23rd World Gas Conference, Amsterdam 2006

A DIRECT ASSESSMENT MODULE FOR PIPELINE INTEGRITY

MANAGEMENT AT GASUNIE

Main author
M.T. van Os

N.V. Nederlandse Gasunie

Gasunie Engineering & Technology
Groningen, The Netherlands
 ABSTRACT
This paper describes the development of a Direct Assessment (DA) module for unpiggable
pipelines for the Gasunie Pipeline Integrity Management software system PIMSlider. This Direct
Assessment module is currently under construction.

The use of External Corrosion Direct Assessment (ECDA), as proposed by the NACE, has
been recognized by Gasunie as a promising method to assess the threat of corrosion to the integrity of
unpiggable pipelines. In addition, Structural Reliability Analysis (SRA) in combination with Bayesian
statistics enables quantification of the results from Direct Assessments. The heart of the method has
been developed over the past five years by Advantica and AFAA. Through an intensive collaboration
between Gasunie and AFAA within the past two years, the method has been modified and expanded
further to fit the needs of Gasunie.

With the expanded method, it is possible to automatically adjust the number of coating defects
and the number of corrosion defects based on the results from two aboveground surveys; one for the
detection of coating defects and one to assess possible corrosion activity.

Furthermore, the method uses the results from excavations to automatically adjust the
following variables:
• The probability of detection and the probability of false indication of each survey technique
• The time of initiation of corrosion defects and the defect density
• The corrosion rate and the defect depth

Based on these updates, the failure probability of the pipeline under investigation can be
calculated. This updating process can be performed after each excavation if required, until the
probability of failure is sufficiently low. This will minimize the total number of excavations and therefore
results in a substantial reduction of cost.

The developed method will be implemented in the Pipeline Integrity Management system
PIMSlider in the course of 2006. The DA module will become an integral part of the system, which will
give the operator the benefits of easy access to relevant data from the pipeline under investigation, as
well as to data (e.g. Direct Assessment results) from other pipelines.
TABLE OF CONTENTS
Abstract

1. Historical background

2. The PIMSlider system

Slider (including Archives and Arclib)
CP Expert
GDLI
Inpipe
Rehab Expert
PSL
Risk Expert

3. The Direct Assessment module

4. Failure Modes
Classification of Failure Modes
Important failure modes at Gasunie
Modelling of Corrosion

5. Pre-assessment
Data collection and visualization
Identification of ECDA regions
The prior condition of the pipeline

6. Indirect Inspections

7. Direct Examination and Post-Assessment

Direct examination
Model
Other ECDA regions
Post-Assessment

8. Conclusions

9. Acknowledgements

References

List of Tables
Table 1: Overview of failure modes
Table 2: Relevant ECDA data elements for Gasunie

List of Figures
Figure 1: Examples from PIMSlider; geographical position of a pipeline (left) and pipeline
related and operational data (right)
Figure 2: An example from the PSL module; the relative societal risk
Figure 3: A typical dGPS/DCVG/CIPS survey team in action for Gasunie
Figure 4: An example of the number of coating defects before and after the surveys and after
excavations
Figure 5: The effect of an ECDA on the probability of failure of a pipeline
Figure 6: Flowchart of the most important calculations that are carried out in the different
ECDA steps.
1. HISTORICAL BACKGROUND
Gasunie owns ± 12.000 km of high-pressure pipeline in the Netherlands. The major part of this
grid has been constructed in the period 1960-1980. Aging of the grid results, to an increasing extent,
in spots with coating degeneration and reduced wall thickness, caused by corrosion as well as
mechanical damage. On top of this, Microbiologically Influenced Corrosion (MIC) was detected on the
high-pressure grid in 1999. This strengthened the opinion of the Asset Management department that
reconsideration of the prevailing policy on pipeline management was required in order to maintain the
high standard on risk- and integrity control within Gasunie. As a result, it was concluded that the policy
on pipeline management had to change from verification of preventive measures to verification of the
actual condition of the pipelines. This new ambition resulted in the following strategies:

1. Integrity Management: To comply with prescribed governmental requirements and

restrictions of the integrity of the assets and to prioritize, to
perform preventive activities and to monitor the actual condition of
pipelines
2. Risk Management: To realize and preserve environmental risk of the pipeline within
acceptable (or agreed) level and to prioritize and perform
mitigating activities

Consequently the In-Line Inspection (ILI) program was intensified (5-10 ILI-runs per year),
which of course resulted in a significant increase of the amount of inspection data to analyze.
Simultaneously a new IT solution was implemented to enable efficient and reliable data processing
and to support all processes on pipeline integrity management (PIMSlider).

Since only 50% of the pipelines of Gasunie are conventionally piggable, it was decided in
2005 to develop a computerized Direct Assessment module in PIMSlider which enables the integrity
analyses of unpiggable pipelines, meeting the requirements of the ECDA process described in NACE
RP0502-2002.1 This paper focuses on the Direct Assessment module of PIMSlider, which is currently
being developed by Gasunie in corporation with AFAA and ATP Neftegazsystema.

2. THE PIMSLIDER SYSTEM

2
The PIMSlider system consists of a number of modules, of which the heart is formed by
Slider. The modules cover the whole spectrum of data management (pipeline-, environmental- and
incident data), CP system monitoring data, analyses of ILI data, defect assessments and quantitative
risk calculations with consideration of the economics involved. A brief description of the modules is
given in the following paragraphs.

Slider (including Archives and Arclib)

This module can be seen as the heart of the system and is used for storage of all pipeline-
related data concerning the position of the pipeline, equipment, crossings, operational data, ILI data,
maps, photographs, population density along the pipeline etc. It is mainly used for information
retrieval. The operator can track the relationships between various figures, as illustrated in Figure 1,
and schedule actions accordingly (surveys, repair, maintenance etc).

CP Expert

This module enables the operator to analyze the effectiveness and the efficiency of an existing
CP system. A modeling function supports the CP engineer in the design of the CP system in case of
construction or modification of a pipeline. CP Expert utilizes data from Slider. It also allows calculation
of the optimum operation mode for CP stations, to ensure reliable and effective protection of the
pipeline.
 Figure 1: Examples from PIMSlider; geographical position of a pipeline (left) and pipeline related and
operational data (right).

GDLI

All the pipeline incidents that have occurred on the Gasunie grid in the past have been stored
in the GDLI database. The GDLI module is designed for the analysis and visualization of these
incidents.

Inpipe

Inpipe enables the analysis of any kind of pipeline defect and other features based on the data
provided by ILI tools. This involves the linking of the features to map coordinates and an accurate
positioning of the in-line data along a 3-dimensional model of the pipeline. The software supports the
calculation of the remaining strength of the pipeline using the methods ASME B31G and RSTRENG.

Figure 2: An example from the PSL module; the relative societal risk

Rehab Expert

This module enables the operator to assess the significance of defects in the pipeline and to
define the most appropriate repair program. Defects can be assessed by the use of defect-geometry
data as reported by the ILI contractor, or by the use of the raw data from the inspection tool (such as
individual sensor signals). When more than one ILI has been performed, the same defect can be
compared at different stages of its lifetime. This enables the operator to optimize the economics of his
inspection and repair program.

PSL

This module is the core element of the system with respect to risk management of gas
transmission pipelines. It is a hazard- and risk-assessment package, which enables automatic
quantitative risk calculations to be made at any moment for any pipeline in the Slider database. In
addition, it enables the engineer to calculate the effect of risk mitigating measures on an existing
pipeline. PSL is based on approaches and assumptions used in PIPESAFE, a risk-assessment model
for gas transmission pipelines that was developed by a group of international gas transmission
companies.

Risk Expert

This module, the ranking tool for operational pipelines, enables the operator to carry out a
relative risk assessment of the pipeline. It is a tool for prioritization of maintenance and inspection
programs. This data-based method uses a model that identifies and quantifies the major threats and
consequences of pipeline objects and the pipeline environment. The likelihood of all threats is
quantified through the use of operational experience, opinions of subject-matter experts or on industry
experience. The calculations are performed for all pipeline sections, defined here as parts of the
pipeline with unchanged conditions. This makes it possible to identify local high-risk areas.

3. THE DIRECT ASSESSMENT MODULE

The DA-module of PIMSlider that is currently being developed enables computerized storage,
retrieval and processing of all appropriate pipeline data, which is stored in the PIMSlider database,
and therefore guarantees highly accurate, reproducible and time saving integrity analyses of the whole
Gasunie grid.
1
The module is based on the NACE Recommended Practice for ECDA in combination with
SRA. The ECDA process integrates information on the pipeline’s physical characteristics including
operating history (pre-assessment) with data from multiple field examinations (indirect inspections)
and pipe surface evaluations (direct examinations). SRA in combination with Bayesian statistics allows
one to quantify the effect of inspections and exavations on the integrity level of the pipeline and, as a
3-7
consequence, supports the integrity manager in the definition of the required inspection program.
The increase in reliability that can be achieved by application of SRA and Bayesian statistics, can
result in substantial savings on cost inspection cost.

The failure mode External Corrosion is modeled in detail. Other failure modes are represented
8
by a constant. The model as developed by Gasunie originally in 2004 has been improved significantly
since, through ongoing research at Gasunie Engineering & Technology and a highly valuable
9
contribution from Andrew Francis. The module consists of the following parts, which will each be
explained further onwards in this paper:
• Pre-Assessment
• Indirect Inspections
• Direct Examination and Post-Assessment

The Pre-Assessment includes data collection and visualization, the identification of so-called
ECDA regions and the calculation of the a priori probability of failure of the pipeline. The objective of
the Indirect Inspections step is to identify and define the severity of coating faults, other anomalies and
areas at which corrosion activity might be occurring. The Direct Examinations are used to determine
which indications from the indirect inspections are most severe, to collect data to assess corrosion
activity and to repair critical defects. Finally, the objectives of the Post-Assessment are to define
reassessment intervals and to assess the overall effectiveness of the ECDA process. In this step all
ECDA-regions are integrated and the failure frequencies of other failure modes are included.

4. FAILURE MODES
When the integrity management of a pipeline is considered, all the potential threats to this
pipeline (failure modes) have to be considered. For the Direct Assessment module the scheme of the
ASME is used. In the DA module, the integrity of a pipeline is evaluated by assessing the failure rate,
expressed as the probability of failure per km per year. An overview of the different failure modes with
the highest failure rates for Gasunie is given below.

The Direct Assessment of external corrosion (ECDA) is modeled in detail. When models
become available for other relevant failure modes, these can be added to the DA module in the future
if required. For now, the contributions of other failure modes to the overall integrity of the pipeline are
modeled in a simple, straightforward manner, namely as a constant.

Classification of Failure Modes

A first step in managing the pipeline integrity is to identify the potential existing threats. All
threats to the integrity of the pipeline need consideration. The PRCI (Pipeline Research Committee
International) has analyzed and classified a large number of gas pipeline incidents in twenty-two
10
different causes. Subsequently, the ASME has grouped these causes in nine different categories.
The following table contains a list with the different failure modes, in which details have been omitted.

Main category Subcategory Failure mode

External corrosion
Time-Dependent Internal corrosion
Stress Corrosion Cracking (SCC)
- Defective pipe seam
Manufacturing related defects
- Defective pipe
Time-Independent
- Defective girth weld
Internal cause Welding/fabrication related
- Defective fabrication weld (fittings a.o.)
Equipment
- Immediate failure
Third party damage
- Previously damaged pipe
Time-Independent Incorrect operations
External cause - Cold weather
(incl. Operations) - Lightning
Weather related and outside force
- Heavy rains or floods
- Ground movements
Table 1: Overview of failure modes

In the PIMSlider/DA-module, a division in failure modes to the level of the subcategories

mentioned in the table above is used.

Important failure modes at Gasunie

All incidents that have occurred in the past on the Gasunie grid have been recorded in the
Gasunie Database for Pipeline Incidents. The incidents include leaks and ruptures, but also coating
damage and damage caused by corrosion.

The main failure modes are:

• External interference (third party) 71%
• External corrosion 6%
• Material and welding defects 4% (mostly welding defects)
• Ground movement (sinking) 2%

Therefore it can be concluded that for Gasunie the most important failure modes are external
corrosion and external interference. External interference and its consequences are already treated in
detail by the PIMSlider/PSL module. External corrosion will be modeled in the PIMSlider/DA module.

Since the failure mode external corrosion will become more important due to the ageing of the
pipeline grid and the growth of corrosion defects with time, it is important to model this failure mode.
11-13
When looking at other studies (by EGIG, UKOPA or PRCI), it appears that for other
pipeline operators material and welding defects and internal corrosion are more important failure
modes than for Gasunie.

With respect to external corrosion, Gasunie holds the opinion that external corrosion should
not contribute significantly to the total incident frequency. At present, a criterion for the maximum
failure rate has yet to be determined.
Modeling of Corrosion

For External Corrosion Direct Assessment (ECDA) a NACE-standard is present and a

complete model has been worked out, based on the work of Francis et al. This will be implemented in
the DA-module of PIMSlider.

For other corrosion Direct Assessments, such as Internal Corrosion (ICDA) and Stress
Corrosion Cracking (SCCDA), standards have been developed by the NACE, but no comprehensive
models are available as yet. The outlines of these standards are largely comparable to that of External
Corrosion, but inclusion of these forms of corrosion into PIMSlider will be postponed until
corresponding probabilistic models have been developed. Both of these mentioned failure modes are
not considered to be relevant threats to the integrity of the Gasunie grid at present.

A variety of causes is known to lead to external corrosion. When a coating defect is present
(however small) and the Cathodic Protection applied is not sufficient (or even too high), corrosion
reactions may occur under influence of stray-currents, AC-interference, bacteria (MIC), tensile stress
(SCC), shielding of CP or overprotection. MIC, SCC, shielding of CP and overprotection are not
covered by ECDA. Since Microbiologically Influenced Corrosion (MIC) is very often associated with
disbonded coating (i.e. shielding of CP), ECDA should not be applied in areas where MIC is known to
occur.

Other forms such as general corrosion, AC corrosion or stray current corrosion are in principle
covered by ECDA. Finding MIC using ECDA techniques may not be ruled out completely, since
experience with MIC at Gasunie has turned out that this form of corrosion often occurs in combination
with (or leads to) degenerated coating. This implies that an initially shielded defect suffering from MIC
may eventually become accessible to CP currents due to degeneration of the coating, thereby
enabling aboveground detection of the defect.

5. PRE-ASSESSMENT
In this section the Pre-Assessment step will be described in general. In addition to this it will
be discussed which data are considered to be relevant for Gasunie to implement in PIMSlider and how
this can be done. Finally, it will be described how the relevant data can be used to quantify the
integrity of a pipeline, using the structural reliability based approach developed by Francis et al.

There are essentially three purposes of the Pre-Assessment step, namely:

• Data collection and visualization
• To identify ECDA regions
• To establish the prior condition of the pipeline

Data collection and visualization

Consistent with the NACE, this first part of the Pre-Assessment requires a sufficient amount of
data collection, integration and analysis. All parameters that impact the selection of the Indirect
Inspection tools and the definition of the ECDA regions shall be considered for initial ECDA
applications. Specifically for the Gasunie grid, the elements listed in the table below have been
identified as essential information in order to assess the prior condition of the pipeline.
Category Data elements
Pipe-related - Material and grade
- Diameter
- Wall thickness
Construction-related - Year installed
- Year of route change/modification
- Locations of valves, clamps, supports, taps, couplings, insulating joints etc.
- Locations and material of casings
- Depth of cover
- Underwater sections; river crossings
- Locations of river weights and anchors
Soils/environmental - Soil characteristics/types, including
• Type of soil (e.g. presence of peat, anaerobic soils)
• Electrical influence (e.g. proximity of DC railways, AC power lines)
• Pollution
• Drainage
• Land use (e.g. paved roads)
Corrosion control - Locations of rectifiers
- Test point locations
- Type of coating
- CP survey data/history
Operational data - Pipe inspection reports, excavation reports
- Repair history/record (e.g. repair sleeves, repair locations)
- Data from previous aboveground surveys
- Inline inspection data
Table 2: Relevant ECDA data elements for Gasunie

Identification of ECDA regions

In the second part of the Pre-Assessment, the gathered information is used to define a
number of ECDA regions. An ECDA region consists of one (or more) section(s) of the pipeline with
similar physical characteristics, corrosion histories, expected future corrosion conditions and where the
same indirect inspection tools can be used. After the user has defined the parameters that are to be
considered for identification of ECDA regions, the regions are calculated by the DA-module. These
ECDA regions can then assist the user throughout the Direct Assessment process to interpret results,
to decide which indirect inspection tools can be used and where direct examinations should be
performed.

The prior condition of the pipeline

In the third part of the Pre-Assessment, the prior condition of the pipeline under investigation is
quantified, by assessing the failure probability of each ECDA region. In order to calculate this
probability of failure, estimations must be made for, among other things, the following key parameters:
• Time of initiation of corrosion defects; during the time period between the year of
construction and the time of initiation, the conditions for corrosion are being established
and the first instances of early corrosion growth take place.
• Defect density (both for coating and corrosion defects); this firstly consists of a starting
value representing the damages originating from the transportation and construction
phases of the pipe(line). Secondly, this consists of a term representing the rate of
introduction of new defects, starting from the time of initiation.
• Defect depth; a certain initial distribution for the defect depth is assumed at the time of
initiation, after which it gradually increases, depending on the corrosion rate.
• Corrosion rate; the rate at which the defect depth grows. This is in general a major cause
of uncertainty and is likely to vary considerably between pipelines.

These estimations will be based on the information that was collected during the first part of
the Pre-Assessment. Data regarding factors such as the age of the pipeline, coating type, level of CP,
soil conditions etc are appropriately combined to determine prior distributions of the densities of
coating and corrosion defects and prior distributions defining the geometry of corrosion defects. If
relevant information is not available for a specific pipeline, the user can fall back on the complete
database of all pipelines for the required analysis. This functionality (automatic retrieval of data from
pipelines with similar specifications or environmental conditions) results in a huge increase of
accessible data (now and especially in future) and makes it possible to reduce the cost of inspections
substantially by applying statistics. In practice, parameters concerning the geometry of the pipeline
(e.g. wall thickness) or material properties (e.g. flow stress) are also associated with uncertainties,
especially in the case of older pipelines. In the DA-module, these quantities are treated as probability
density functions as well, rather than using constant values.

In addition to this, the probability that a single defect will fail, depending on the age of the
pipeline and the uncertainties described earlier, can be calculated by the model. To predict the failure
pressure of part-wall corrosion defects in pipelines, guidelines from a Joint Industry Project, the
14
Linepipe Corrosion Group Sponsored Project, have been used, developed through a combination of
analysis and full scale testing. By combining the probability of failure for a single defect with the a
priori defect density, the probability of failure for each ECDA region can be calculated.

6. INDIRECT INSPECTIONS
The objective of the Indirect Inspection step is to identify and define the severity of coating
faults, other anomalies, and areas at which corrosion activity may have occurred or may be occurring.
The NACE requires the use of at least two aboveground inspections over the entire length of each
ECDA region. Within Gasunie, indirect inspections for ECDA purposes usually consist of the following
measurements:
1. Direct Current Voltage Gradient (DCVG) survey: to detect and pinpoint coating defects
along the pipeline
2. Close Interval Potential Survey (CIPS): to measure the on- and off-potentials over the
pipeline, as well as the on- and off-potential gradients to remote earth. These
measurements are used to determine whether a possible coating defect is sufficiently
protected by the Cathodic Protection (CP) system
3. Wenner measurements: to measure the soil resistivity at regular intervals along the route
4. dGPS-measurements: to measure the position of coating defect indications, soil resistivity
measurements and characteristic features along the pipeline

Figure 3 shows a photograph of a typical survey team. In this approach, CIPS is not used as a
method to detect coating defects, but to assess corrosion activity at possible coating anomalies
detected by DCVG. For this reason, CIPS is treated as a corrosion survey rather than a coating
survey. However, complete derivations have also been made by A. Francis for the methods to
determine the number of coating defects based on the results from a single coating survey or two
coating surveys.

Since no inspection technique is 100% reliable, it is important to take the performance of the
used inspection tools into account, possibly depending on factors such as type of coating, soil
conditions, depth of cover etc. This performance is characterized by the following two variables:
• Probability of Detection (PoD): the probability that a defect present is detected by the
survey technique
• Probability of False Indication (PoFI): the probability that the survey technique gives an
indication where no actual defect is present
 Figure 3: A typical dGPS/DCVG/CIPS survey team in action for Gasunie

The Indirect Inspection step can be divided in the following four parts:
1. Data storage: the data from the aboveground surveys are stored in a database for
processing and for future reference.
2. Data processing, interpretation and visualization: if required, the data from the surveys can
be corrected for factors such as the depth of cover or currents through the pipeline.
Furthermore, the DA module calculates the so-called IR-free potential (the potential of the
steel at the exact point where the surface of the steel meets the surrounding environment,
not distorted by the soil resistance between reference electrode and the pipeline) and its
uncertainty. The DA module presents the required data in a clear and comprehensive
manner by plotting (combinations of) graphs on the screen, to simplify interpretation of
complex data by the operator.
3. Generation of a priority list for direct examinations (excavations): This list will be generated
based on certain criteria for parameters chosen by the operator, such as coating defect
size, expected corrosion, soil resistivity etc or combinations thereof.
4. To establish the condition of the pipeline after the indirect inspections: based on the
findings from the aboveground surveys, the distributions of the numbers of coating and
corrosion defects are updated using Bayesian statistics. Obviously, the performance
characteristics of the applied survey techniques play an important role in these
calculations. The prior distributions of the PoD and the PoFI of a survey technique can be
constructed from previous experience with the technique or from recommendations from
the manufacturer.

7. DIRECT EXAMINATION AND POST-ASSESSMENT

Direct examination

In the Direct Examination step all ECDA-regions that have been inspected are evaluated
sequentially, starting with the ECDA-region with the highest risk according to the Pre-Assessment
step. This step is set up such that after each number of excavations the pipeline integrity can be
evaluated. Based on the outcome of that evaluation, additional excavations may need to be carried
out, after which the pipeline integrity must be evaluated again.

Based on the priority list from the two aboveground surveys, locations for bell hole excavations
are selected. In general excavations will take place at locations:
• Where both surveys (coating and corrosion) have given an indication; usually to determine
the size of corrosion defects and repair critical defects (if any)
• Where only one of the surveys has given an indication; usually to check the survey
characteristics and to determine the size of any found corrosion defects
 • Where no indications were given; these excavations are generally referred to as blind
digs, and may be used to assess the confidence in the probability of detection of the
survey technique(s).

Based on information found at the excavations, the information used in the Pre-Assessment
and the Indirect Inspection step will be updated. This includes an updating of the survey
characteristics, the number of defects, the corrosion rate, the time of initiation, the defect length and
the critical defect depth. Based on these updates, new values are calculated for the probability of
failure of a defect and per km of the ECDA-region.

The following information has to be gathered from the Bell Hole excavations: location of the
coating and (active) corrosion defects, dimensions of the corrosion defects (defect depth, defect length
in axial direction) and the clock position of the corrosion defects (not relevant to the model at present).
Every excavation should be included, also the excavations where no or only small defects were found.

Model

Model calculations have to be performed for all ECDA-regions. In the Direct Examination step
the following groups of calculations (updates) are made for each ECDA-region:
• The survey characteristics
• The number of defects (coating and corrosion) as a function of time
• The corrosion rate and the time of initiation
• The defect length and the critical defect depth
• The probability of failure (of a defect and per km of the ECDA-region)

To illustrate the updating process, an example is given in Figure 4 for the number of coating
defects, represented by a normal distribution. The prior estimate can be based on information
collected during the Pre-Assessment step (e.g. coating condition, age of the pipeline), but may also be
deducted from previous surveys on other, similar pipelines. As soon as new information becomes
available from the Indirect Inspections, the prior distribution is updated using Bayesian statistics. The
excavations will allow an update of the survey characteristics, after which the number of coating
defects can be updated once more.

0,30

0,25
Prior
Probability

0,20 estimate
Updated after
0,15
surveys
0,10 Updated after
excavations
0,05

0,00
0 5 10 15 20 25 30 35 40

Number of coating defects per km

Figure 4: An example of the number of coating defects before

and after the surveys and after excavations

Figure 5 shows the effect of an ECDA on the probability of failure of a pipeline. Firstly, the
probability of failure increases due to the growth of corrosion defects and the increasing number of
defects. Secondly, the calculated probability of failure decreases as a result of the Indirect Inspections
and Direct Examinations. Thirdly, the probability of failure will increase again due to ongoing corrosion
processes.
 1,0E+00

Probability of failure [km ]

-1
1,0E-02
Updated after
1,0E-04 aboveground surveys

1,0E-06
Updated after
1,0E-08 excavations
1,0E-10
ECDA
1,0E-12 applied

1,0E-14
0 5 10 15 20 25 30 35 40 45 50 55 60

Number of years since commissioning

Figure 5: The effect of an ECDA on the probability of failure of a pipeline

For ECDA it is assumed that all found corrosion defects are repaired (or recoated). The repair
of coating defects will not have a significant effect on the ECDA-calculations, but for consistency in the
ECDA calculations it is assumed that all coating defects that are found, are repaired.

The updated probability of failure of a defect at the time of the excavations depends upon the
updated value of the distribution of the defect depth, the distribution of the critical defect depth and the
expected number of corrosion defects. The updated probability of failure/km at the time of excavations
depends on the updated distribution of failure of a defect and the length of the ECDA-region.

The probability of failure for a single defect will change with each excavation due to change in
the distribution of the corrosion rate and the time of initiation. For the probability of failure a confidence
interval will be calculated. As the number of excavations increases, the probability of failure/km will
normally decrease, until the criterion for the probability of failure is met. In that case sufficient
excavations for the specific ECDA-region have been carried out.

Other ECDA regions

The ECDA-regions are treated sequentially, starting with the region with the highest initial risk.
The results of the calculations for an ECDA-region can be used as a starting point for the next ECDA-
region. The initial values for the corrosion rate, the time of initiation and the defect density for that
ECDA-region can be replaced by the user with the values calculated for the ECDA-region with the
higher initial risk. This is repeated until all ECDA-regions have been covered.

Post-Assessment

According to the NACE, the objectives of the Post-Assessment are to define reassessment
intervals and to assess the overall effectiveness of the ECDA process. For the developed SRA model,
this step consists of the following calculations:
• Calculate the probability of failure/km due to external corrosion for each ECDA-region for
future years, based on the Indirect Inspections and Direct Examinations
• Calculate the probability of failure/km for each ECDA-region for future years for all other
failure modes considered
• Depending on the calculated probability of failure/km, calculate the time interval until the
next Direct Assessment is required.

When new aboveground surveys are carried out on the same pipeline, the pipeline will be
subject to a new Direct Assessment. Obviously, the results of a Direct Assessment are available for
use in the Pre-Assessment step of the next Direct Assessment for that pipeline.

Finally, Figure 6 gives a summary of the most important routines that the operator will go
through when using the DA module for PIMSlider, in the form of a flowchart. Only the steps relevant to
the model are included in this figure.
Pre-
Pre-Assessment
Pipe data, Defect data,
Soil data

Construct prior Construct prior Calculate prior

distributions defect distribution distribution critical
dimensions number of defects defect depth

Calculate prior probability of

failure

Indirect Inspections Aboveground

surveys

Survey Number of defect

characteristics indications

Updated
distribution
number of defects

Direct Examination & Post Assessment

Examination(s ) Results from Examination(s)

Update survey
characteristics
Update distribution defect
and critical defect depth
Update distribution
number of defects

Calculate probability of failure

no
Probability of failure low enough?
yes
Calculate future probability of
failure and re-assessment interval

Figure 6: Flowchart of the most important calculations that are

carried out in the different ECDA steps.
8. CONCLUSIONS

For unpiggable pipelines, the use of ECDA as proposed by the NACE has been recognized by
Gasunie as a valuable method to assess the threat of corrosion. SRA in combination with Bayesian
statistics makes it possible to quantify the results obtained during an ECDA process. The main part of
the method has been developed by Advantica and AFAA. This method has now been modified and
expanded further by Gasunie in cooperation with AFAA.

With the expanded method it is possible to automatically adjust the number of coating defects
and the number of corrosion defects based on the results of two aboveground surveys (one for the
detection of coating defects and one for the detection of corrosion defects).

The method makes it possible to use the results of excavations to adjust the following
variables:
• The probability of detection and the probability of false indication of each survey technique
• The time of initiation of corrosion defects and the defect density
• The corrosion rate and the defect depth

Based on these updates, the probability of failure of the pipeline under investigation (per km)
can be calculated. This updating process can be performed after each number of excavations, until
the probability of failure is sufficiently low. This allows the integrity manager to minimize the number of
excavations required.

The developed method will be implemented in the Pipeline Integrity Management system
PIMSlider in the course of 2006. It will be an integral part of the system, which will bring the benefits of
easy access to all relevant data of the pipeline under investigation and to all the results of Direct
Assessments that were performed on other pipelines.

9. ACKNOWLEDGEMENTS
The authors greatly acknowledge Andrew Francis from AFAA, for the many valuable
contributions he has made to this project over the past two years. Wytze Sloterdijk, Giorgio
Achterbosch, Rob Bos, Kees Dijkstra and Henk Horstink, all from Gasunie, are kindly acknowledged
for their input throughout the duration of the project. The company ATP Neftegazsystema is
acknowledged for their critical questions and feedback during the process of developing the functional
specification of the DA module for PIMSlider.

REFERENCES

1. NACE Standard Recommended Practice RP0502-2002 (2002). Pipeline External Corrosion Direct
Assessment Methodology, Houston, Texas.
2. Bos, R. (2005). The PIMS Implementation Program for Sasol’s Gas Operations, Journal of
Pipeline Integrity, Quarter 1.
3. Francis, A., Edwards, A.M., Espiner, R.J. and Senior, G. (1999). Applying Structural Reliability
Methods to Ageing Pipelines, Paper C571/011/99, IMechE Conference on Ageing Pipelines,
Newcastle, UK.
4. Francis, A. and McCallum, M.A. (2003). Integrity Management Using Direct Assessment and
Structural Reliability Analysis, Workshop at Gasunie Research, Groningen, The Netherlands.
5. Andrew Francis & Associates Limited (2004). A Robust Methodology for External Corrosion Direct
Assessment, Report AFAA-R0003-04.
6. Andrew Francis & Associates Limited (2004). A Robust Methodology for External Corrosion Direct
Assessment-Supplement, Report AFAA-R0004-04.
7. Francis A., Harris, J., McCallum, M.A., McQueen, M., Sansom, A. and Ward, C.R. (2005).
Structural Reliability Analysis for ECDA, Prepared for Gas Research Institute, GRI-04/0093.2.
8. Os, M.T. van, Mastrigt, P. van, Horstink, G.H., Achterbosch, G.G.J., Stallenberg, G.A.J. and Dam,
A.M. (2005). A Structural Reliability Based Assessment of Non-Piggable Pipelines, Paper 05151,
NACE International, Corrosion 2005, Houston, Texas.
9. Andrew Francis & Associates Limited (2005). A Detailed Methodology for External Corrosion
Direct Assessment Taking Account of Two Above Ground Surveys, AFAA-R0013-05.
10. ASME B31.8S (2001). Managing System Integrity of Gas Pipelines, Supplement to ASME B31.8.
th
11. EGIG (2002). Gas Pipeline Incidents (1970-2001), 5 Report of the European Gas Pipeline
Incident Data Group, EGIG 02.R.0058.
12. UKOPA (2002). Pipeline Product Loss Incidents (1961-2000), 2nd Report of the UKOPA Fault
Database Management Group, Advantica Report R 4798.
13. PRCI (2002). Analysis of DOT Reportable Incidents for Gas Transmission and Gathering System
Pipelines, 1985 through 1997, Report PR-219-9801.
14. Batte, A.D., Fu, B., Kirkwood, M.G. and Vu, D. (1997). New Methods for Determining the
th
Remaining Strength of Corroded Pipelines, 16 International Conference on Offshore Mechanics and
Arctic Engineering, Yokohama, Japan.