Thermal insulation material for subsea pipelines b

See discussions, stats, and author profiles for this publication at: />
Thermal Insulation Material for Subsea Pipelines: Benefits of Instrumented
Full-Scale Testing To Predict the Long-Term Thermomechanical Behaviour
Article · January 2007
DOI: 10.4043/18679-MS

CITATIONS

READS

3

694

6 authors, including:
Nadège Bouchonneau

Valérie Sauvant-Moynot

Federal University of Pernambuco

IFP Energies nouvelles

17 PUBLICATIONS   168 CITATIONS   

45 PUBLICATIONS   857 CITATIONS   

SEE PROFILE

Francois Grosjean
IFP Energies nouvelles

46 PUBLICATIONS   289 CITATIONS   
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Finite Element modeling and Analysis of corroded pipelines View project

PhD Thesis View project

All content following this page was uploaded by Francois Grosjean on 07 September 2015.

The user has requested enhancement of the downloaded file.

SEE PROFILE


OTC 18679
Thermal Insulation Material for Subsea Pipelines: Benefits of Instrumented Full-Scale
Testing To Predict the Long-Term Thermomechanical Behaviour

N. Bouchonneau, Ifremer, IFP , Franche-Comté U.; V. Sauvant-Moynot and F. Grosjean, IFP; D. Choqueuse, Ifremer; and
E. Poncet and D. Perreux, Franche-Comté U.
Copyright 2007, Offshore Technology Conference
This paper was prepared for presentation at the 2007 Offshore Technology Conference held in
Houston, Texas, U.S.A., 30 April–3 May 2007.
This paper was selected for presentation by an OTC Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Offshore Technology Conference and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Offshore Technology Conference, its officers, or members. Papers presented at

OTC are subject to publication review by Sponsor Society Committees of the Offshore
Technology Conference. Electronic reproduction, distribution, or storage of any part of this
paper for commercial purposes without the written consent of the Offshore Technology
Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not
more than 300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, OTC, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
External coating systems of flowlines and risers ensure both
structural and thermal insulation functions which should be
efficient throughout the design life in-service, typically 25
years. In that context, the long term behaviour of thermal
insulation materials is difficult to predict due to the coupled
effects of three factors: hydrostatic pressure up to 300 bar,
thermal gradient over 120°C between internal effluents and
external sea water and the water absorption of constitutive
materials. In addition, laboratory data collected on small size
specimens of insulation materials are normally used to predict
the thermo-mechanical behaviour of full scale systems, but
laboratory testing simply do not properly simulate the service
conditions, in particular the complex loading existing through
the coating thickness. This paper covers the background to the
development of both test facilities and models to study the
thermo-mechanical behaviour of production coated steel pipe
in ultra deep water conditions. This original work was
launched to provide both experimental and computed data to
better understand and predict the thermo-mechanical
behaviour of insulation materials whilst considered as a full
scale system. On the one hand, experimental data obtained on

instrumented insulated pipes immersed in large scale facilities
simulating ultra deep water are presented in both steady and
transient states. On the other hand, a finite element model
dedicated to the abovementioned insulated pipes was
developed to predict their thermo-mechanical behaviour.
Correlation between full scale experimental data and related
model predictions are discussed to validate the predictive
model taking into account the coupling between hydrostatic
pressure and temperature gradient. Additional modelling
developments to include the water absorption are planned to
reach a suitable prediction of the whole service life.

Introduction
Optimistic estimations of oil reserves in deep water and
current oil & gas prices sustain the increasing interest towards
offshore deepwater field production. The ultra deep water
(3000m water depth) is one of the next issues. Indeed, 4% of
the world offshore surface with WD>1500m includes
sedimentary areas with hydrocarbon potential (minimum
sediment thickness of 2000m) [1]. Those ultra-deepwater
fields, between 100 to 500 [1], are expected to be located in
the Gulf of Mexico, in the Atlantic off Brazil, Nigeria and
Angola, and also near Aegypta in the Nil delta. It is worth
noting that the hydrocarbon reserves identified and to be
identified in both onshore and conventional offshore
sedimentary basins represent 19% of the world surface. In
comparison to onshore and conventional offshore
hydrocarbons, the partial exploitation of ultra-deep reserves,
about 1% of the world surface, would correspond to 30 billion
to 100 billion of barrels equivalent petrol [1]. As a

consequence, the ultra-deep offshore production representing
10% of the offshore production in 2005 is expected to grow to
25% in 2025 [2].
In that context, flow assurance continues to be a critical part of
system design and operations, with lower seabed temperatures
- typically in the 1 to 4°C range at 1500m-3000m depth - and
rising insulation costs in deepwater [3]. Among others, the
heat management in normal (steady state) and dynamic
(transient) operations relies on the selection of proper
insulation materials and designs for subsea flowlines and
risers to meet the increasing demand for deeper waters. Pipein-pipe configurations are under study to optimize their
performance but their heavy weight may be a limitation.
Advanced insulation materials and coatings are also being
developed and designed for subsea use to offer both
appropriate thermal and mechanical properties in ultra-deep
water applications [4, 5]. Moving towards ultra-deep water
applications also emphasises the need to have test methods
and facilities that establish whether a given coating system is
fit-for-purpose - either new insulation materials/systems or
existing materials/systems to be subjected to conditions for
which there are no available data. Full-scale thermal testing
protocoles and facilities have been developed since the late
1980's to study the behaviour of thermal insulation coating
systems on lengths of pipe under simulated service conditions
[6, 7]. Such full-scale tests are becoming part of the


2

qualification testing since laboratory testing on small size

specimen of insulation materials simply do not properly
simulate the service conditions, in particular the complex
loading existing through the coating thickness. The
development and validation of a test protocol and
instrumentation which will permit the study of the thermal
behavior of a coated insulation pipe when subjected to
conditions (300 bar) simulating ultra-deep water (nearly
3000m water depth) is the first issue of this paper.
In parallel, the modelling of insulation systems at a global
level is a necessary basis to address their long term behaviour
in service. In particular, water absorption effects have to be
considered at design stage for syntactic foams which are
widely used as rigid and flexible insulation materials [8, 9].
Indeed the long-term response of syntactic foams while
exposed to seawater at high pressure is highly non-linear,
particularly at elevated temperatures, revealing the occurrence
of complex degradation mechanisms leading to glass
microsphere filling [10-14]. The modelling of water
absorption for syntactic foams has been recently performed on
small size specimens and validated on a large data base (4
syntactic materials with 4 geometries and aged under 18
conditions from 4°C, 1bar up to 130°C, 300 bar) [15]. But the
long term behaviour of thermal insulated structures is difficult
to predict due to the coupled effects of three combined factors:
ƒ hydrostatic pressure up to 300 bar, inducing a
stress gradient within the coating material,
ƒ thermal gradient over 120°C between internal
effluents and external sea water,
ƒ the water absorption of constitutive materials.
The development of a finite element model to satisfactorily

predict the thermo-mechanical behavior of a production
insulated steel pipe in ultra deep water conditions (transient
and steady states) is the second issue of this paper.
The study work presented in this paper is an expansion of the
work given in [16] and was part of a PhD study.
Experimental set-up
Insulated pipe configuration
Both experimental and computational developments
achieved in this work were dedicated to 1.2m length steel pipe
sections industrially coated with an externally applied fivemultilayer insulation system mainly based on syntactic
polypropylen (PP) material. Test pipe sections were taken
from a length of pipe with internal diameter of 180mm, coated
under normal production conditions. The geometry and
composition of the 61mm thick insulation system are
summarised in Figure 1. The outer diameter is 338mm.
Simulated service test
Pressure vessel
The pressure vessel used in this work (1m diameter and
2m height) is located in Ifremer. Both water temperature and
pressure (up to 1000bar) are monitored and regulated from the
external data acquisition unit. The pressure is monitored using
a pressure transducer mounted at the top of the pressure
vessel. A schematic view of the coated test pipe in vertical

OTC 18679

position during testing in the pressure vessel is given in the
Figure 2. The pressure vessel flange allows the connection of
the inner instrumentation to the outer data acquisition unit
where readings obtained from all sensors are recorded.

Equipment of coated test pipe
The design of the prototype equipment is a crucial point in
the experimental set up of full-scale tests. The insulated pipe
section was machined at both ends to adapt two metallic steel
caps (Stainless steel APX4) equipped with connectors that
resist external pressure. Three 10-channels connectors were
necessary to allow the electrical supply of the inner heating
system and to collect inner sensor data. Steel caps were
covered by 100mm thick Polytetrafluoroethylene (PTFE)
insulating caps to limit the axial heat flow as much as
possible. The equipment of the insulation prototype is shown
in progress on Figure 3 and the schematic representation of the
fully instrumented pipe section is given in Figure 4.
Instrumentation is detailed hereafter.
Heating system
An original heating system was developed in this study
instead of the classical circulating oil to limit the convection
effects inside of the pipe. The heating system consisting of
heating elements (NiCr) embedded in a thin silicone layer was
placed on the internal diameter of the steel pipe and kept in
place by brushes. It is worth noting that such an inner 'dry'
configuration with no pressure and no liquid is also very
beneficial for the instrumentation used inside the pipe and will
simplify the modelling of the inner heat flux boundary
condition. In addition, the electric power applied to the system
during steady state phase will provide an indirect monitoring
of the radial heating flux through the coating.
Temperature sensors
The insulated pipe section was instrumented with six
commercial temperature sensors (Pt100) specified up to 200°C

minimum (precision of about 0.3 % at 100 °C), located in both
inner and outer parts along the pipe length and on the caps
(Figure 4):
- Ti (°C): inner temperature of the steel surface in the
center of the pipe (one measurement);
- Te (°C): outer temperature of the coating surface in
the center of the pipe (one measurement);
- Tb (°C): inner temperature of the steel surface in the
center of one cap (one measurement);
- T100 (°C): inner temperature of the steel surface along
the pipe 100mm distant from cap (one measurement);
- T50 (°C): inner temperature of the steel surface along
the pipe 50mm distant from cap (one measurement);
- TPTFE (°C): outer temperature of the PTFE surface in
the center of the cap (one measurement).
Besides, the outer temperature of the water in the vessel,
Twater (°C), was also measured using a platinium sensor.
Heat flux sensors
The insulated pipe section was also instrumented with four
commercial heat flux sensors located in both inner and outer
parts along the pipe length and the caps:
- φi (W.m-2): inner thermal flux density on the steel
surface in the center of the pipe (one measurement


OTC 18679

with soft circular fluxmeter of 5µV.m2.W-1 sensitivity
specified up to 200°C);
- φe (W.m-2): outer thermal flux density on the coating

surface in the center of the pipe (three measurements
with a semi-rigid fluxmeter of 50µV.m2.W-1
sensitivity specified up to 100°C and 100bar, one
measure with a rigid fluxmeter of 30µV.m2.W-1
sensitivity specified up to 250°C and 150bar);
- φcap (W.m-2): inner thermal flux density on the steel
surface in the center of the steel cap (one
measurement with rigid rectangular fluxmeter of
36µV.m2.W-1 sensitivity specified up to 200°C).
It should be pointed out that heat flux sensors with soft flat
form were selected for the internal and external pipe surfaces
to reduce errors related to the difficulties in mounting rigid flat
sensors.
Test programs
Two cases were considered both considering a length of
steel pipe coated with an externally applied multilayer
insulation material as described previously:
1. No external pressure was applied in the pressure vessel
(external pressure was 1 bar).
2. Hydrostatic pressure simulating in-service conditions in
ultra-deep water was applied to the test pipe section.
The first case was performed as a basis of comparison for the
numerical simulation to the experimental results. The second
case was conducted because it was the main objective to
evaluate the thermal performance of the insulation coating
system in steady and transient conditions while immersed in
seawater and subjected both to a temperature gradient across
its full thickness and a hydrostatic pressure identical to those
which it would experience when in-service in ultra-deep
water.

Both test programs are shown in Figure 5 and Figure 6.
Each step of the testing program of case 2 simulating ultradeep water immersion is described hereafter. A preliminary
test of the instrumented insulated pipe section referred to as
step 1 was performed in fresh water under ambient
temperature and pressure prior to the simulated service test.
This first step was necessary to check that the prototype
equipment and related instrumentation had been properly
installed. In step 2 (about 30 minutes duration), the outer
pressure was increased to 300bar to simulate in-service on the
seabed at 3000m depth. Heating power values of 120W and
240W were applied in steps 3 and 4 to reach representative
thermal gradients through the coating thickness, respectively
around 50°C and 110°C, given the outer water temperature
thermoregulated around 15°C, the OHTC of the insulation
coating under consideration and its surface area. The duration
of each step 2, 3 and 4 was typically around 3 days. On
completion of the test, the heating circuits were switched off
allowing the internal temperature of the coated test pipe to
decrease down to temperature of the sea water surrounding the
pipe in the pressure vessel (step 5). When the response of the
test pipe to the removal of the temperature gradient had
stabilised, the hydrostatic pressure in the pressure vessel was
rapidly reduced to atmospheric (step 6).

3

In both cases, it should be emphasised that a test lasting
approximately 10 days cannot be used to predict the long term
evolution of the insulation coating systems for which water
uptake and creep cannot be neglected.

Numerical model of simulated service test
Numerical model
A two-dimensional axisymetric numerical model of the
insulated pipe section was developed using finite element
multiphysics commercial software where mechanical and
thermal aspects are coupled under the following assumptions:
- Coating and pipe materials are simulated as solids with linear
elastic behaviour (no creep).
- The thermal conduction of coating and pipe materials is
simulated using Fourier's law.
- The natural convection between the coating surface and the
external water is simulated using equations derived from
Newton's law. The convection coefficients were calculated
from experimental temperatures Te and Twater according to
[17].
Material properties and boundary conditions
Thermal and mechanical properties of each constitutive
material of the test section are reported, respectively, in Table
1 and Table 2. Values collected from experimental
measurements performed at 1 bar on small samples [15] or
from the literature were used as input data in the simulation.
Boundary conditions used were conduction along the inner
surface of the steel pipe, insulation at both inner end caps and
convection along the external surfaces in contact with water.
For tests under hydrostatic pressure, a stress condition is
applied on the external surfaces of the structure. The
displacements perpendicular to the symmetry conditions are
also blocked. The initial conditions (temperature or pressure)
depend on the different sequences of the tests performed on
the structure.

The geometry of the computational domain, the numerical
mesh and the boundary conditions are shown in Figure 7. The
mesh was locally refined near interfaces and sensors to
enhance the resolution.
Applications of the numerical model
Simulation at design stage
At the conception stage of the pipe equipment, the thermomechanical simulation of the insulated pipe section performed
in steady and transient states on the basis of material
geometries and physical properties can provide:
- the insulation coating thickness shrinkage / swelling
under increasing pressure / temperature along the length of the
insulated pipe immersed in water under pressure;
- the external surface temperatures and heat flux
distributions along the insulated pipe section.
Numerical results help for the design of end caps for example
but such simulations remain only indicative since physical
properties used as input data are obtained through material
testing under laboratory conditions which does not reflect full
scale conditions.
Investigation of insulation properties
During the testing stage of one insulation pipe section
under 1 bar (case 1), simulations were performed under


4

OTC 18679

transient and steady state conditions. The comparison between
experimental and simulation results contributed to validation

of the model and to assessing the prototype instrumentation.
During the testing stage of one insulation pipe section
under 300bar (case 2), there were no direct measurements of
the radial heat fluxes since commercial sensors are limited to a
lower pressure range. Hence the thermo-mechanical model
was used to determine the heat transfer coefficient and the
material properties from the experimental data.
Steady state conditions.
The radial heat flux was determined by simulation in order
to fit the experimental temperature distribution. Then the heat
transfer coefficient and the apparent thermal conductivity of
the insulation material (apparent meaning that the thermal
conductivity was averaged on the insulation cross section
under thermal gradient) were calculated using classical
analytical expressions.
The analytical expression for the radial heat flux under steady
state conditions for a one-dimensional conduction problem in
a composite cylinder, evaluated on the inner surface of the
structure taken as reference, is given by:
(1)
Q = -U.S. (Text - Tint)
In a case of a multilayer structure, and by assuming that the
thermal contact resistance between each layer can be
neglected, the heat transfer coefficient U of the structure can
be expressed in terms of constitutive material thermal
conductivities with the following relation:
1
U=
⎡ ⎛ Di +1 ⎞ ⎤
(2)

⎟⎟ ⎥
ln⎜
n ⎢ ⎜
⎢ ⎝ Di ⎠ ⎥
S×
⎢ 2πLλi ⎥
i =1
⎢
⎥
⎣⎢
⎦⎥

∑

With these hypotheses, and according to the characteristics
given by manufacturers, the overall heat transfer coefficient
also defined as "U value" is 4.2W.m-2.K-1 at 20°C. This
coefficient is a representative thermal characteristic of the
entire system: steel pipe and insulation coating.
Transient state conditions.
Both thermal conductivity and heat capacity can be
determined directly by an optimisation program developed
with a commercial software. This routine allowed the
evolution of the inner temperature of the steel pipe Ti to be
adjusted during the testing time simulated by an analytical
relation to the experimental data by optimisation (square root
method) of the thermal parameters of the material studied,
here the syntactic foam.
In the case of a one-dimensional radial transfer in a onelayer structure limited by radii r=rint and r=rext and of length l,
the heat equations (temperature and heat flux) are:

1 ∂ ⎛ ∂T ⎞ 1 ∂T
for rint < r < rext
⎜r
⎟=
r ∂r ⎝ ∂r ⎠ a ∂t
with T=T0 for t=0, and

(3)

Φ = −λS

∂T
∂r

(4)

Applying a Laplace transform to the variable t, these equations
lead to:

d 2 θ 1 dθ p
+
− θ
dr 2 r dr a

(5)

and

φ = −λS

∂θ
∂r

(6)

The quadrupoles notation [18] is well suited to relate the
Laplace transforms of the temperatures and fluxes at inner and
external boundaries obtained by solving the above equations:

⎡θ int ⎤ ⎡ A B ⎤ ⎡θ ext ⎤
⎢φ ⎥ = ⎢C D⎥ ⎢φ ⎥
⎦ ⎣ ext ⎦
⎣ int ⎦ ⎣

(7)

θint et θext correspond to the transforms of the inner and outer
surface temperatures of the cylindrical structure respectively,
and A, B, C and D are analytical relations involving Bessel
functions and the geometrical characteristics of the structure.
This development is applied to the multilayered structure
under test (6 layers including steel pipe), submitted to outer
convective losses. Equation (7) becomes:
1 ⎤
⎡
6
⎡θ int ⎤
⎡ Ai Bi ⎤ ⎢1
⎥ ⎡ θ water ⎤

(8)
=
h
⎥
⎢φ ⎥
⎢C D ⎥ ⎢
ext Sext ⎥ ⎢
i⎦ 0
⎣φconvective ⎦
⎣ int ⎦ i =1 ⎣ i
1
⎣
⎦
with θwater and φconvective the Laplace transforms of the water
temperature and the convective heat flux respectively.

∏

Temperatures and heat fluxes evolutions are obtained in
the time space by means of the numerical inversion of each
Laplace transform. Transient model hypotheses are:
- One-dimensionnal axisymmetric.
- Constant convective heat transfer coefficient and
water temperature .
- Constant heat flux equal to the value of the steady
state heat flux measured in the structure
(experimental values at 1 bar and simulated values
for the tests under 300 bar).
- Initial temperature of the entire structure supposed to
be stabilized at the water temperature;

- The external surface area is constant (no thermal
expansion and no pressure effect).
Results
Main experimental and computational results are summarised
in this section relative for both cases under consideration.
Case 1- 1bar
Steady state
Experimental heat fluxes measured in the steady state
under 1bar and 120W, then 1bar and 240W, are reported in the
Table 3. One can notice that inner heat fluxes are smaller than
the heating mat power values, due to thermal losses at both
ends and possibly in the inside of the pipe. But the particularly


OTC 18679

low value measured under 240W (even smaller than the
external heat flux) reveals that the soft heat flux sensor used at
the interface between the heating mat and the steel pipe does
not sustain the high temperature environment which it is
subjected to.
Three methods are proposed to determine the heat transfer
coefficient U.
Method A.
The heat transfer coefficient U is determined using
equation (1) directly from the experimental external heat
fluxes and temperatures measured. The precision of the U
value is 3.1% based on the sensor measurement errors given
by manufacturers. The apparent thermal conductivity values of
the syntactic PP were then derived from equation (2). Results

reported in the Table 3 show that U values and apparent
thermal conductivities are not significantly influenced by the
thermal gradients (differences are within incertainty ranges).
U values measured are smaller than the one given by the
insulated pipe manufacturer and apparent thermal
conductivities of syntactic PP are also slightly lower than
values measured on small specimens (Table 1).
Method B.
In the absence of any external heat flux sensor, the external
heat flux could be approached to first approximation by
substracting the heat losses measured at both caps from the
power of the heating mat. Values of 115W and 226W were
obtained for, respectively, 120 W and 240W heating power.
Internal heat flux densities were calculated from the
aforementioned heat fluxes by dividing flux values by 0.57m2
(internal heating mat surface area). The simulation of thermomechanical behaviour of the insulated pipe section was
performed with those calculated internal flux densities. The
comparison of experimental and simulated temperature values
are presented in the Figure 8. Inner and external temperature
measurements agree quite well but one can note that other
simulated temperature are overestimated, suggesting that the
internal heat flux density was also overestimated due to
underestimated heat losses at both ends. This method would
be improved by a better estimation of end losses.
Method C.
The internal heat flux is optimised to fit test temperatures
with the numerical simulation results. Results obtained with
the simulation and experimental temperatures are presented
Figure 9. The temperature distribution simulated within
insulated pipe section is shown in Figure 10. Calculated U

values and apparent thermal conductivities of syntactic PP are
given in the Table 3. Thermal properties estimated by method
C are comparative to those given by method A, which
validates the use of the thermo-mechanical numerical
simulation to determine the OHTC and the apparent thermal
conductivity of insulation materials.
Transient state
Since the analysis of experimental data with the transient
model approach requires that initial temperatures are fully
stabilised, the evolution with time of the inner temperature
was simulated only for the 120W thermal gradient. The
experimental values and simulated curve are compared on
Figure 11. Optimisation input data and results are reported in
the Table 4. The very similar apparent thermal conductivity

5

values obtained compared to the steady state approaches
(methods A and C) validate the transient state analysis. The
apparent heat capacity of syntactic PP is lower than the value
determined experimentally, but this later value should be
considered with caution since heat capacity values are
extremely difficult to measure.
Case 2- 300bar
Steady state
Experimental heat fluxes measured in the steady state
under 300bar and 120W, then 300bar and 240W, are reported
in the Table 3. No external heat fluxes were measured in
simulated ultra deep water conditions, therefore U value and
apparent thermal conductivity of syntactic PP cannot be

determined using method A. Therefore, method C was used to
evaluate the heat transfer coefficient of the structure.
The internal heat flux used in the thermomechanical
modelling was optimised to fit the temperatures more closely.
Results obtained with the numerical simulation agree quite
well with the experimental temperatures (Figure 12). U values
and apparent thermal conductivities of syntactic PP calculated
using analytical expressions (1) and (2) are reported in the
Table 3. One can observe that values obtained under 300bar
120W are similar to those obtained under 1 bar. But this is no
longer true for 240W experiments. The significant increase of
the U value (+10%) and related increase of the apparent
thermal conductivity could be explained by damage occurring
in the foam microstructure, in particular in the vicinity of the
pipe where material is subjected to coupled effect of high
temperature and complex stress distribution.
Transient state
The evolution with time of the inner temperature was also
simulated during the establishment of the temperature
distribution with a heating power of 120W. The experimental
values and simulated curve are compared on Figure 13.
Optimisation of input data and results, reported in the Table 4,
led to very similar apparent thermal conductivity value
compared to the steady state approach and validate once again
the transient state analysis. From apparent thermal
conductivity and heat capacity values obtained under 120W
1bar and 120W 300bar, there is no significant difference as
stated previously from steady state analysis. Thus, one can
conclude that the coupling effect of pressure and thermal
gradient induces no short term consequences on syntactic PP

provided the inner effluents have a temperature around 60°C.
But in the case of temperatures of 80°C and above coupled
with ultra deep service pressure, short term phenomena may
occur in the syntactic PP leading to a lowering of insulation
performance.
Conclusion
Very demanding in service conditions in ultra deep water
require specific test conditions and experimental equipment to
perform full scale test. Tests have been performed on insulated
structures under service conditions (P=300bar, Tint=95°C). An
original heating system was developed instead of the classical
circulating oil to limit the convection effects inside of the pipe
and simplify the boundary conditions of heat flux modelling.
Novel instrumentation has been developped in order to
monitor different test parameters (internal and external


6

temperatures, heat fluxes, …). These tests allow pertinent
results to be obtained.
When there is no available external heat flux sensor, one
effective way to determine the OHTC and the insulation
material thermal properties is to perform numerical
simulations and fit the temperature distributions in both steady
and transient states. Satisfactory agreements between the twodimensional numerical simulation results including thermal
and mechanical coupling and tests results given by
conventional instrumentation were obtained at 1bar.
Numerical simulations, on the other hand, may be used to
guide the design of an insulated flowline test systems.

In the near future, the water diffusion into the insulation
material will be taken into account in order to predict the long
term insulation behaviour.
Nomenclature
U = heat transfer coefficient of the structure relative to a
reference surface [W.m-2.K-1].
S = inner surface area, expressed as S=πLD1 [m²].
Sext = external surface area [m²].
Text = external surface temperature in steady state
conditions [°C].
Tint = internal surface temperature in steady state
conditions [°C].
Di = inner diameter of the layer i of the structure [m].
Di+1 = external diameter of the layer i of the structure [m].
D1 = inner diameter of the steel pipe [m].
L = steel pipe length [m].
λi = thermal conductivity of the layer i [W.m-1.K-1].
hext = convective heat transfer coefficient at the interface
between insulation coating and water [W.m-2.K-1].
a = thermal diffusivity [m2.s-1].
T0 = initial temperature [°C].
T= temperature [K].
Acknowledgments
The authors wish to thank Socotherm for providing the
insulated coated pipes, in particular G. P. Guidetti for his
interest to this work and N. Lacotte and A. Deuff for the
performing of hyperbaric tests.
References
1. MATHIEU, Y., IFP technical note, October 2006.
2. ROBERTSON, S., MACFARLAN, G., et SMITH, M., "Deep

water expenditures to reach $20 billion/year by 2010", Offshore
Magazine, 2005.
3. McMULLEN N.D., "Flow-Assurance Field Solutions", Offshore
Technology Conference - OTC 18381, Houston, Texas U. S. A.,
1-4 May 2006.
4. BOYE HANSEN A., JACKSON A., 'High performance
polypropylene thermal insulation for high temperature and deep
water applications", 16th International Conference on Pipeline
Protection, Paphos, Cyprus, 2-4 November 2005.
5. BERTI, E., "Syntactic polypropylene coating solution provides
thermal insulation for Bonga risers", Offshore Magazine, 2004.
6. HALDANE D., GRAAF F.v.d. et LANKHORST A.M., "A direct
measurement system to obtain the thermal conductivity of
pipeline insulation coating systems under simulated service
conditions", Offshore Technology Conference - OTC 11040,
Houston, Texas U. S. A., 3-6 May 1999.

OTC 18679

7. MELVE B., RYDIN C. et BOYE HANSEN A., "Long term
testing of high temperature thermal insulation for subsea
flowlines at simulated seabed conditions", 15th International
Conference on Pipeline Protection, Aachen, Germany, 29-31
October 2003.
8. DAVALATH J., "Cool-down thermal performance of subsea
systems based on Gulf of Mexico Field Experience", Offshore
Technology Conference - OTC 17972, Houston, Texas U. S. A.,
1-4 May 2006.
9. CHALUMEAU A., FELIX-HENRY A., "Water absorption effect
on syntactic foam thermal insulation of a flexible pipe", 25th

International Conference on Offshore Mechanics and Arctic
Engineering (OMAE), Hamburg, Germany, 4-9 June 2006.
10. CHOQUEUSE D., CHOMARD A. et BUCHERIE C.,
"Insulation materials for ultra deep sea flow assurance:
Evaluation of the material properties", Offshore Technology
Conference - OTC 14115, Houston, Texas (U.S.A.), 6-9 May
2002.
11. CHOQUEUSE D., CHOMARD A. et CHAUCHOT P., "How to
provide relevant data for the prediction of long term behavior of
insulation materials under hot/wet conditions ?", Offshore
Technology Conference - OTC 16503, Houston, Texas U.S.A.,36 May 2004.
12. GIMENEZ N., SAUVANT-MOYNOT V. et SAUTEREAU H.,
"Wet ageing of syntactic foams under high pressure / high
temperature in deionized and artificial sea water", 24th
International Conference on Offshore Mechanics and Artic
Engineering, Halkidiki, Greece, 12-17 June 2005.
13. HALDANE D., SCRIMSHAW K.H., "Development of an
alternative approach to the testing of thermal insulation
materials for subsea applications", 14th International Conference
on Pipeline Protection, Barcelona, Spain, 29-31 October 2001.
14. SAUVANT-MOYNOT V., GIMENEZ N. et SAUTEREAU H.,
"Hydrolytic ageing of syntactic foams for thermal insulation in
deep water: degradation mechanisms and water uptake model",
Journal of Materials Science, 2006, 41 (13), p. 4047-4054.
15. LEFÈBVRE X., SAUVANT-MOYNOT V., CHOQUEUSE D. et
CHAUCHOT P., "Durabilité des matériaux syntactiques
d'isolation thermique et de flottabilité: des mécanismes de
dégradation à la modélisation des propriétés long terme",
Matériaux 2006, Dijon, France, 13-17 November 2006.
16. BOUCHONNEAU N. et al., "Multilayer systems for thermal

insulation: thermomechanical behaviour of prototypes for deep
sea applications", Oilfield Engineering with Polymers, 29-31
March 2006.
17. EYGLUNENT B., "Manuel de thermique - Théorie et pratique";
HERMES Science Publications, Paris, 1997.
18. MAILLET D., ANDRÉ S., BATSALE J.-C., DEGIOVANNI A.
and MOYNE C., “Thermal quadrupoles : solving the heat
equation through integral transforms”; John Wiley & Sons, Inc.,
2000.


Table 1-Thermal properties

Thermal conductivity a)
Heat capacity a)
(W.m-1.K-1)
(J.kg-1.K-1)
Steel pipe
45
475
Fusion bonded Epoxy
0.3
2000
Adhesive PP
0.22
2090
PP
0.22
2000
Syntactic PP

0.165 + 10-4 * T
1506.6 + 6.26 * T
Steel (cap) - APX4
19
460
PTFE (insulating end cap)
0.24
1050
a) Values are given at 20°C when the temperature dependence is not specified
Material

Table 2- Mechanical and thermomechanical properties

Expansion coefficient
Material
(between 10 and 100 °C)
(°C-1)
Steel pipe
7850
218
0.33
1×10-5
Fusion bonded Epoxy
1200
3
0.4
5.3×10-5
Adhesive PP
900
1.3

0.4
1.6×10-4
PP
900
1.3
0.4
1.6×10-4
Syntactic PP
640
E = -0.94 * 10-3 T + 1.1
0.32
5×10-5
Steel (cap) - APX4
7700
211
0.33
1×10-5
PTFE (insulating end cap)
2200
0.4
0 .46
1.3×10-4
a) Values are given at 20°C when the temperature dependence is not specified
Density
(kg.m-3)

Elastic modulus a)
(GPa)

Poisson

coefficient

Table 3- Heat fluxes and temperatures in the steady state and related thermal properties (simulation results are in italics)

Pressure
(bar)

Heating
mat power
(W)

1
1
1
1
1
1
300
300

120
120
120
240
240
240
120
240

Method

Internal
radial heat
flux
(W)

External
axial heat
flux
(W)

External
radial
heat flux
(W)

Ti
(°C)

Te
(°C)

U
(W.m-2.K-1)

A
B
C
A
B

C
C
C

99.7
114.6
109
177
226.3
223.7
97.5
195

5.4
13.7
7
15.5

89
96.6
91.9
178
191
188.8
84.3
172

56.4
56.4
56.4

95.8
95.8
95.8
55.6
88.5

16.4
16.4
16.4
17.6
17.6
17.6
18.3
20

3.88±0.12
4.21±0.13
4.00±0.12
3.97±0.12
4.26±0.13
4.21±0.13
3.94±0.12
4.37±0.13

Apparent
thermal
conductivity
of syntactic PP
(W.m-1.K-1)
0.152±0.005

0.166±0.006
0.157±0.005
0.155±0.006
0.168±0.006
0.166±0.006
0.154±0.005
0.173±0.006

Table 4- Thermal properties of syntactic PP determined by the transient state analysis

Pressure
(bar)

Heating
mat power
(W))

External radial
heat flux in
steady state
(W)

Water
temperature
(°C)

Mean external
convection
coefficient
(W.m-2.K-1)

1
300

120
120

89
84.3

15.3
16.7

125
170

Apparent
thermal
conductivity of
syntactic PP
(W.m-1.K-1)
0.150
0.154

Apparent
heat
capacity of
syntactic PP
(J.kg-1.K-1)
1510

1519


8

OTC 18679

PP (2,5 mm)
Syntactic PP (55 mm)

PP (3 mm)
Adhesive PP (0,25 mm)
Fusion bounded Epoxy (0,25 mm)
Steel pipe wall (18,26 mm)
Figure 1-Section of the 5-multilayer insulation coating on steel pipe (thickness values are given in brackets)

Electric connection to the data
acquisition unit

Pressure vessel
Pmax = 1000 bar
Ø=1m
h=2m

Coated pipe section

Figure 2-Schematic representation of the insulated pipe section during test in the pressure vessel.

1

5

4

2

3

Figure 3-Instrumentation in progress of the insulated pipe section
(1) PP multilayer insulation coating, (2) Steel cap with connector, (3) PTFE cap, (4) Heating mat, (5) Outer flux meter.


OTC 18679

9

Te

T50
Tb
φb

TPTFE

T100

Ti

φe

Heating
system
power input

φi

Tb

Figure 4-Configuration of the fully instrumented test section

Heating power (W)
240

120

0
Time
1

3

2

4

Figure 5-Testing program under 1 bar pressure

Heating power (W)

Pressure (bar)

240
300
120

1

0
Time
1

2

3

4

5

6

Figure 6-Testing program simulating ultra-deep water


10

OTC 18679

z

Displacement
(r) blocked
Thermal
convection :
Teau, h

Thermal
insulation

Hydrostatic
pressure
Thermal
heat flow

O
Displacement (z)
blocked

r

Figure 7-Boundary conditions and mesh of insulated pipe section

(a)

Ti

110

Te

100

Tb

Temperature (°C)

90

T50

80

T100

70

TPTFE

60

Ti-simulation

50

Te-simulation

40

Tb-simulation

30

T50-simulation

20

T100-simulation
TPTFE-simulation

10
0

50000

100000

150000

200000

Time (s)

(b)

110

Ti

100

Te

Temperature (°C)

90

Tb

80

T50

70

T100

60

Ti-simulation

50

Te-simulation

40

Tb-simulation

30

T50-simulation

20

T100-simulation

10
0

40000

80000

120000

160000

Time (s)

Figure 8-Comparison between experimental and simulated temperatures during test at 1 bar
(a) 120W, (b) 240W (method B)


OTC 18679

11

(a)

110

Ti

100

Te

Temperature (°C)

90

Tb

80

T50

70

T100

60

Ti-simulation

50

Te-simulation

40

Tb-simulation

30

T50-simulation

20

T100-simulation

10
0

50000

100000

150000

200000

Time (s)

Temperature (°C)

(b)

110
100

90
80
70
60
50
40
30
20
10

Ti
Te
Tb
T50
T100
Ti-simulation
Te-simulation
Tb-simulation
T50-simulation
T100-simulation
0

40000

80000

120000

160000

Time (s)

Figure 9-Comparison between experimental and simulated temperatures during test at 1 bar
(a) 120W, (b) 240W (method C)

Figure 10-Temperature distribution within insulated pipe section in the steady state during test at 1 bar 120W


12

OTC 18679

100

Temperature (°C)

90

Ti experience
proto2 1bar
120W

80
70
60
50

Ti matlab
optimisation
proto2 1bar

120W

40
30
20
10
0

50000

100000

150000

200000

Time (s)

Figure 11-Comparison between experimental and simulated temperatures with transient model during test at 1 bar

(a)
110

Ti

100

Te

Temperature (°C)

90

Tb

80

T50

70

T100

60

Ti-simulation

50

Te-simulation

40

Tb-simulation

30

T50-simulation

20

T100-simulation

10
0

50000

100000

150000

200000

250000

Time (s)

(b)
110

Ti

100

Te

Temperature (°C)

90

Tb

80

T50

70

T100

60

Ti-simulation

50

Te-simulation

40

Tb-simulation

30

T50-simulation

20

T100-simulation

10
0

60000

120000

180000

240000

Time (s)

Figure 12-Comparison between experimental and simulated temperatures during test at 300 bar
(a) 120W, (b) 240W (method C)


OTC 18679

13

Temperature (°C)

110
Ti experience
proto2 300bar
120W

90

70
50

Ti matlab
optimisation
proto2 300bar
120W

30
10
0

50000

100000

150000

200000

250000

Time (s)
Figure 13-Comparison between experimental and simulated temperatures with transient model during test at 300 bar

View publication stats