Discretisation, characterisation, and complexification of

multiphase pipeline elevation profiles

E Zakarian, J Morgan
Woodside Energy Ltd, Australia
H Holm
Statoil, Norway
D Larrey
Total, France

1 ABSTRACT

Two discretisation methods have been proposed recently for the simplification of detailed
gas-condensate pipeline elevation profiles (1). Both methods preserve the hydrodynamic
behaviour of the original detailed profile and produce a suitable simplified profile for the
dynamic simulation of multiphase flows. A number of operators and engineering
companies have been evaluating and, in some cases, using these methods for the design
of world-class multiphase transport systems. The first part of this paper summarises the
key lessons learned from these first applications and concludes with a set of
recommendations for the simplification of detailed pipeline elevation profiles in the
context of dynamic simulation of multiphase flows.

A comprehensive analysis of seabed surveys and as-laid bottom-of-pipe profiles has

revealed that non-dimensional large-scale indicators can be used to develop a sensible
basis for the characterisation and complexification of elevation profiles. The second part
of this paper introduces a new complexification methodology which proposes to increase
the roughness of coarse elevation profiles by means of non-dimensional indicators.
Correlations for the determination of roughness indicators are proposed for various
geographical areas.

2 INTRODUCTION

In the early stages of offshore field developments, the potential routes for production
pipelines are commonly assessed from limited bathymetric data whilst detailed data from
geophysical surveys of the seabed may be available for the last phases of engineering
design. When pipelines are designed for multiphase flow transport, the complexification
of coarse bathymetric contours (to model the influence of the seabed roughness on the
hydrodynamic behaviour of multiphase flows) or the simplification of detailed bottom-
of-pipe and seabed profiles (to reduce the computation time of multiphase flow
simulations) must be exercised with caution because multiphase flows are known to be
sensitive to pipe inclination. However that sensitivity is more or less pronounced,
depending on the driving fluid (gas or liquid) and number of fluid phases involved (two
or three).

Four common multiphase flow configurations in offshore production are discussed below
on the basis of steady-state simulations performed with a thoroughly validated,

© BHR Group 2013 Multiphase 16 267

commercial simulator, using fluid compositions and conditions from operating assets or
on-going field developments. The liquid holdup is the main focus here. The sensitivity of
pressure drop to the roughness of elevation profiles is discussed later in this paper (see
Section 5).

The steady-state condensate holdup in a pipe transporting a flow of dry (two-phase) gas-
condensate with a low liquid loading is known to be very sensitive to the gas velocity in
slightly upwardly inclined pipes (2). In Figure 1 (top left), a sharp transition from low to
high holdup occurs as the gas superficial velocity is gradually reduced, for any small,
positive angle to the horizontal. The critical gas superficial velocity, at which an abrupt
transition from low to high holdup (or holdup discontinuity) occurs, depends on the
inclination of the pipe: the higher the inclination, the higher the critical gas superficial
velocity. Although the steady-state condensate holdup may never be reached in practice
in large diameter, long gas-condensate pipelines, it is anticipated that the simplification
of a detailed elevation profile in the context of dry gas-condensate transport with a low
liquid loading would likely result in a significant underestimation of liquid inventories
should the angle distribution of the original profile be poorly preserved through the
simplification process; e.g. significant loss of high angles. Note: the angle distribution
refers here to the distribution of accumulated pipe lengths for a set of predetermined
angle groups (see Section 5 for an example).

The liquid holdup in wet gas-condensate systems is also sensitive to the pipe inclination
although the presence of a second (aqueous) liquid phase may reduce the sharpness of the
transition from low to high holdup (2). As shown in Figure 1 (top right), the sensitivity of
the total (condensate and water) liquid holdup to pipe inclination is less significant at
angles greater than 2 degrees as opposed to the sensitivity of the total liquid holdup to
smaller, positive angles. In this example a sharp drop of the liquid holdup is observed at
low gas superficial velocities and ca. 2 degrees. At this critical angle a flow regime
transition from stratified to intermittent flow takes place. This transition amplifies the
mixing between gas and liquid phases, resulting in a reduced slip, and subsequently a
decrease in liquid holdup.

In two-phase oil-water flows, the sensitivity of the aqueous phase holdup to inclination at
low oil superficial velocities is primarily driven by the sign of the pipe inclination. As
shown in Figure 1 (bottom left), this holdup varies sharply from low to high values as the
angle varies from negative values to a small, positive angle. Above this critical
inclination close to zero, the water holdup decreases gradually as the inclination
increases. The preservation of the angle distribution for positive inclinations would be
recommended for this fluid system should a detailed elevation profile be transformed into
a simpler geometry.

Similarly a variation from low to high water holdup is observed at low oil superficial
velocities in three phase oil systems when the angle approaches zero from negative
values (see Figure 1, bottom right). However, unlike two-phase oil-water systems, the
presence of a light vapour phase reduces the sharpness of this gravity-driven transition.

The black solid line shown in Figure 1 is a mathematical function modelling the typical
S-shaped curve of the liquid holdup with respect to pipe inclination; see Eq. 1, Section 3.
Although this function was originally introduced for the characterisation of gas-
condensate pipeline elevation profiles, through the concept of Profile Indicator (3), a
similar (S-shaped) behaviour can be observed across different types of multiphase flow
systems; see examples in Figure 1. As reminded elsewhere the Profile Indicator is more
an engineering tool for the characterisation of pipeline elevation profiles than a predictive

268 © BHR Group 2013 Multiphase 16

model (1) (3). On the basis of a smooth S-shaped mathematical function, the Profile
Indicator provides the Engineer with a practical tool (in combination with other
indicators) to simplify detailed multiphase pipeline elevation profiles, without the
requirement for the predefinition of pipe inclination classes or angle groups (1).

Dry gas-condensate Wet gas-condensate

Condensate holdup vs. pipe inclination Total liquid holdup vs. pipe inclination
and gas superficial velocity (Usg) and gas superficial velocity (Usg)
0.7 0.7

0.6 0.6
Condensate holdup [-]

Total liquid holdup [-]

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0 0
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Pipe inclination [deg] Pipe inclination [deg]
Usg=1.05 m/s Usg=1.58 m/s Usg=2.12 m/s Usg=1.06 m/s Usg=1.60 m/s Usg=2.14 m/s
Usg=2.65 m/s Usg=3.19 m/s Usg=3.73 m/s Usg=2.67 m/s Usg=3.22 m/s Usg=3.76 m/s
Usg=4.27 m/s Usg=4.81 m/s Usg=5.35 m/s Usg=4.30 m/s Usg=4.83 m/s Usg=5.34 m/s
Holdup (Eq. 1) Holdup (Eq. 1)

Oil & water Oil, water & gas

Water holdup vs. pipe inclination Water liquid holdup vs. pipe inclination
and oil superficial velocity (Uso) and oil superficial velocity (Uso)
0.9 0.6

0.8
0.5
0.7
Water holdup [-]

0.6
Water holdup [-]

0.4

0.5
0.3
0.4

0.3 0.2

0.2
0.1
0.1

0 0
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Pipe inclination [deg] Pipe inclination [deg]
Uso=0.07 m/s Uso=0.10 m/s Uso=0.13 m/s Uso=0.25 m/s Uso=0.41 m/s Uso=0.58 m/s
Uso=0.20 m/s Uso=0.27 m/s Uso=0.40 m/s Uso=0.75 m/s Uso=0.91 m/s Uso=1.08 m/s
Uso=0.67 m/s Uso=0.93 m/s Uso=2.00 m/s Uso=1.24 m/s Uso=1.41 m/s Uso=2.08 m/s
Holdup (Eq. 1) Holdup (Eq. 1)

Figure 1: Liquid holdup vs. inclination and superficial velocity for various
multiphase flow systems

3 DISCRETISATION OF DETAILED ELEVATION PROFILES

The discretisation of pipeline elevation profiles consists of transforming detailed datasets

(typically several tens of thousands of coordinates) into simplified geometries. The
discretisation process shall preserve the hydrodynamic behaviour of the original, detailed
profile and produce a suitable, simplified profile for the simulation of multiphase flows.
The following requirements are commonly recommended (1):

© BHR Group 2013 Multiphase 16 269

 The overall length and horizontal span of the original elevation profile should be
preserved through the discretisation process.

 The Total Climb, also referred to as the “sum of uphill flow elevation changes” (4),
should be conserved to predict the same overall liquid content in steady-state flow
conditions

 The angle distribution of the discretised profile should be as close as possible to the
original distribution. Note that a perfect match would implicitly satisfy the above
requirement for preservation of the Total Climb.

If the discretised profile is used for dynamic flow simulations, the following
requirements should also be pursued:

 The simplified geometry should have the same overall shape (large and small scale
undulations) to preserve the transient behaviour of the original profile; e.g.
propagation of level slugs upon restart, formation of liquid surge waves during
production ramp-ups or sustained production at low flow rates (gravity dominated
flow), production of slugs during routine pigging operations or inline inspection, etc.

 The distribution of segment lengths should be uniform to avoid severe time step
limitation. In commercial dynamic simulators the calculation segments are defined
by the elevation profile and are not subdivided during simulation as they would be in
steady-state simulators where the calculation engine would automatically subdivide
long segments as needed.

Two discretisation methods satisfying the above requirements have been proposed
recently in the literature (1). A number of operators and engineering companies have
been evaluating and, in some cases, using these methods for the design of world-class,
multiphase transport systems. The next section describes the fundamentals of the first
method. Then some recommendations are made for the discretisation of elevation
profiles.

3.1 Discretisation method

Using two indicators to quantify the propensity for a pipeline to accumulate liquids in
steady-state flow conditions, namely the Profile Indicator and Total Climb, the first
discretisation method presented in (1) is based on a three-step algorithm consisting of
simplifying, splitting and complexifying a set of sub-profiles extracted from a detailed
elevation profile.

The Profile Indicator (PI) is a dimensionless parameter based on a simple mathematical

function, hereafter referred to as , which aims to reproduce the behaviour of the
liquid holdup with respect to pipe inclination in gas-condensate systems (3):

0.49/ . 1.9 0.66 0.25 (Eq. 1)

The argument of the function, , symbolises the slope in % of the pipe to the horizontal.
A representation of the function can be seen in Figure 1 where is converted
into degrees. The PI is formulated as follows (3):

1000
0 (Eq. 2)

270 © BHR Group 2013 Multiphase 16

where is the inclination in % of the ith pipe to the horizontal; is the length of the ith
pipe; is the total pipeline length [m]: ∑ ; and N is the number of pipes
comprising the elevation profile.

The Total Climb (TC) is defined as the “sum of uphill flow elevation changes” (4):

,0 (Eq. 3)

where Δz is the elevation change in metres of the ith pipe in the direction of the flow.

From these definitions the indicators PI and TC can be computed for any portion of an
elevation profile, hereafter referred to as sub-profile. From our experience a uniform
distribution of sub-profiles with a constant span of 10 km is adequate in most cases for
the discretisation of detailed profiles. For example a pipeline covering a total distance of
135 km should be split in 13 or 14 sub-profiles.

For each sub-profile the discretisation process consists of (1):

1. Extracting data points from the original (detailed) bottom-of-pipe elevation

profile at a constant distance step;

2. Adding intermediate segments at a fixed distance step to generate a uniform

distribution of segment lengths; and

3. Complexifying the simplified profile by changing the elevation of the

intermediate segments randomly. The complexification is complete when the
PI and TC indicators for each complexified sub-profile are almost identical to
their target value determined from the original (detailed) sub-profile.

When only detailed seabed survey data are available for the derivation of a pipeline
profile, it is advised to filter any noise from the data to remove unphysical roughness
from the seabed profile. It is also recommended to calculate roughness indicators from
the detailed seabed profile (see Section 4.1) and reduce the value of these indicators to
take account of the smoothing effect of pipeline laying; see Section 4.4. Then use these
reduced roughness indicators to calculate PI and TC indicators for the complexification
step above; see Eq. 4 and 5.

3.2 Recommendations
The following recommendations should be pursued for the discretisation of detailed
elevation profiles:

 The discrepancy between the original and discretised TC shall not exceed a relative
difference of 1% for each sub-profile. A larger relative difference for the PI may be
acceptable although a relative difference below 1% is feasible in most cases for each
sub-profile.

 Undefined or poorly defined portions of the pipeline route should be removed from
the original profile prior to discretisation which would be performed on distinct,
detailed sub-profiles. The latter can be used to characterise the original pipeline
profile (see Section 4) and generate indicators for the complexification of the coarse

© BHR Group 2013 Multiphase 16 271

 portions (see Section 5). As a rule of thumb the maximum segment length in the
original (detailed) profile should not exceed 50 m or so.

 Vertical or steep geometries exceeding angles of +/- 45° like risers and wells should
be removed from the original profile prior to discretisation. However a manual
simplification of these geometries is recommended prior to reconnection to the
discretised sub-profiles to avoid severe time step limitation during dynamic
simulation should their minimum section length be much smaller than the minimum
section length in the discretised sub-profiles.

 Unrealistic offsets in bathymetric or bottom-of-pipe profile data should be deleted

from the original profile prior to discretisation. Geometric singularities or steep
elevation changes across items such as: subsea manifolds, templates, Christmas trees,
subsea processing stations, riser base spools, entries to buried/trenched portions of a
pipeline route, etc. may be lost through the discretisation process. If the influence of
geometric singularities on the hydrodynamic behaviour of the production system
needs to be quantified, these singularities should be removed from the original profile
data prior to their reassembly to adjacent discretised sub-profiles. Adjustment of the
segment lengths along the final profile is recommended to avoid exceeding a
maximum factor of 2 in length ratio between adjacent sections.

 Following the simplification of the original profile (see Step 1 in Section 3.1) a
splitting based on 3 intermediate points per simplified section, or 4 segments, is
recommended. Whether a larger number of points in the final discretised profile is
deemed necessary to capture terrain undulations and to preserve the angle
distribution more precisely it is preferable to increase the number of original points in
the simplified profile (i.e. reduce the span between consecutive data points in Step 1)
rather than increasing the number of intermediate segments in Step 2.

 Discretisation can be limited to a simplification process (i.e. no complexification

required) for pipelines that are laid on a smooth terrain, or more formally, when the
Terrain Indicator of all original sub-profiles is close to zero (see Section 4.1).

4 CHARACTERISATION OF ELEVATION PROFILES

This section introduces a new methodology for the characterisation of pipeline elevation
profiles. Working as a sort of fingerprinting this methodology was developed through a
comprehensive analysis of detailed seabed surveys and as-laid bottom-of-pipe profiles
covering long distances (more than 100 km in most cases).

4.1 Roughness indicators

The characterisation of the terrain roughness is a fundamental input to the
complexification of elevation profiles. A comprehensive analysis of detailed seabed
surveys and bottom-of-pipe profiles from a substantial number of offshore developments
has revealed that non-dimensional, large-scale indicators could be used to provide a
sensible basis for the characterisation of elevation profiles.

The Profile Indicator and Total Climb are two practical indicators to quantify the
propensity for a pipeline to accumulate liquids in steady-state flow conditions (see
Section 3). However these indicators are inappropriate to quantify the roughness of an
elevation profile. For example, a pipeline laid on a rugged, horizontal terrain and a

272 © BHR Group 2013 Multiphase 16

pipeline of same length running uphill on a steep, smooth continental slope can have the
same PI and TC. To eliminate the dependence of the PI and TC indicators on large-scale
elevation changes, we propose to use the following non-dimensional indicators, namely
the Terrain Indicator (TI) and Rough Profile Indicator (RPI):

,0
(Eq. 4)
/1000
(Eq. 5)

where SPI symbolises the Smooth Profile Indicator of an elevation profile:

0 . 1000 (Eq. 6)

. 100 (Eq. 7)

Δ is the horizontal span of the profile and Δ is the corresponding elevation change; see
Figure 2 for an illustration of these variables.

Original bottom-of-pipe profile Smooth profile

-120
-122
-124

z
-126
Elevation [m]

-128
-130
-132
-134 x
-136
-138
-140
0 2,000 4,000 6,000 8,000 10,000
Distance [m]

Figure 2: Elevation profile

The TI indicator is a positive indicator. It is equal to zero for a straight profile (no
undulations). When an elevation profile is horizontal or running downhill (Δ 0 the
TI is equivalent to a “kilometric Total Climb” ( / /1000 .

The SPI indicator quantifies the “smooth fraction” of the PI. It is positive when a pipeline
is upward inclined (Δ 0 and negative if the pipeline is running downhill (Δ 0 .A
profile with an end-to-end inclination less than 0.66% (0.378°) has a SPI below 150 (see
Figure 3). A pipeline laid on a continental shelf is an example of such profile. A pipeline
laid on a steep continental slope has a SPI above 150. A flowline sloping downwards to
the base of a deep water riser or the downhill section of a pipeline crossing an oceanic
trench are two examples of profiles with negative SPI.

The RPI indicator measures the “rough fraction” of the PI. For SPI values between ∞
and 150, the RPI is positive. For steeply inclined pipelines (SPI > 150), the RPI is
negative: see Figure 4.

© BHR Group 2013 Multiphase 16 273

 Smooth profile indicator (SPI) vs. sub-profile inclination
SPI vs. inclination [deg] SPI vs. inclination [%]
400
350
Smooth Profile Indicator [-]

300
250
200
150
100
50
0
-50
-100
-10 -5 0 5 10
Sub-profile inclination [deg or %]

Figure 3: SPI vs. sub-profile inclination

PI and RPI vs. sub-profile inclination

Profile Indicator (PI) [-] Rough Profile Indicator (RPI) [-]
400 250
350 Terrain Indicator [-] = 10 200

Rough Profile Indicator [-]

300 150
Profile Indicator [-]

250 100
200 50
150 0
100 -50
50 -100
0 -150
-50 -200
-100 -250
-10 -5 0 5 10
Sub-profile inclination [deg]

Figure 4: PI and RPI vs. sub-profile inclination

4.2 Regression of roughness indicators

Five areas are analysed in this paper: the North Sea, North West Shelf (Australia), Timor
Sea, Persian Gulf, and Barents Sea. For each of the 12 pipeline profiles characterised in
this paper, a detailed bottom-of-pipe elevation profile was cut into a uniform distribution
of sub-profiles. Then the TI and RPI indicators were calculated for each sub-profile. The
results are presented in Figure 5 for positive RPIs only.

A linear dependence of the TI on the RPI is visible for each geographical area, which
indicates that the TI and RPI indicators are adequately defined to quantify the roughness
of the seabed. Long, steep routes with PIs exceeding 150 are characterised by negative
RPI values as per Figure 4. The regressions presented in Figure 5 cannot be used for the
complexification of pipelines laid on steep continental slopes. However TIs can be
calculated from analogues in the same area and used for the complexification of steep
profiles (see Section 5).

274 © BHR Group 2013 Multiphase 16

 North Sea North West Shelf
Pipeline A Pipeline B Pipeline C Pipeline A Pipeline B
9 6

8
5
7
y = 0.0848x - 0.3796 y = 0.0739x + 0.1084
Terrain Indicator [-]

Terrain Indicator [-]

6 4

5
3
4

3 2

2
1
1

0 0
0 20 40 60 80 100 0 10 20 30 40 50 60
Rough Profile Indicator [-] Rough Profile Indicator [-]

Timor Sea Persian Gulf

Pipeline A Pipeline B Pipeline A Pipeline B Pipeline C
14 3.5

12 3
y = 0.0652x + 0.1076
10 2.5
Terrain Indicator [-]
Terrain Indicator [-]

8
y = 0.0896x - 0.2207 2

6 1.5

4 1

2 0.5

0 0
0 20 40 60 80 100 0 10 20 30 40 50
Rough Profile Indicator [-] Rough Profile Indicator [-]

Barents Sea
Pipeline A Pipeline B
20

18

16

14 y = 0.1546x - 4.2183
Terrain Indicator [-]

12

10

-2
20 40 60 80 100 120

Rough Profile Indicator [-]

Figure 5: TI vs. RPI for various offshore areas

© BHR Group 2013 Multiphase 16 275

4.3 Seabed feature analysis
Non-dimensional indicators can also be used for the characterisation of seabed features
with respect to the unevenness of their profiles. From detailed seabed survey information,
typically available in the form of alignment sheets and geophysical reports, a list of
seabed features was identified for each (10km-long) sub-profile extracted from the
bottom-of-pipe profile of two pipelines characterised in the previous section. The results
are reported in Figure 6 and Figure 7.

North West Shelf Pipeline A

Terrain Indicator vs. Rough Profile Indicator
Bottom-of-Pipe profile at normal operating conditions
6
Terrain Indicator [-]

5 y = 0.0755x + 0.1778

0
0 10 20 30 40 50 60
Rough Profile Indicator [-]
Flat Seabed + Megaripples + Scarps + Palaeochannels + Ridges + Sandwaves + Low Relief
Flat Seabed + Megaripples + Scarps + Palaeochannels + Ridges
Flat Seabed + Megaripples + Scarps + Palaeochannels + Sandwaves + Depressions
Flat Seabed + Megaripples + Scarps + Palaeochannels + Sandwaves + Low Relief
Flat Seabed + Megaripples + Palaeochannels + Sandwaves + Low Relief
Flat Seabed + Megaripples + Palaeochannels + Sandwaves
Flat Seabed + Megaripples + Scarps + Ridges
Low Relief
Flat Seabed + Megaripples + Sandwaves + Low Relief
Flat Seabed + Megaripples + Sandwaves
Flat Seabed + Palaeochannels
Megaripples + Sandwaves
Flat Seabed + Low Relief
Flat Seabed + Megaripples
Flat Seabed

Figure 6: TI vs. RPI for various seabed features (North West Shelf Pipeline A)

We observe that sub-profiles sharing common geomorphic features have similar values
of TIs and RPIs. Sub-profiles with no apparent features (flat seabed) or sharing a
common benign feature are characterised by low values of TIs and RPIs: megaripples,
low relief or sandwaves in Figure 6; gentle iceberg scours and low density, elongated
pockmarks in Figure 7. Profile portions combining hilly features are characterised by
high values of roughness indicators: scarps and ridges in Figure 6; high density, deep,
iceberg scours and pockmarks in Figure 7.

4.4 The smoothing effect of pipeline laying

The smoothing effect of pipeline laying on the seabed is another key parameter to
consider when the seabed roughness of a geographical area is regressed with indicators.
Figure 8 and Figure 9 present a comparison between the bottom-of-pipe elevation profile
(with a content density based on normal operating conditions) and the underneath seabed
profile for the two pipelines analysed in the previous section. These two examples show

276 © BHR Group 2013 Multiphase 16

how the value of roughness indicators determined from detailed seabed surveys may be
adjusted to model the smoothing effect of pipeline laying.

Barents Sea Pipeline A

Terrain Indicator vs. Rough Profile Indicator
Bottom-of-pipe profile at normal operating conditions
20

15
Terrain Indicator [-]

y = 0.1523x - 3.897

10

0
20 30 40 50 60 70 80 90 100 110 120
Rough Profile Indicator [-]
High Density Deep Iceberg Scours + Pockmarks
High Density Deep Iceberg Scours
High Density Iceberg Scours + Depressions + Pockmarks
High Density Elongated Pockmarks
High Density Iceberg Scours + Depressions
High Density Pockmarks
Iceberg Scours (steep wall inclinations: 6° to 37.5°)
High Density Iceberg Scours + Pockmarks
Iceberg Scours + Depressions + Pockmarks
Iceberg Scours + High Density Pockmarks
Iceberg Scours + Pockmarks
Elongated Pockmarks
Low Density Elongated Pockmarks
Iceberg Scours (gentle wall inclinations: 1.5° to 4°)

Figure 7: TI vs. RPI for various seabed features (Barents Sea Pipeline A)

North West Shelf Pipeline A

Seabed profile Bottom-of-pipe profile (normal operating conditions)
8
y = 0.0855x - 0.2389
7

6
Terrain Indicator [-]

4
y = 0.0755x + 0.1778
3

0
0 10 20 30 40 50 60 70 80 90
Rough Profile Indicator [-]
Figure 8: TI vs. RPI (North West Shelf Pipeline A)

© BHR Group 2013 Multiphase 16 277

 Barents Sea Pipeline A
Seabed profile Bottom-of-pipe profile (normal operating conditions)
25

20
y = 0.199x - 5.8333
Terrain Indicator [-]

15
y = 0.1523x - 3.897

10

0
20 30 40 50 60 70 80 90 100 110 120
Rough Profile Indicator [-]
Figure 9: TI vs. RPI (Barents Sea Pipeline A)

5 COMPLEXIFICATION OF ELEVATION PROFILES

This section introduces a complexification methodology which consists of increasing the

roughness of coarse elevation profiles by means of non-dimensional indicators, namely
the Terrain Indicator (TI) and Rough Profile Indicator (RPI) defined in Section 4:

 When detailed profile data from pipelines laid in the same area are available, the
regression of a correlation between the TI and RPI indicators can be used to generate
representative values of complexification indicators for a specific geographical area;
see Section 4.2.

 When geomorphic features of the seabed are identified along the potential routes for
a production pipeline, roughness indicators can be determined for the
complexification of coarse bathymetric profiles; see Section 4.3.

 Region-specific databases of roughness indicators can be used to generate a

representative range of complexified profiles when no detailed information on the
seabed is available; see Figure 11.

The use of roughness indicators for the complexification of elevation profiles is

explained and illustrated through an example. Figure 10 presents a dummy elevation
profile comprising seven straight portions for a total distance of 310 km. Let us assume
that this pipeline could be laid on the North West Shelf in Australia. From the
characterisation of bottom-of-pipe elevation profiles presented in Section 4.2, the RPI
could vary between 0 and 60 and the TI would be determined from the following
correlation as per Figure 5:

0.0739 0.1084 (Eq. 8)

Rather than considering random values between 0 and 60, fixed RPI values are assumed
here for illustration purposes. Assuming a RPI of 10 to model a low seabed roughness
outcome and 50 for a high outcome, the TI would be 0.85 and 3.81 according to Eq. 8,

278 © BHR Group 2013 Multiphase 16

respectively. From these values, we calculate a Total Climb target for each of the straight
portions shown in Figure 10, using Eq. 4. We also note that the SPI, as defined in Eq. 6,
is equivalent to the PI of each straight sub-profile in Figure 10 since 1 and ;
see Eq. 2.

0
-20
Elevation [m]

-40
-60
-80
-100
-120
0 50 100 150 200 250 300 350
Distance [km]
Figure 10: Dummy pipeline elevation profile

As for the discretisation process described in Section 3.1, a complexification process

consists of 3 steps:

1. Dividing each straight sub-profile into segments of same distance (e.g. 1 km).
2. Adding intermediate segments at a fixed distance step for a uniform
distribution of segment lengths.
3. Complexifying each sub-profile by changing the elevation of the intermediate
segments randomly. The complexification is complete when the Total Climb
(TC) of each complexified sub-profile is close to the value determined from the
TI (see Eq. 8), using Eq. 4.

The results of the complexification process for both RPIs of 10 and 50 are presented in
Table 1. The overall characteristics of the original and complexified profiles are reported
in Table 2. Angle distributions and various profiles are shown in Figure 11.

Table 1: Dummy pipeline complexification: results

Sub-profile: first KP 0 50 100 150 200 250 300
Sub-profile: last KP 50 100 150 200 250 300 310
Elevation change [m] 20 -3 28 20 0 15 17
PI: original value [-] 4.78 -0.69 6.80 4.78 0 3.55 23.07
TC: original value [m] 20 0 28 20 0 15 17
Inclination [%] 0.04 -0.006 0.056 0.04 0 0.03 0.17
RPI [-] 10 10 10 10 10 10 10
Complexified

Terrain Indicator [-] 0.85 0.85 0.85 0.85 0.85 0.85 0.85
PI: target value [-] 14.78 9.31 16.80 14.78 10.00 13.55 33.07
PI: final value [-] 13.17 4.97 16.38 12.62 4.58 10.79 43.63
TC: target value [m] 62.50 42.50 70.50 62.50 42.50 57.50 25.50
TC: final value [m] 62.30 42.71 71.20 62.67 42.38 57.02 25.54
RPI [-] 50 50 50 50 50 50 50
Complexified

Terrain Indicator [-] 3.81 3.81 3.81 3.81 3.81 3.81 3.81
PI: target value 54.78 49.31 56.80 54.78 50.00 53.55 73.07
PI: final value [-] 61.48 50.52 62.67 60.76 54.32 60.14 69.34
TC: target value [m] 210.50 190.50 218.50 210.50 190.50 205.50 55.10
TC: final value [m] 212.38 190.48 216.94 210.80 190.28 205.47 55.65

© BHR Group 2013 Multiphase 16 279

 Table 2: Dummy pipeline complexification: overall characteristics
Complexified profile Complexified profile
Original profile
(RPI = 10) (RPI = 50)
Pipeline Indicator (PI) [-] 3.85 11.49 58.67
Total Climb (TC) [m] 100 363.82 1,282
Total length [m] 310,000.03 310,001.06 310,015.42
Number of sections 310 1,240 1,240
Total distance [m] 310,000 310,000 310,000

Dummy pipeline: angle distribution

Total pipe length per angle group [m]

Complexified profile (RPI = 50) Complexified profile (RPI = 10) Original profile
250,000

200,000

150,000

100,000

50,000

Pipe angle group [deg]

Dummy pipeline: elevation profile
Complexified profile (RPI = 50) Complexified profile (RPI = 10) Original profile
0
Elevation [m]

-20
-40
-60
-80
-100
-120
0 50 100 150 200 250 300 350
Distance [km]
Dummy pipeline: Total Climb profile
Complexified profile (RPI = 50)Complexified profile (RPI = 10) Original profile
1500
Total Climb [m]

1000

500

0
0 50 100 150 200 250 300 350
Distance [km]
Dummy pipeline: inclination profile
Complexified profile (RPI = 50) Complexified profile (RPI = 10) Original profile
2
Inclination [deg]

1
0
-1
-2
0 50 100 150 200 250 300 350
Distance [km]

Figure 11: Dummy pipeline complexification: angle distributions and profiles

280 © BHR Group 2013 Multiphase 16

 Dry gas-condensate flow Dry gas-condensate flow
Condensate accumulation vs. total mass flow rate Pressure drop vs. total mass flow rate
20,000 100
Condensate accumulation [m3] 18,000
RPI = 50
90
RPI = 50
RPI = 10 RPI = 10
16,000 80
Original profile Original profile

Pressure drop [bar]

14,000 70
12,000 60
10,000 50
8,000 40
6,000 30
4,000 20
2,000 10
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Total mass flow rate [kg/s] Total mass flow rate [kg/s]

Wet gas-condensate flow Wet gas-condensate flow

Liquid accumulation vs. total mass flow rate Pressure drop vs. total mass flow rate
20,000 140
RPI = 50 RPI = 50
18,000
Liquid accumulation [m3]

RPI = 10 120 RPI = 10

16,000 Original profile
Original profile
Pressure drop [bar]
14,000 100
12,000 80
10,000
8,000 60
6,000 40
4,000
20
2,000
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Total mass flow rate [kg/s] Total mass flow rate [kg/s]

Two-phase oil-water flow Two-phase oil-water flow

Water accumulation vs. total mass flow rate Pressure drop vs. total mass flow rate
30,000 100
RPI = 50 RPI = 50
90
Water accumulation [m3]

25,000 RPI = 10 RPI = 10

80
Original profile Original profile
Pressure drop [bar]

20,000 70
60
15,000 50
40
10,000
30
5,000 20
10
0 0
0 50 100 150 200 0 50 100 150 200
Total mass flow rate [kg/s] Total mass flow rate [kg/s]

Three-phase oil-water-gas flow Three-phase oil-water-gas flow

Water accumulation vs. total mass flow rate Pressure drop vs. total mass flow rate
40,000 160
RPI = 50 RPI = 50
38,000 140
Water accumulation [m3]

RPI = 10 RPI = 10
36,000
Pressure drop [bar]

Original profile 120 Original profile

34,000
32,000 100
30,000 80
28,000
60
26,000
40
24,000
22,000 20
20,000 0
0 50 100 150 200 0 50 100 150 200
Total mass flow rate [kg/s] Total mass flow rate [kg/s]

Figure 12: Dummy pipeline complexification: steady-state simulation results

© BHR Group 2013 Multiphase 16 281

For each of the multiphase flow systems described in Section 2, a set of parametric
steady-state analyses was performed with a commercial multiphase flow simulator to
assess the sensitivity of both liquid holdup (condensate, aqueous phase, or both) and
pressure drop to the total mass flow rate. A fixed inner diameter of ca. 18 in. (assuming a
nominal pipe diameter of 20 in.) was assumed for every run. The simulation results are
presented in Figure 12. As expected an increase in profile roughness leads to more liquid
accumulation (mainly in gas systems) and higher pressure drops. Particularly:

 Note the presence of gradient discontinuities (4 in total) on the condensate

accumulation curve for the original profile with a dry two-phase fluid (top left).
These discontinuities are caused by the simplistic angle distribution of the original
dummy profile which comprises 4 different positive angles only (see Table 1). The
combination of this oversimplified distribution and the spectacular sensitivity of dry
two-phase flows to (positive) pipe inclinations (see Section 2) explain this outcome.
To a lesser extent gradient discontinuities can be observed for the other fluids as
well.
 A dramatic increase in pressure drop is noticed for the three-phase oil system when
the RPI is increased from 10 to 50 (bottom right). A detailed analysis of the
simulation results suggests that this abrupt increase is caused by a change in flow
regime from stratified to slug flow in a large portion of the pipeline, which is
conceivable for complexified profiles with broader angles distributions (see Figure
11). Furthermore this change of flow regime explains the reduction in water
accumulation resulting from an increased mixing and reduced slip between the
phases when the roughness of the profile is increased (bottom left).

6 CONCLUSION

The methodologies presented in this paper for the discretisation, characterisation, and
complexification of pipeline elevation profiles are relevant at every stage of field
developments. Complexification is recommended in the early stages of engineering
design where the potential routes for production pipelines are commonly assessed from
limited bathymetric data. Discretisation may be required in the last phases of design if
detailed data from geophysical surveys of the seabed are available. Finally, non-
dimensional indicators can be used to: quantify the roughness of the seabed and detailed
pipeline elevation profiles; anticipate the severity of phenomena such as liquid build-up
and slugging; and improve the interpretation of multiphase flow simulations.

7 REFERENCES

(1) Zakarian, E., Holm, H., and Larrey, D. (2009). “Discretization Methods for
Multiphase Flow Simulation of Ultra-Long Gas-Condensate Pipelines”, 14th
International Conference on Multiphase Production Technology, BHR Group,
Cannes, France, 16-19 June, 2009.
(2) Langsholt, M. and Holm, H. (2007). “Liquid accumulation in gas-condensate
pipelines – an experimental study”, 13th International Conference on Multiphase
Production Technology, BHR Group, Edinburgh, UK, 13-15 June, 2007.
(3) Barrau, B. (2000). “Profile indicator helps predict pipeline holdup, slugging”,
Oil & Gas Journal, Vol. 98, Issue 8, p. 58-62, Feb 21, 2000.
(4) Jackson, D. (2008). “Filtering Elevation Profile Data To Improve Performance
of Multiphase Pipeline Simulations”, CIPC/SPE Gas Technology Symposium
2008 Joint Conference, 16-19 June 2008, Calgary, Alberta, Canada.

282 © BHR Group 2013 Multiphase 16