Oct. 29, 1990
J. S. Mandke Southwest Research Institute San Antonio Corrosion is the leading cause of failures of subsea pipelines in the U.S. Gulf of Mexico. Third-party incidents, storms, and mud slides are additional principal causes of offshore pipeline failures. These are among the major conclusions of an analysis of 20-year pipeline-failure data compiled by the U.S. Minerals Management Service. 1
J. S. Mandke
Southwest Research Institute
San Antonio

Corrosion is the leading cause of failures of subsea pipelines in the U.S. Gulf of Mexico. Third-party incidents, storms, and mud slides are additional principal causes of offshore pipeline failures.

These are among the major conclusions of an analysis of 20-year pipeline-failure data compiled by the U.S. Minerals Management Service. 1

For small size lines, additionally, failures due to external corrosion were more frequent during the period studied than internal corrosion. In medium and large-size lines, failures due to internal corrosion were more frequent than those due to external corrosion.

Also, the majority of corrosion failures occurred on or near the platform and among the small-size pipelines.

The motivation for the study described here was to perform a more in-depth evaluation of the pipeline failure data for the Gulf of Mexico than reported earlier, using an extended data base for the period 1967-87, and to compare the results with those reported earlier.

In their overall study for failures of both onshore and offshore pipelines, Andersen and Misund included a brief review of the MMS data for the years 1967-77 and for pipe sizes larger than 10 in. 2 Starting reported in 1981 on a stud of pipeline failures that occurred in the North Sea during installation and start-up phases. 3

The study results presented here provide an improved basis for assessment of safety of pipelines and for further improvements to current pipeline design, inspection, maintenance, and construction procedures.


The significant components of a typical offshore pipeline system transporting hydrocarbons are: Platform risers, expansion loops or thermal offsets, subsea valves and fittings, tie-in spools, and the main trunk line or the infield flow line.

An understanding of the varying risks of damage and their consequences associated with these components can be developed from an evaluation of the historical data on the reported pipeline failures.

Failure data on offshore pipelines are not readily available for all regions of the world. Most of the reported information is on the pipelines in the Gulf of Mexico and the North Sea.

In the U.S., the Department of Interior's MMS has kept a record of offshore pipeline failures since 1967. 1 No other data source with comparable details is available in the public domain on failures of offshore pipelines.

Failure data published by the MMS' for about 690 failures that occurred during 1967-87 was compiled into a personal-computer data base.

Although the MMS data on pipeline failures are the most comprehensive source of information available, the information for some of the failures reported is either insufficient or unclear. In those instances, some judgment and assumptions had to be exercised during compilation of these data.

This did not affect the actual results, however, because the emphasis of this study has been on detecting the overall failure trends for offshore pipelines rather than the absolute numbers on failures.

Fig. 1 shows the histogram on the number of pipeline failures that occurred during the 20-year period since 1967. The distinctly high number of failures during 1985 can be attributed to the arrival of four hurricanes--Elena, Danny, Juan, and Kate--in the Gulf of Mexico during that year.

Out of the 87 failures reported during 1985, about 56 directly resulted from hurricane-induced storms or the resulting mud slides.

The significant increase in failures since 1975 can be attributed to the increase in the pipeline population, aging of the pipelines installed earlier, and the increased offshore construction activity.


Fig. 2 shows the total number of failures associated with each pipe size category. This figure shows that for pipelines in sizes greater than 12 in., the failure rate drops significantly.

Fig. 3 shows the distribution of failures among three pipe size (OD) groups:

  • Small, 2-6 in.

  • Medium, 8-16 in.

  • Large, 18-36 in.

Throughout this article, the pipe-size groups have been referred to in the context of the preceding definition. Only 3% of the failures were associated with large-diameter lines, whereas 59% were associated with small-diameter lines.

This confirms the general awareness in the industry that the smaller-size pipelines are more susceptible to damage than the larger-size pipelines.

Fig. 4 shows the total percentage of pipeline failures grouped under each failure-cause category. As shown, the various causes of pipeline failures have been grouped into five principal categories:

  1. Material or equipment failure

  2. Operational

  3. Corrosion

  4. Storm/mud slides

  5. Mechanical damage/third-party incidents.

These failure-cause categories are discussed in detail in later sections.

Fig. 4 shows that, next to corrosion (50% of failures), mechanical damage due to third-party incidents (20%) and storm/mud slide-induced damages (12%) are the major sources of pipeline failures.

Fig. 5 shows the distribution of failures based on the type of product transported.

The majority of failures occurred on oil lines (51%).

Although gas pipelines have a larger population than oil lines, in the Gulf of Mexico their reported failure rate was only 28%. This may be due to relative ease of leak detection in oil lines as compared to gas lines.

Other possible reasons for the disproportionate number of failures associated with oil lines have been given by Andersen and Misund .2 They need further evaluation.

Multiphase lines and other types of lines, such as NGL, condensate, test lines, etc., grouped under the miscellaneous category, each had 3% of the total failures.


Failures were also categorized on the basis of their location in the transportation system.

Fig. 6 shows that a majority of failures, about 64% (439), took place at Location 1 which includes the platform deck, the riser section, and the seabed within 500 ft of a platform.

The failures reported at Location 1 included 266 incidences on the riser sections, 128 incidences on the pipeline segment on the seabed in the vicinity of platform, and 45 incidents on the deck piping or above the splash zone.

Out of these 439 cases, about 75% were on small-size lines, 21% on medium-size lines, and 2% on large-size lines.

In comparison, about 200 incidents occurred on pipelines on the seabed (Location 2), of which 38% were on small-size lines, 56% on medium-size lines, and 5% on large lines.

Thus, on and close to the platform, the small-size lines have the highest failure rate, whereas away from the platform, the medium-size lines have the highest failure rate.

About 12 failures occurred around other facilities such as wellhead, storage barge, loading buoy, etc., and 5 failures occurred near the shore.

The correlation study between the causes of failures and their locations in the pipeline system has shown that the two leading sources of riser failures have been corrosion (68%) and storms (15%).

In comparison, for the pipelines close to platforms, the leading sources of failure have been corrosion (43%) and third-party incidents (34%).

Corrosion and third-party incidents have also been the leading sources of damage to pipelines away from the platform with associated failure rates of 35% and 32%, respectively.


From the 690 failures considered, only 290 cases resulted in measurable pollution. Of these, 234 failures were for oil lines, and the remaining were for multiphase and condensate lines.

From the reported 290 cases of pollution, about 274 (94%) cases resulted in pollution of less than 100 bbl, 9 cases resulted in pollution between 100 and 1,000 bbl, 5 cases resulted in pollution between 1,000 and 10,000 bbl, and 2 cases resulted in pollution more than 10,000 bbl.

The two cases with the largest pollution occurred during 1967 and 1974 and dumped, respectively, 161,000 and 20,000 bbl of oil in the sea. These cases can be considered aberrations because the technology has improved significantly since these failures.

As expected, the majority of the cases resulting in pollution of up to 10 bbl was associated with small-size lines damaged mostly from corrosion, whereas the majority of cases resulting in pollution greater than 10 bbl was associated with medium-size lines.

All cases of pollution greater than 500 bbl resulted from anchor damage to pipelines.

Fig. 7 gives the distribution among the 290 cases of pollution based on the amount of pollutant discharged in each case. These data clearly demonstrate that submarine pipelines offer the safest mode of transporting hydrocarbons from the offshore locations.

The MMS data also provide information regarding the corrective action or the repair procedure implemented in these reported failures. The distribution of failures based on the type of repair solution implemented is given in Fig. 8.

In 181 cases out of the total of 690, the damaged pipe was repaired by installation of a clamp. Most (149) of these failures were due to corrosion.

Spool piece repair was implemented in 369 cases, out of which 273 cases needed a spool piece of length less than 100 ft. Use of mechanical connectors as compared to welding was explicitly identified in 35 cases of spool-piece repairs.

In 23 instances, the pipeline was abandoned after the discovery of damage. About 61 instances required solutions other than replacement of pipe sections, such as repairing valves, flanges, tightening of bolts, etc.


The following section presents a detailed discussion of the different causes of pipeline failure.


Material failures include instances where the pipe material ruptured or the weld cracked and failed. Equipment failures were primarily due to leakages or malfunctioning of fittings such as flanges, clamps, valves, etc.

Out of the 60 total failures that were grouped under this category, about 23% were attributed to material failure, and the remaining 77% were attributed to equipment failure.


Only seven failures were attributed to operational problems. These were mostly the result of lines being overpressured either during the normal operation or the pigging operation.


Three subcategories comprise corrosion failures.

In the first two cases, the failure was clearly identified as the result of either internal or external corrosion. In the third case, the origin of the corrosion was not clearly identified. We will refer to this as general corrosion.

Out of the 343 total cases of corrosion failures, 15% resulted from internal corrosion, 46% from external corrosion, and 39% from general corrosion.

Fig. 9 shows the number of total corrosion failures per pipe size. This figure shows that the corrosion failures have been highest among the small-size pipelines.

Further evaluation of these data showed that for the smaller-sized pipe, external corrosion failures were more common, whereas for medium and larger-sized pipe internal corrosion was more common. This latter observation is consistent with the observation made by Andersen and Misund. 2

About 78% of the total corrosion failures occurred on the platform, in the riser section or its vicinity on the seabed, and 20% occurred on pipelines on the seabed away from the platform.

Regarding the distribution of corrosion failures, based on the pipeline product, 27% occurred in gas lines, 49% occurred in oil lines, and in 17% of the cases, the product was not identified.


There were 63 incidents of pipeline failures as a result of storm loading. From these, about 87% were among small-size lines.

The majority of storm damage incidents (83%) occurred on or near platforms. There were 19 incidents of damage due to mud slides which mostly resulted from storms. Most of the failures due to mud slides were on medium-sized pipe (74%) and on the seabed away from the platforms.


The principal sources of mechanical damage to pipelines are anchors and anchor lines, work and supply boats, construction vessels, and trawlers. Fig. 10 shows the distribution of third-party incidents among these sources.

There were 70 incidents of damage due to anchors, wire ropes, etc. From these, 34 failures occurred near platforms, and 33 failures occurred on the seabed away from platforms. Most of the anchor-damage incidents occurred on small (30) and medium (37) size pipelines.

The majority of failures needed spool-piece repair, and only in seven cases could clamp repair be implemented.

Damage incidents due to work boats or supply boats totalled 14, out of which 9 were on small-size lines and 5 were on medium-size lines. Ten failures occurred on or near the platforms. The majority of these incidents resulted from boats colliding with riser pipes during severe weather conditions.

The current practice of routing risers inside the jacket in the splash zone should reduce these types of failures in the future.

Construction-vessel failures included mishaps during pipe laying, trenching with jet sleds, erroneous setting of jack ups, impact from dropped objects, and movement of heavy objects on the seabed.

Out of the 20 reported failures under this category, 16 were associated with small-sized pipes. Eleven of these incidents occurred around the platform and 8 on the seabed away from the platform.

The potential for damage to pipelines from impact with trawling gear in the Gulf of Mexico is not as severe as in the North Sea. Nevertheless, there were 10 reported cases of damage to pipelines from trawling gear, 5 on small sizes, 3 on medium sizes, and 2 on large-sized pipelines. Most of these incidents occurred away from the platform.


The analysis of the failure data presented here has indicated significant trends in pipeline failures.

It is customary to convert the failure data to probability of failure or the failure rate per km-year or mile-year of the pipeline. Because the appropriate actuarial details on these failures were not available, probabilistic analysis of the failure data could not be performed.

However, the results summarized below provide an important input for assessment of risk of damage to offshore pipelines. Such an assessment should form the basis from which the contingency plans for emergency repair of pipelines are developed.

  • Corrosion is the leading cause of pipeline failures. It is followed by third-party incidents and storms and mud slides as the other principal causes of offshore pipeline failures in the Gulf of Mexico.

  • For small-size lines, failures due to external corrosion were more frequent than internal corrosion. In medium and large-size lines, failures due to internal were more frequent than those due to external corrosion.

  • Also, the majority of corrosion failures occurred on or near the platform and among the small-size pipelines.

  • A disproportionately higher number of failure incidents have been reported on the oil lines than on the gas lines.

  • The risk of damage to pipelines generally decreases with the increase in pipe size. Within the riser sections and segments close to platforms, the small-size lines had the largest number of failures. Away from the platform, however, the medium-size lines had the largest number of failures.

  • Pollution from most of the pipeline failures has been insignificant. About 94% of the incidents with measurable discharge of hydrocarbons resulted in less than 100 bbl of pollution.

  • Clamp repair was implemented in 26% of the total failures, mostly to correct pipe damaged due to corrosion. Spool repair was implemented in 53% of failures.

    Use of mechanical connectors in spool repair was explicitly identified in 5% of the total failures.

  • Severe storm loadings resulted in mostly damaging small-sized pipes on or near the platform jacket, whereas mud slides resulted in damaging mostly medium-sized pipes away from platforms.

  • Anchors and anchor lines have been the leading source of mechanical damage to pipelines.

    Other major sources of mechanical damage to pipelines have been mishaps during laying and trenching (jetting), operation of jack ups, and anchoring of supply and work boats near platforms.

  • The significant majority of pipeline failures occurred in the riser section and on the seabed close to the platform. Therefore, potential improvements to pipeline design and inspection practices should focus on platform risers and the pipeline segment within 500 ft from the platform.


  1. "Pipeline Leaks and Breaks," Minerals Management Service, U.S. Department of Interior, Metairie, La.

  2. Andersen, T., and Misund, A., "Pipeline Reliability: An Investigation of Pipeline Failure Characteristics and Analysis of Pipeline Failure Rates for Submarine and Cross-Country Pipelines," Journal of Petroleum Technology, April 1983.

  3. Starting, J., "A Survey of Pipelines in the North Sea: Incidents During Installation, Testing and Operation," Offshore Technology Conference, Paper No. 4069, 1981.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.