GAS PIPELINE INCIDENTS

11th Report of the European Gas Pipeline Incident Data Group

(period 1970 – 2019)

Comprising:

Gas Networks Ireland (Ireland)

DGC (Denmark)
ENAGAS, S.A. (Spain)
EUSTREAM (Slovak Republic)
Fluxys (Belgium)
Gasgrid Finland (Finland)
GRTgaz (France)
National Grid (UK)
Gasunie (Netherlands / Germany)
NET4GAS (Czech Republic)
Gasconnect (Austria)
Open Grid Europe (Germany)
REN Gasodutos S.A. (Portugal)
Snam Rete Gas (Italy)
Swedegas A.B. (Sweden)
SWISSGAS (Switzerland)
Teréga (France)

Doc. number VA 20.0432

17TH of December 2020
 Copyright © December 2020 by
European Gas Pipeline Incident Data Group (EGIG):

Gas Networks Ireland (Ireland)

DGC (Denmark)1
ENAGAS, S.A. (Spain)
EUSTREAM (Slovak Republic)
Fluxys (Belgium)
Gasgrid Finland (Finland)
GRTgaz (France)
National Grid (UK)2
NET4GAS (Czech Republic)
Gasunie (Netherlands / Germany)
Gasconnect GmbH (Austria)
Open Grid Europe (Germany)
REN Gasodutos S.A. (Portugal)
Snam Rete Gas (Italy)
Swedegas A.B. (Sweden)
SWISSGAS (Switzerland)3
Teréga (France)

All rights reserved. No part of this publication may be reproduced, stored in any retrieval
system or transmitted in any or by any means, electronically, mechanical photocopying,
recording or otherwise, without prior written permission from the copyright owners. In
case of any citation, this document must be mentioned as the document of origin.
Unlicensed multiple copying of the contents of this publication is illegal. Inquiries should
be addressed to the Secretariat of EGIG.

Comments or questions to this publication can be directed to the EGIG secretariat:

N.V. Nederlandse Gasunie N.V. Nederlandse Gasunie

dr. ir. M.T. Dröge mr. R. Kenter
Project Manager EGIG Secretariat EGIG
P.O. Box 19 P.O. Box 19
NL - 9700 MA GRONINGEN NL - 9700 MA GRONINGEN

E-mail: [email protected]
Website: https://www.EGIG.eu

EGIG encourages the reader, who would like specific information not available in the report, to make
direct contact with the companies involved. Company addresses are available on the EGIG website.

1
Representing Ørsted, Energinet and Evida.
2
Representing National Grid, Scotia Gas Networks, Wales and the West Utilities and Northern Gas Networks.
3
Representing Swissgas, Erdgas Ostschweiz, Erdgas Zentralschweiz, GAZNAT, Gasverbund Mittelland and
Aziende Industriali di Lugano (AIL)

11th EGIG-report VA 20.0432

1970-2019 Page 2 of 56
CONTENT

CONTENT .......................................................................................................................... 3

List of appendices ............................................................................................................ 4

List of tables..................................................................................................................... 4

List of figures ................................................................................................................... 4

Summary .......................................................................................................................... 6

Conclusions ...................................................................................................................... 6

1 Introduction ............................................................................................................... 7

2 EGIG database ............................................................................................................ 9

2.1 Objective ............................................................................................................ 9

2.2 Criteria ............................................................................................................... 9

2.3 Contents ............................................................................................................. 9

2.4 Definitions ........................................................................................................ 11

2.5 The use of EGIG data ........................................................................................ 11

3 Analyses and results ................................................................................................. 13

3.1 Trends gas transmission system ...................................................................... 13

3.1.1 Total length ...................................................................................................... 13
3.1.2 Exposure .......................................................................................................... 17

3.2 Trends of the number of incidents .................................................................... 19

3.3 Failure frequencies analyses ............................................................................ 20

3.3.1 Primary failure frequencies ................................................................................. 20
3.3.2 Primary failure frequencies per leak size ............................................................... 23
3.3.3 Secondary failure frequencies.............................................................................. 27
3.3.3.1 Relationship between diameter class and size of leak ...................................................... 28
3.3.3.2 Relationship between external interference, size of leak and design parameter................... 29
3.3.3.3 Relationship between corrosion, size of leak and design parameter ................................... 33
3.3.3.4 Relationship between construction defect/material failures, leak size and design parameter . 38
3.3.3.5 Relationship between hot tap made by error, size of leak and design parameter ................. 41
3.3.3.6 Ground movement ...................................................................................................... 42
3.3.3.7 Other and unknown .................................................................................................... 45

3.4 Other analyses ................................................................................................. 45

3.4.1 Relationship between corrosion and age ............................................................... 45
3.4.2 Ignition of releases ............................................................................................ 47
3.4.3 Injuries and fatalities ......................................................................................... 48
3.4.4 Detection of incidents ......................................................................................... 50

4 Conclusions .............................................................................................................. 52

5 Bibliography ............................................................................................................. 53

11th EGIG-report VA 20.0432

1970-2019 Page 3 of 56
LIST OF APPENDICES

APPENDIX 1: Statistics ........................................................................................................ 54

APPENDIX 2: Poisson law .................................................................................................... 56

LIST OF TABLES

Table 1: Primary failure frequencies...................................................................................... 20

Table 2: Primary failure frequencies per cause (confidence intervals are given in APPENDIX 1) .... 22
Table 3: Primary failure frequency per leak size of the period 2015-2019 .................................. 23
Table 4: Primary failure frequency, cause and size of leak (2010-2019) .................................... 27
Table 5: Secondary failure frequency pipeline diameter and size of leak (2000-2019) ................. 29
Table 6: Secondary failure frequency, pipeline diameter and size of leak (2010-2019) ................ 29
Table 7: Ignition of releases per leak type ............................................................................. 48
Table 8: Detection of incidents ............................................................................................. 50
Table 9: Primary failure frequencies and confidence intervals over different time intervals........... 54
Table 10: Primary failure frequencies and confidence intervals per leak size (period 2015–2019) . 54
Table 11: Primary failure frequencies and confidence intervals per cause (1970-2019) ............... 54
Table 12: Primary failure frequencies and confidence intervals per cause (2000-2019) ............... 55
Table 13: Primary failure frequencies and confidence intervals per cause (2010-2019) ............... 55
Table 14: Primary failure frequencies and confidence intervals per cause (2015-2019) ............... 55

LIST OF FIGURES

Figure 1: Total length of the European gas transmission system in EGIG ................................... 13
Figure 2: Total length per diameter ...................................................................................... 14
Figure 3: Total length per year of construction ....................................................................... 14
Figure 4: Total length per type of coating .............................................................................. 15
Figure 5: Total length per depth of cover (cd) ........................................................................ 15
Figure 6: Total length per wall thickness (wt) ........................................................................ 16
Figure 7: Total length per grade of material ........................................................................... 16
Figure 8: Total length per Maximum Operating Pressure (p) .................................................... 17
Figure 9: Evolution of the exposure ...................................................................................... 18
Figure 10: Average age of the pipeline system ....................................................................... 18
Figure 11: Number of incidents per year ............................................................................... 19
Figure 12: Cumulative number of incidents ............................................................................ 19
Figure 13: Primary failure frequencies ................................................................................... 21
Figure 14: Distribution of incidents (2010–2019).................................................................... 21
Figure 15: Primary failure frequencies per cause (five year moving average) ............................. 22
Figure 16: Primary failure frequency (five year moving average) per leak size ........................... 23
Figure 17: Distribution for incidents with leak size pinhole/crack (2010-2019) ........................... 24
Figure 18: Distribution for incidents with leak size hole (2010-2019) ........................................ 24
Figure 19: Distribution for incidents with leak size rupture (2010-2019) .................................... 25
Figure 20: Relationship primary failure frequency, cause and size of leak (1970-2019) ............... 26
Figure 21: Relationship primary failure frequency, cause and size of leak (2010-2019) ............... 26
Figure 22: Number of incidents per cause in the period of 2010 to 2019 ................................... 27
Figure 23: Secondary failure frequency, pipeline diameter and size of leak (2000-2019) ............. 28

11th EGIG-report VA 20.0432

1970-2019 Page 4 of 56
Figure 24: Relationship external interference, leak size and diameter (d) (1970-2019) ............... 30
Figure 25: Relationship external interference, leak size and diameter (d) (2010-2019) ............... 30
Figure 26: Relationship external interference, leak size and depth of cover (cd) (1970-2019) ...... 31
Figure 27: Relationship external interference, leak size and depth of cover (cd) (2010-2019) ...... 31
Figure 28: Failure frequency of external interference (five year moving average) and depth of cover
.................................................................................................................................. 32
Figure 29: Relationship external interference, leak size and wall thickness (wt) (1970-2019) ....... 32
Figure 30 : Relationship external interference, size of leak and wall thickness (wt) (2010-2019) .. 33
Figure 31: Relationship corrosion, size of leak and year of construction (1970-2019) .................. 34
Figure 32: Relationship corrosion, size of leak and year of construction (2010-2019) .................. 34
Figure 33: Relationship corrosion, size of leak and most common type of coating (1970-2019) .... 35
Figure 34: Relationship corrosion, size of leak and most common type of coating (2010-2019) .... 35
Figure 35: Relationship corrosion, size of leak and wall thickness (wt) (1970-2019) ................... 36
Figure 36: Relationship corrosion, size of leak and wall thickness (wt) (2010-2019) ................... 36
Figure 37: Breakdown of corrosion incidents on basis of location and appearance (1970-2019) .... 37
Figure 38: Breakdown of corrosion incidents on basis of location and appearance (2010-2019) .... 38
Figure 39: Relationship construction defect, size of leak and year of construction (1970-2019) .... 39
Figure 40: Relationship construction defect, size of leak and year of construction (2010-2019) .... 39
Figure 41: Relationship material failure, size of leak and year of construction (1970-2019) ......... 40
Figure 42: Relationship material failure, size of leak and year of construction (2010-2019) ......... 40
Figure 43: Relationship hot tap made by error, leak size and diameter (1970-2019) ................... 41
Figure 44: Relationship hot tap made by error, leak size and diameter (2010-2019) ................... 42
Figure 45: Relationship ground movement, size of leak and diameter (1970-2019) .................... 43
Figure 46: Relationship ground movement, size of leak and diameter (2010-2019) .................... 43
Figure 47: Distribution of the sub-causes of ground movement (1970-2019) ............................. 44
Figure 48: Distribution of the sub-causes of ground movement (2010-2019) ............................. 44
Figure 49: Relation wall thickness and failure frequency of incidents caused by lightning. ............ 45
Figure 50: Failure frequency (five year moving average) of corrosion incidents and year of
construction ................................................................................................................. 46
Figure 51: Relationship failure frequency of corrosion incidents and the age at the time of failure 46
Figure 52: 3D plot of failure frequency of corrosion incidents and the age at the time of failure.... 47
Figure 53: Percentages ruptures that ignited subdivided in diameter and pressure (1970-2019) .. 48
Figure 54: Percentage of accidents of groups involved in pipeline incidents (1970-2019) ............. 49
Figure 55: Percentage fatalities of accidents of groups as a function of leak size (1970-2019)...... 50
Figure 56: Detection of incidents per leak size (2010-2019) .................................................... 51

11th EGIG-report VA 20.0432

1970-2019 Page 5 of 56
SUMMARY

In 1982 six European gas transmission system operators took the initiative to gather data on the
unintentional releases of gas in their transmission pipeline systems. This cooperation was formalised
by the setting up of EGIG (European Gas pipeline Incident data Group). Presently, EGIG is a
cooperation of seventeen gas transmission system operators in Europe and it is the owner of an
extensive database of pipeline incident data collected since 1970.

The EGIG database is a valuable and reliable source of information that is used to establish pipeline
failure frequencies and analyse causes of failures in the gas transmission pipeline systems.

CONCLUSIONS

• The EGIG database is a valuable source of information on European gas pipelines and pipeline
incidents.
• EGIG has maintained and expanded the European Gas pipeline incident database. Seventeen
gas transmission system operators in Europe now collect incident data on 142,711 km of
pipelines every year. The total exposure, which expresses the length of a pipeline and its
period of operation, is 4.84 million km·yr.
• In the EGIG database 1,411 pipeline incidents are recorded in the period from 1970-2019.
• The history of incidents collected in the database gives reliable failure frequencies. The
overall failure frequency over the period 1970-2019 is equal to 0.29 incidents per year per
1,000 km.
• The five year moving average failure frequency in 2019, which represents the average failure
frequency over the past 5 years, equals 0.126 per year per 1,000 km.
• The five year moving average and overall failure frequency have reduced over the years,
although it has tended to stabilise over recent years.
• Incidents caused by external interference and ground movement are characterised by
potentially severe consequences. This emphasises the importance of measures to prevent
these incidents taken by pipeline operators and authorities.
• Corrosion as a primary cause has now the same frequency rate as external interference,
although consequences are much less severe.
• Over the last ten years, external interference, corrosion, construction defects and ground
movement, represent 27%, 27%, 16% and 16% respectively of the pipeline incidents
reported.

11th EGIG-report VA 20.0432

1970-2019 Page 6 of 56
1 INTRODUCTION

The use of pipelines for the transport of large quantities of natural gas to industry and to commercial
and domestic consumers represents a reliable mode of transport of energy.
In 1982, six European gas transmission system operators took the initiative to gather data on the
unintentional releases of gas in their transmission pipeline systems. This cooperation was formalised
by the setting up of EGIG (European Gas pipeline Incident data Group). The objective of this initiative
was to provide a broad basis for the calculation of safety performance of pipeline systems in Europe,
thus providing a reliable picture of the numbers and frequencies of incidents. Nowadays, EGIG is a
cooperation of seventeen gas transmission system operators in Europe and it is the owner of an
extensive database of pipeline incident data collected since 1970. The participating companies are
now:

Gas Networks Ireland (Ireland)

DGC (Denmark)1
ENAGAS, S.A. (Spain)
EUSTREAM (Slovak Republic)
Fluxys (Belgium)
Gasgrid Finland (Finland)
GRTgaz (France)
National Grid (UK)2
NET4GAS (Czech Republic)
Gasunie (The Netherlands / Germany)
Gasconnect GmbH (Austria)
Open Grid Europe (Germany)
REN Gasodutos S.A. (Portugal)
Snam Rete Gas (Italy)
Swedegas A.B. (Sweden)
SWISSGAS (Switzerland)
Teréga (France)

Considering the number of participants, the extent of the pipeline systems and the exposure period
involved, the EGIG database is a valuable source of information on European gas pipelines and
pipeline incidents. The results of the database present an average of all participating companies and
do not highlight the geographical differences.

Definitions have been used consistently over the entire period. Consequently, provided that the data
is correctly used and interpreted, the EGIG database gives useful information about trends which
have developed over the years. Nevertheless, particular care must be given to the use and
interpretation of the statistical data. The EGIG report gives the failure frequency per design
parameter (diameter, pressure, wall thickness) and conclusions about combination of design
parameters cannot be drawn.

This report describes the structure of the EGIG database and presents different analyses and their
results. The results of the analyses are commented on and give the most interesting information that
can be extracted from the database. Linking of results of different analyses is provided where

1
Representing Ørsted, Energinet and Evida.
2
Representing National Grid, Cadent, Scotia Gas Networks, Wales and the West Utilities and Northern Gas
Networks.

11th EGIG-report VA 20.0432

1970-2019 Page 7 of 56
possible. Anyone who would like to combine different results should be very careful before drawing
conclusions.

11th EGIG-report VA 20.0432

1970-2019 Page 8 of 56
2 EGIG DATABASE

The EGIG database is a database for pipeline data and pipeline incident data. Pipeline data and
incident data of natural gas transmission pipelines are in the database from 1970 on.

2.1 Objective
The objective of EGIG is to collect and present data on loss of gas incidents in order to present the
safety performance of the European gas transmission network to the general public and authorities.

2.2 Criteria
The required criteria for an incident to be recorded in the EGIG database are the following:

• The incident must lead to an unintentional gas release.

• The pipeline must fulfil the following conditions:
▪ To be made of steel.
▪ To be onshore.
▪ To have a Maximum Operating Pressure higher than 15 barg.
▪ To be located outside the fence of a gas installation.

Incidents on production lines or involving equipment or components (e.g. valve, compressor) are not
recorded in the EGIG database.

2.3 Contents
The EGIG database contains general information about the European gas transmission pipelines
system as well as specific information about the incidents.

Every year the length of the pipeline system is collected for the following parameters:

• Diameter
• Pressure
• Year of construction
• Type of coating
• Depth of cover
• Grade of material
• Wall thickness.

Specific information about incidents comprises:

• The characteristics of the pipeline on which the incident happened, namely the general
information listed above.
• The leak size:
▪ Pinhole/crack: the effective diameter of the hole is smaller than or equal to 2 cm
▪ Hole: the effective diameter of the hole is larger than 2 cm and smaller than or equal
to the diameter of the pipe
▪ Rupture: the effective diameter of the hole is larger than the pipeline diameter.
• The initial cause of the incident
▪ External interference
▪ Corrosion
▪ Construction defect/material failure
▪ Hot tap made by error

11th EGIG-report VA 20.0432

1970-2019 Page 9 of 56
 ▪ Ground movement
▪ Other and unknown.
• The occurrence (or non-occurrence) of ignition.
• The consequences.
• Information on the way the incident has been detected (e.g. contractor, landowner, patrol).
• A free text for extra information.

Additional information is also given for every cause:

• External interference:
▪ The activity having caused the incident (e.g. digging, piling, ground works).
▪ The equipment involved in the incident (e.g. anchor, bulldozer, excavator, plough).
▪ The installed protective measures (e.g. casing, sleeves).
• Corrosion:
▪ The location (Internal, External, Unknown).
▪ The appearance (General, Pitting, Cracking).
▪ In line inspected (yes, no, unknown).
• Construction defect/material failure:
▪ The type of defect (construction or material).
▪ The defect details (hard spot, lamination, material, field weld or unknown).
▪ The pipeline component type (straight, field bend, factory bend).
• Ground movement:
▪ The type of ground movement (dike break, erosion, flood, landslide, mining, erosion
of riverbed, erosion of the riverbank or unknown).
• Other and unknown:
▪ The sub-causes such as design error, lightning, maintenance error.

This information has been used for the analyses given in this report. EGIG is always considering
whether changes in the information would be useful to enhance these analyses.

11th EGIG-report VA 20.0432

1970-2019 Page 10 of 56
2.4 Definitions

Failure frequency: The failure frequency is calculated by dividing the number of incidents by the
exposure. The EGIG report presents two kinds of failure frequencies, the primary and the
secondary. They refer to the notions of total and partial exposure respectively. These notions are
defined below.
• Exposure is the length of a pipeline multiplied by its exposed duration and is expressed in
kilometres-years [km·yr]. Example: a company has a constant length of transmission
pipelines over 5 years of 1,000 km. Its exposure is then 5 times 1,000 km, so 5,000 km·yr.
• The total system exposure is the exposure as defined above, calculated for the complete
system.
• The partial system exposures are the exposures calculated per class of a certain design
parameter, e.g. per diameter class or per depth of cover class.

Five year moving average: In order to illustrate trends, a five year moving average has been
introduced. The five year moving average for the year in question means that the calculations have
been performed over the five previous years.

Confidence interval: A confidence interval gives an estimated range of values which is likely to
include an unknown population parameter, the estimated range being calculated from a given set of
sample data. In this report, a confidence interval of 95% is calculated for the failure frequencies.

2.5 The use of EGIG data

The objective of the EGIG group is to show the incident data of gas transmission pipelines, registered
by a European group of operators which in general follow similar design, construction, inspection and
maintenance practices.

Within EGIG, all data collected, reported and analysed is data of the group as a whole and no
distinction can and will be made per operator.

EGIG publishes statistics over different time intervals. In this report, the statistics of the whole
database (covering the period 1970-2019), but also the most important statistics over the last 40,
30, 20, 10 and 5 years are reported. It must be noted that given the theory of statistics, the
confidence interval of the mean values of the failure frequencies over five years is larger than for a
longer period (for instance 20 years). The user of EGIG data must consider the statistical reliability
of the data when deciding how it is to be used (see also APPENDIX 2).

Graphs
Some of the graphs presented in this report will cover the whole period of the EGIG database (1970-
2019). To demonstrate developments and trends over more recent periods, the EGIG report also
shows graphs that cover the last ten years (2010-2019) or represent the five year moving average.

The report aims to interpret the information contained in the data in order to draw conclusions from
the sample or the population from which the sample is taken. The statistical analyses are based on
the calculation of indicators such as failure frequency and the percentage of the releases that ignited.

The EGIG database offers an overview of the failure frequencies of the European gas transmission
pipelines system. It gives information on the failure frequencies in relation to one pipeline parameter
(e.g. diameter, pressure, wall thickness), but does in general not offer the possibility of making

11th EGIG-report VA 20.0432

1970-2019 Page 11 of 56
correlation analyses. For example, with the EGIG database it is possible to establish the failure
frequency of ≥42-inch pipelines or to establish the failure frequency of pipelines with a wall thickness
of >15 mm, but it is not possible to calculate the failure frequency of ≥42-inch pipelines with a wall
thickness of >15 mm.

11th EGIG-report VA 20.0432

1970-2019 Page 12 of 56
3 ANALYSES AND RESULTS

3.1 Trends gas transmission system

This paragraph gives information on the trends in the European gas transmission system. It not only
shows the evolution of the exposure, but also which design characteristics tend to be more or less
used in today’s construction. This paragraph gives a picture of the European gas transmission system
from 1970 up to the present.

3.1.1 Total length

The total length of the European gas transmission pipelines system in EGIG has remained at
approximately the same level since the last six years. The evolution of the total length of the system
is shown in Figure 1 and is also given per class in Figure 2 to Figure 8 for several pipeline parameters
(diameter, pressure, etc.).

Figure 1: Total length of the European gas transmission system in EGIG

Figure 1 shows the increase in the length of the European gas transmission system in EGIG, which
has significant step changes in the years 1975, 1989, 1991, 1998, 2003, 2007 and 2011. These
changes correspond to (data of) new members joining EGIG. The pipeline length stabilizes from the
year 2011.

11th EGIG-report VA 20.0432

1970-2019 Page 13 of 56
Figure 2: Total length per diameter

Figure 2 demonstrates that the 5’’≤diameter<11’’ and the 11’’≤diameter<17’’ classes are still the
most commonly used.

Figure 3: Total length per year of construction

Figure 3 shows that more pipelines were built in the period 1964-1973 than in other periods. No
significant drop can be observed, which means that most of these pipelines are still in operation. Also
new pipelines continue to be constructed.

11th EGIG-report VA 20.0432

1970-2019 Page 14 of 56
Figure 4: Total length per type of coating

Figure 4 shows that coal tar, bitumen and polyethylene are the most common coatings in the
database, with a clear predominance of the last one. In the most recent decades the vast majority
of new pipelines have been coated with polyethylene.

Figure 5: Total length per depth of cover (cd)

Figure 5 shows that the vast majority of the pipelines have a depth of cover greater than 80 cm.

11th EGIG-report VA 20.0432

1970-2019 Page 15 of 56
Most companies and design codes recognise depth of cover as an important factor in reducing
exposure to external interference.

Figure 6: Total length per wall thickness (wt)

Figure 6 shows that the most commonly used pipeline wall thicknesses are 5 to 10 mm. The figure
also shows that the pipeline length for every wall thickness class increases constantly over time
except for the ≤ 5 mm class, which has remained more or less constant in length since 2001.

Figure 7: Total length per grade of material

11th EGIG-report VA 20.0432

1970-2019 Page 16 of 56
Line pipe grade designations come from different specifications. The EGIG database is arranged
according to equivalent API 5L grades, i.e. line pipe can have grade A, B or a higher grade with
designation X followed by a number specifying the yield strength (in pounds per square inch) of the
pipe steel. Grade A is used for older pipelines. Grade B is still used for new pipelines, especially for
pipelines with relative small diameters.
Figure 7 demonstrates that four grades of material are predominant, namely: Grade B, X52, X60 and
X70.

Figure 8: Total length per Maximum Operating Pressure (p)

Figure 8 shows a predominance of Maximum Operating Pressure of 65 bar and higher.

3.1.2 Exposure
Figure 9 shows the increase of exposure over the years. As discussed in paragraph 2.4, exposure is
the length of a pipeline multiplied by its exposed duration and is expressed in kilometres-years
[km·yr]. In 2019, the total system exposure was equal to 4.84 million km·yr. Figure 10 shows the
average age of the pipeline system over the years.

11th EGIG-report VA 20.0432

1970-2019 Page 17 of 56
Figure 9: Evolution of the exposure

Figure 10: Average age of the pipeline system

11th EGIG-report VA 20.0432

1970-2019 Page 18 of 56
3.2 Trends of the number of incidents
In the tenth EGIG report, which covers period 1970-2016, a total of 1,366 incidents were recorded.
In the last three years, 45 incidents were reported by the EGIG members, which brings the total
number of incidents to 1,411 for the period 1970-2019. Figure 11 shows the number of incidents per
year. Figure 12 shows the cumulative number of incidents.

Figure 11: Number of incidents per year

Figure 12: Cumulative number of incidents

11th EGIG-report VA 20.0432

1970-2019 Page 19 of 56
3.3 Failure frequencies analyses
This paragraph deals with the calculation of safety indicators, namely the primary and secondary
failure frequencies.

3.3.1 Primary failure frequencies

As explained in paragraph 2.4, the primary failure frequency is the result of the number of incidents
(Figure 12) within a period divided by the corresponding total system exposure (Figure 9). Depending
on the period considered, the number of incidents varies and so does the total system exposure.

EGIG has compared the primary failure frequencies of different periods, namely the total period
(1970-2019), periods corresponding to the previous EGIG reports and of periods of the last 40, 30,
20, 10 and 5 years.

The primary failure frequencies of these periods are given in Table 1. The 95% confidence limits of
the failure frequencies of these periods are given in APPENDIX 1. For the statistical analysis the
assumption is made that the number of incidents follows Poisson’s law (see APPENDIX 2).

Primary failure
Number of Total system exposure
Period Interval frequency per 1,000
incidents ·106 km·yr
km·yr
1970 – 2007 7th report, 38 years 1,173 3.15 0.372
1970 – 2010 8th report, 41 years 1,249 3.55 0.351
1970 – 2013 9th report, 44 years 1,309 3.98 0.329
1970 – 2016 10th report, 47 years 1,366 4.41 0.310
1970 – 2019 11th report, 50 years 1,411 4.84 0.292
1980 – 2019 40 years 1,050 4.36 0.241
1990 – 2019 30 years 663 3.63 0.183
2000 – 2019 20 years 388 2.64 0.147
2010 – 2019 10 years 184 1.42 0.129
2015 – 2019 5 years 90 0.71 0.126

Table 1: Primary failure frequencies

In 2019, the primary failure frequency over the entire period (1970-2019) was equal to 0.29 per
1,000 km·yr. This is slightly lower than the failure frequency of 0.31 per 1,000 km·yr reported in the
10th EGIG report (1970-2016).

The primary failure frequency over the last five years was equal to 0.13 per 1,000 km·yr, showing
an improved performance over recent years.

Figure 13 illustrates the steady drop of the primary failure frequencies. The primary failure frequency
over the entire period decreased from 0.87 per 1,000 km·yr in 1970 to 0.29 per 1,000 km·yr in 2019.
The five year moving average primary failure frequency decreased by a factor 6 (0.86 to 0.13 per
1,000 km·yr).

11th EGIG-report VA 20.0432

1970-2019 Page 20 of 56
Figure 13: Primary failure frequencies

In Figure 14, the incident distribution per cause over the last 10 years is given. Corrosion and external
interference incidents occurred in about the same rate. However, corrosion incidents tend to have
smaller leak sizes (see Figure 20 and Figure 21).

Figure 14: Distribution of incidents (2010–2019)

Figure 15 illustrates the decreasing failure frequencies per cause over the years. The decrease may
be explained by technological developments, such as: welding, inspection, condition monitoring using
in-line inspection and improved procedures for damage prevention and detection. Improvements in

11th EGIG-report VA 20.0432

1970-2019 Page 21 of 56
the prevention of external interference incidents may be explained by a more stringent enforcement
of land use planning and the application of one-call systems for the digging activities of external
parties. In several countries, there is now a legal requirement to report digging activities. Companies
have adopted appropriate actions, like supervision or marking of the pipeline in the direct
neighbourhood of the digging activities.

Figure 15: Primary failure frequencies per cause (five year moving average)

Primary failure frequency

Cause 1970-2019 2000-2019 2010-2019 2015-2019

per per per per
1,000 km·yr 1,000 km·yr 1,000 km·yr 1,000 km·yr
External interference 0.134 0.054 0.035 0.036
Corrosion 0.050 0.033 0.034 0.032
Construction defect / Material failure 0.048 0.020 0.020 0.015
Hot tap made by error 0.013 0.005 0.002 0.001
Ground movement 0.025 0.020 0.020 0.017
Other and unknown 0.022 0.015 0.017 0.024

Table 2: Primary failure frequencies per cause (confidence intervals are given in APPENDIX
1)

To demonstrate failure frequencies over a more recent period, Table 2 also presents, in addition to
the frequencies for the whole period, frequencies over a time span of the last 5, 10 and 20 years. As
far as the cause external interference and corrosion is concerned, its associated primary failure
frequencies over the five year moving average has levelled off between 0.03 and 0.04 per 1,000
km·yr.

11th EGIG-report VA 20.0432

1970-2019 Page 22 of 56
3.3.2 Primary failure frequencies per leak size

Not all leaks result in severe consequences. The EGIG database distinguishes between incidents with
different leak size (ruptures, holes and pinholes/cracks). Figure 16 demonstrates the five year
moving average failure frequency per leak size.

Figure 16: Primary failure frequency (five year moving average) per leak size

Figure 16 shows that the failure frequencies for holes and ruptures are smaller than the failure
frequencies for pinhole/cracks. Also a decrease over the years of the five year moving average can
be seen for all leak sizes. From the year 2000 on this trend seems to stabilise. For the year 2019
these values are given in Table 3.

Primary 5 year mov.

Leak size failure frequency
per 1,000 km·yr
Unknown 0.003

Pinhole/crack 0.088

Hole 0.022

Rupture 0.013

Table 3: Primary failure frequency per leak size of the period 2015-2019

Figure 17, Figure 18 and Figure 19 show the distribution of the incidents per leak size. From these
figures it can be seen that pinholes are mainly caused by corrosion, holes are mainly caused by
external interference and the main causes of ruptures are ground movement and external
interference.

11th EGIG-report VA 20.0432

1970-2019 Page 23 of 56
 Other and unknown External interference
16% 17%

Ground movement
9%

Hot tap made by error

1%

Construction defect /
Material failure Corrosion
19% 38%

Figure 17: Distribution for incidents with leak size pinhole/crack (2010-2019)

Other and unknown

3%

Ground movement
22%

Hot tap made by error

3%

Construction defect / Material failure

6%
External
interference
66%

Figure 18: Distribution for incidents with leak size hole (2010-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 24 of 56
 Other and unknown
5%

External interference
38%

Ground movement
48%

Construction defect / Material failure

9%

Figure 19: Distribution for incidents with leak size rupture (2010-2019)

Figure 20 (period 1970-2019), Figure 21 (period 2010-2019) and Table 4 show the failure frequency
per leak size and per incident cause. Although the failure frequency decreased over the years, the
general trend in the distribution of the leak sizes remain the same: holes and ruptures were mainly
caused by external interference. For pinhole/crack leak sizes, corrosion remains the main cause.

Figure 20 and Figure 21 show that corrosion in the vast majority of incidents has led to pinhole/crack
type of leak. Very few holes were observed and only one rupture occurred on a pipeline. This rupture,
on a pipeline constructed before 1954, was caused by internal corrosion of a pipeline originally used
for the transportation of coke oven gas and is not representative for normal corrosion incidents.

11th EGIG-report VA 20.0432

1970-2019 Page 25 of 56
Figure 20: Relationship primary failure frequency, cause and size of leak (1970-2019)

Figure 21: Relationship primary failure frequency, cause and size of leak (2010-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 26 of 56
 Failure frequency per 1,000 km·year
Leak size Construction
External Hot tap made Ground Other and
Corrosion defect / Mat.
interference by error movement unknown
Failure
Rupture 0.006 0.000 0.001 0.000 0.007 0.001

Hole 0.015 0.000 0.001 0.001 0.005 0.001

Pinhole/crack 0.015 0.033 0.017 0.001 0.008 0.014

Unknown 0.000 0.001 0.001 0.000 0.001 0.001

Table 4: Primary failure frequency, cause and size of leak (2010-2019)

Figure 22 shows the number of incidents per cause that occurred in the last 10 years. What can be
seen is that corrosion incidents together with external interference incidents nowadays are the largest
cause of the incidents. However, external corrosion incidents have never lead to ruptures of pipelines.
Ground movement incidents also can lead to ruptures because of the large forces released onto the
pipelines in this type of incidents.

Figure 22: Number of incidents per cause in the period of 2010 to 2019

3.3.3 Secondary failure frequencies

The secondary failure frequencies are calculated by dividing the number of incidents by a partial
system exposure. Partial system exposure means, for example, the exposure related to one diameter
class or one year of construction class.

The calculation of secondary failure frequencies is done to consider the influence of design
parameters (pressure, diameter, depth of cover, etc.) on the failure frequencies per incident cause
and per type of leak size. The calculations are performed for the whole database and for a more
recent time period of the last 10 years (2010-2019).

11th EGIG-report VA 20.0432

1970-2019 Page 27 of 56
For six incident causes, the secondary failure frequencies have been calculated according to the
following design parameters:

• External interference: the diameter of the pipeline, the depth of cover and the wall thickness.
• Corrosion: the year of construction, the type of coating and the wall thickness.
• Construction defect/material failure: the year of construction.
• Hot tap made by error: the diameter of the pipeline.
• Ground movement: the diameter of the pipeline.
• Other and unknown: main causes.

For “Ground movement” and “other and unknown” causes other more relevant considerations are
reported.

3.3.3.1 Relationship between diameter class and size of leak

Figure 23 demonstrates the relationship between the secondary failure frequency, the leak size and
diameter of the pipeline. The secondary frequencies are given for a time period of 20 years as this
is considered more representative for the current operating practises than taking the whole period.

Figure 23: Secondary failure frequency, pipeline diameter and size of leak (2000-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 28 of 56
 System
Secondary failure frequency per 1,000 km·yr
exposure
Nominal diameter
·106 km·yr
Unknown Pinhole/crack Hole Rupture

diameter < 5'' 0.274 0.007 0.332 0.099 0.066

5" ≤ diameter < 11" 0.651 0.006 0.126 0.060 0.029

11" ≤ diameter < 17" 0.433 0.005 0.062 0.030 0.014

17" ≤ diameter < 23" 0.283 0.004 0.046 0.025 0.007

23" ≤ diameter < 29" 0.269 0.000 0.056 0.022 0.011

29" ≤ diameter < 35" 0.165 0.000 0.018 0.000 0.006

35" ≤ diameter < 41" 0.274 0.000 0.011 0.000 0.004

41" ≤ diameter < 47" 0.106 0.000 0.009 0.000 0.000

diameter ≥ 47" 0.178 0.000 0.006 0.000 0.006

Table 5: Secondary failure frequency pipeline diameter and size of leak (2000-2019)

System
Secondary failure frequency per 1,000 km·yr
Nominal diameter exposure
·106 km·yr Unknown Pinhole/crack Hole Rupture
diameter < 5'' 0.142 0.007 0.367 0.049 0.049

5" ≤ diameter < 11" 0.340 0.006 0.117 0.050 0.023

11" ≤ diameter < 17" 0.230 0.009 0.070 0.017 0.013

17" ≤ diameter < 23" 0.150 0.007 0.040 0.007 0.000

23" ≤ diameter < 29" 0.148 0.000 0.047 0.020 0.013

29" ≤ diameter < 35" 0.094 0.000 0.011 0.000 0.011

35" ≤ diameter < 41" 0.149 0.000 0.013 0.000 0.000

41" ≤ diameter < 47" 0.058 0.000 0.000 0.000 0.000

diameter ≥ 47" 0.110 0.000 0.009 0.000 0.000

Table 6: Secondary failure frequency, pipeline diameter and size of leak (2010-2019)

Figure 23, Table 5 and Table 6 illustrate that ruptures mainly occurred at pipelines with smaller
diameters and that failure frequencies decrease with increasing diameter.

3.3.3.2 Relationship between external interference, size of leak and design parameter
Figure 24 to Figure 30 show the failure frequencies for the incident cause “external interference” for
different pipeline design parameter classes and leak sizes. The design parameters considered are:
pipeline diameter, depth of cover and wall thickness. For the design parameters diameter and wall
thickness the graphs are presented for both the whole period 1970-2019 and the last ten years
(2010-2019). For depth of cover a graph is presented for the period 1970-2019 and a graph is
presented with the development of the five year moving average failure frequency per depth of cover.
Although the graphs are presented separately, it must be noticed that the design parameters are
correlated. No quantitative correlations between parameters have been studied.

11th EGIG-report VA 20.0432

1970-2019 Page 29 of 56
Figure 24: Relationship external interference, leak size and diameter (d) (1970-2019)

Figure 25: Relationship external interference, leak size and diameter (d) (2010-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 30 of 56
Figure 26: Relationship external interference, leak size and depth of cover (cd) (1970-
2019)

Figure 27: Relationship external interference, leak size and depth of cover (cd) (2010-
2019)

11th EGIG-report VA 20.0432

1970-2019 Page 31 of 56
Figure 28: Failure frequency of external interference (five year moving average) and depth
of cover

Figure 29: Relationship external interference, leak size and wall thickness (wt) (1970-
2019)

11th EGIG-report VA 20.0432

1970-2019 Page 32 of 56
Figure 30 : Relationship external interference, size of leak and wall thickness (wt) (2010-
2019)

From these figures, some general conclusions can be drawn:

• Large diameter pipelines are less vulnerable to external interferences than smaller diameter
pipelines (Figure 24 and Figure 25). There might be several explanations for this: small
diameter pipelines can be more easily hooked up during ground works than bigger pipelines,
their resistance is often lower due to thinner wall thickness and they might be found more
frequently in urban areas where third party activity is generally higher.
• The depth of cover is one of the leading indicators for the failure frequencies of pipelines.
Pipelines with a larger depth of cover have a lower failure frequency. This can be seen from
Figure 26.
• Figure 28 shows that the external interference failure frequencies of all depth of cover classes
have decreased over the years.
• Pipelines with a larger wall thickness have a lower failure frequency for external interference
(Figure 29 and Figure 30).
• No External Interference incidents occurred with wall thicknesses above 15 mm.

3.3.3.3 Relationship between corrosion, size of leak and design parameter

Figure 31 to Figure 36 show the failure frequencies for the incident cause “corrosion” for different
pipeline parameter classes and leak sizes. The parameters considered are year of construction, type
of coating and wall thickness. For each design parameter two graphs are constructed: one for the
period 1970-2019 and one for the period 2010-2019.

11th EGIG-report VA 20.0432

1970-2019 Page 33 of 56
Figure 31: Relationship corrosion, size of leak and year of construction (1970-2019)

Figure 32: Relationship corrosion, size of leak and year of construction (2010-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 34 of 56
Figure 33: Relationship corrosion, size of leak and most common type of coating (1970-
2019)

Figure 34: Relationship corrosion, size of leak and most common type of coating (2010-
2019)

From these figures, it appears that older pipelines, with predominantly tar coatings, have higher
failure frequencies. Nowadays, most transmission operators use modern coatings like polyethylene
coatings.

11th EGIG-report VA 20.0432

1970-2019 Page 35 of 56
Different protective measures are undertaken by pipeline owners to prevent leakage due to corrosion.
These measures are for example cathodic protection and pipeline coating. In-line inspections and
pipeline surveys also allow corrosion to be detected at an earlier stage.

Figure 35: Relationship corrosion, size of leak and wall thickness (wt) (1970-2019)

Figure 36: Relationship corrosion, size of leak and wall thickness (wt) (2010-2019)

From these figures some general conclusions can be drawn:

• The failure frequency decreases with increasing year of construction.

11th EGIG-report VA 20.0432

1970-2019 Page 36 of 56
 • The failure frequency decreases with increasing wall thickness. Corrosion is a time dependent
phenomenon of deterioration of the pipelines. Corrosion takes place independently of the
wall thickness, but the thinner the corroded pipeline wall, the sooner the pipeline fails.
Corrosion on thicker pipelines takes longer before causing an incident and therefore has more
chance to be detected by inspection programs.
• Pipelines coated with a polyethylene coating or epoxy coatings have a far lower failure
frequency than pipelines with other types of coating.

For the corrosion incidents, two other types of data are registered:

• the location of corrosion (Internal, External, Unknown),

• the appearance of corrosion (General, Pitting, Cracking).

Figure 37: Breakdown of corrosion incidents on basis of location and appearance (1970-
2019)

11th EGIG-report VA 20.0432

1970-2019 Page 37 of 56
Figure 38: Breakdown of corrosion incidents on basis of location and appearance (2010-
2019)

Figure 37 to Figure 38 demonstrate that pitting is the most common form of corrosion. Almost all
corrosion incidents with pitting occur on the external surface of the pipelines.

Corrosion appearing as cracks is the second corrosion form to be found. These cracks are found on
both the inner and the external surface of the pipelines. For the more recent period of 2010-2019 all
cracks were found on the external surface.

General corrosion takes place evenly over the surface of the metal. This type of corrosion defects are
almost always on the external surface of the pipeline.

3.3.3.4 Relationship between construction defect/material failures, leak size and design
parameter
EGIG recognizes construction defects / material failures as one of the causes of pipeline incidents.
During the last ten years, they represented about 16% of the pipeline incidents and are ranked on a
third position together with ground movement in the causes of incidents (Figure 14).

The EGIG database makes it possible to distinguish between construction defect and material failures.

Figure 39 to Figure 42 show the failure frequencies for the incident cause “construction defect” and
“material failure” in relation to construction year and leak size for the periods 1970-2019 and 2010-
2019.

From these figures, some general conclusions can be drawn: failure frequencies for “construction
defects” and “material failure” generally decrease with increasing year of construction. New pipelines
are less vulnerable to construction defects due to technical improvements.

11th EGIG-report VA 20.0432

1970-2019 Page 38 of 56
 Years 1970 - 2019

Figure 39: Relationship construction defect, size of leak and year of construction (1970-
2019)

Years 2010 – 2019

Figure 40: Relationship construction defect, size of leak and year of construction (2010-
2019)

11th EGIG-report VA 20.0432

1970-2019 Page 39 of 56
 Years 1970 – 2019

Figure 41: Relationship material failure, size of leak and year of construction (1970-2019)

Years 2010 -2019

Figure 42: Relationship material failure, size of leak and year of construction (2010-2019)

Figure 41 and Figure 42 show the failure frequencies for the incident cause “material failure” for
different classes of material grade and leak sizes for the periods: 1970-2019 and 2010-2019. Grade
A material has the highest failure frequency for “material failure” in the period 1970-2019, 2010-
2019

11th EGIG-report VA 20.0432

1970-2019 Page 40 of 56
3.3.3.5 Relationship between hot tap made by error, size of leak and design parameter
The term “hot tap made by error” means that a connection has been made by error to the gas
transmission pipeline, assuming it was another pipeline.

Figure 43 and Figure 44 show the failure frequencies for the incident cause “hot tap made by error”
for different pipeline diameter classes and leak sizes. The first graph present the failure frequency
for the period 1970-2019 and the second graph for the period 2010-2019.

From these figures, some general conclusions can be drawn: the failure frequency for “hot tap made
by error” decreases with increasing pipeline diameter. The same trend is true for every leak size.
The failure frequency of “hot tap made by error” has decreased over the years.

Figure 43: Relationship hot tap made by error, leak size and diameter (1970-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 41 of 56
Figure 44: Relationship hot tap made by error, leak size and diameter (2010-2019)

3.3.3.6 Ground movement

Ground movement is responsible for 16% of the incidents over the last ten years (see Figure 14).
Figure 45 and Figure 46 show the failure frequencies for the incident cause “ground movement” for
different pipeline diameter classes and leak sizes.

Both graphs present the failure frequency per pipeline diameter class, one for the period 1970-2019,
the second for the period 2010-2019.

From these figures some conclusions can be drawn:

For the period 1970-2019 failure frequencies for “ground movement” generally decrease with
increasing pipeline diameter. The failure frequency for the diameter  47’’ is caused by one ground
movement incident.

11th EGIG-report VA 20.0432

1970-2019 Page 42 of 56
Figure 45: Relationship ground movement, size of leak and diameter (1970-2019)

Figure 46: Relationship ground movement, size of leak and diameter (2010-2019)

There are many types of “Ground movement” incidents. Figure 47 and Figure 48 give more details
on the different types of ground movements that caused a pipeline incident. Landslides are by far
the most common type causing a ground movement incident. Failing of pipelines caused by flooding
is the second cause of pipeline incidents within this group.

No incidents were recorded that were caused by earthquakes.

11th EGIG-report VA 20.0432

1970-2019 Page 43 of 56
Figure 47: Distribution of the sub-causes of ground movement (1970-2019)

Figure 48: Distribution of the sub-causes of ground movement (2010-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 44 of 56
3.3.3.7 Other and unknown
29.3% of the incidents in the category “other and unknown” are caused by lightning. Within the
period 1970-2019, 32 incidents caused by lightning have been recorded in the EGIG database, which
represents a failure frequency equal to 0.0066 per 1,000 km·yr. EGIG examined the distribution of
the consequences of lightning in terms of leak sizes. Out of 32 incidents, 30 were pinholes/cracks
and 2 resulted in a hole.

Figure 49 shows the failure frequency for the incidents caused by lightning in relation to the wall
thickness. Here it can be seen that the failure frequencies for the incidents caused by lightning
decreases with increasing wall thickness. This might be explained by the higher energy density of a
lightning for a lower wall thickness.

Figure 49: Relation wall thickness and failure frequency of incidents caused by lightning

3.4 Other analyses

3.4.1 Relationship between corrosion and age

In this analysis, the failure frequency of corrosion incidents has been studied as a function of
construction year and the age of the pipeline at the moment of the incident.

11th EGIG-report VA 20.0432

1970-2019 Page 45 of 56
Figure 50: Failure frequency (five year moving average) of corrosion incidents and year of
construction

The increase of the failure frequency of pipelines with an age older than 60 constructed before 1954
was caused by two incidents in 2015 and a decreasing population in the years after 2015.

Figure 51: Relationship failure frequency of corrosion incidents and the age at the time of
failure

Explanation Figure 51:

in Figure 51 for instance a pipeline constructed before 1954: the failure frequency 15 to 20 years
after construction is 0.062 per 1,000 km·yr, whereas it is 0.013 per 1,000 km·yr after 35-40 years.

11th EGIG-report VA 20.0432

1970-2019 Page 46 of 56
EGIG started data collection from 1970 on, therefore no data is available for failure frequencies at
the early life stage of pipelines constructed before 1954 or pipelines constructed between 1954 and
1964.

Figure 52: 3D plot of failure frequency of corrosion incidents and the age at the time of
failure

Figure 52 shows the dependency of failure frequency with the age of the pipelines and the year of
construction.

The first conclusion of Figure 51 and Figure 52 is that early constructed pipelines (before 1964) had
higher failure frequencies than recently constructed pipelines at the same age. Pipelines constructed
in the last 50 years do not show a dependency between the failure frequency of corrosion and their
age or construction year class.

3.4.2 Ignition of releases

Fortunately, not every gas release ignites, which limits the consequences of the incidents. In the
period 1970-2019, only 5.2% of the gas releases recorded in the EGIG database ignited. Pipeline
ruptures with ignition can cause severe societal consequences. This is especially the case for pipelines
with larger diameters. Figure 53 shows that gas releases from large diameter pipeline ruptures at
high pressure have ignited more frequently than smaller diameter pipeline ruptures at lower
pressure. This data is based on only a few ruptures. Care should be taken when using it as an ignition
probability, as the uncertainty is high. In the paper (Michael R. Acton, 2008) ’’Ignition Probability for

11th EGIG-report VA 20.0432

1970-2019 Page 47 of 56
High Pressure Gas update to 2019” an analysis is made of ignition probabilities. This paper shows
that even ruptures of large diameter pipelines and high pressure not always ignite.

Information on ignited releases is presented in Table 7 as a function of size of leak and pipeline
diameter.

Size of leak % of releases with ignition

Pinhole-crack 4.7
Hole 2.2
Rupture (all diameters) 4.7
Rupture ≤ 16 inches 9.8
Rupture > 16 inches 40.7

Table 7: Ignition of releases per leak type

1 of 1 1 of 1

100%
3 of 4
90%

80% 2 of 3

70%
1 of 2
60%
4 of 11
2 of 6
5 of 13
50%

40% 1 of 3

30%
0 of 6
20% 2 of 26 0 of 14
1 of 30 p > 55 bar
2 of 24 1 of 6
0 of 6
10% 1 of 26 35 < p ≤ 55 ba r
0 of 1
p ≤ 35 ba r
0%
d<5'' 5"≤d<11" 11"≤d<17" 17"≤d<23" 23"≤d<29" 29"≤d<35" 35"≤d<41" 41"≤d< 47" d≥47"

Figure 53: Percentages ruptures that ignited subdivided in diameter and pressure (1970-
2019)

3.4.3 Injuries and fatalities

EGIG database also registers qualitative information about the consequences of incidents, amongst
other injuries and fatalities that, unfortunately, occurred in some of them.

EGIG studied the injuries and fatalities among different groups involved in pipeline incidents. These
groups are:
• employees or contractors of the transmission system operator;
• third party directly involved in causing the incidents (for example digger drivers in the case
of external interference incidents);
• emergency services (firefighters, medical assistance);
• the general public.

11th EGIG-report VA 20.0432

1970-2019 Page 48 of 56
The EGIG database contains a total of 1,411 incidents, but as is shown in Figure 54 only a small
percentage leads to injuries and fatalities. The highest fatality and injury rate can be found among
the people who are directly involved in causing the incident. In 6 cases (0.43%) these incidents
caused fatalities among the people causing the incident. Two incidents (0,14%) involved fatalities
among the public. In Figure 55 it can be seen that the fatalities mainly occurred when the incident
was a pipeline rupture.

Figure 54: Percentage of accidents of groups involved in pipeline incidents (1970-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 49 of 56
Figure 55: Percentage fatalities of accidents of groups as a function of leak size (1970-
2019)

Although the occurrence of injuries and fatalities is low, safety remains the highest priority for the
gas transmission companies.

3.4.4 Detection of incidents

Incidents are detected in different ways. Table 8 shows the distribution per detection type. People
directly involved with the transmission networks, like patrol, contractors and staff, are the most
common detector (approximately 42% of the incidents). In the period 1970-2019, 15% of the
incidents were detected by the patrols, 15% by contractors and 10% by staff.
Public also detect a significant part of the incidents. In the period between 1970-2019 public detected
35% of the incidents. In the last 10 years (2010-2019) this number decreased to approximately
17%. The percentage of incidents detected by landowners and others has increased in the last ten
years.

Incident distribution Incident distribution

Detection 1970 – 2019 2010–2019
[%] [%]

Public 34.6 16.8

Gas company 42.2 50.5

Unknown 6.7 0.5

Landowner 5.5 14.1

Distribution company 4.8 4.3

Other 6.2 13.6

Table 8: Detection of incidents

11th EGIG-report VA 20.0432

1970-2019 Page 50 of 56
Figure 56: Detection of incidents per leak size (2010-2019)

Figure 56 shows that most pinhole/cracks are detected by the gas company and public. Holes and
ruptures are mainly detected by the gas company, landowners and public.

11th EGIG-report VA 20.0432

1970-2019 Page 51 of 56
4 CONCLUSIONS

• The EGIG database is a valuable source of information on European gas pipelines and
incidents.
• EGIG has maintained and expanded the European Gas pipeline incident database. Seventeen
gas transmission system operators in Europe now collect incident data on more than 142,711
km of pipelines every year. The total exposure, which expresses the length of a pipeline and
its period of operation, is 4.84 million km·yr.
• In the EGIG database 1,411 pipeline incidents are recorded in the period from 1970-2019.
• The history of incidents collected in the database gives reliable failure frequencies. The
overall failure frequency over the period 1970-2019 is equal to 0.29 incidents per year per
1,000 km.
• The five year moving average failure frequency in 2019, which represents the average failure
frequency over the past 5 years, equals 0.126 per year per 1,000 km.
• The five year moving average and overall failure frequency have reduced consistently over
the years, although it has tended to stabilise over recent years.
• Incidents caused by external interference and ground movement are characterised by
potentially severe consequences. This emphasises their importance to pipeline operators and
authorities.
• Corrosion as a primary cause has now the same frequency rate as external interference,
although consequences are much less severe.
• Over the last ten years, external interference, corrosion, construction defects and ground
movement, represent 27%, 27%, 16% and 16% respectively of the pipeline incidents
reported.

11th EGIG-report VA 20.0432

1970-2019 Page 52 of 56
5 BIBLIOGRAPHY

API. (2012, December). 5L Specification for Line Pipe 45th edition.

EGIG. (2008). 7th Report of the European Gas Pipeline Incident Data Group, 1970-2007.
https://www.egig.eu/.
EGIG. (2011). 8th Report of the European Gas Pipeline Incident Data Group, 1970-2010.
https://www.egig.eu/.
EGIG. (2015). 9th Report of the European Gas Pipeline Incident Data Group, 1970-2013.
https://www.egig.eu/.
EGIG. (2018). 10th Report of the European Gas Pipeline Incident Data Group, 1970-2016.
https://www.egig.eu/.
Michael R. Acton, P. J. (2008). 7th International Pipeline Conference. Ignition Probability for High
Pressure Gas Transmission Pipelines. Calgary, Alberta, Canada, September 29–October 3,
2008: ASME.

11th EGIG-report VA 20.0432

1970-2019 Page 53 of 56
APPENDIX 1: Statistics

Primary failure frequencies over different time intervals

95% LL 95% UL
Primary
Primary Primary
Number Total system failure
failure failure
Period Interval [years] of exposure frequency
frequency per frequency
incidents ·106 km·yr per 1,000
1,000 km·yr per 1,000
km·yr
km·yr
1970 – 2007 7th report, 38 years 1,173 3.15 0.372 0.351 0.394
1970 – 2010 8th report, 41 years 1,249 3.55 0.352 0.333 0.372
1970 – 2013 9th report, 44 years 1,309 3.98 0.329 0.311 0.347
1970 – 2016 10th report, 47 years 1,366 4.41 0.310 0.294 0.327
1970 – 2019 11th report, 50 years 1,411 4.84 0.292 0.277 0.307
1980 – 2019 40 years 1,050 4.36 0.241 0.226 0.256
1990 – 2019 30 years 663 3.63 0.183 0.169 0.197
2000 – 2019 20 years 388 2.64 0.147 0.133 0.163
2010 – 2019 10 years 184 1.42 0.129 0.111 0.150
2015 – 2019 5 years 90 0.714 0.126 0.101 0.155

Table 9: Primary failure frequencies and confidence intervals over different time intervals

95% LL 95% UL
Primary failure Primary failure Primary failure
Leak size frequency per 1,000
km·yr frequency per frequency per
1,000 km·yr 1,000 km·yr
Unknown 0.003 0.000 0.010

Pinhole/crack 0.088 0.068 0.113

Hole 0.022 0.013 0.036

Rupture 0.013 0.006 0.024

Table 10: Primary failure frequencies and confidence intervals per leak size (period 2015–
2019)

95% LL 95% UL
Primary failure
Primary failure Primary failure
Cause frequency per
frequency per frequency per
1,000 km·yr
1,000 km·yr 1,000 km·yr
External interference 0.134 0.124 0.145
Corrosion 0.050 0.044 0.057
Construction defect / Material failure 0.048 0.042 0.055
Hot tap made by error 0.013 0.010 0.016
Ground movement 0.025 0.021 0.030

Table 11: Primary failure frequencies and confidence intervals per cause (1970-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 54 of 56
 95% LL 95% UL
Primary failure
Primary failure Primary failure
Cause frequency per
frequency per frequency per
1,000 km·yr
1,000 km·yr 1,000 km·yr
External interference 0.054 0.046 0.064
Corrosion 0.033 0.026 0.041
Construction defect / Material failure 0.020 0.015 0.026
Hot tap made by error 0.005 0.002 0.008
Ground movement 0.020 0.015 0.027

Table 12: Primary failure frequencies and confidence intervals per cause (2000-2019)

95% LL 95% UL
Primary failure
Primary failure Primary failure
Cause frequency per
frequency per frequency per
1,000 km·yr
1,000 km·yr 1,000 km·yr
External interference 0.035 0.026 0.046
Corrosion 0.034 0.025 0.046
Construction defect / Material failure 0.020 0.014 0.029
Hot tap made by error 0.002 0.000 0.006
Ground movement 0.020 0.014 0.029

Table 13: Primary failure frequencies and confidence intervals per cause (2010-2019)

95% LL 95% UL
Primary failure
Primary failure Primary failure
Cause frequency per
frequency per frequency per
1,000 km·yr
1,000 km·yr 1,000 km·yr
External interference 0.036 0.024 0.053
Corrosion 0.032 0.020 0.048
Construction defect / Material failure 0.015 0.008 0.028
Hot tap made by error 0.001 0.000 0.008
Ground movement 0.017 0.009 0.029

Table 14: Primary failure frequencies and confidence intervals per cause (2015-2019)

11th EGIG-report VA 20.0432

1970-2019 Page 55 of 56
APPENDIX 2: Poisson law

A confidence interval is made to take uncertainty into account. To calculate a confidence interval the
population is assumed to have a known distribution. The assumption is made that the number of
incidents follows Poisson’s law, also called law of rare events.

Exact Poisson confidence limits for the estimated rate are found as the Poisson means, for
distributions with the observed number of events and probabilities relevant to the chosen confidence
level, divided by time at risk. The relationship between the Poisson and chi-square distributions is
employed here

 2 
 
 2Y ,  
 
=  2 
Yi 2

 2 
 
 2(Y +1),1−  
 
=  2 
Yu 2

where Y is the observed number of events, Yl and Yu are lower and upper confidence limits for Y
respectively,  is the chi-square quantile for upper tail probability on  degrees of freedom.

REFERENCE
Ulm K. A simple method to calculate the confidence interval of a standardized mortality ratio.
American Journal of Epidemiology 1990;131(2):373-375.

11th EGIG-report VA 20.0432

1970-2019 Page 56 of 56