Proceedings of the Twenty-first (2011) International Offshore and Polar Engineering Conference

Maui, Hawaii, USA, June 19-24, 2011

Copyright 2011 by the International Society of Offshore and Polar Engineers (ISOPE)
ISBN 978-1-880653-96-8 (Set); ISSN 1098-6189 (Set); www.isope.org

Top Tensioned Riser Layout Design Optimization

Yongjun Chen, Peimin Cao
SBM Atlantia, Inc.
Houston, TX USA

at the deck wellbay and at the seafloor should be maximized. However,

the riser spacing at deck wellbay is restrained by the overall deck design
and layout, and is normally pre-defined. If it is not pre-defined, this
methodology still can be used to evaluate various deck layout options.
The riser spacing at the seafloor is restrained by the riser vertical angle
required by the installation and well operations. In addition, in order to
reduce TTR stroke, the offset of well at seafloor level should be
minimized from the riser slot at the deck level in any direction, i.e. the
riser vertical angle should be minimized.

ABSTRACT
Dry tree production platform becomes more and more popular due to its
low OPEX and ease of well intervention. For TTR layout design, there
are minimum spacing requirements for both deck wellbay and seafloor.
There are also constraints on the deck spacing and the riser offset at
seafloor relative to the wellbay location, i.e. riser vertical angle. For
fields with substantial reserve, the number of TTRs may become large;
and thus the TTRs may cluster together, which makes the riser layout
design very complex, and may require exhaustive design iterations.

The most preferable layout is that all TTRs being installed without
offset, which is easy for installation, well operation, and induces no
geometric stroke. However, for clustered and large fields with 20+
wells, running all TTR straightly vertical down is difficult, or even
impossible, to achieve and the TTR layout design becomes very
complex and may require extensive design iterations.

In order to solve the problem, this paper presents a novel optimization

approach called Equilateral Triangle Grid method. Based on the
predefined deck wellbay layout, the method guarantees the minimum
spacing requirement at seafloor while tries to minimize the riser vertical
angle.

This paper presents a novel optimization method Equilateral Triangle

Grid for TTR layout design for clustered wells. For a predefined deck
wellbay layout, the method first puts all the seafloor wellheads on an
equilateral-triangle grid, which maintains the wellhead seafloor spacing
as the required; it then shifts and rotates the grid to minimize the riser
vertical angle. This method has been applied to a case study, and its
results are compared to a commercial optimization program.

KEY WORDS: TTR Layout; Dry-Tree System; Optimization

INTRODUCTION
Dry tree production systems, such as TLPs and Spars, are continuously
being designed and installed for deepwater developments worldwide,
especially for reservoirs with large reserves. This is because dry tree
systems offer direct access to the development wells, and thus reduces
the cost associated with drilling and well intervention. Top tensioned
risers (TTRs) installed on the dry tree production system provides the
shortest flow path, which reduces flow assurance problem and increases
recovery rate from the reservoir.

PHYSICAL MODEL DESCRIPTION

The TTR layout design includes both deck wellbay layout and seafloor
well layout. The deck wellbay layout is generally governed by the deck
space, equipment handling, jumper/tree clashing requirements, and
operational limitations. Therefore the deck wellbay layout is normally
predefined. The seafloor well layout is generally governed by the riser
clashing requirements, deployment and stab-in requirements, and ROV
access. The seafloor well layout is normally determined and optimized
by the engineering design.

The TTR layout design includes both deck wellbay layout and seafloor
well layout. The deck wellbay layout is generally governed by the deck
space, equipment handling, jumper/tree clashing requirement, and
operational limitations. The seafloor well layout is generally governed
by the riser clashing requirements, riser deployment and stab-in
requirements, and ROV access.

Under sufficient tension, each TTR can be simplified as a straight linepipe connecting two nodes, one at the deck (tensioner ring, air can, or
guide structure), and the other at the seafloor (wellhead tie-back

For equipment handling and riser clashing, the distances between risers

57

For the seafloor well layout design, the following constraints should be
considered:

connector). Figure 1 shows a typical TTR and connections (Rainey,

2002); Figure 2 shows a typical deck wellbay layout (Jordon, Otten,
Trent, and Cao, 2004), with tensioner, jumper, and trees are shown; and
Figure 3 shows a typical riser layout, both at deck wellbay and at
seafloor.

1)

2)

3)

The spacing between any pair of wellheads should be not less than
a threshold value set by the installation, drilling and ROV access
requirements;
The offset of riser wellhead relative to the riser slot at the deck in
any direction, i.e. riser vertical angle, should be limited to no
greater than a certain value set by deployment and final stab-in
angle requirements;
There is no clashing between risers under extreme environment
with the consideration of wake effects.

Normally constraint 1) is a mandatory requirement which can not be

relaxed or design around; while constraints 2) and 3) can be achieved
through engineering design and optimization.
Constraints 1) and 3) require large seafloor spacing; on the contrary,
constraint 2) requires small seafloor spacing. This creates an
optimization problem minimize riser the vertical angle while satisfy
the seafloor spacing and no clashing constraints.

MATHEMATICAL MODEL DESCRIPTION

There are two sets of nodes, A and B, each has N nodes. The positions
of A nodes (A1, A2 ) are predefined (deck); while the positions of B
nodes (B1, B2 ) need to be determined (seafloor). Nodes Ai and Bi
are considered as one pair, and connect them forms the ith riser.

Figure 1. Typical TTR and Connection

Definitions:
DAB(i) horizontal distance between nodes Ai and Bi, offset
for the ith riser.
DBB(i,j) - distance between any two nodes (Bi, Bj) within set B
Known Variables:
Ai(x), Ai(y) - position of all nodes within set A
DBB0 - a preset threshold value
Decision Variables:
Bi(x), Bi(y) - position of all nodes within set B
Optimization Target:
Minimize the maximum value of all N DAB values

Figure 2. Typical Deck Wellbay Layout

Constraints:
DBB(i,j) DBB0

Seabed
Deck
C06

C18

C12

C05

C24

(for all i and j, i j)

Figure 4 illustrates the optimization problem.

C17

C11

C04

C23

C16

C10

C03

C22

C15

C09

C02

C21

C14

C08

C01

C20

C13

0
2

C07

C19

Figure 3. Typical TTR Layout

Figure 4. Mathematical Model Illustration

58

1.
This optimization problem has 2N decision variables and N*(N-1)/2
constraints. When the number N becomes large, such as 20+, the
problem becomes very complex and needs special program to handle.

2.
3.

EQUILATERAL TRIANGLE GRID METHOD

4.

In order to simplify the above problem, the first task is to reduce the
number of decision variables and/or constraints.

5.
6.

Random select a starting point (normally one of the center nodes)

and initial heading angle (Figure 6)
Generate an equilateral triangle grid which covers all the A nodes
(Figure 7)
Pick up B nodes from the grid, and pair/connect them with the
closest A node (Figure 8)
Calculate the distance for all pairs of AB nodes, and then calculate
the maximum distance
Minimize the maximum distance DAB by shifting/rotating the grid
Report the positions of B nodes

Since the distance between any two nodes within set B should be not
less than a certain value; if all the B nodes are chosen from an
equilateral triangle grid (grid spacing = DBB0) as shown Figure 5, then
all the constraints (DBB(i,j) DBB0) are satisfied.

Figure 6. Step 1 - Select Starting Point and Initial Heading

Figure 5. Equilateral Triangle Grid

By choosing B nodes from an equilateral triangle grid (grid spacing =
DBB0), the distance between any pair of B nodes is guaranteed not less
than the required value of DBB0.
In addition, for any equilateral triangle grid, the location of all nodes
can be determined based on the following three variables (as shown in
Figure 5).

(BX1, BY1) - the position of the starting node

- the initial heading of the first grid line

i.e.

BXi = f (BX1, BY1, )

BYi = f (BX1, BY1, )

Figure 7. Step 2 - Generate Equilateral Triangle Grid

Therefore, the original optimization can be simplified as:

Minimize the maximum value of all N DAB values by

adjusting decision variables (BX1, BY1) - shift the grid
and adjusting decision variable - rotate the grid

This simplified optimization problem only has three decision variables

without constraints, and thus is much easier to handle than the original
problem.
Please bear in mind that the solution space for the simplified problem is
only a subset of the original problems solution space, and thus the
optimal solution for the simplified problem may not be the optimal
solution for the original problem.
The general procedure for design/optimize the TTR layout design is:

Figure 8. Step 3 - Pick B Nodes and Connect to A Nodes

59

S e abe d - Un k n o wn
B Y (ft )
B X (ft )
-24.06
-55.88
-24.76
-35.89
-25.46
-15.90
-26.16
4.09
-26.86
24.08
-27.56
44.06
-7.10
-45.28
-7.80
-25.29
-8.50
-5.30
-9.20
14.69
-9.90
34.68
-10.60
54.66
10.56
-54.66
9.86
-34.68
9.16
-14.69
8.46
5.30
7.76
25.29
7.06
45.28
27.52
-44.06
26.82
-24.08
26.12
-4.09
25.42
15.90
24.72
35.89
24.02
55.88

Ris e r ID

IMPLEMENTATION USING EXCEL

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

For the simplified optimization problem, Microsoft Excel is the best

tool to implement the Equilateral Triangle Grid method. The
implementation includes 4 sections.
Section 1 provides user interface for input design requirements, decision
variables, and target values; as shown in Figure 9.

TTR Layout Design Optimization

- Using Equilateral Triangle Grid Method
Design Requirement:
Required Dist. DBB0 =
Calculated DBB-min =
Meet Requirement?
Decision Variables
Initial Heading =
Starting Node BX1
Starting Node BY1
Target Value - Minimize
Max of DAB

20.00
20.00
Y
0.175
0.000
0.000
27.04

DAB
S -play
10.34
10.01
9.83
9.79
9.90
10.16
2.48
3.68
4.89
6.10
7.30
8.51
8.49
7.28
6.07
4.87
3.66
2.46
10.12
9.87
9.76
9.80
9.99
10.31

De c k - Kn o wn
AX (ft )
AY (ft )
-18.00
-47.50
-18.00
-28.50
-18.00
-9.50
-18.00
9.50
-18.00
28.50
-18.00
47.50
-6.00
-47.50
-6.00
-28.50
-6.00
-9.50
-6.00
9.50
-6.00
28.50
-6.00
47.50
6.00
-47.50
6.00
-28.50
6.00
-9.50
6.00
9.50
6.00
28.50
6.00
47.50
18.00
-47.50
18.00
-28.50
18.00
-9.50
18.00
9.50
18.00
28.50
18.00
47.50

Figure 10. Implementation in Excel Coordinates Table

Figure 9. Implementation in Excel Users Input
DBB Table

Section 2, as shown in Figure 10, is the table for TTR node coordinates,
which includes both the deck nodes (set A, known) and the seabed
nodes (set B, unknown). The coordinates of set B are calculated based
on BX1, BX2, and equilateral equation. The distance between each pair
of nodes (DAB) is also calculated in the table, and the max DAB is
highlighted.

D BB

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

19

20

21 22 23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

20

40

20

60

40

20

80

60

40 20

6 100

80

60 40

20

20

35 53 72

92

35

20

20 35 53

72

20

53

35

20 20 35

53

40

The distance between any two nodes within set B (DBB) are calculated
and listed in Section 3 (DBB table) as shown in Figure 11. The minimum
distance values are highlighted.

10

72

53

35 20

20

35

60

40

20

11

92

72

53 35 20

20

80

60

40

20

12 111

92

72 53 35

20

100

80

60

40

13

40

53 69

87

106

20

35

53

72

Section 4 provides the plot for user to view the layout results.
After the above initial set up, Solver add-in in Excel can be used to
search the best solution. The Solver set up is:
Set Target Cell:
Equal To:

Max of DAB
Min

By Changing Cells:

35

20

20

20
92

111

14

40

35

40 53

69

87

20

20

35

53

72

92

20

15

53

40

35 40 53

69

35

20

20

35

53

72

40

20

16

69

53

40 35 40

53

53

35

20

20

35

53

60

40

20

17

87

69

53 40 35

40

72

53

35

20

20

35

80

60

40

20

18 106

87

69 53 40

35

92

72

53

35

20

20

100

80

60

40

19

53

53

60 72

87

104

35

40

53

69

87

106

20

20

35

53

72

92

20

60

53

53 60 72

87

40

35

40

53

69

87

35

20

20

35

53

72

20

21

72

60

53 53 60

72

53

40

35

40

53

69

53

35

20

20

35

53

40

20

22

87

72

60 53 53

60

69

53

40

35

40

53

72

53

35

20

20

35

60

40

20

23 104

87

72 60 53

53

87

69

53

40

35

40

92

72

53

35

20

20

80

60

40

20

24 122

104 87 72

53

106

87

69

53

40

35

111

92

72

53

35

20

100

80

60

40

60

20

Figure 11. Implementation in Excel DBB Table

Initial Heading , Starting Node BX1, BY1

Subject to the Constraints:

It is unnecessary to add constraints on DBB DBB0
since the equilateral triangle grid method
guarantees this requirement.
In order to speed-up the search process, user may
add constraints on the values of , BX1, and BY1
Figure 12 shows the Solver setup in Excel.
Figure 12. Implementation in Excel Solve Setup

60

20

CASE STUDY
The model of the problem consists of 48 nonlinear decision variables,
276 constraints. Both Non-Linear Programming and Global Solver are
employed. It took several hours to find a global optimal solution.

A deepwater dry tree facility is designed with 24 well slots. The well
pattern is 6 rows in North-South direction with 4 wells each row (EastWest). The well spacing is 19-ft along both North-South direction and
12-ft along East-West direction. At seabed, the minimum distance
between wells is 20-ft. To facilitate riser installation, drilling and well
intervention, the riser vertical angle with respect to the vertical direction
for the bottom connection shall be not more than 2.0 degrees, and this
includes the wellhead inclination tolerance that is 1.0 degree. So, for
installation, the maximum angle with respect to the vertical direction is
1.0 degree. The water depth at the field is 1500-ft.

Figure 14 shows the solution from LINGO. The optimized/minimized

maximum distance of all AB pairs (DAB) is 9.69-ft.
TTR Layout Optim ization

Seabed

Deck

60
50

The problem was modeled in Microsoft Excel using Solve feature. It

took less than one minute to find an optimal solution.

40

Figure 13 shows the optimal solution from the Equilateral Triangle

Grid method with three decision variables having values:

30
20

BX1=-8.50; BY1=-5.30; =2.0 degrees

Y (S-N)

10

The optimized/minimized maximum distance of all AB pairs (DAB) is

10.34-ft, this occurs at both the bottom-left node and the top-right node.

0
-10
-20

TTR Layout Optim ization

Seabed

Deck

-30

60

-40

50

-50

40

-60

30

-60

-40

-30

-20

-10

10

20

30

40

50

60

Figure 14. Case Study Solution from LINGO

10
Y (S-N)

-50

X (W-E)

20

By comparing Figure 13 with Figure 14, it can be found that the optimal
solution from the Equilateral Triangle Grid method is very close to
the global optimal solution from LINGO, for both seafloor layout
pattern and minimized DAB value.

0
-10
-20
-30

CONCLUSION AND DISCUSSION

-40

An Equilateral Triangle Grid method was proposed to optimize the

TTRs layout design for the dry tree facilities with clustered risers. The
method provides the most compact grid for searching the solution. A
case study was examined and the result demonstrates that the solution
from the proposed method could be close to the global optimal solution.

-50
-60
-60

-50

-40

-30

-20

-10

10

20

30

40

50

60

X (W-E)

Figure 13. Case Study Solution

Due to the extra constraints introduced into the simplified problem, i.e.
all nodes are on the equilateral triangle grid; the solution space of the
simplified problem is only a subset of the solution space for the original
problem. Therefore the optimal solution from the method may not be
the global optimal solution of the problem. However, the method is still
a very useful tool because of its ease to model, quick to find solution,
and the solution can be very close to the global optimal solution.

To show effectiveness of the method, the problem was also solved using
the commercial optimization software LINGO from LINDO
SYSTEMS.
LINGO is a comprehensive tool (modeling language) designed to make
building and solving Linear, Nonlinear (convex and non-convex /
global), Quadratic, Quadratically Constrained, Second Order Cone,
Stochastic, and Integer optimization models faster, easier and more
efficient. User does not need to specify which solver to use; the LINGO
automatically selects the appropriate solver based on users model and
formulation.

For projects with special requirements on the deck and seafloor well
layout to assist a certain installation or drilling sequence, such as all
wells are required to be on concentric circles or on straight lines, the
proposed method may not provide a good feasible solution. However,
by employing the same principle and procedure, Concentric Circle

61

Grid, Rhombus Grid, or Square Grid method can be used to search

the best solution meeting the project requirements.

REFERENCES

For projects with strict requirements on the riser vertical angle and
spacing, multiple grid methods can be used simultaneously and select
the best solution among all sub-optimal solutions.

Rainey RM, (2002), Brutus Project Overview: Challenges and

Results, Proc of Offshore Tech Conf, Houston, USA
Jordon R, Otten J, Trent D, Cao P, (2004), Matterhorn TLP Dry Tree
Production Risers, Proc of Offshore Tech Conf, Houston, USA
LINDO SYSTEMS website: http://www.lindo.com

ACKNOWLEDGMENTS
The authors would like to acknowledge the management of SBM
Atlantia for allowing this paper to be published. The valuable comments
and advice from S. Schuurmans of SBM are also greatly appreciated.

62