report no. 4/07

I

Performance of European cross-country oil pipelines Statistical summary of reported spillages in 2005 and since 1971 Prepared by the CONCAWE Oil Pipelines Management Group’s Special Task Force on oil pipeline spillages (OP/STF-1)

P.M. Davis J. Dubois A. Olcese F. Uhlig J-F. Larivé (Technical Coordinator) D.E. Martin (Consultant) Reproduction permitted with due acknowledgement © CONCAWE Brussels May 2007

report no. 4/07

II

ABSTRACT

CONCAWE has collected 35 years of performance data on Western European cross-country oil pipelines, which currently comprise 34.8 thousand km transporting 789 million m3 per year of crude oil and oil products through a pipeline inventory which was very similar to that in 2004. This report covers the performance of these pipelines in 2005 and also shows how the pipeline system reported on has developed. Incidents are analysed by cause and the effectiveness, cost and time for clean-up are recorded. The 2005 performance was better than average for the last 35 years. As in 2004, fewer accidents were caused by third party activities than previous years with mechanical failures the most common cause. The performance over the whole 35 years is analysed and comparisons made of the different causes of failure. The data on safety-related incidents are reported and the levels and trends of spillage incidence, gross and net spillage volumes and the significant features of individual cause categories: mechanical failure, operational, corrosion, natural hazard and third party. Most European pipeline spillages are shown to have been small and effects were generally localised and temporary. Moreover, integrity is shown to be on an improving trend with spillage frequency over the period reduced from 1.2 to 0.32 spillages per 1000 km of pipeline. The 2005 usage of intelligence pigs is also reported and compared to previous years.

KEYWORDS

Clean-up, CONCAWE, intelligence pig, oil spill, performance, pipeline, safety, soil pollution, spillage, statistics, trends, water pollution

INTERNET

This report is available as an Adobe pdf file on the CONCAWE website (www.concawe.org).

NOTE Considerable efforts have been made to assure the accuracy and reliability of the information contained in this publication. However, neither CONCAWE nor any company participating in CONCAWE can accept liability for any loss, damage or injury whatsoever resulting from the use of this information. This report does not necessarily represent the views of any company participating in CONCAWE.

report no. 4/07

III

CONTENTS Page

SUMMARY V

1. INTRODUCTION 1

2. PIPELINE INVENTORY, THROUGHPUT AND TRAFFIC 2 2.1. CRITERIA FOR INCLUSION IN THE SURVEY 2 2.2. REPORTING COMPANIES 2 2.3. INVENTORY DEVELOPMENTS 1971-2005 3 2.3.1. Pipeline service, length and diameter 3 2.3.2. Age distribution 4 2.4. THROUGHPUT AND TRAFFIC 5

3. PIPELINE SAFETY 7 3.1. FATALITIES AND INJURIES 7 3.2. FIRES 8

4. SPILLAGE PERFORMANCE IN THE LAST 5 YEARS 9 4.1. 2005 SPILLAGE INCIDENTS 9 4.1.1. Mechanical Failure 9 4.1.1.1. Construction Fault 9 4.1.1.2. Materials Fault 10 4.1.2. Operational 11 4.1.2.1. Systems Malfunction 11 4.1.2.2. Human Factors 11 4.1.3. Corrosion 11 4.1.3.1. External Corrosion 11 4.1.3.2. Internal Corrosion 12 4.1.4. Third party activity 12 4.1.4.1. Direct Damage - Accidental 12 4.1.4.2. Direct Damage - Malicious 13 4.1.4.3. Direct Damage - Incidental 13 4.2. 2001-2005 SPILLAGE OVERVIEW 13

5. HISTORICAL ANALYSIS OF SPILLAGES 1971-2005 15 5.1. NUMBERS AND FREQUENCY 15 5.2. SPILLAGE VOLUMES 18 5.2.1. Aggregated annual spilled volumes 18 5.2.2. Spillage volume per event 21 5.3. HOLE SIZE 23 5.4. ENVIRONMENTAL IMPACT 25 5.4.1. Location of spillages 25 5.4.2. Ground area affected 25 5.4.3. Impact on water bodies 26 5.5. SPILLAGE DISCOVERY 26

6. DETAILED ANALYSIS OF SPILLAGE CAUSES 28 6.1. MECHANICAL FAILURE 28 6.2. OPERATIONAL 29 6.3. CORROSION AND IMPACT OF AGEING 29 6.4. NATURAL HAZARD 31 6.5. THIRD PARTY 31 6.5.1. Accidental damage 32

report no. 4/07

IV

6.5.2. Intentional damage 35 6.5.3. Incidental damage 36

7. INTELLIGENCE PIG INSPECTIONS 37 7.1. INTELLIGENCE PIG INSPECTION ACTIVITY 37 7.2. ACTIVITY IN 2005 37 7.3. ACTIVITY SINCE 1971 37 7.4. REPEAT INSPECTIONS 39

8. REFERENCES 41

APPENDIX 1 DEFINITIONS 42

APPENDIX 2 SPILLAGE SUMMARY 43

report no. 4/07

V

SUMMARY

CONCAWE has collected data over a 35-year period on the performance of cross-country oil pipelines in Western Europe with particular regard to spillages of 1 m3 or more, the clean-up carried out and the environmental consequences. The results have been published in annual reports since 1971. This report covers both the results for the year 2005 and the analysis of the accumulated results for the whole 35 year period from 1971 to 2005.

Approximately seventy companies and other bodies operating oil pipelines in Europe currently provide statistics for the CONCAWE annual report on the performance of cross-country oil pipelines. These organisations operate some 250 different service pipelines, which, at the end of 2005, had a combined length of 34,826 km, slightly shorter than that for 2004 although the difference is mainly due to corrections to the reported data. The volume transported in 2005 was 789 Mm3 of crude oil and refined products, which is 7% less than in 2004. Total traffic volume in 2005 amounted to 127 x 109 m3 x km, 10.6% less than in 2004.

There were 11 reported oil spillages from pipelines during 2005. There were no associated fires, fatalities or injuries. The gross spillage was 511 m3, equivalent to 0.65 parts per million (ppm) of the total volume transported. A total of 407 m3, i.e. 80% of the spillage was recovered or safely disposed of. The net oil loss into the environment amounted therefore to 105 m3, or 0.13 ppm. Of the spillages, five resulted from mechanical failure, two from operational causes, two from corrosion and the last two resulting from third party activities.

This report also provides comparative data for the five-year period between 2001 and 2005 and for all reported incidents since 1971. In terms of numbers of spillages, the 2005 performance was slightly better than average with eleven spillages compared to the long-term average of 12.5 and 11.2 for the period 2001 to 2005. Moreover, the system length is now much longer than in earlier years (reported length in 1971 of 12,800 km). This means that the spillage frequency of 0.32 spillages per 1000 km per year was the same as the average over the last five years but less than the long term average frequency of 0.52. The performance was also good in terms of volume spilled. The gross spillage volume per 1000 km of pipeline was 14.7 m3 per 1000 km compared to the long-term average of 89 m3 per 1000 km. Also included is the record of intelligence pig inspection activity in 2005 and the records since the technique was first used. In 2005, 109 inspections were reported using some sort of intelligence pigs, covering over 6000 km of pipeline, the greatest length of reported inspections in any year.

Pipelines constitute one of the main means of oil transport in Europe and are considered to be one of the safest. Whereas major and sometimes repeated accidents with large media exposure have occurred with road, rail and sea transportation, nothing similar has happened with oil pipelines. Almost inevitably though, with such a massive undertaking operating for 35 years, a handful of incidents has occurred that have resulted in a small number of fatal injuries and fires.

The system is ageing. Whereas in 1971 70% of it was 10 years old or less, by 2005 only 7% was 10 years old or less and 35% was over 40 years old. However, this so far does not appear to have led to any increase in spillages.

report no. 4/07

VI

Most pipeline spillages are very small and just over 5% of the spillages are responsible for 50% of the gross volume spilled. Pipelines carrying hot oils such as fuel oil have in the past suffered very severely from external corrosion due to design and construction problems. Many have been shut down or switched to cold service. The great majority of pipelines now carry unheated petroleum products and crude oil.

The two most important causes of spillages have been third party incidents and mechanical failure, with corrosion well back in third place and operational and natural hazards making minor contributions. Third party accident frequency has been significantly reduced progressively over the years. However, after having made great progress reducing mechanical failure frequencies during the first 20 years, by the mid ‘90s it appeared that something of an upward trend could be setting in.

Overall there is no evidence to show that the ageing of the pipeline system poses any greater level of risk. The development and institution of new techniques, such as internal inspection using intelligence pigs, hold out the prospect that pipelines can continue reliable operations for the foreseeable future. Future monitoring of CONCAWE pipeline performance statistics will be necessary to confirm the position.

report no. 4/07

1

1. INTRODUCTION

The CONCAWE Oil Pipelines Management Group (OPMG) has collected data on the safety and environmental performance of oil pipelines in Europe since 1971. Information on annual throughput and traffic, spillage incidents and intelligent pig inspection activities are gathered yearly by CONCAWE via questionnaires sent out to the pipeline operating companies early in the year following the reporting year.

CONCAWE has set a minimum spillage size at 1 m3 for reporting purposes (unless there are exceptionally serious safety or environmental consequences to be reported for a < 1 m3 spill). Direct comparison of different spillage data sets therefore need to include careful consideration of the spill cut-off size.

The results have been analysed and published annually in a series of annual reports [1,2] and in two summary reports [3,4] covering the years 1971 to 1995 and 1971 to 2000 respectively. CONCAWE has also held a number of seminars to disseminate information throughout the oil pipeline industry on the developments in techniques available to pipeline companies to help improve the safety, reliability and integrity of their operations. These seminars have included reviews of spillage and clean-up performances to cross communicate experiences so that all can learn from each other’s incidents.

Aggregation and statistical analysis of the performance data provides objective evidence of the trends and focuses attention on existing or potential problem areas which helps operators to set priorities for future efforts.

The format and content of this report has been altered to include not only the 2005 performance, but also a full statistical analysis of the last 5 years and of the whole 35 years period from 1971 as well as essential information on all spillage incidents in the database. From this year on, the format will be kept the same and each yearly report will supersede the previous one. As usual, a general map of European land-based oil pipelines is attached to this report. Link to map of Refineries and Oil Pipelines in Western Europe.

Section 2 gives details of the pipeline inventory covered by the survey (length, diameter, type of product transported) and how this has developed over the years. Throughput and traffic data is also included.

Section 3 focuses on safety performance i.e. the number of fatalities and injuries associated with pipeline failure incidents.

Section 4 gives a detailed analysis of the spillage incidents in 2005 and of all incidents over the last 5 years. Section 5 analyses spillage incidents for the whole reporting period since 1971 while section 6 provides a more detailed analysis of the causes of spillage.

Finally section 7 gives an account of intelligence pig inspections.

report no. 4/07

2

2. PIPELINE INVENTORY, THROUGHPUT AND TRAFFIC

2.1. CRITERIA FOR INCLUSION IN THE SURVEY

The definition of pipelines to be included in the CONCAWE inventory has remained unchanged since 1971. These are pipelines

• Used for transporting crude oil or petroleum products,

• With a length of 2 km or more in the public domain,

• Running cross-country, including short estuary or river crossings but excluding under-sea pipeline systems. In particular, lines serving offshore crude oil production facilities and offshore tanker loading/discharge facilities are excluded.

• Pump station and intermediate storage facilities are included but origin and destination terminal facilities and tank farms are excluded.

The geographical region covered was originally consistent with CONCAWE’s original terms of reference i.e. OECD Western Europe, which then included 19 member countries. However, Turkey was never covered. From 1971 to 1987, only pipelines owned by oil industry companies were included but from 1988, non-commercially owned pipeline systems (essentially NATO) were brought into the inventory. Following the reunification of Germany, the pipelines in the former East Germany (DDR) were added to the database from 1991. This was followed by crude and product lines in the Czech Republic and Hungary and crude lines in Slovakia in 2001 and finally Slovakian product lines in 2003.

Although CONCAWE cannot guarantee that every single line meeting the above criteria is actually covered, it is believed that most such lines operated in the reporting countries are included.

It should be noted that all data recorded in this report and used for comparisons or statistical analysis relate to the inventory reported on in each particular year and not to the actual total inventory in operation at the time. Thus, year on year performance comparisons must be approached with caution and relative numbers (e.g. per 1000 km of line) are more meaningful than absolute ones.

2.2. REPORTING COMPANIES

In 2005, 70 operating companies reported results. Some affiliates and other operating entities of certain large companies are counted individually in these numbers. The number of companies / non-industry bodies reporting data to CONCAWE in 1971 was unrecorded but by 1980 approximately 70 companies participated in the CONCAWE survey. This number has remained more or less constant although several new companies have taken over pipelines, but others have closed down or amalgamated with others.

report no. 4/07

3

2.3. INVENTORY DEVELOPMENTS 1971-2005

2.3.1. Pipeline service, length and diameter

Currently, there are some 250 pipeline systems recorded in the CONCAWE database, reported in terms of some 664 discrete sections. The sections are further classified according to their service i.e. the type of product transported, for which we distinguish crude oil, white products, black products transported in heated lines and other products. A few pipelines transport both crude oil and products. Although these are categorised separately in the database they are considered to be in the crude oil category for aggregation purposes. Finally some lines may be out of service in a certain year without being permanently retired.

The CONCAWE survey covered 34,841 km in 2005. As shown in Figure 1, this figure has been fairly stable in recent years. Most of the major pipelines were built in the 60s and 70s. The two historical step increases in the "CONCAWE" inventory occurred when systems previously not accounted for in the survey were added. In the late 80s the majority of the NATO pipelines were included and in the early part of this decade a number of former Eastern bloc systems joined the survey. The increase was mostly in the "products" category, the main addition in the crude oil category being the Friendship or "Druzba" system that feeds Russian crude into Eastern European refineries

Figure 1 CONCAWE pipeline inventory and main service categories

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005

'000

km

TotalCrudeWhite productsHOTx10

report no. 4/07

4

There are three pipeline service populations: crude oil, petroleum products (white oils), both transported at ambient temperature and oils transported at elevated temperature comprising hot crude oil, lubricating oils and heavy fuel oils (black oils). These three populations are referred to as crude, product and "hot" in this report.

Figure 1 shows that the first two categories represent the bulk of the total inventory. Between 1971 and 2005 some 166 pipeline sections totalling 7150 km were permanently shutdown. Of these, 24 sections totalling 360 km were in hot service. This is a much larger proportion of the hot inventory than for the other services and reflects the decline in the heavy fuel oil business since the mid 1970s as well as specific action taken by operating companies because of the corrosion problems and generally poor reliability experienced with several of these pipelines (see section 5.1).

Figure 2 shows the diameter distribution in 2005 for each category of use. In general, the crude pipelines are significantly larger than the other two categories. Some 88% of the crude pipelines are 16” (400 mm) or greater up to a maximum of 48” (1200 mm) whereas around 85% of the product and some 98% of the hot pipelines are less than 16”. The smallest diameter product pipelines are typically 6” (150 mm) although a very small number go down to 3” (75 mm).

Figure 2 Pipeline diameter distribution and service in 2005

0%

20%

40%

60%

80%

100%

Crude White products HOT

>=3024 - <3016 - <2412 - <168 - <12>8

2.3.2. Age distribution

When the CONCAWE survey was first performed in 1971, the pipeline system was comparatively new with some 70% being 10 years old or less. Although the age distribution was quite wide, the oldest pipelines in the 26-30 year age bracket represented only a tiny fraction of the inventory.

report no. 4/07

5

Over the years, a number of new pipelines have been commissioned while older ones were taken out of service. As mentioned above existing lines were also added to the inventory at various stages, contributing their specific age profile. Although some short sections may have been renewed, there has been no large scale replacement of existing lines. The development of the overall age profile is shown in Figure 3.

The system has clearly been progressively ageing. By 2005, only some 2500 km, i.e. 7.2% of the total, was 10 years old or less and some 12,300 km (36%) was over 40 years old. The impact of age on spillage performance is discussed in section 6.3.

Figure 3 Pipeline age distribution

0%

20%

40%

60%

80%

100%

1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 20050.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

'000

km

1-1011-2021-3031-4041-5051-6060+

2.4. THROUGHPUT AND TRAFFIC

In total, 789 Mm3 of crude oil and refined products was transported through the pipeline system in 2005, a 7% decrease from 2004. It should be realised however, that this figure is only indicative. Large volumes of both crude and products pass through more than one pipeline, and whilst every effort is made to only count the flow once, the complexity of some pipeline systems is such that it is often difficult to estimate what went where. Indeed, there are a few pipelines where the flow can be in either direction. A more meaningful figure is the traffic volume which is the flow-rate times the distance travelled. This is not affected by how many different pipelines each parcel of oil is pumped through. In 2005, the total traffic volume was 127 x 109 m3 x km, a 10.6% decrease from 2004. For products alone however, the throughput was 260 Mm3 and the traffic volume 37.4 x 109 m3 x km, both of which are very similar to the 2004 figures of 257 Mm3 and 37.6 x 109 m3 x km.

report no. 4/07

6

Throughput and traffic are not significant factors for pipeline failures or spillages. Although higher flow rates may lead to higher pressure, line deterioration through metal fatigue is more directly related to pressure cycles than to the absolute pressure level (as long as it remains within design limits). These figures are, however, useful as a divider to express spillage volumes in relative terms (e.g. as a fraction of throughput, see section 4), providing figures that can be compared with the performance of other modes of oil transportation.

report no. 4/07

7

3. PIPELINE SAFETY

The CONCAWE database includes records of fatalities, injuries and fires related to spillages. There were no such occurrences in 2005, neither were there any incidents involving intentional damage.

3.1. FATALITIES AND INJURIES

Over the 35 reporting years there have been a total of 14 fatalities in five separate incidents in 1975, 79, 89, 96 and 99. All but one of the fatalities occurred when people were caught up in a fire following a spillage.

In three of these four fire cases the ignition was a delayed event hours or days after the detection and demarcation of the spillage area had taken place. In one incident involving a spillage of chemical feedstock naphtha 3 bystanders were engulfed in fire, having themselves probably been the cause of ignition. In another incident ignition of spilled crude oil occurred during attempts to repair the damaged pipeline. The repairers escaped but the spread of the fire caught 4 people who had entered inside the marked spillage boundary some distance away. The third incident also involved a maintenance crew (5 people) carrying out repair activities following a crude oil spill, none of whom escaped. These fatalities all occurred after the spillage flows had been stemmed, during the subsequent incident management and reinstatement period. It appears that the spillages themselves did not cause the fatalities. Stronger management of spillage area security and working procedures might well have prevented the fires and fatalities.

In just one case, fire ensued almost immediately when a bulldozer doing construction work hit and ruptured a gasoline pipeline. A truck driver engaged in the works received fatal injuries.

The single non-fire fatality was a person engaged in a theft attempt who was unable to escape from a pit which he had dug to expose and drill into the pipeline. This caused a leak that filled the pit with product in which the person drowned.

It is apparent that the casualties were not members of the general public going about their normal activities in locations where they should have been allowed to be at the time. Thus these occurrences should not be used out of context for any assessment of societal risk inherent to oil pipeline operations.

Two spillage reports recorded a single non-fatal injury. Both resulted from inhalation / ingestion of oil spray/aerosol.

There has been no reported fatality or injury since 1999.

report no. 4/07

8

3.2. FIRES

Apart from those mentioned above, five other fires are on record:

• A large crude oil spill near a motorway probably ignited by the traffic.

• A gasoline theft attempt in an untypical section of pipeline located on a pipe bridge. The thieves may have deliberately ignited it.

• A slow leak in a crude production line in a remote country area found to be burning when discovered. It could have been ignited purposely to limit the pollution.

• A tractor and plough that had caused a gasoline spill caught fire, which also damaged a house and a railway line.

• A mechanical digger damaged a gasoline pipeline and also an electricity cable, which ignited the spill.

There were no casualties reported in any of these incidents.

report no. 4/07

9

4. SPILLAGE PERFORMANCE IN THE LAST 5 YEARS

4.1. 2005 SPILLAGE INCIDENTS

A total of 11 spillage incidents were recorded in 2005. Table 1 gives a summary of the main causes and spilled volumes and environmental impact. For definition of categories of causes and gross/net spilled volume, see Appendix 1.

Table 1 Summary of causes and spilled volumes for 2005 incidents

Event Location Line size Product Injury Fire(") spilled Fatality Gross Net loss Ground area Water

(1) (2) m2 (3)

Mechanical failure426 Pump station 12 Diesel fuel - - 18.7 18.7 NA427 Line 12 Jet fuel - - NA NA NA G 428 Line 20 Crude oil - - 350.0 10.0 15000 G 429 Line 6 Jet fuel - - 20.0 0.0 58 S 430 Line 6 Jet fuel - - 38.0 0.0 42 S

Operational431 Line 10 Crude oil - - 30.0 4.2 1000 G 432 Line 8 Jet fuel - - 15.0 0.0 1000

Corrosion433 Line 10 Jet fuel - - 3.0 0.6 50 S 434 Line 24 Crude oil - - 64.0 63.0 150 G

Third party activity435 Line 8 Jet fuel - - 15.0 8.0 1000 G 436 Line 24 White prod. - - 0.5 0.0 3000 S G

Total 554.2 104.5(1) Spillage events are numbered from the beginning of the survey in 1971(2) I = Injury, F = Fatality(3) S = Surface water, G = Groundwater, P = Potable water

Spilled volume Contamination

(m3)

The circumstances of each spill including information on consequences remediation and cost are described in the next section according to cause. Further details are available in Appendix 2 which covers all spillage events recorded since 1971.

4.1.1. Mechanical Failure

There were five incidents resulting from mechanical failure, two caused by a construction fault and three resulting from materials faults.

4.1.1.1. Construction Fault

Event 426: A branch leading to a pressure relief valve on a 12” diesel fuel pipeline in a pump station fractured although this piece of pipe was only one and a half years old. The fracture was caused by vibration in a section of pipe which was not adequately supported. Some 18.7 m3 of diesel fuel was spilled in an area of vines and orchards. The leak was detected almost immediately by the pipeline operator and the pipeline shut in. The spilled oil was absorbed using sepiolite so that the final state of the ground was similar to that before the failure. Some 1500 € damage was caused to nearby vines. The costs of repairs and clean-up were minimal at only 2000 €.

report no. 4/07

10

Event 427: A 12” pipeline transporting jet fuel was exposed by a third party for them to carry out construction work in an industrial area. They reported that the pipeline was leaking and it was immediately shut in. The leak was found to be in a welded joint and was caused by a defective weld. The length of time that the pipeline had been leaking is unknown as is the volume of oil spilled although it is believed to be more than 1 m3. The pipeline was shut in for 2 days while repairs were carried out and discussions are ongoing as to the best way to clean up the pollution caused including to groundwater.

4.1.1.2. Materials Fault

Event 428: During normal operation of a 20” pipeline transporting crude oil, a sudden pressure drop was observed. The automatic leak detection system reacted within three minutes and a full manual shutdown was performed 22 minutes later. By the time the system had de-pressured, some 350 m3 of oil had been spilled in a corn field. The failed section of pipe was excavated, removed and replaced. Subsequent investigation revealed a crack of 36 mm by 1 mm caused by a hidden metallurgical defect from the pipe manufacture. To clean up the site, pits were dug down to the groundwater table and crude oil collected from the water surface. In all, some 15,000 m2 of ground was affected. 220 m3 of oil was collected as liquid and it is estimated that a further 120 m3 of oil was removed with contaminated soil leaving about 10 m3 net loss. Clean-up took one year and a cost of 2,800,000 €, most of which (2,500,000 €) was for the disposal of contaminated soil. The repairs to the pipeline cost 200,000 € and took 6 days.

Events 429/430: Two leaks occurred in the same pipeline, 7 km apart, within 3 months of each other. Both incidents were caused by cracks forming in the defective pipe. In the first, the leak was detected by routine monitoring by the operator and the line rapidly shutdown. A crack 12 mm long was found and it was estimated that 20 m3 of diesel had leaked into arable land and contaminated about 60 m2 of land. The clean up involved the removal of nearly 4000 tonne of soil and took 75 days at a cost of 420,000 €. Repairs to the pipeline cost a further 7,500 €. There was slight contamination of surface water. In the second incident, diesel oil was reported by a third party to be leaking into a meadow, a ditch and two ponds. On investigation, a crack 7 mm long was found and it was estimated that 38 m3 had leaked out and contaminated some 42 m2 of land. The clean up involved the removal and disposal of 2200 tonne of soil, took 21 days at a cost of 160,000 €. Repairs to the pipeline cost 6,000 €. The pipeline had been inspected by a metal loss pig a month before the first incident. This had revealed a large number of anomalies and a programme to remediate these is under way.

report no. 4/07

11

4.1.2. Operational

There were two incidents caused by operational factors, both attributable to human factors.

4.1.2.1. Systems Malfunction

There were no incidents in this category in 2005.

4.1.2.2. Human Factors

Event 431: A 10” crude oil pipeline was shut in for maintenance operations. The pipeline was not completely emptied of crude oil so that when work was being carried out on a valve, oil spilled out. This came out of the valve pit through the concrete walls surrounding the valve. Around 15 m3 of oil was spilled of which 10 m3 was recovered as liquid, the remaining 5 m3 being removed with contaminated soil for disposal. Around 1000 m2 of ground was affected. Repairs to the pipeline were not necessary. Clean-up took nearly four months at a cost of 150,000 € with a further 400,000 € for disposal of contaminated soil.

Event 432: A trench was being dug to access an 8.5” pipeline transporting jet fuel for maintenance purposes. The site was near a road crossing where the pipeline was at 5 m below ground. So as to make the excavation safe, sheet steel piling was being driven to support the trench. The position and depth of the pipeline had been identified and the equipment operator had also been made aware of these and measures had been agreed to protect the pipe. Nevertheless, the pipeline was holed by the sheet piling. The incident occurred in agricultural land and the piling made a 350 mm x 2 mm gash in the pipe. The incident was detected by the automatic detection system but by the time the pipeline had been shut in, some 30 m3 of oil had been spilled. This impacted groundwater but not surface or drinking water supplies and some 1000 m2 of soil was contaminated. As well as removal of contaminated soil, wells have been drilled and pumping undertaken to depress the groundwater surface so that oil can be recovered. This process was still in operation after 6 months, by which time the majority of the oil had been recovered, i.e. 25 m3 by pumping, and 0.8 m3 in contaminated soil. The net loss is thus some 4.2 m3. Costs so far have been 10,000 € for repairs to the pipeline, 50,000 € for the bore hole pumping and 940,000 € for disposal of contaminated soil making a total cost of 1,000,000 €.

4.1.3. Corrosion

There were two incidents resulting from corrosion, one each from internal and external corrosion.

4.1.3.1. External Corrosion

Event 433: While pressure testing a 10” pipeline with jet fuel, a flange blew out in a depot and this was presumed to be the cause of the pressure drop. The line was depressured to repair the flange but 10 days later, a third party reported oil coming to the surface in a derelict oil depot. It was estimated that approximately 3 m3 of oil was spilt. The

report no. 4/07

12

area of ground affected by oil was about 50 m2 and some 15 m3 of oily soil was removed for disposal. Ten days after the spill, oil was observed on the surface of water in a ditch adjacent to the pipeline and this was reported to the authorities. Absorbents were used in the ditch and drainage sumps were dug between the pipeline and ditch and oil was collected from these by vacuum tanker. By these means, about 2.4 m3 was recovered as liquid leaving a net loss of 0.6 m3. Excavation revealed a 3mm by 3mm hole caused by external corrosion under the coal tar coating. A clamp has been fitted to the pipeline as a temporary repair but full replacement of the defective section is planned. The temporary repair cost 100,000 €, the clean-up cost 120,000 € and took three and a half months; disposal of contaminated soil cost 5,000 € giving a total cost of 225,000 €.

4.1.3.2. Internal Corrosion

Event 434: A leak of crude oil from a pipeline in a terminal was discovered by the pipeline staff when oil appeared on the surface of the ground. The leak was on a length of 24” pipe connecting two main pipelines. However, it had been out of service since 1986 when one of the main pipelines had been mothballed. The routine annual pressure testing of the whole system the day before the leak was detected presumably caused the failure of a corrosion defect which had been forming for some time. In the area of the leak, the pipeline is actually lying in the water table. Therefore, the oil could not soak down into the ground. As a result, almost all of the 64 m3 spilled could be recovered by pumping and 62.5 m3 was collected in this way. It is estimated that a further 0.5 m3 was disposed of with contaminated soil leaving perhaps only 1 m3 lost to the ground, all within the confines of the terminal. About 150 m2 of ground was affected. Repairs to the pipeline cost 120,000 € and took one day. The clean-up cost 50,000 € with a further 200,000 € for disposal of contaminated soil. Total costs were 370,000 €.

4.1.4. Third party activity

There were two incidents resulting from third party activity, one in the accidental damage category and the other in the incidental damage category.

4.1.4.1. Direct Damage - Accidental

Event 436: An excavator operating without authority over the route of an 8” pipeline, and despite near-by pipeline markers, dug a trench just over the pipeline. The blade carved a gouge 10 cm long by 1 cm wide in the top of the pipe. The pipeline immediately ruptured, but fortunately, the point of impact was only 100 m upstream of a non-return valve. The automatic oil spill protection system immediately shut down the pipeline thus limiting the spillage to 15 m3 of jet fuel. However, the spillage was in a very sensitive area inside a National Park. Of the spilled oil, 7 m3 was recovered giving a net loss of 8 m3 to soil and groundwater. The area of ground affected was 1000 m2. The pipeline was repaired and returned to service within one day. This cost 200,000 €, the initial clean-up cost 1,850,000 €, disposal of contaminated soil cost 406,000 € giving a total cost of 2,456,000 €. However, the final clean-up remains to be done and the plans for this have to be approved by the Authorities. It is proposed that this will include a skimmer system and “bio-sparging” to recover pollution from the water table and further removal of contaminated soil.

report no. 4/07

13

4.1.4.2. Direct Damage - Malicious

There were no incidents in this category in 2005.

4.1.4.3. Direct Damage - Incidental

Event 437: A landowner observed oil in a meadow and reported this to the pipeline operator. The pipeline was shut in and the pipeline exposed when it was found that there was a drain which had been installed crossing over the pipeline. The drain laying machine had cut a notch in the pipe and this had developed into a crack some 800 mm long. It was estimated that only 0.45 m3 of product had escaped. This would normally not be included in the CONCAWE statistics but in this case, there were significant environmental impacts. Because of the high water table, both ground and surface water were affected and some 3000 m2 of ground contaminated. 1600 tonnes of soil were removed and 300 tonnes of water removed and a further 4400 m3 of water treated in an oil water separator. As a result, all traces of oil were removed. Repairs to the pipeline took 35 hours and cost 120,000 €. The clean-up cost 170,000 € and took 80 days.

4.2. 2001-2005 SPILLAGE OVERVIEW

2005 was only slightly better than average for spillages with eleven spillages reported compared to the average since CONCAWE records began of 12.5 per year and 11.5 over the five year period 2001 to 2005.

Table 2 shows the spillage performance for the 2001-2005 5-year period. Of the 57 spillages recorded for the period, 54 caused some temporary environmental pollution. 10 spillages affected surface waters and 9 affected groundwater but only one of these had any impact on potable water supplies.

In terms of spillage volumes, 2005 was much better than the average for 2001 to 2005 with a total of 511 m3 gross spillage (2001-2005 average 1363 m3) and a net loss of 105 m3 (2001-2005 average 457 m3) although volumes for one spillage were not known. The gross spillage volume was 14.7 m3 per 1000 km (long term average 89 m3 per 1000 km) and represents 0.65 ppm of the annual throughput. Compared to the average for these five years, 2005 had more spillages caused by mechanical failures and operational failures but less caused by third parties.

Although some cost data is included for most spill events, we do not feel it is appropriate to try and compare such figures. Cost data, particularly remediation costs, are very case-specific depending on local circumstances, topography, geo/hydro-logy as well as local legislation. They are not always fully reported for e.g. confidentiality of legal reasons. Also remediation costs can occur over long periods of time and are not always fully known for recent spills. Attempting to draw statistical conclusions from such data would be futile and possibly misleading.

report no. 4/07

14

Table 2 Five-year comparison by cause, volume and impact: 2001 – 2005

2001 2002 2003 2004 2005 2001‑ 2005Combined Length km x 103 34.9 34.8 35.4 34.6 34.9 34.9Combined Throughput m3 x 106 708 724 817 847 789Combined traffic volume m3 x km x 109 131 125 143 142 127Spillage incidents 15 14 12 5 11 57MECHANICAL FAILURE Construction 2 1 2 2 7 Material 3 1 1 3 8OPERATIONAL System 2 2 HumanCORROSION External 2 5 1 1 9 Internal 1 1 2 Stress corrosion cracking 1 1NATURAL HAZARD Subsidence Flooding Other 1 1THIRD PARTY ACTIVITY Accidental 3 4 4 2 1 14 Malicious 4 1 4 9 Incidental 1 2 1 4Volume spilled m3 * * AverageGross spillage 1150 2185 2834 138 554 1372Net loss 180 318 1309 26 105 388Average gross loss / incident 77 156 236 34 55 120Average net loss / incident 12 23 109 7 10 34Average gross loss/1000 km 33 63 80 4 16 `Average net loss/1000 km 5 9 37 1 3 11Gross spillage per causeMechanical failure 853 10 30 48 427 273Operational 0 0 0 0 45 9Corrosion 113 493 2 0 67 135Natural hazard 0 250 0 0 0 50Third party activity 184 1432 2802 90 15 905Net loss distribution(No of incidents)

< 10 11 7 4 3 2 2711 -100 4 6 7 1 7 25

101- 1000 1 1 2> 1000 m3 1 1

Environmental impactNONE 2 1 3SOIL < 1000 m2 13 5 10 1 5 34 > 1000 m2 2 7 2 4 5 20WATER BODIES Surface Water 3 4 3 10 Groundwater 1 1 1 6 9POTABLE WATER 1 1 2* Volume from one spillage incident not reported

report no. 4/07

15

5. HISTORICAL ANALYSIS OF SPILLAGES 1971-2005

5.1. NUMBERS AND FREQUENCY

Over the 35 year survey period there have been 436 spillage incidents i.e. an average of 12.5 per year. 67 of these spillages occurred in "hot" pipelines, a disproportionately large proportion in regard of the share of such pipelines in the total inventory.

Figure 4 shows the number of spillages per year, moving average and 5-year average trends over the 35 years since 1971 and for all pipelines. There is a clear long-term downward trend which bears witness to the industry improved control of pipeline integrity. The overall moving average has reduced about 18 spillages per year in the early 1970s to 12.5 by 2005. The moving average increases in the late ‘80s to early ‘90s and again in the early 2000 are partly linked to the additions to the pipeline inventory monitored. The largest number of spillages recorded in any one year was 21 in 1972 and the smallest number was 5 in 2004.

Figure 4 35-year trend of the annual number of spillages (all pipelines)

0

5

10

15

20

25

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Spill

ages

per

yea

r

YearlyRunning average5-year moving average

The large changes in the inventory monitored by CONCAWE over the year clearly make the absolute numbers difficult to interpret. The spillage frequency i.e. number of spills per unit length of pipeline is therefore a better metric. Figure 5 shows the same data as figure 4, now expressed in spillages per 1000 km of pipeline and the steady downward trend appears much more clearly. The 5-year frequency moving average has been reduced from around 1.1 in the mid 70s to around 0.3 spills per year and per 1000 km of pipeline today.

report no. 4/07

16

Figure 5 35-year trend of the spillage frequency (all pipelines)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Spill

ages

per

yea

r per

'000

km

YearlyRunning average5-year moving average

These overall figures mask the poor performance of hot pipelines (related to corrosion issues, see section 5.1), particularly in the early part of the period. This is illustrated in Figure 6 which compares the spillage frequency for hot and cold pipelines.

Figure 6 5-year moving average of spillage frequency (hot and cold pipelines)

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Spill

ages

per

yea

r per

'000

km

HOTCold x 10

report no. 4/07

17

Clearly, the cold and the hot pipelines have demonstrated entirely different behaviours. Figures 7 & 8 show the evolution over 5-year periods of the spillage frequency for hot and cold pipelines respectively, now broken down into main cause.

The hot pipeline spillage frequency starts from a much higher base than is the case for the cold pipelines, with a very large proportion of failures due to corrosion. In the 1970s and early ‘80s, due to design and construction deficiencies several hot pipelines suffered repeated external corrosion failures and they were shutdown or switched to clean (cold) product service. These actions have greatly contributed to the performance improvement which has been remarkable. The recent hot pipelines spillage frequency is still about on a par with what the product pipelines achieved back in 1971-75. It has to be noted that the statistical data have become less significant in recent years as the inventory of hot pipelines has steadily decreased. There was just one hot corrosion spillage in 2001-2005 from a now low total length of hot pipelines.

When the hot pipeline data are excluded, the cold pipelines show a somewhat slower improvement trend than the all pipelines data. Still the incidence of spillages has been reduced by two thirds over the last 35 years. This statistic best represents the performance improvement achieved by the operators of the bulk of the pipeline system.

Figure 7 Hot pipelines spillage frequencies by cause

0%

20%

40%

60%

80%

100%

1971 -1975

1976 -1980

1981 -1985

1986 -1990

1991 -1995

1996 -2000

2001 -2005

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0Sp

ills

per y

ear p

er '0

00 k

m

3rd partyNaturalCorrosionOperationalMechanicalAll causes

Albeit with fluctuations, the analysis by cause (Figure 8) shows that corrosion is a much less prevalent cause of failure for cold pipelines. There is a slight relative decrease of all causes except third party activities which has been somewhat increasing and is the most important cause of spillage. A more complete analysis of causes is given in section 6.

report no. 4/07

18

Figure 8 Cold pipelines spillage frequencies by cause

0%

20%

40%

60%

80%

100%

1971 -1975

1976 -1980

1981 -1985

1986 -1990

1991 -1995

1996 -2000

2001 -2005

0.0

0.2

0.4

0.6

0.8

1.0

Spill

s pe

r yea

r per

'000

km

3rd partyNaturalCorrosionOperationalMechanicalAll causes

5.2. SPILLAGE VOLUMES

5.2.1. Aggregated annual spilled volumes

Figure 9 shows the total gross spillage volume over the complete period, year by year and in terms of running and 5-year moving average. The same data is shown per 1000 km of pipeline in Figure 10 and as a proportion of throughput in Figure 11. Although there are fairly large year-to-year variations mostly due to a few very large spills that have occurred randomly over the years, the long-term trend is clearly downwards. Over the last 5 years, the gross pipeline spillage has averaged about 2 parts per million (ppm), or 0.0002%, of the oil transported.

report no. 4/07

19

Figure 9 Gross spillage volume

0

1000

2000

3000

4000

5000

6000

7000

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

m3

YearlyRunning average5-year moving average

Figure 10 Gross spillage volume per 1000 km

0

50

100

150

200

250

300

350

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

m3

YearlyRunning average5-year moving average

report no. 4/07

20

Figure 11 Gross yearly spillage volume as a proportion of throughput

0

2

4

6

8

10

12

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

ppm

Yearly total 5-year moving average

It might be expected that the trend in the differences between the annual gross volume spillage and the net volume spillage, i.e. the recovered spillage, would indicate the degree of success in improving clean-up performance. In practice this is not a very sound proposition. For one thing, maximum removal by excavation of spilled oil, which is biodegradable, is not necessarily the correct response to minimise environmental damage and this is now better understood than it once was. Another compounding factor is that the growth in the pipeline inventory has been predominantly for refined product pipelines and it can be assumed that less invasive recovery techniques are justified for white oil products than for fuel oil or crude oil to achieve a given visual and environmental standard of clean-up. The development of annual recovery percentages (gross-minus-net / gross) shown in Figure 12 indicates no significant trend. Over the whole period, the average recovery of the spilled oil is 56% leaving an average net loss of oil to the environment of 73 m3 per spill.

report no. 4/07

21

Figure 12 Spilled oil recovery

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

5.2.2. Spillage volume per event

The gross volume released is a measure of the severity of a spillage incident. Figure 13 shows that, beyond the large year-by-year variations, there is no clear long term trend in the average spill size by incident. In other words, the gradual reduction of the annual total spilled volume is related to the reduction of the number of spillage incidents rather than their severity. This is partly due to the mix of spillage causes changing over the years, e.g. the proportion of corrosion spillages, which on average are smaller ones, have decreased relative to third party spillages which are among the largest (see Figure 14).

The average annual figure in the last 5 years has consistently been around 100 m3

per spill compared to the long term average of 166 m3 per spill. It remains to be seen whether this improvement will continue but it can be expected that improved monitoring of pipelines and the generalised use of automated leak detection systems should lead to a reduction in spill sizes. There is insufficient data on record to establish any trend in the speed of detection or the response time to stem leakages.

Figure 14 shows the average spill size for each cause category. The largest spillages on average have resulted from mechanical failure, third party activities and natural hazards whereas operational problems and corrosion have caused smaller spills. As a rule of thumb, on average the three ‘largest spills’ categories result in spillages that are twice the size of the two ‘smallest spills' categories.

report no. 4/07

22

Figure 13 Yearly average gross spillage volume per event

0

50

100

150

200

250

300

350

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

m3

Figure 14 35-year average gross spillage volume per event by cause

0

50

100

150

200

250

Mechanical Operational Corrosion Natural 3rd party

m3

report no. 4/07

23

Figure 15 shows the distribution of spillage sizes, demonstrating that less than 20% of all spillages account for 80% of the cumulative volume spilled. Clearly a majority of the spillages recorded in the CONCAWE database were so small that they have only had a very limited and localised impact. This also highlights the importance of considering the cut-off spillage size before comparing data sets taken from different sources.

Figure 15 Distribution of Gross and net spillage sizes (over 35 years)

0%

20%

40%

60%

80%

100%

0% 20% 40% 60% 80% 100%

% of spillage events

% o

f cum

ulat

ive

volu

me

spill

ed/lo

st

Gross spillage volumeNet loss

5.3. HOLE SIZE

The following arbitrary definitions have been adopted for classifying hole size:

• Pinhole = less than 2 mm x 2 mm, • Fissure = 2 to 75 mm long x 10% max wide, • Hole = 2 to 75 mm long x 10% min wide, • Split = 75 to 1000 mm long x 10% max wide, • Rupture = >75 mm long x 10% min wide.

Out of the 436 pipeline spillages, hole size data is only available for 176 (40%). The corresponding statistics are shown in Table 3.

report no. 4/07

24

Table 3 Distribution of spillages by hole size

Pinhole Fissure Hole Split Rupture OverallNumber 22 26 74 27 44 193

11% 13% 38% 14% 23% 100%Surface area mm2

Average 0.8 43 460 8,419 128,405 30,634 min 0.01 0.5 2 16 1,600 Max 1 350 3600 81,000 3,150,000Average Gross spillage volume m3 65 274 97 130 284 164Spillage volume / Hole area m3/mm2 84 6.4 0.21 0.015 0.002 0.005Hole cause category Mech. Failure 9% 27% 12% 26% 11% 16% Operational 0% 4% 1% 11% 5% 4% Corrosion 77% 27% 28% 33% 10% 30% Natural hazard 5% 4% 0% 7% 2% 3% Third party 9% 38% 58% 22% 74% 48%Hole type by cause category Mech. Failure 7% 23% 30% 23% 17% 100% Operational 0% 14% 14% 43% 29% 100% Corrosion 29% 12% 36% 16% 7% 100% Natural hazard 20% 20% 0% 40% 20% 100% Third party 2% 11% 46% 6% 34% 100%

As expected, pinholes result in the smallest spillages and ruptures in the largest. For the other three categories, other factors are clearly more important as determinants of the spillage outcome.

Pinholes are nearly always caused by corrosion. Mechanical incidents often result in ruptures whilst operational and natural hazard incidents tend to cause more than their share of splits. Otherwise hole types follow similar patterns to the cause incidences.

A majority of mechanical, operational and natural hazard incidents cause the largest two types of hole whereas third party is equally divided and the corrosion preponderance is with the smaller hole types.

It would be expected that the larger the hole the larger on average the spillage would be, on the proviso that the pipeline was pumping i.e. not static at the time of the incident. The two rather obvious reasons for this are that higher leakage rates come out of larger holes and the hole sizes are to an extent related to the pipeline diameter which in turn tends to set the potential flow rate available for leakage. However, there are many other factors involved including the pressure in the pipeline, the length of time between the start of leakage and the leak being detected and the pipeline shut in, and the volume of pipe available to leak after shut in. The table above shows that there is indeed a weak relationship between the average gross spillage size and the hole size.

report no. 4/07

25

5.4. ENVIRONMENTAL IMPACT

5.4.1. Location of spillages

We differentiate between failures occurring either in a pipeline proper or in pumping stations and also record the type of land use in the area. Not surprisingly, most incidents (86%) occur in the pipeline themselves. The type of location has been reported for a total of 353 spillages.

Table 4 Location of spillage incidents

Number % Number %Commercial/Industrial 69 24% 52 78%Residential 15 5% 2 3%Agricultural 198 69% 13 19%Forest/Mountain 5 2% 0 0%Surface water 0 0% 0 0%Sub-total 287 67Unspecified 76 6Total 363 73

Pump station / manifoldPipeline

Whereas we do not have statistics of the length of pipeline installed for each land use type it is clear that the number of spillages in commercial and industrial areas is much higher than would be expected from consideration of installed length alone. Evidently, the vulnerability of the pipelines is significantly increased in such areas by a factor of possibly as much as ten compared to other areas. The bulk of the spillages from pump stations occur in industrial areas simply because their location is mostly classified as such.

5.4.2. Ground area affected

The current CONCAWE performance questionnaire, in use with minor changes since 1983, requests reporting of the area of ground (m2) affected by the spillage. Before that date, area data were reported infrequently. Out of the 436 recorded spillages, area data is available for 232 (53%). For these events, the percentages that fall within the area ranges are shown in Figure 16 together with the average spill size for each category.

If we exclude the one spillage that affected more than 100,000 m2, and for which the gross spillage was relatively modest, there is a direct relationship between spill size and area affected. Bigger spillage volumes affect larger areas.

This relationship is, however, to some extent fortuitous. There are two ways in which small spillage volumes can affect larger areas of ground. Fine sprays directed upwards can be spread around by winds. This factor tends to be more prevalent in the smaller area ranges. Other smaller spillages can be spread over larger areas by the influence of groundwater or surface water flows. This is the main mechanism by which relatively small spillages can affect very large areas. Conversely, comparatively large spillages, particularly those that occur over extended periods of

report no. 4/07

26

time and in the lower quadrants of the pipeline circumference, can have their main effect underground with relatively little impact on the surface. Porous ground and hot arid conditions can also lead to the surface consequences being limited.

Figure 16 Ground area (m2) affected by spillages (% of number reporting)

9.4%

17.6%

36.5%

24.9%

11.2%0.4%

<1010-99100-999 1000-9999 10,000-99,999 >=100,000

25 m3

198 m3

77 m3

42 m3

172 m3

698 m3Size of contaminated

area (m2)

5.4.3. Impact on water bodies

The spillage reports record the incidents where oil pollution of the water table and underground aquifers and surface watercourses has had consequences for the abstraction of potable water. Some 14 spillages, representing 3.2% of the total, have had some effect. It is believed that all of these effects have been temporary. For the last five years, impacts on other types of water have been reported. In the years 2001 to 2005, of the 57 reported spillages, 10 have affected surface water, 8 have affected ground water but only 2 have impacted potable water supplies.

5.5. SPILLAGE DISCOVERY

The way in which the occurrence of a spillage was detected is reported in nine categories (Table 5). The pattern for spillages from pump stations differs from that from pipelines.

The most common means of detection of pipeline spillages was third party passer-by (45%) who warned of spillages that on average were about 60% of the average size. Pipeline instrumentation, measurement and control systems were involved in detecting only 28% of the spillages.

Pipeline company resources detected some 84% of the pump station spillages. When third party have detected spillages, 16% of the total, the spills have tended on

report no. 4/07

27

average to be the smaller ones; presumably those that are below the warning capabilities of the instrumentation.

Table 5 Discovery of spillages

Number % Number %

m3 m3

Right of way survey by p/l staff 32 9% 232 0 0% 0Routine monitoring by p/l staff 69 19% 364 24 33% 111Maintenance staff 17 5% 187 21 29% 33Pressure testing 23 6% 129 1 1% 30Automatic detection system 34 9% 167 10 14% 49Third party 185 51% 127 17 23% 36Pipeline internal inspection survey 3 1% 6 0 0% 0Total 363 187 73 62

Pipeline Pump station / manifoldAverage

gross spillageAverage

gross spillage

report no. 4/07

28

6. DETAILED ANALYSIS OF SPILLAGE CAUSES

CONCAWE classifies spill causes into five major categories: mechanical failure, operational, corrosion, natural hazard and third party, themselves divided into sub-categories. Definitions are given in Appendix 1. The survey returns provide more details information on the actual cause and circumstances of failure and these are analysed in this section.

As already discussed in section 5, the main causes of failure are very different for hot and cold pipelines and this further illustrated in Figure 17. Whereas 82% of hot pipeline failures are related to corrosion, the figure is only 19% for cold pipelines for which mechanical and mostly third party related failures are the most prevalent.

Figure 17 Distribution of major spillage causes

Hot pipelines

6% 1%

82%

4%7%

MechanicalOperationalCorrosionNatural3rd party

Total: 67 incidents

Cold pipelines

28%

8%

19%3%

42%MechanicalOperationalCorrosionNatural3rd party

Total: 369 incidents

6.1. MECHANICAL FAILURE

There have been 106 mechanical failures, 24% of the total of 436 spillage events. This is an average of about 3 spillages per year. 40 failures were due to construction faults and 66 to material faults.

The most common causes of mechanical failures are illustrated in Figure 18.

It should be noted that by definition the dents that cause mechanical failure have to have been made during the pipeline’s construction. Dents made subsequently that eventually lead to spillages are categorised as third party.

report no. 4/07

29

Figure 18 Causes of mechanical failures

0

10

20

30

40

50

60

Pipeline Line Valves Flanges / Fittings Pump Station

No.

of S

pilla

ges Gasket / Gland

OtherFaulty Material

Weld Faults

Dents

OtherAbove Ground

Although there is no available figure on it, by far the greatest part of the material in place in the pipeline system is the underground pipe itself. The fact that only 53 out of the 106 spillages occurred in these underground sections indicates that the most vulnerable features from a mechanical standpoint are pipeline valves, flanges and other fittings and the pump stations. Adding seemingly useful features such as more section block valves, instrument connections, sampling systems can therefore potentially have a negative impact on spillage frequency. Wherever possible, these more vulnerable features should be designed out of the pipeline system.

There is no evidence of any increase in those mechanical failures that are potentially age-related, such as metal fatigue failures of pipelines under cycling pressure conditions. If any such pipelines exist, they are only a very small part of the inventory and the zero spillage record shows that no pipeline has reached an age where repeat failures are being experienced.

6.2. OPERATIONAL

There have been 31 spillages in this category (21 human errors and 10 system failures), i.e. just under 1 per year or 7% of the total. Except for their propensity to cause smallish sized spillages, there is no general trend apparent.

6.3. CORROSION AND IMPACT OF AGEING

There have been 125 spillages in this category i.e. 3.6 per year and 29% of the total. As noted earlier though, 54 of these occurred in the more vulnerable hot pipelines and in the early years. For cold pipelines corrosion causes represent only 19% of the total. The events have been subdivided into external and internal corrosion and, 10

report no. 4/07

30

years ago, stress corrosion cracking (SCC) was introduced as an extra category. The number of spillages in each sub-category and the resulting average spilled volumes are shown in Table 6.

Table 6 Corrosion-related spillages

Hot Cold AllGross Net loss

External 53 46 99 76 179Internal 1 21 22 179 546SCC 0 4 4 546 151

Average spilled volume

m3Number

Internal corrosion is much less prevalent than external corrosion. Some 73% of the cold pipeline internal corrosion incidents occurred in crude oil service although crude pipelines only account for less than a third of the cold pipeline inventory. Thus crude pipelines appear to be much more vulnerable to internal corrosion than product pipelines. Only one of the pipelines suffering a spill reported that inhibitor was used, one did not report and the other 20 did not use inhibitors.

External corrosion resulted in smaller sized spills than any of the other causes except for operational. The figure is higher for internal corrosion but it is heavily weighted by a single event where 2000 m3 were spilled.

Although there have only been four SCC-related spillages so far (including one re-categorised from external corrosion), they have been relatively large spillages, possibly as a result of the more severe failures resulting from this type of corrosion.

Out of all the possible combinations of pipeline service and corrosion categories three are particularly useful to highlight specific aspects: cold pipelines external and internal corrosion, hot pipelines external corrosion; and all pipelines internal corrosion.

In a gradually ageing pipeline inventory, increased occurrence of corrosion is a concern which is addressed by pipeline operators through the use of increasingly sophisticated inspections techniques. As already mentioned in section 5.1 the failure frequency of hot pipelines, mostly related to corrosion, has fallen dramatically over the years. Figure 19 shows no evidence of any increasing trend in corrosion failures of cold pipelines. If anything, the rate has decreased.

Out of the 71 corrosion-related failures in cold pipelines, 24 were related to special features such as road crossings, anchor points, sleeves, etc. which therefore appear particularly vulnerable.

It is anticipated that inspections using intelligence pigs should improve this situation. This should prevent any occurrence of ‘end of life takeoff’ in spillage numbers. Indeed, there is the strong prospect of further reducing corrosion spillage incidents by catching the corrosion before it gets too far advanced.

report no. 4/07

31

Figure 19 Corrosion-related spillage frequency (all types) for cold pipelines (5-year moving average)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Spill

ages

per

yea

r per

'000

km

There is therefore no evidence as yet to suggest that generalised corrosion is becoming a problem. There is, of course no guarantee that this will not start to happen at some point and thus there is a need for continued monitoring of performance on this basis. Inspection methods involving intelligence pigs are now available to monitor pipeline condition and early identification of the onset of corrosion. These techniques together with the general adoption of integrity management systems by all EU pipeline companies should ensure that any upturn in age-related spillages is prevented or delayed for many years.

6.4. NATURAL HAZARD

Natural hazards have caused only 15 spillages, 10 of which were due to landslides or subsidence, 2 to flooding and 3 to other hazards. This category contributes 3.44% of the total number of spillages. The resulting spillage volume was 2921 m3 gross, 104 m3 net, 3.4% of the gross spillage and 3.2% of the net spillage totals from all causes. The natural hazard- caused spill sizes are 195 m3 gross, 68 m3 net per spill, i.e. very close to the overall average spill sizes.

No less than 10 of the natural hazards spills have occurred in the same country. This appears to be a direct consequence of the difficult terrain and hydrological conditions that apply to a significant part of that country’s pipeline network.

6.5. THIRD PARTY

Third parties have caused the largest number of spillages with 159 events, an average of 4.5 per year and over 36% of the total. 114 events were accidental, 19

report no. 4/07

32

were intentional (mostly theft attempts) and 26 were incidental i.e. resulted from prior damage inflicted to the pipeline by a third party at some point in the past. As discussed in section 5 third party activities also result in relatively large spills and account for the largest total volume spilled of all causes.

6.5.1. Accidental damage

The most common causes of accidental third party spills are shown in Figure 20 and their sizes are shown in Figure 21.

Figure 20 Causes of accidental third party spills

0

5

10

15

20

25

30

35

40

Road C

onstr

uctio

n

Farming

Trench

ing

Pipelin

e Main

tenan

ce

Other In

dustr

y

Pipelin

e Con

struc

tion

Other

No.

of S

pilla

ges

OtherPloughingDiggingBulldozingDrilling

Only two events were not caused by direct damage from some form of digging or earth moving machinery. In one case an electrical earthing deficiency had arisen on a pipeline with no previous problem as a consequence of the electrification of an adjacent electric railway line. In the other, an electricity pylon fell over and one of the arms punctured a pipeline.

Damage by machinery occurs as a combination of lack of communication and awareness and lack of care or skill. Pipeline operators are not always made aware of impending ground working jobs and cannot therefore supply appropriate advice on exact pipeline location and working procedures and exercise adequate supervision of the work. Even when good communication has been established between the pipeline operator and the third party company, the actual machinery operator may be left partially or completely unaware of a pipeline's existence or fail to apply the requisite care or skill.

report no. 4/07

33

Figure 21 Average spillage volumes per spill by type of third party activity causing spillage

0

50

100

150

200

250

300

Road C

onstr

uctio

n

Farming

Trench

ing

Pipelin

e Main

tenan

ce

Other In

dustr

y

Pipelin

e Con

struc

tion

Other

m3 Recovered

Net loss

Figure 22 shows the awareness data (reported for 79% of the third party-related spillages) as the percentage of cases where each party was aware of either the impending activity (pipeline operator) or the presence of a pipeline (machinery operator). It should be noted that there are no instances where the pipeline operator was aware of the works but the machinery operator was not aware of the pipeline.

Lack of awareness by pipeline operating companies is an almost universal factor behind spillages caused by farming activities and in 60-80% of all other not pipeline-related works. Overall some 65% of the third party accidental spillages would most probably have been prevented by proper communication to pipeline operators by the third parties. Lack of care or skill by the third party works management or machinery operators is responsible for 35% of the spillages.

report no. 4/07

34

Figure 22 Awareness of impending works and of pipeline location

0%

20%

40%

60%

80%

100%

RoadConstruction

Farming Trenching PipelineMaintenance

OtherIndustry

PipelineConstruction

Aw

aren

ess

Both Equipment Operator Pipeline operator Neither

An analysis has been made of the relationship between the vulnerability to third party damage and various physical attributes. The strongest relationship, i.e. with pipeline diameter is shown in Figure 23. In this figure the frequencies of spillages caused by accidental damage by third parties have been calculated for the average length of each group of diameters for the periods 1971 to 1987, 1988 to 2005 and 1971 to 2005. These periods have been chosen because of the major change in the reported pipeline inventory between 1987 and 1988 following the inclusion of the non-commercially owned pipelines. Although the absolute values vary, the relative values are very clear and show that the larger the pipeline, the less likely is it to be damaged severely enough by third parties to lead to a spillage. Taking the overall average figures, the below 8” size range is fourteen times more vulnerable than the 30”+ population. A number of possible reasons for this could be postulated but there is no way of determining from the available data what each risk-increasing factor might contribute. Neither is there sufficient data on depth below surface to indicate how much the risk is reduced by deeper coverage. It is not recorded if larger pipelines have greater coverage than small ones.

The prevention of third party accidental spillages is of the highest priority due to its place in the spillage cause league. It is also the most amenable to improvement by sharing experiences and comparing operating and work control practices between pipeline operators from different companies and countries.

report no. 4/07

35

Figure 23 Third party accidental damage frequencies against pipeline diameter

0

0.1

0.2

0.3

0.4

0.5

>8" 8 to <12" 12 to <16" 16 to <24" 24 to <30" >30"

Pipeline diameter

Spill

age

freq

uenc

y

1971-19871988-20051971-2005

6.5.2. Intentional damage

There have been 19 spillages caused by intentional damage by third parties.

Table 7 Intentional damage by third party

Cause Gross spillage Net loss

Terrorism 2 920 710Vandalism 5 612 568Theft 12 1377 606Totals 19 2909 1884

m3Number of spills

None of the terrorist or vandalism incidents was from underground piping; one was from an above ground section of pipeline, all the rest were at valves or other fittings at pump stations or road / river crossings, etc. Since 1999, theft attempts by drilling into pipes have become a regular feature of the spillage statistics although there have been no such incidents in 2004 or 2005. However, a number of theft attempts have been discovered which fortunately did not lead to spillages.

This category of spillages represents 4.4% of the total number of spillages and has been responsible for about 4% of the total gross spillage loss and 6% of the total net loss.

report no. 4/07

36

6.5.3. Incidental damage

This category is somewhat of a catchall and includes those incidents where damage was done at some unknown point in a pipeline’s lifetime, which subsequently suffers deterioration over time resulting eventually in a spill. In general they result from unreported damage done after the original construction when a pipeline has been knowingly or unknowingly hit during some or other third party’s groundwork activities.

There have been 19 incidental damage incidents. They all started off from dents, scrapes and such like. Thus they share the characteristic that they may well be detectable by intelligence pig inspections.

report no. 4/07

37

7. INTELLIGENCE PIG INSPECTIONS

7.1. INTELLIGENCE PIG INSPECTION ACTIVITY

CONCAWE has been collecting data on intelligent pig inspection activity for the past fifteen years, including a one-off exercise to collect back data from the time intelligence pigs were first used back in 1977. Separate records are kept for metal loss pig, crack detection pig and for geometry (calliper) pig inspections. Each inspection may entail one or more passes of a pig along a piggable pipe section.

Leak detection pigs are also sometimes used but their function is quite different. They can reduce the consequences from a leak that has already started by helping to catch it earlier. They do nothing to help prevent the leak occurring in the first place.

7.2. ACTIVITY IN 2005

The total length of pipelines inspected by any type of intelligence pig in 2005 was 6226 km or 18% of the total length of the inventory. 109 sections out of a total of 663 were inspected by one or more pig. Inspections were split as follows amongst the individual classes of pig:

• Metal loss pig 5860 km, 105 sections • Crack detection pig 2132 km, 34 sections • Geometry pig 2643 km, 50 sections From this it can be seen that most inspection programs involved the running of more than one type of pig.

60 inspections used metal loss and / or crack detection pigs only, while 49 used one or other of these in association with geometry pigs. There was only one inspection where a geometry pig was used without one or other of metal loss or crack detection pigs.

7.3. ACTIVITY SINCE 1971

As shown in Figure 24, the growth in intelligence pig use for internal inspection of pipelines was spectacular up to 1994, but then reduced to levels that maintained inspection integrity. There has been a further increase over the last five years, 2005 being the highest ever year in terms of total length. This can only partly be explained by the increase in the reported length of the pipeline inventory.

In the 28 years of use of the technique the proportion of the pipeline inventory surveyed grew from nothing to peak at 19% of the total system in 1995. After many pipelines had been inspected once, the rate of inspections fell to around 10% to 15% of the inventory annually but was up to 18% in 2005.

report no. 4/07

38

Figure 24 Growth in the use of intelligence pigs

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

Tota

l ann

ual i

insp

ecte

d le

ngth

(km

) Geometry

Cracks

Metal loss

At the end of 2005 there was a total of 5934 km (17.0%) in 172 sections (25.9%) for which CONCAWE has no record of them having been inspected (Figure 25). This is a considerable reduction on the 2004 figures of 6,966 km (19.7%) in 185 sections (28.2%). The difference between the percentages of non-inspected km and section number indicates that a majority of the not-inspected sections are short. There are certainly some pipeline sections (mainly very old ones) which were not designed to be pigged and because of small size or tight bends or lack of suitable pig launchers or receivers cannot be intelligence pigged. The relatively recent introduction of pigs to inspect 150 mm (6 inch) diameter pipelines means that small diameter is no longer a bar to pig inspections. Less than 150 mm diameter pipelines represent a very small percentage of the pipeline inventory.

The main reason for the still significant length of apparently non-inspected pipelines is probably under-reporting particularly as we only started collecting information on intelligence pigging in 1995 and replies to this part of the questionnaire have never been as complete as for other sections. In addition, a number of pipeline companies in Eastern Europe have joined the survey in recent years and previous pigging records have not on the whole been provided. The length of un-inspected pipelines is therefore certainly less than the above figure and should continue to decrease in future years.

report no. 4/07

39

Figure 25 Pig-inspected pipelines as per end 2005

0

100

200

300

400

500

600

Sections Total length

Num

ber o

f sec

tions

or t

otal

leng

th ('

00 k

m)

Geometry onlyMetal loss & geometryMetal loss & crack only

7.4. REPEAT INSPECTIONS

Many pipelines have been inspected a number of times. Indeed, for some pipelines, regular intelligence pig inspections are required by the authorities. Some 142 pipelines have been reported as inspected by metal loss pigs twice, 11 pipelines by crack pigs twice, and 98 pipelines by geometry pig twice. The number of repeat inspections is shown in Figure 26 which shows that 2 pipelines have been inspected no fewer than fourteen times.

report no. 4/07

40

Figure 26 Repeat Inspections

0

10

20

30

40

50

60

70

3 4 5 6 7 8 9 10 11 12 13 14

Number of repeat inspections

Num

ber o

f sec

tions

Metal lossCrackGeometry

The intelligence pig inspection technique only finds flaws, corrosion and other sorts of damage in or on the pipe inner or outer walls. Over the past 35 years, as shown in the table below, there have been 126 spillages, some 3.6 per year, where the trouble might have been discovered by internal inspection before the failure had occurred.

Table 8 Historical spillages possibly preventable by internal inspections

Mechanical failures 45(line pipe welds, pipe material faults)Corrosion 59(excluding excess historic hot incidents)Third party incidental 22(non-construction scrapes and dents)Total 126

These categories will all tend to increase with age at some point in the future. Internal inspections should ensure that repairs will be made before they become spillages.

report no. 4/07

41

8. REFERENCES

1. CONCAWE (1972) Spillages from oil industry cross-country pipelines in Western Europe. Statistical summary of reported incidents 1966 - 1970. Report No. 2/72. Brussels: CONCAWE

2. CONCAWE Performance of oil industry cross-country pipelines in Western Europe. Statistical summary of reported spillages. Reports No. 2/73, 1/74, 5/74, 7/75, 7/76, 9/77, 3/78, 6/79, 10/80, 2/82, 11/82, 9/83, 12/84, 9/85, 7/86, 8/87, 8/88, 9/89, 6/90, 4/91, 4/92, 2/93, 5/94, 4/95, 4/96, 7/97, 6/98, 3/99, 3/00, 4/01, 1/03, 7/04, 3/05, 3/06. Brussels: CONCAWE

3. CONCAWE (1998) Western European cross-country oil pipelines – 25-year performance statistics. Report No. 2/98. Brussels: CONCAWE

4. CONCAWE (2002) Western European cross-country oil pipelines – 30-year performance statistics. Report No. 1/02. Brussels: CONCAWE

report no. 4/07

42

APPENDIX 1 DEFINITIONS

Spillage volume

Gross spilled volume: the estimated total quantity, expressed in m3, of hydrocarbons released from the pipeline system as a result of the incident

Recovered oil: the estimated quantity, expressed in m3, recovered during the clean-up operation, either as oil or as part of the contaminated soil removed

Net loss: the difference between gross spilled volume and recovered oil.

Categories of spillage causes

CONCAWE classifies spill causes into five major categories: mechanical failure, operational, corrosion, natural hazard and third party.

Mechanical: a failure resulting from either a material fault (e.g. metallurgical defect) or a construction fault (e.g. defective weld, inadequate support etc). This also includes failure of sealing system (gasket, pump seal etc).

Operational: a failure resulting from operational upsets, malfunction or inadequacy of safeguarding systems (e.g. instrumentation, mechanical pressure relief system) or from operator errors.

Corrosion: a failure resulting from corrosion either internal or external of either a pipeline or a fitting.

Natural hazard: a failure resulting from a natural occurrence such as flooding, land movements, lightning strike etc.

Third party: a failure resulting from an action by a third party either accidental or intentional. This also includes "incidental" damage, undetected when it occurred and resulting in a failure at some later point in time.

These main categories are subdivided to give a total of 13 subsets shown in Table 1.1.

Table 1.1 Categories of spillage causes

A B CA Mechanical Failure Construction Fault Material FaultB Operational System Malfunction Human ErrorC Corrosion External Internal Stress CorrosionD Natural Hazard Landslide / Subsidence Flooding OtherE Third Party Activity Accidental Intentional Incidental

SecondaryMain

Detailed reporting in Appendix 2 further identifies, within each category a primary cause.

report no. 4/07

43

APPENDIX 2 SPILLAGE SUMMARY

Key to table Service Discovery

1 Crude oil 1 Right of way survey2 White product 2 Routine monitoring3 Fuel oil 3 Maintenance4 Crude/white pro. 4 Pressure testing5 Others 5 Automatic detection system

6 Third party7 Pipeline internal inspection survey

System part Land use1 Underground pipe 1 Commercial2 Above ground pipe 2 Industrial3 Road/rail crossing 3 Residential4 River crossing 4 Agricultural5 Line valve 5 Forest6 Line fitting 6 Mountains7 Manifold 7 Barren land8 Terminal 8 Surface water9 Pump station10 Fitting (other)

Primary cause1 Dent2 Faulty weld (undetected)3 Faulty weld (repaired)4 Faulty heat treatment5 Pipe fitting6 Gasket7 Lamination8 Faulty material9 Temperature compensation10 Gland11 Traffic12 Mining13 Overpressrue14 Vibration15 Temperature variations16 Bolt/screw/plug17 Design18 Bypass/deadleg19 Electric current20 Burner/welder damage

report no. 4/07

44

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

1 1971 11 2 1 1 3 1 3 4 AA 62 1 4 3 9 AA 163 11 2 0 6 4 6 AA 34 20 1 40 5 5 9 5 AB 6 60,0005 1 350 3 9 9 2 BA6 1 25 3 9 BB7 5 3 3 6 1 8 CA8 8 2 6 6 2 1 20 CA9 20 1 300 50 3 1 5 EA 1,00010 34 1 2000 3 1 9 EA11 8 2 2 2 6 1 20 EB12 1972 16 2 5 3 1 4 AB 513 28 1 800 150 2 9 12 2 AB 214 12 2 70 39 6 6 5 4 AB 615 9 1 10 5 6 3 29 CA16 9 1 40 35 6 3 29 CA17 10 1 1 1 3 2 39 2 CA18 10 1 1 1 3 2 39 2 CA19 12 3 500 6 1 12 1 CA20 12 3 5 1 6 1 12 1 CA21 10 2 150 50 2 1 7 CA22 4 3 0 6 1 15 2 CA23 6 3 1 0 6 1 15 CA24 20 1 200 60 3 1 8 1 EA25 20 1 250 100 3 1 8 EA26 28 1 60 12 6 1 16 EA27 10 1 90 6 1 6 EA28 8 1 7 6 1 8 4 EA29 10 2 30 6 1 9 EA30 8 2 400 350 2 1 2 4 EA31 10 2 99 96 6 1 6 4 EA32 12 3 0 6 1 5 EC33 1973 5 3 4 1 1 8 AA 1534 20 1 25 3 6 9 1 2 AA 635 16 1 0 3 9 3 2 AB 636 1 4 3 9 11 2 AB 1437 24 2 25 3 9 2 2 AB 1038 18 1 11 1 2 9 13 2 AB 639 6 2 12 6 6 1 1 2 AB 1440 9 1 12 12 1 1 32 CA41 5 3 15 1 1 8 CA42 5 3 15 1 1 8 CA43 12 3 200 2 6 1 13 CA44 12 3 12 2 3 2 13 CA45 12 3 250 5 6 2 13 CA46 12 3 150 2 1 2 13 CA47 12 3 310 10 6 1 13 1 CA 30,00048 28 1 100 40 6 1 16 DA49 10 3 8 6 1 9 4 EA50 12 3 0 6 1 6 EC 1151 12 3 1 6 1 6 EC 1152 12 3 0 1 1 6 EC 1153 1974 1 1 0 3 9 4 2 AA 1154 1 3 2 2 9 5 2 AA 14 1,00055 6 1 20 6 6 15 AA 1556 9 1 10 1 1 33 CA57 2 2 2 3 10 6 CA58 10 3 1 2 1 9 2 CA59 12 3 5 6 1 8 CA60 13 3 5 6 1 8 CA61 4 3 1 6 1 17 2 CA62 6 3 0 6 1 16 CA63 16 3 1 6 1 9 4 CB P64 7 1 1 6 1 8 4 CB65 16 1 500 6 3 10 EA66 5 2 1 0 6 1 21 EA67 8 2 30 4 2 4 22 EA68 8 2 200 2 6 1 22 EA69 10 2 668 668 2 1 18 EA70 10 2 489 405 2 1 18 4 EA

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

report no. 4/07

45

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

71 1975 20 2 4 30 10 4 10 11 4 AB 272 34 1 30 2 6 1 12 AB 273 10 3 3 2 10 5 3 AB 674 1 10 2 3 9 2 BA75 2 4 5 9 2 BA76 8 2 20 10 3 9 4 2 BB77 1 5 3 9 2 BB78 10 3 50 2 3 11 CA79 12 3 3 6 1 9 CA80 6 3 25 1 1 9 CA81 10 3 1 0 2 9 6 2 CA82 4 3 1 6 1 18 CA83 8 3 0 7 1 6 CA84 8 3 0 1 1 6 4 CA85 12 3 0 3 9 6 2 CA86 6 1 15 0 6 1 23 4 EA87 18 1 5 0 2 1 12 EA88 8 1 120 3 2 1 9 EA89 8 2 60 60 2 1 23 EA90 6 1 15 6 6 1 4 EA91 1976 8 2 6 1 9 AA 392 8 3 6 5 13 4 AA 893 1 9 2 5 13 2 AB 694 24 2 17 1 6 10 17 2 AB 895 16 1 1322 433 2 1 13 AB 1396 10 3 80 2 1 11 CA97 4 2 90 90 6 1 16 CA98 24 1 200 2 1 10 DA99 10 3 50 25 2 1 DA

100 10 1 40 2 6 1 13 4 EA101 8 2 44 14 2 1 24 4 EA102 18 1 802 606 6 1 7 4 EA103 8 2 153 153 2 1 4 EA104 14 2 358 358 6 4 23 4 EC 12105 1977 2 32 3 9 9 2 AB 6 150106 2 28 3 9 9 2 AB 6 140107 20 2 2 6 1 8 4 AB 8108 36 1 3 5 3 2 AB 10109 1 50 2 9 19 2 BB110 1 1 3 9 7 2 BB111 12 2 350 220 4 1 10 4 CA112 10 3 315 90 2 1 8 3 CA113 1 6 3 9 9 2 CB114 12 2 103 6 1 19 DA115 20 1 550 500 1 4 13 4 DB116 24 1 600 25 5 4 11 4 DC117 10 1 160 2 1 12 4 EA 1,500118 18 1 80 2 1 5 4 EA 400119 8 2 3 3 2 1 25 4 EA120 8 2 3 1 2 1 13 4 EA121 12 2 191 2 1 19 4 EA122 8 2 269 6 1 19 4 EA123 20 2 2530 2500 2 4 9 4 EC 12124 1978 34 1 2000 300 6 4 16 4 AB 8125 8 2 235 205 2 5 16 4 AB 7126 22 1 19 6 1 7 4 AB 8 1,800127 6 2 12 6 6 1 18 1 CA128 10 2 100 10 2 4 14 4 CA129 12 3 2 6 3 14 4 CA130 8 3 120 60 4 1 7 4 CA131 8 3 80 40 4 1 7 4 CA132 12 3 2 1 1 12 2 CA133 18 3 4 1 6 1 6 2 CA134 16 4 400 250 2 1 14 4 DA135 11 2 3 0 6 1 10 4 EA136 12 2 5 58 40 4 1 10 4 EA137 24 1 1 6 5 4 EA138 16 1 255 245 2 1 15 4 EA 5,865139 1979 22 1 100 40 4 1 8 4 AA 1 16,000140 24 1 100 1 6 1 5 AA 1 2,700141 9 2 50 6 1 17 4 CA 350142 12 2 300 200 1 1 23 4 CA143 18 3 20 1 1 12 1 CA 500144 18 3 5 1 1 12 1 CA 100145 18 1 50 1 6 1 16 4 EA 2,500146 12 2 90 50 6 1 23 4 EA147 8 1 245 150 6 1 23 4 EA148 11 2 950 380 2 2 15 2 EB P 6,400

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

report no. 4/07

46

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

149 1980 13 2 8 1 2 9 12 2 AB 6150 40 1 4800 400 6 1 9 4 AB 8 10,000151 10 3 80 6 3 10 4 CA152 10 3 10 1 1 10 4 CA153 7 3 1 1 1 15 4 CA 10154 12 3 111 12 6 3 15 4 DA P 10,000155 10 4 762 135 2 1 15 4 EA 10,000156 12 2 270 6 1 EA157 8 2 313 2 1 EA158 1 30 6 9 2 EB159 1981 34 4 10 2 6 5 6 AB 10160 40 1 10 6 7 5 2 AB 6 80161 10 2 600 150 2 1 AB 8162 20 1 19 1 6 1 17 4 CA163 8 3 5 4 5 12 4 CA164 8 3 19 4 5 12 4 CA165 12 3 5 2 6 1 15 2 CA 50166 10 2 92 58 2 1 25 4 CA167 20 1 5 3 6 1 15 2 CA168 10 2 10 6 3 CA169 26 2 125 45 6 1 18 4 DA170 24 3 30 10 4 5 14 2 DC171 7 1 132 132 2 1 15 4 EA172 8 2 322 317 2 1 24 4 EA173 5 1 96 6 1 EA174 28 1 5 0 1 3 16 2 EC 11175 1982 8 2 12 12 6 2 20 4 AA 1 P176 24 1 9 6 1 18 4 AB 8 1,000177 8 1 2 1 1 20 4 CA178 12 3 8 6 1 16 2 CA 30179 10 3 400 16 6 1 19 4 CA180 5 1 20 6 9 10 2 CB181 7 1 140 140 6 1 16 4 CB 3,000182 22 1 15 5 6 3 18 3 CB183 6 1 31 6 1 20 4 EA184 8 2 7 1 2 1 30 2 EC 1185 1983 4 5 10 2 1 22 4 AA 9 100186 4 5 1 5 1 22 4 AA 9 9187 4 5 4 6 6 22 4 AB 9 80188 16 4 442 111 4 1 18 4 BB189 6 2 12 4 3 15 2 CA 3,600190 7 1 182 120 2 1 17 4 CB 20,000191 7 1 148 110 6 1 17 4 EA 18,000192 10 2 213 171 6 1 29 4 EA193 14 2 675 470 6 3 3 4 EB194 12 1 1 0 6 1 20 2 EC 1 15195 1984 28 1 4363 3928 1 1 10 4 AA 1 6,500196 24 1 141 6 1 18 4 AA 4 4,500197 28 1 3 5 10 11 4 AB 8 120198 8 2 16 3 6 10 17 4 AB 8 720199 34 1 5 2 2 9 13 2 BA 1,000200 16 1 10 2 9 18 4 BA 50201 1 10 10 3 1 21 4 BB 50202 12 3 2 1 3 17 2 CA203 6 1 20 16 6 1 24 2 CA 250204 16 2 5 1 6 9 11 2 CA 10205 9 2 236 236 6 1 11 4 CB 200206 10 1 150 1 6 1 23 5 EA 100207 11 2 244 240 5 1 21 EB208 1985 24 1 1 1 1 1 14 4 AA 1 18209 20 1 25 4 6 9 9 2 BA210 10 2 16 5 9 17 2 BA211 10 2 7 5 9 17 2 BA212 6 2 4 5 9 17 2 BA213 16 1 1100 756 2 1 9 4 CC 2 13,000214 8 2 211 195 2 1 33 4 EC 1,000215 1986 16 2 160 6 5 9 17 4 AB 6 200216 20 1 53 6 2 1 12 4 AB 6 3,000217 24 2 292 4 3 5 26 4 AB 7 3,000218 16 3 20 5 6 3 38 3 CA219 20 2 2 2 6 1 22 3 CA220 8 3 10 4 1 25 4 CA 20221 9 1 10 10 6 1 45 4 CB 180222 34 1 7 7 1 1 14 2 CB 84223 8 2 192 95 6 1 15 4 EA 1,500224 14 2 280 56 5 1 18 4 EA 100225 6 2 52 41 5 1 13 4 EA 10226 8 2 11 6 5 4 19 4 EB 3

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

report no. 4/07

47

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

227 1987 20 2 1000 120 4 1 20 2 AA 3228 26 4 2 1 6 1 25 4 AA 4 1,000229 9 1 25 2 6 1 46 4 AB 5 200230 16 3 550 150 2 3 39 4 CA 200231 9 1 8 1 6 1 46 3 CB 280232 12 2 12 10 6 1 21 4 DA P 2,000233 22 2 3 1 3 1 20 2 EA 10234 16 2 300 115 6 1 18 2 EC 1 P235 1988 34 1 10 1 6 6 26 2 AB 6 200236 12 2 90 42 6 1 30 3 AB 8 P 1,500237 8 2 97 21 2 9 28 4 AB 6 500238 34 1 81 1 6 3 17 2 CA 5,000239 11 2 80 80 2 1 35 3 CA240 28 1 5 1 6 10 31 3 CA 400241 10 2 305 5 2 4 23 4 DA 5,000242 20 2 40 10 6 1 24 2 EA 30243 3 1 2 1 6 1 28 4 EA 100244 10 1 1 14 1 6 1 23 4 EA 100245 8 2 3 1 6 3 35 3 EA 20246 16 2 3 1 6 1 16 4 EA 150247 16 1 650 650 5 3 23 3 EA 550248 4 2 2 1 6 1 26 4 EA 9249 6 2 63 56 6 1 33 4 EA 1,200250 6 2 18 1 6 1 33 4 EA 1,800251 1989 26 1 3 2 6 1 26 4 AA 2 100252 12 3 1 6 6 2 AA 2 6253 1 2 25 7 6 2 1 4 AA 6 10,000254 26 1 155 5 6 1 26 4 AB 2 P 2,000255 10 2 1 66 16 2 1 27 4 BB256 9 1 25 5 4 1 48 4 CA 50257 12 3 240 150 2 4 17 2 CA258 10 2 400 90 5 1 24 4 CB 2,000259 16 2 253 253 6 1 22 4 EA 500260 16 2 660 472 5 1 20 4 EA P261 10 2 82 4 5 2 24 4 EA 200262 12 2 298 298 2 1 32 4 EA 6,000263 6 2 52 27 6 1 33 4 EA 2,000264 8 2 3 3 1 32 4 EA 66265 8 2 3 186 126 6 1 29 4 EA266 40 1 40 5 6 1 17 4 EC 1 4,000267 11 1 2 6 4 26 4 EC 17268 1990 13 2 105 105 3 5 4 BB 30269 10 2 252 221 6 9 33 4 BB 1,500270 8 2 9 3 10 48 4 BB 10271 11 3 325 11 2 3 22 2 CA272 11 2 225 194 6 1 11 4 EA 3273 6 2 3 1 6 1 34 4 EA 324274 10 2 189 34 6 1 24 4 EA275 1991 20 2 275 118 5 1 24 4 AA 17 14,000276 2 50 38 6 5 10 4 AA 16 1,200277 20 1 20 13 6 1 24 4 AA 1 4,500278 12 2 25 7 2 9 20 2 AA 16 150279 12 2 5 2 6 5 21 4 AA 17 320280 12 2 29 29 6 1 38 4 AB 5 600281 2 4 1 5 9 31 2 AB 8 250282 2 172 68 5 9 11 2 AB 8 100,000283 2 2 6 10 4 AB 6284 10 2 80 4 6 1 26 4 CA 1,500285 7 1 20 6 6 30 4 CB 300286 8 2 100 60 4 3 17 4 CB 10,000287 8 2 15 10 4 1 17 2 CB 25288 8 2 4 6 1 49 4 EA 6289 6 2 21 13 6 1 34 4 EA 500290 6 2 1 6 1 37 4 EA 2291 2 84 75 5 9 1 4 EB292 13 2 485 485 2 9 24 4 EB 7,000293 8 2 10 1 6 1 24 4 EC 1 30294 1992 8 2 1000 400 2 1 34 2 AA 1295 2 128 98 2 5 4 AB 2 5,400296 2 113 8 3 9 12 2 AB 8297 8 2 30 15 3 7 33 2 AB 6298 8 2 5 5 7 1 13 5 AB 8 10299 2 275 248 2 9 2 BB 1,100300 2 5 1 2 7 22 2 BB 1,350301 10 2 2 3 5 30 BB302 8 3 200 6 1 25 4 CA 300303 24 2 13 1 6 1 27 2 CA 250304 6 2 3 3 4 1 49 4 CA 2305 12 2 75 75 6 1 28 4 DB306 8 2 50 50 4 1 25 4 EC 1 20307 8 2 25 25 4 1 25 4 EC 1 60

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

report no. 4/07

48

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

308 1993 34 1 248 18 4 1 31 4 AA 8 45,000309 2 3 6 9 2 2 AB 16 80310 12 2 2 1 1 5 23 2 AB 6 400311 18 2 14 13 7 3 27 1 CA 400312 13 2 580 500 2 4 26 4 CB 800313 20 1 2000 500 2 3 19 4 CB 25,000314 26 2 10 7 6 1 31 6 DA P315 9 2 8 6 6 1 30 4 EA 50316 24 2 49 39 6 1 33 4 EA 40,000317 8 2 3 1 6 1 37 4 EA 100318 12 2 101 19 6 4 31 4 EA319 20 2 3050 1450 2 1 29 2 EC 1320 7 2 3 3 6 1 13 3 EC 1 6321 1994 16 1 200 160 5 1 31 4 AB 8 6,000322 16 1 1350 1295 2 1 31 4 AB 8 25,000323 6 2 250 14 2 9 16 2 AB 6 50324 6 2 1 1 1 1 16 2 AB 8 25325 11 2 5 5 6 10 9 4 AB 6 100326 1 2 2 6 9 2 BA 100327 12 3 90 60 6 1 24 4 CA328 32 1 10 5 2 8 21 2 CB 18 500329 10 2 285 285 6 4 26 4 EA330 9 2 195 170 5 1 37 4 EA P 8,000331 8 2 46 6 1 36 4 EA 1,150332 1995 2 280 80 2 7 22 1 AA 8 10,000333 10 2 30 30 6 3 35 4 AA 2 750334 2 53 41 6 6 5 4 AB 8335 6 2 115 1 1 36 4 AB 8 500336 16 1 132 82 5 1 30 4 BB 6,500337 10 2 1000 270 1 1 31 2 CA 55,000338 9 2 48 18 5 1 28 4 EA 1,500339 9 2 20 20 5 1 39 2 EA 100340 13 2 139 113 6 1 5 4 EA 300341 6 2 12 5 1 37 4 EA 30342 1996 9 2 165 99 2 9 5 2 AB 6 40343 14 2 292 209 6 1 40 3 BB 300344 12 3 1 6 1 30 1 CA 16345 9 2 437 343 2 1 40 2 EA 20346 7 2 1 19 19 6 1 40 4 EA 350347 10 2 500 62 6 1 64 2 EC 20 23,000348 1997 12 2 19 3 1 1 27 4 CA 2,800349 10 1 2 0 1 1 7 2 CB 20350 12 2 422 341 2 1 30 4 CC 2351 12 2 435 267 2 1 30 3 CC 2 P352 8 2 13 2 2 1 33 4 EA 150353 12 2 40 1 6 1 24 2 EC354 1998 1 30 4 3 9 30 2 AB 5 400355 6 3 0 0 6 1 34 4 BB356 13 2 486 247 2 1 42 4 BB 100357 16 2 250 20 6 1 30 2 CA358 10 2 340 313 5 1 6 3 EA 500359 10 2 15 14 1 1 4 4 EA 600360 9 2 176 67 5 1 42 4 EA 160361 2 30 2 5 6 4 EA 650362 8 2 0 6 1 25 4 EA 4363 1999 1 7 2 9 2 BB 200364 1 3 30 2 3 32 2 CA 300365 11 2 167 64 2 1 32 4 CA 60366 6 2 1 1 5 1 25 4 CA 5367 4 1 1 1 6 9 35 2 CA368 8 2 80 20 6 1 48 4 EA 500369 13 2 84 13 5 1 10 2 EA370 6 2 29 14 6 1 40 4 EA371 8 2 1 80 30 6 1 35 4 EB 1,000372 11 2 36 28 5 1 5 4 EB 100373 12 2 1 2 1 36 2 EC 1374 2000 2 175 3 6 10 24 2 AB 5 60375 12 1 10 7 6 1 30 2 CB 150376 12 2 8 8 6 1 31 4 EA377 11 2 159 64 5 1 8 4 EA 5,000378 12 2 7 1 6 3 26 3 EA379 24 2 1 1 6 1 41 4 EC 1 150

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

report no. 4/07

49

Year Service Fatalities Injuries Discovery AgeGross Net loss Years Category Primary Water

bodiesContaminated

land area" m2

380 2001 1 800 8 6 2 4 AA 2 10,000381 10 2 1 1 6 3 39 4 AA 2 10382 10 2 5 5 6 1 38 4 AB 8 500383 6 2 37 7 4 1 27 4 AB 8 900384 12 2 10 2 6 1 15 2 AB 8 120385 34 1 6 1 5 1 29 2 CA 500386 12 2 4 4 6 1 26 4 CA 1,000387 13 1 103 50 2 9 23 2 CB 225388 11 2 55 51 6 1 9 4 EA389 10 2 10 1 6 1 11 4 EA390 6 2 5 5 6 1 47 3 EA 400391 12 1 10 7 6 1 30 4 EB 250392 12 1 17 12 6 1 30 4 EB 400393 16 2 2 2 6 1 18 4 EB 350394 8 2 85 24 2 1 47 4 EB P 404395 2002 8 2 10 10 6 1 47 4 AB 5 325396 20 1 100 2 1 36 2 CA 500397 10 2 80 20 6 1 38 2 CA 10,000398 10 3 1 6 1 28 4 CA 14,000399 6 2 17 2 7 33 2 CA 400400 8 2 70 2 6 ? 2 CA401 13 2 225 58 5 1 46 4 CC 2 400402 24 2 250 20 6 5 39 2 DA 5,000403 30 1 2 3 10 40 2 EA 40404 8 2 170 120 4 1 57 4 EA405 16 1 750 45 1 1 39 4 EA 20,000406 20 1 280 30 6 1 40 4 EA 12,000407 12 1 40 15 6 1 33 4 EB 6,000408 8 2 190 5 1 2 EC 1409 2003 14 2 30 30 5 1 AA410 20 4 2 2 1 52 2 CA S 2411 12 2 2 6 1 32 2 EA S 5412 11 2 83 74 5 1 46 4 EA 1,800413 11 2 45 31 6 1 46 1 EA 600414 6 2 2 5 1 EA415 11 2 74 49 5 1 46 4 EB 500416 16 1 5 5 1 1 41 5 EB 120417 16 2 28 10 6 1 29 4 EB 400418 16 2 52 3 4 1 29 4 EB 400419 12 2 11 7 4 1 45 2 EC 20 800420 20 2 2500 1100 6 1 31 7 EC P 80,000421 2004 16 2 2 0 1 1 32 4 AA 4,000422 10 2 26 18 2 10 40 4 AA 1 6,000423 22 1 20 6 2 9 5 2 AB 200424 8 2 90 50 6 1 5 4 EA 1,500425 10 2 5 1 29 2,3 EA 2,000426 2005 12 2 19 19 2 9 4 AA 14427 12 2 3 1 2 AA 2 G 428 20 1 350 10 5 1 45 4 AA G 15,000429 6 2 20 2 1 28 4 AB 7 S 58430 6 2 38 6 1 28 4 AB 7 S 42431 9 1 30 4 5 3 14 4 BB G 1,000432 10 1 15 6 10 22 4 BB 1,000433 10 2 3 1 6 1 25 2 CA S 50434 24 1 64 63 2 1 40 2 CB 18 G 150435 8 2 15 8 6 1 41 4 EA G 1,000436 24 2 0 6 1 46 EC 1 S G 3,000

Spillage ID

Impact

m3

Spillage volume Pipe diameter

CauseSystem part

Land use

LefkosiaLefkosa