one line cannot push through the other, but the actual value is not
important in this case.
To turn the clashing routine on, select Clash Check on the Structure tab for both 'Ropeway' and
'Hanger'. Note that the Clash check will result in slower runs so should only be applied where
required. Links A and B lift the Hanger so that it forms a loop around the Ropeway during statics,
then releases at the start of Stage 0.
Results
Look at the animation through the whole simulation. The scout slides along the line then flies off
the end to crash on the ground. This example is just a bit of fun, but it does illustrate how the line
clashing feature can be used to model sliding contact.
W Orcina

Miscellaneous: Z03 Vortex Tracking Page 1 of 2
Z03 Vortex Tracking
X
Z
0.06 m
X
Z
OrcaFlex 9.2a26: H03 Vortex Tracking.sim(modified 11:41 on 03/04/2008 by OrcaFlex 9.2a26) (azimuth=270; elevation=0)
Replay Time: 8.70s

Two lines are placed horizontally, perpendicular to a slab current. The lower line is held fixed at
both ends, while the upper line has one end spring mounted and the other fixed.
The OrcaFlex Vortex Tracking (1) model is applied to both lines, as a demonstration of how VIV
can be assessed in OrcaFlex using this tool.
Building the model
The lines are both single-segment lines, with the same properties. The scale of the whole model is
very small look at the scale bar in the default view.
The vortex tracking model is set up on the VIV tab on the line data form. Highlight some part of
the data on this tab and press F1 for a detailed explanation of the parameters used.
The Filter Period variable is worth remarking upon in this note. VIV is caused by fluid flow past
the lines. Subsequent VIV motion of the lines then introduces changes to the relative flow
velocities. The VIV models are not calibrated to deal with this feedback, the calibration is against
relatively steady flow normal to the line, and so we filter out the VIV-driven changes to relative
velocity.
Typically, the flow conditions will have components at different periods. A constant current
varies with infinite period, while the period of VIV oscillations is typically quite small. Wave
periods would normally be somewhere between the two.
The filter will try to remove rapid variations, like those from the VIV response itself, from the
fluid flow pattern used to calculate the VIV motion of the lines. For a model that includes wave
action, we want to retain the contribution of the wave loading to the VIV and so we obtain the
recommended relationship:
VIV Period << Filter Period << Wave Period
The Filter period is set to include wave period and greater variation as contributing to the VIV,
but will exclude more rapid variations, such as the VIV itself. Ideally the three periods should be
far apart but this is often not possible, and the filter period is placed centrally between the VIV
period and wave period.
W Orcina

Miscellaneous: Z03 Vortex Tracking Page 2 of 2
Because there are no waves in this example, the filter period is set to include only the current flow
as causing VIV, using the fact that current variation has infinite period.
The spring-mounted cylinder is held using very long links. The link length is long enough so that
cylinder motion causes no significant change in the spring force angle.
It should be noted that the vortices from the fixed cylinder will not interact with the vortices shed
from the spring-mounted cylinder. The lines are not coupled through any fluid interaction.
Results
Some of the available results specific to VIV motion are shown in the associated workspace. The
Stagnation point results are only available when the Vortex Tracking model is used. For the
spring-mounted cylinder we show the inline force due to VIV, the transverse VIV offset (which is
the motion due to VIV) and finally an estimated transverse A/D value. The estimate of A/D is
calculated by dividing the transverse VIV offset by the line VIV diameter. The VIV diameter is
set on the VIV tab of the line data form. In this case the VIV diameter is set to ~, which means
that it will be set to the line type outer diameter. It is typically assumed that VIV motion is self-
limiting, and that the A/D value should be somewhere between 1 and 2 at the most.
The inline vortex force on the spring-mounted cylinder is quite variable, especially compared to
the same result for the fixed cylinder. This variation is mostly due to the spring-mounted
cylinders motion, which leads to vortex shedding at larger angles relative to the cylinder,
changing the amount of vortex force considered to be in the inline VIV direction.
For the fixed cylinder, we show the angle of the VIV stagnation point around the line node. The
stagnation point is where the flow onto the face of the line separates in order to pass round the
line. The angles of the separation points (where the flow field releases a vortex by separating
from the line surface) are also available.
For this model, the shaded view is not relevant, as vortices are not drawn in the shaded mode.
Watching the replay on the default wire frame view gives an interesting look at the release of the
vortices responsible for VIV.

Combined Orcaflex Example

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Combined Orcaflex Example

Uploaded by

Copyright:

Available Formats

W Orcina

Production Risers: A01 Catenary and Wave Systems Page 1 of 14

You might also like